Crystallization and preliminary X-ray crystallographic analysis of bacterial tRNA^Sec in complex with seryl-tRNA synthetase

Bacterial selenocysteine tRNA was crystallized as the heterologous complex with archaeal seryl-tRNA synthetase. X-ray diffraction was improved by introducing point mutations and heavy-atom labeling, and a 3.2 Å diffraction data set for phase determination was obtained from a platinum-labeled crystal.

Abstract

Selenocysteine (Sec) is translationally incorporated into proteins in response to the UGA codon. The tRNA specific to Sec (tRNA^Sec) is first ligated with serine by seryl-tRNA synthetase (SerRS). To elucidate the tertiary structure of bacterial tRNA^Sec and its specific interaction with SerRS, the bacterial tRNA^Sec from Aquifex aeolicus was crystallized as the heterologous complex with the archaeal SerRS from Methanopyrus kandleri. Although X-ray diffraction by crystals of tRNA^Sec in complex with wild-type SerRS was rather poor (to 5.7 Å resolution), the resolution was improved by introducing point mutations targeting the crystal-packing interface. Heavy-atom labelling also contributed to resolution improvement. A 3.2 Å resolution diffraction data set for phase determination was obtained from a K₂Pt(CN)₄-soaked crystal.

1. Introduction  

Selenocysteine (Sec), the 21st amino acid, contains selenium and is translationally incorporated into proteins (Böck et al., 1991 ▶). The Sec-­specific tRNA (tRNA^Sec) is the longest tRNA and contains an anticodon complementary to the stop codon UGA (Leinfelder et al., 1988 ▶). The secondary structures of tRNA^Sec species from bacteria and eukarya/archaea are quite different from the canonical structure and even from each other (Sturchler et al., 1993 ▶; Baron et al., 1993 ▶; Jühling et al., 2009 ▶). In contrast to the canonical amino acids, Sec is synthesized on tRNA^Sec through multiple steps, in which tRNA^Sec is first ligated with serine by seryl-tRNA synthetase (SerRS; Leinfelder et al., 1988 ▶) and serine is then converted to Sec. SerRS is also responsible for serine ligation to the canonical tRNA^Ser, which has been analyzed using crystal structures of the SerRS–tRNA^Ser complex (Biou et al., 1994 ▶).

Although crystal structures of eukaryal and archaeal tRNA^Sec species have been reported (Itoh et al., 2009 ▶; Chiba et al., 2010 ▶), that of bacterial tRNA^Sec has not been solved. Furthermore, structural information on the SerRS–tRNA^Sec complex is required in order to understand the tRNA-specificity mechanism of SerRS, which recognizes distinct tRNA species. In this communication, we report the crystallization of the bacterial tRNA^Sec from Aquifex aeolicus in complex with the heterologous SerRS from the archaeon Methanopyrus kandleri. Strategies for improvement of diffraction resolution and phase determination are also discussed.

2. Materials and methods  

2.1. Overexpression and purification of M. kandleri SerRS  

The M. kandleri SerRS gene was cloned into pET26b vector (Novagen). Point mutations were introduced with a QuikChange site-directed mutagenesis kit (Stratagene). Escherichia coli strain Rosetta 2 (DE3) (Stratagene) was transformed with the expression plasmids and the proteins were overexpressed in LB medium at 310 K. The selenomethionine-substituted SerRS mutant was overexpressed in E. coli Rosetta 2 (DE3) transformed with the same vector as that encoding the native protein. The cells were grown in M9 minimal medium at 310 K to an OD₆₀₀ of 0.25; 60 mg l⁻¹ l-selenomethionine (SeMet), 100 mg l⁻¹ each of l-threonine, l-lysine and l-phenyl­alanine and 50 mg l⁻¹ each of l-leucine, l-isoleucine and l-valine were then added to the medium. After a 15 min pre-incubation, gene expression was induced by adding 0.2 mM isopropyl β-d-1-thio­galactopyranoside. The cells were cultured at 298 K for a further 22 h prior to harvesting.

