Crystal structure of Sphingobacterium multivorum serine palmitoyltransferase complexed with tris(hydroxymethyl)aminomethane

Serine palmitoyltransferase (SPT) is the rate-limiting key enzyme in the sphingolipid biosynthetic pathway. Here, the structure of SPT from S. multivorum complexed with tris(hydroxymethyl)aminomethane is reported at 1.65 Å resolution, which is an improvement on the previously reported structure at 2.3 Å resolution.

Abstract

Serine palmitoyltransferase (SPT) catalyses the first reaction in sphingolipid biosynthesis: the decarboxylative condensation of l-serine (l-Ser) and palmitoyl-CoA to form 3-ketodihydrosphingosine. SPT from Sphingobacterium multivorum has been isolated and its crystal structure in complex with l-Ser has been determined at 2.3 Å resolution (PDB entry 3a2b). However, the quality of the crystal was not good enough to judge the conformation of the cofactor molecule and the orientations of the side chains of the amino-acid residues in the enzyme active site. The crystal quality was improved by revision of the purification procedure and by optimization of both the crystallization procedure and the post-crystallization treatment conditions. Here, the crystal structure of SPT complexed with tris(hydroxymethyl)aminomethane (Tris), a buffer component, was determined at 1.65 Å resolution. The protein crystallized at 20°C and diffraction data were collected from the crystals to a resolution of 1.65 Å. The crystal belonged to the tetragonal space group P4₁2₁2, with unit-cell parameters a = b = 61.32, c = 208.57 Å. Analysis of the crystal structure revealed C4—C5—C5A—O4P (77°) and C5—C5A—O4P—P (–143°) torsion angles in the phosphate-group moiety of the cofactor pyridoxal 5′-phosphate (PLP) that are more reasonable than those observed in the previously reported crystal structure (14° and 151°, respectively). Furthermore, the clear electron density showing a Schiff-base linkage between PLP and the bulky artificial ligand Tris indicated exceptional flexibility of the active-site cavity of this enzyme. These findings open up the possibility for further study of the detailed mechanisms of substrate recognition and catalysis by this enzyme.

1. Introduction

Serine palmitoyltransferase (SPT) is a key enzyme in sphingolipid biosynthesis and catalyzes the pyridoxal 5′-phosphate (PLP)-dependent decarboxylative condensation of l-serine (l-Ser) with palmitoyl-CoA (PalCoA) to generate 3-keto­dihydrosphingosine (Hanada, 2003 ▸). Eukaryotic SPT is a membrane-bound enzyme composed of two core subunits named SPTLC1 and SPTLC2/SPTLC3 (Buede et al., 1991 ▸; Nagiec et al., 1994 ▸; Hornemann et al., 2006 ▸) that are associated with the regulatory components ssSPTa/ssSPTb and ORMDL3 (Parthibane et al., 2021 ▸; Chauhan et al., 2016 ▸; Han et al., 2009 ▸; Gable et al., 2000 ▸; Breslow et al., 2010 ▸). SPT is essential for cell viability in eukaryotes, and alterations in SPT activity caused by mutations of either the SPTLC1 or the SPTLC2 gene are linked to neurodegenerative diseases such as hereditary sensory and autonomic neuropathy type I (HSAN1) in humans (Bejaoui et al., 2001 ▸; Dawkins et al., 2001 ▸; Hornemann et al., 2009 ▸; Rotthier et al., 2010 ▸). HSAN1-related SPT variants utilize l-alanine or glycine rather than l-Ser as the substrate to produce the corresponding atypical deoxysphingolipids, which lack a critical hydroxyl moiety, cannot be efficiently degraded in vivo and cause toxicity (Penno et al., 2010 ▸; Ernst et al., 2015 ▸; Bode et al., 2016 ▸; Gable et al., 2010 ▸).

Sphingolipid-producing prokaryotes such as Sphingomonas paucimobolis and S. multivorum contain a water-soluble homodimeric SPT (Ikushiro et al., 2001 ▸, 2007 ▸). We have studied the catalytic function of SPT using bacterial enzymes with a series of l-Ser analogs and S-(2-oxoheptadecyl)-CoA, a nonreactive analog of PalCoA (Ikushiro et al., 2004 ▸, 2008 ▸; Shiraiwa et al., 2009 ▸; Ikushiro & Hayashi, 2011 ▸). Spectroscopic and site-directed mutagenesis studies indicated the importance of a unique histidine residue located at the re side of PLP in catalytic function; i.e. the histidine residue anchors l-Ser in the correct orientation to prevent unwanted side reactions and acts as the acid catalyst promoting both Claisen-type condensation and decarboxylation in later steps. The crystal structures of bacterial SPTs support interaction between the histidine residue and the carboxyl group of the l-Ser substrate (Ikushiro et al., 2009 ▸; Raman et al., 2009 ▸).

The previously reported crystal structure of SPT from S. multivorum at 2.3 Å resolution was not good enough to judge the orientation of the bound substrate in the SPT active site (Ikushiro et al., 2009 ▸). In order to improve the resolution, the entire procedures of protein purification, crystallization and cryoprotection were re-examined. Here, the crystal structure of SPT complexed with tris(hydroxymethyl)­aminomethane (Tris), a buffer component that was also found to be essential to obtain suitable crystals, was determined at 1.65 Å resolution.

2. Materials and methods

2.1. Macromolecule production

The expression of the SPT gene derived from S. multivorum and the purification of the gene product were performed as follows. The gene encoding the sequence for SPT, cloned into the pET-21b expression vector (Novagen), was introduced by thermal shock into Escherichia coli BL21(DE3) pLysS cells (Table 1 ▸). The E. coli cells were cultivated in LB Broth medium (Becton, Dickinson & Co., Sparks, Maryland, USA) containing 50 µg ml⁻¹ ampicillin at 37°C. During the exponential growth phase (OD₆₀₀ of 0.6–0.8), SPT expression was induced by adding 0.1 mM isopropyl β-d-1-thiogalacto­pyranoside. After 4 h of further incubation at 37°C, the cell pellet was harvested by centrifugation at 5000g for 5 min and stored at −80°C. The cell pellet was suspended in buffer A (50 mM Tris–HCl buffer pH 7.5, 0.1 mM EDTA) and the cells were disrupted by sonication using a Sonifier model 450 (Branson Ultrasonic, Connecticut, USA). Cell debris was removed by centrifugation at 20 000g for 30 min. The supernatant was applied onto a HiPrep DEAE FF Crude 5 ml column equilibrated with buffer A using an ÄKTA FPLC system (GE Healthcare). The proteins were eluted with a linear gradient of 0–500 mM NaCl in buffer A. The fractions containing SPT were collected, ammonium sulfate was added to 30% saturation and the solution was applied onto a HiPrep Butyl FF Crude 5 ml column equilibrated with buffer A containing 30% saturated ammonium sulfate. SPT was eluted with a decreasing linear gradient of ammonium sulfate concentration (30–0%) in buffer A. The pooled fractions were concentrated and were desalted using a HiPrep Desalting 16/10 column equilibrated with buffer A. The SPT fractions were then again applied onto a HiPrep DEAE FF Crude 5 ml column equilibrated with buffer A and eluted with a linear gradient of 0–500 mM NaCl in buffer A. The pooled SPT was desalted using a HiPrep Desalting 16/10 column equilibrated with 20 mM potassium phosphate buffer pH 7.5 containing 0.1 mM EDTA. Purified SPT was concentrated to 20 mg ml⁻¹, filter sterilized and stored at 4°C. The homogeneity of the purified protein was confirmed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis, which showed a single band corresponding to the molecular mass estimated from the amino-acid sequence of SPT. The concentration of the purified SPT subunit in solution was determined spectrophotometrically using a molar extinction coefficient of 26 780 M ^–1 cm^–1 at 280 nm for the PLP form of the enzyme, which was calculated on the basis of the number of tryptophan and tyrosine residues in SPT (Ikushiro et al., 2007 ▸). Information pertaining to the production of the macromolecule is summarized in Table 1 ▸.