The harvested cells expressing wild-type and mutant SerRSs were resuspended in 50 mM Tris–HCl buffer pH 8.5 containing 500 mM NaCl, 500 mM MgCl₂, 10 mM β-mercaptoethanol (β-ME) and were disrupted using an ultrasonic homogenizer. The supernatant of the lysate was incubated at 358 K for 30 min to denature and remove most of the E. coli proteins. After centrifugation, ammonium sulfate was added to the supernatant (final concentration 2 M), which was applied onto a Toyopearl Phenyl-650 column (Tosoh). The column was washed with 20 mM Tris–HCl buffer pH 7.5 containing 2.0 M ammonium sulfate and 10 mM β-ME and the proteins were eluted with a linear gradient of 2.0–0 M ammonium sulfate. They were dialyzed against 20 mM Tris–HCl buffer pH 7.5 containing 0.2–0.4 M NaCl and 10 mM β-ME and concentrated to 12–15 mg ml⁻¹.

2.2. Lysine methylation  

Methylation of the lysine residues in wild-type and mutant SerRS was performed using formaldehyde and dimethylamine–borane complex as described by Walter et al. (2006 ▶). Prior to methylation, the proteins were diluted with 20 mM HEPES–NaOH buffer pH 7.5 containing 0.4–0.8 M NaCl and 10 mM β-ME. The methylation reagents were removed by repeated dialyses against Tris–HCl buffer pH 7.5 containing 0.2–0.4 M NaCl and 10 mM β-ME.

2.3. Preparation of the tRNA transcript  

The gene encoding A. aeolicus tRNA^Sec was cloned into the pUC119 vector together with the T7 promoter sequence. The template DNA for in vitro transcription was amplified by PCR using the plasmid as the template. In the PCR reaction, the M13 forward primer and the primer complementary to the 3′ end of the target tRNA were used. The tRNA was transcribed in vitro with T7 RNA polymerase, as described previously (Sekine et al., 1996 ▶), and was purified by chromatography on a Resource Q column (GE Healthcare). The purified tRNA was dissolved in 10 mM Tris–HCl buffer pH 8.0 containing 10 mM MgCl₂ and was heated to 338 K for 5 min for refolding.

2.4. Crystallization and diffraction data collection  

The complex of wild-type M. kandleri SerRS and A. aeolicus tRNA^Sec was prepared by mixing the components at final concentrations of 80 and 120 µM, respectively, in 20 mM Tris–HCl buffer pH 7.5 containing NaCl (100 mM for wild-type SerRS and 400 mM for the mutants), 10 mM MgCl₂, 3 mM β-ME, 500 µM 5′-O-[N-(l-seryl)sulfamoyl]adenosine (Ser-SA). Ser-SA is a nonhydrolyzable analogue of seryl-adenylate.

Crystallization was performed at 293 K using the sitting-drop vapour-diffusion method by mixing the SerRS–tRNA^Sec solution and reservoir solution in a 1:1 ratio. Four screening kits, Crystal Screen 2, Natrix (Hampton Research) and Wizard I and II (Emerald Bio­Systems), were used for initial screening of the crystallization conditions. The initial crystallization conditions were refined by optimizing the reservoir-solution composition, including the precipitant concentration, buffer pH and additives.

Prior to the X-ray diffraction experiment, the crystals were transferred into a cryoprotective solution and flash-cooled in a cryocooled N₂ stream (∼90 K). For all experiments, an optimized cryoprotective solution (100 mM trisodium citrate–HCl buffer pH 5.6 containing 2.0 M ammonium sulfate, 1.7 M lithium sulfate, 75 mM potassium sodium tartrate and 10 mM MgCl₂) was used. Diffraction data were collected under a cryocooled N₂ stream on BL41XU at SPring-8 (Hyogo, Japan), and BL5A, NE3A and NW12A at the Photon Factory (PF; Tsukuba, Japan). The data were processed with HKL-2000 (Otwinowski & Minor, 1997 ▶).