Table 1. Macromolecule-production information.

Source organism	Sphingobacterium multivorum
Gene	Serine palmitoyltransferase (UniProtKB accession code A7BFV6)
DNA source	Genomic DNA
Forward primer†	GACTGTACGCATATGAGTAAAGGAAAGTTAGG
Reverse primer‡	GACTGTACGGAATTCTTATATTAATGTTTCAACTT
Expression vector	pET21_bSPT
Plasmid construction method	Restriction–ligation
Expression host	Escherichia coli BL21(DE3)
Complete amino-acid sequence of the construct produced	MSKGKLGEKISQFKIVEELKAKGLYAYFRPIQSKQDTEVKIDGRRVLMFGSNSYLGLTTDTRIIKAAQDALEKYGTGCAGSRFLNGTLDIHVELEEKLSAYVGKEAAILFSTGFQSNLGPLSCLMGRNDYILLDERDHASIIDGSRLSFSKVIKYGHNNMEDLRAKLSRLPEDSAKLICTDGIFSMEGDIVNLPELTSIANEFDAAVMVDDAHSLGVIGHKGAGTASHFGLNDDVDLIMGTFSKSLASLGGFVAGDADVIDFLKHNARSVMFSASMTPASVASTLKALEIIQNEPEHIEKLWKNTDYAKAQLLDHGFDLGATESPILPIFIRSNEKTFWVTKMLQDDGVFVNPVVSPAVPAEESLIRFSLMATHTYDQIDEAIEKMVKVFKQAEVETLI

^†The NdeI site is in italics.

^‡The EcoRI site is in italics.

2.2. Crystallization

Crystals of SPT were obtained using the sitting-drop vapor-diffusion method in 24-well plates at 20°C. 2 µl aliquots of 20.0 mg ml⁻¹ protein solution were mixed with 4 µl reservoir solution consisting of 100 mM Tris–HCl pH 8.5, 200 mM sodium acetate, 19–23%(w/v) poly(ethylene glycol) 4000 (PEG 4000). The drops were equilibrated against 500 µl reservoir solution for two days and the microbridges were then transferred to a new 24-well plate, in which the fresh reservoir solution consisted of 100 mM Tris–HCl pH 8.5, 200 mM sodium acetate, 15–19%(w/v) PEG 4000. Small plate- or cubic-shaped crystals were confirmed by visual observation within 2–5 days. Plate-shaped crystals reproducibly grew to a large enough size within about one month of incubation at 20°C. Crystallization information is summarized in Table 2 ▸.

Table 2. Crystallization.

Method	Vapor diffusion, sitting drop
Plate type	24-well crystallization plate
Temperature (K)	293
Protein concentration (mg ml−1)	20
Buffer composition of protein solution	20 mM potassium phosphate buffer pH 7.5, 0.1 mM EDTA
Composition of first reservoir solution	100 mM Tris–HCl pH 8.5, 200 mM sodium acetate, 22%(w/v) PEG 4000
Composition of second reservoir solution	100 mM Tris–HCl pH 8.5, 200 mM sodium acetate, 16%(w/v) PEG 4000
Volume and ratio of drop	6 µl, 1:2 ratio of protein:reservoir solution
Volume of reservoir (µl)	500
Composition of the cryoprotectant	20%(v/v) ethylene glycol in the second reservoir solution
Drop setting	Manual
Seeding	No

2.3. Data collection and processing

A crystal with dimensions of about 400 × 400 × 20 µm was used for diffraction experiments. The crystal was cryoprotected by rapid soaking for a few seconds in mother liquor supplemented with 20%(v/v) ethylene glycol and was then flash-cooled in liquid nitrogen or under a N₂ gas cryostream (100 K). X-ray diffraction data were collected on a Rigaku FR-X rotating-anode X-ray source using Cu Kα radiation (λ = 1.5418 Å) equipped with a Rigaku R-AXIS VII image plate as a detector (Rigaku Corporation, Osaka, Japan). The data set was collected to 1.65 Å resolution on the in-house Rigaku system with a crystal-to-detector distance of 120 mm using an oscillation angle of 0.5° and an exposure time of 60 s. A total of 270 diffraction images were used in data processing with the XDS program package (Kabsch, 2010 ▸). Data-collection statistics are detailed in Table 3 ▸.

Table 3. Data collection and processing.

Values in parentheses are for the highest resolution shell.

Diffraction source	Rigaku FR-X rotating anode
Wavelength (Å)	1.5418
Temperature (K)	100
Detector	Rigaku R-AXIS VII image plate
Crystal-to-detector distance (mm)	120
Rotation range per image (°)	0.5
Total rotation range (°)	135
Exposure time per image (s)	60
Space group	P41212
a, b, c (Å)	61.32, 61.32, 208.57
α, β, γ (°)	90, 90, 90
Mosaicity (°)	0.229
Resolution range (Å)	50.00–1.65 (1.75–1.65)
Total No. of reflections	425548 (32429)
No. of unique reflections	47616 (6313)
Completeness (%)	96.8 (81.0)
Multiplicity	8.9 (5.1)
〈I/σ(I)〉	23.4 (6.4)
CC1/2	0.999 (0.956)
R meas	0.063 (0.222)
R p.i.m.	0.024 (0.128)
Overall B factor from Wilson plot (Å2)	21.7

2.4. Structure solution and refinement

Initial phases for the structure were determined by the molecular-replacement method with MOLREP (Vagin & Teplyakov, 2010 ▸) in the CCP4 suite (Winn et al., 2011 ▸) using PDB entry 3a2b as the search model after the removal of all water molecules. The model was refined using REFMAC5 (Murshudov et al., 2011 ▸) in the CCP4 suite and manual adjustment and rebuilding of the model was performed using Coot (Emsley et al., 2010 ▸). The quality of the structure was determined by MolProbity (Chen et al., 2010 ▸). Refinement statistics are summarized in Table 4 ▸.

Table 4. Structure solution and refinement.

Values in parentheses are for the highest resolution shell.

Resolution range (Å)	46.03–1.65 (1.69–1.65)
Completeness (%)	96.96 (73.95)
σ Cutoff	None
No. of reflections, working set	42685 (2362)
No. of reflections, test set	4916 (270)
Final R cryst †	0.170 (0.254)
Final R free ‡	0.214 (0.323)
Cruickshank DPI	0.1067
No. of non-H atoms
Protein§	3181
PLP–Tris	23
Ethylene glycol	24
Waters§	458
Total	3686
R.m.s. deviations
Bond lengths (Å)	0.011
Angles (°)	1.675
Average B factors¶ (Å2)
Protein	15.0
PLP–Tris	11.8
Ethylene glycol	38.2
Waters	27.6
Ramachandran plot
Favored regions (%)	98
Additionally allowed (%)	2
Unmodelled/incomplete residues (%)	1.3

^† R _cryst is the conventional crystallographic R factor; R _cryst = , where F _obs and F _calc are the observed and calculated structure factors, respectively.

^‡ R _free was calculated as for R _cryst but was calculated for 10% of the reflections that were chosen at random and omitted from the refinement process (Brünger, 1992 ▸).

^§The corresponding two atoms with 50% occupancy in the double conformers were counted separately.

^¶The average temperature factor was calculated based on the number of non-H atoms.

3. Results and discussion

3.1. Improvement in SPT crystal quality

The second anion-exchange column chromatography step using an ÄKTA FPLC system was found to be critical to improve the quality of the crystals of recombinant SPT. Initial crystallization trials with 19–23.5%(w/v) PEG 4000 as a precipitant yielded thin square-planar crystals that were clustered in the droplets after 1–3 weeks: each crystal had a maximum size of 0.1 mm and a maximum diffraction limit of 2.0–2.3 Å (Fig. 1 ▸ a).

Figure 1.

Photographs of the SPT crystals obtained before (a) and after (b) improvement of the purification and crystallization conditions.

We explored crystallization processes according to the ‘backing off’ experiment (Saridakis & Chayen, 2000 ▸, 2003 ▸) using a minor modification of the sitting-drop method in which crystal growth was controlled by separating the nucleation and growth phases by transferring microbatch drops, which had been incubated under spontaneous nucleation conditions for varied periods, to metastable conditions. The best crystals were obtained by transferring droplets incubated for two days under nucleation conditions containing 19–23%(w/v) PEG 4000 to wells under metastable conditions containing 15–19%(w/v) PEG 4000 (Fig. 1 ▸ b).

Although good crystals were occasionally also obtained at lower PEG concentrations (without incubation under the nucleation conditions) after 8–12 weeks, the transfer method reproducibly yielded fewer, larger and better crystals within four weeks. Crystals thus obtained yielded diffraction data to 1.6–1.9 Å resolution, but serious fading or diffuse intensity of the X-ray diffraction patterns during X-ray irradiation often prohibited proper structure determination. After several cryoprotectants, such as glycerol, ethylene glycol and PEGs, had been examined at different concentrations, deterioration of the diffraction quality was found to be controlled well by soaking the crystals in cryoprotectant solution containing 20%(v/v) ethylene glycol, enabling high-redundancy data collection in which the diffraction signals were sufficiently strong.

3.2. Crystal structure of the SPT–Tris complex

3.2.1. Overall crystal structure of the SPT–Tris complex

Diffraction data were collected from a fragment of the largest crystal with original dimensions of about 400 × 200 × 20 µm (Fig. 1 ▸ b). The crystal diffracted to 1.65 Å resolution and the diffraction data were indexed in the tetragonal space group P4₁2₁2, with unit-cell parameters a = b = 61.32, c = 208.57 Å. A summary of the data statistics is presented in Table 3 ▸. Our modifications of the crystallization process did not affect the space group and unit-cell dimensions of the SPT crystal.

The crystal structure of SPT complexed with Tris, a component of the crystallization mother solution, was determined at a resolution of 1.65 Å and refined to R _cryst and R _free values of 0.170 and 0.214, respectively, with a Cruickshank diffraction-component precision index of 0.1067 Å. The refinement statistics are summarized in Table 4 ▸. The final model contains 394 amino-acid residues in the polypeptide chain, one PLP–Tris Schiff base, 456 water molecules and six ethylene glycol molecules per monomer (Fig. 2 ▸). The first observed residue is Ser2 and the last is Val395. Four residues at the C-terminus (Glu396-Thr397-Leu398-Ile399) were not observed in the electron-density maps. One local structural difference involves the main chain of Arg127 and Asn128; in the new SPT structure the amide plane connecting the carbonyl group of Arg127 and the amino group of Asn128 is flipped relative to its conformation in the previously reported SPT structure at 2.3 Å resolution (PDB entry 3a2b). All side chains of the amino-acid residues, except for Phe83* as mentioned below, were well resolved, although 20 residues, mainly located on the surface of the protein molecule and exposed to solvent, were clearly divided into two conformations. The B factors of the protein moiety and the PLP–Tris aldimine moiety were decreased to 15.0 and 11.9 Ų Da⁻¹, respectively (Fig. 3 ▸; the B factor of the protein moiety was 27.2 Ų Da⁻¹ in PDB entry 3a2b).

Figure 2.

A ribbon diagram of the SPT symmetric dimer in complex with Tris viewed down its crystallogaphic twofold axis. SPT crystallized with one monomer per asymmetric unit and a dimer was generated by crystallo­graphic symmetry. The N- and C-termini of subunit A are labeled N and C, respectively. The three structural domains of subunit A are shown in different colors: the N-terminal domain is in brown, the small domain is in blue and the large domain is in green. The PLP–Tris aldimine and ethylene glycol in subunit A are shown as yellow and magenta stick models, respectively. The symmetry partner is shown in gray. This figure was generated by PyMOL (version 2.4.0; Schrödinger; http://www.pymol.org/).

Figure 3.

B-factor diagram of SPT as represented by the B-factor putty program in PyMOL. (a) 1.65 Å resolution structure of SPT complexed with Tris (this work; PDB entry 8guh). (b) 2.3 Å resolution structure of SPT complexed with l-serine (PDB entry 3a2b). The diameter of the tubes is proportional to the B factor. A darker shade of red corresponds to a higher B factor, whereas a darker shade of blue corresponds to a lower B factor.

3.2.2. Crystal structure of the active site of the SPT–Tris complex

The PLP–Tris aldimine fitted well into the electron density (Fig. 4 ▸ a). Electron density for aldimine formation between the C4A atom of PLP and Tris was visible in the map, while it was clear from the electron density that there was no covalent linkage between the PLP cofactor and Lys244 (Figs. 4 ▸ b and 4 ▸ c). Although the C4—C5—C5A—O4P (77°) and C5—C5A—O4P—P (–143°) torsion angles of of PLP in the improved SPT crystal differed from those in PDB entry 3a2b (14° and 151°, respectively; Figs. 4 ▸ a and 5 ▸ a), the architecture of the interaction network surrounding PLP of the active-site cavity was basically conserved as follows (Figs. 4 ▸ b, 4 ▸ c and 5 ▸ a). The pyridine ring of PLP is sandwiched between the side chains of His138 and Ala212 by van der Waals interactions and is held in position by a hydrogen-bond/electrostatic interaction between the side chain of Asp210 and the N1 atom of PLP. The O3 atom of PLP forms a hydrogen bond to N^ɛ2 of His213, and the phosphate moiety of PLP was anchored by polar interactions with active-site residues such as the main-chain N atoms of Phe114 and Ala274*, the side chains of Ser243 and Ser273* and two water molecules. (To distinguish the residues in the two subunits in the biological dimer, we labeled the residues from the opposite subunit with an asterisk.) As shown in Fig. 5 ▸(a), the electron density for the water molecules in the active site of the SPT–Tris complex (PDB entry 8hug) is clearer and more spherically shaped than that in the SPT–l-Ser complex (PDB entry 3a2b), substantiating the accuracy of the improved crystallographic analysis in the present study.

Figure 4.

Close-up view of a portion of the SPT active site. The amino-acid residues and the bound PLP–Tris aldimine molecule are shown in stick representation. The aldimine and each subunit of the dimer are color-coded by C atoms (yellow, green and blue). N, O and P atoms are colored blue, red and orange, respectively. Residues from the opposite subunit are labeled with an asterisk. The dashed lines show hydrogen bonds. (a) Electron density around the PLP–Tris Schiff base (external aldimine). The mesh represents a composite omit map contoured at 1σ. The atoms related to torsion angles of the PLP moiety (C4, C5, C5A, O4P and P) are labeled. (b) Two-dimensional residue-interaction diagrams observed in the SPT active site. Water molecules are shown in gray. This figure was generated with the LigPlot+ program (version 2.2; https://www.ebi.ac.uk/thornton-srv/software/LigPlus/). The calculated 2F _o − F _c electron-density map contoured at the 1σ level is shown in (c), (d) and (e). (c) Stereorepresentation of the SPT active site. (d) Electron density around the side chain of Thr241. (e) Electron density around the side chain of Phe83*. The main-chain structure of each subunit is color-coded in the same way. The figures were generated by PyMOL (version 2.4.0; Schrödinger; http://www.pymol.org/).

Figure 5.

Superimposition of the crystal structures of the complexes with Tris (PDB entry 8guh) and l-Ser (PDB entry 3a2b) (a) and the putative binding mode of 3-ketodihydrosphingosine (KDS) (b). (a) Stereorepresentation of the superimposed active-site structures. The Tris–SPT complex is color-coded as in Fig. 4 ▸. The l-Ser–SPT complex is shown in gray. The amino-acid residues and the bound PLP aldimine molecules are shown in stick representation. Water molecules are shown as spheres and their calculated 2F _o − F _c electron-density maps contoured at the 1σ level are shown in blue for the Tris–SPT complex and gray for the l-Ser–SPT complex. (b) The KDS-bound subunit of the SPT homodimer is shown as a surface model and the opposite subunit is shown as a ribbon model. Each subunit is highlighted by amino-acid hydrophobicity, with red representing the most hydrophobic through pink for neutral to white for the most hydrophilic. The side chain of Phe83* and the reaction product KDS are shown as red and green spheres, respectively. This figure was generated by PyMOL (version 2.4.0; Schrödinger; http://www.pymol.org/).

The side chain of Thr241 was revealed to have two conformers: the hydroxyl group in one conformation is directed towards the ɛ-amino group of Lys244, while the hydroxyl group in the other conformation is directed towards the phosphate group of PLP (Figs. 4 ▸ d and 5 ▸ a). The previous crystal structure of the S. multivorlum SPT–l-Ser complex (PDB entry 3a2b) showed that the hydroxyl group of Thr241 adopts a single conformation directed towards the ɛ-amino group of Lys244 (Ikushiro et al., 2009 ▸). The crystal structures of S. paucimobilis SPT showed that Thr262 (corresponding to Thr241 of S. multivorlum SPT) adopts a single conformation, and its hydroxyl group interacts with the phosphate group of PLP regardless of the presence or absence of the ligand (Yard et al., 2007 ▸; Raman et al., 2009 ▸). The observed conformational disorder of the side chain of Thr241 in the S. multivorlum SPT–Tris complex might be produced by the improved data quality or might be caused by the difference in the bound ligand. One of the hydroxyl groups of Tris forms hydrogen bonds to the phosphate group of PLP and one water molecule and indirectly interacts with the main-chain carbonyl group of Ser243 via this water molecule. Another hydroxyl group of Tris forms a hydrogen bond to N^ɛ2 of His138, and the remaining hydroxyl group of Tris forms hydrogen bonds to N^ɛ2 of His213 and the O3 atom of PLP (Figs. 4 ▸ b and 4 ▸ c).

The hydrophobic side chain of Phe83* of the improved S. multivorlum SPT crystal, which is located in the binding site for the acyl-chain moiety of the acyl-CoA substrate or the reaction product (Fig. 5 ▸ b), was not well resolved (Fig. 4 ▸ e). The orientation of Met104* in the S. paucimobilis SPT crystals (corresponding to Phe83* of S. multivorlum SPT) changed depending on the presence or absence of the ligand (Yard et al., 2007 ▸; Raman et al., 2009 ▸). These observations indicate that there is a space or some flexibility allowing rotation of the bulky side chain of the amino-acid residue at this position, and such leeway may be related to the broad acyl-CoA substrate specificity of bacterial enzymes (Ikushiro et al., 2001 ▸, 2007 ▸).

The K _d value of S. multivorlum SPT for Tris was estimated as 39.8 ± 5.6 mM by the spectroscopic titration method (Ikushiro et al., 2004 ▸) and therefore the observation that SPT co-crystallized with Tris was within expectation. Other buffer bases lacking an amino group were not suitable for the growth of a single crystal of S. multivorlum SPT. Tris binding may promote the S. multivorlum SPT to assemble with a preferred orientation or could fix the structure of SPT by occupying the space in the enzyme active site and stabilizing the packing state of SPT molecules by restraining the domain movements, resulting in better quality of the SPT crystal.

4. Conclusion

We improved the quality of the S. multivorum SPT crystal by adding one more step to the protein-purification procedure, optimizing the nucleation and growth conditions for crystallization and utilizing a suitable cryoprotectant. The crystal structure of SPT complexed with Tris was determined at 1.65 Å resolution. Clearly observed electron density confirmed the presence of a Schiff-base linkage between PLP and Tris, which mimics the external aldimine intermediate formed in the SPT catalytic cycle. This indicates an exceptionally flexible capacity of the active-site cavity of the S. multivorlum SPT. The high-resolution crystal structure that we have presented here will be useful in further studying the mechanisms of ligand binding and catalysis in SPT. The atomic coordinates and crystal structure of SPT have been deposited in the Protein Data Bank with accession code 8guh.

Supplementary Material

PDB reference: serine palmitoyltransferase in complex with Tris, 8guh

Funding Statement

This work was supported by Japan Society for the Promotion of Science KAKENHI Grants No. 25440036 and 22K06153 to HI.