2.5. Soaking of the heavy-atom reagent  

The crystals were washed with the reservoir solution and then transferred into a solution of the heavy-atom reagent diluted in the reservoir solution. After 10–15 h incubation, the crystals were transferred into the cryoprotective solution and flash-cooled. The X-­ray absorption fine structure (XAFS) was measured to monitor the heavy-atom labelling and to estimate the peak and edge wavelengths of the anomalous scattering.

2.6. Phase determination  

Phaser (McCoy et al., 2007 ▶) was used for molecular replacement and autoSHARP (Vonrhein et al., 2007 ▶) was used for phase calculation by the single isomorphous replacement with anomalous scattering (SIRAS) and multiwavelength anomalous dispersion (MAD) methods. DM (Winn et al., 2011 ▶) was used for subsequent density modification.

3. Results  

3.1. Crystallization and diffraction analysis of the SerRS–tRNA^Sec complex with wild-type SerRS  

Initial crystallization-condition screening of the complex of M. kandleri SerRS and A. aeolicus tRNA^Sec yielded cube-shaped crystals using reservoir solution No. 14 of Crystal Screen 2 (100 mM trisodium citrate–HCl buffer pH 5.6 containing 2.0 M ammonium sulfate and 200 mM potassium sodium tartrate). A brief optimization trial of the reservoir solution indicated that the initial conditions were the best. An ∼250 µm crystal obtained by mixing 0.75 µl complex solution and 0.75 µl reservoir solution (Fig. 1 ▶ a) diffracted to 5.7 Å resolution. The crystal belonged to space group I432, with unit-cell parameters a = b = c = 270 Å (Table 1 ▶).

Figure 1.

Crystals of the M. kandleri SerRS–A. aeolicus tRNA^Sec complex. Typical crystals of the tRNA^Sec in complex with the wild-type SerRS (a), Lys-methylated wild-type SerRS (b), Lys-methylated SerRS^Set2 (c), Lys-methylated native SerRS^Set3 (d) and Lys-methylated SerRS^Set3 with SeMet substitution (e) are shown. Each scale bar indicates 200 µm.

Table 1. Data-collection and phasing statistics.

Values in parentheses are for the highest resolution shell. FOM, figure of merit; DM, density modification (solvent flattening).

SerRS mutation	Wild type	Wild type	Wild type	Set2	Set2	Set3	Set3			Set3 [SeMet]
Lys methylation	−	−	+	+	+	+	+			+
Heavy-atom labelling		HgCl2			HgCl2		K2Pt(CN)4			K2Pt(CN)4
Data collection
Beamline	SPring-8 BL41XU	SPring-8 BL41XU	SPring-8 BL41XU	SPring-8 BL41XU	PF NW12A	PF BL5A	PF NE3A			PF BL5A
Wavelength (Å)	0.97917	1.00000	1.00000	0.97896	1.00418 [peak]	1.00000	1.07186 [peak]	1.07227 [edge]	1.05404 [remote]	0.97848
Space group	I432	I432	P432	I432	I432	I432	I432			I432
Unit-cell parameters (Å)	a = b = c = 269.8	a = b = c = 273.2	a = b = c = 271.6	a = b = c = 274.4	a = b = c = 274.7	a = b = c = 274.0	a = b = c = 273.1			a = b = c = 272.8
Resolution (Å)	50.0–5.70 (5.90–5.70)	50.0–6.30 (6.52–6.30)	50.0–6.10 (5.32–6.10)	50.0–4.60 (4.76–4.60)	50.0–4.70 (4.87–4.70)	50.0–3.95 (4.09–3.95)	50.0–3.20 (3.31–3.20)	50.0–3.20 (3.31–3.20)	50.0–3.20 (3.31–3.20)	50.0–3.10 (3.21–3.10)
Unique reflections	5141 (482)	3978 (372)	8667 (837)	10021 (980)	9491 (933)	15660 (1545)	28852 (2825)	28815 (2835)	27446 (2694)	31714 (3126)
Completeness (%)	99.6 (99.8)	99.9 (100.0)	100.0 (99.8)	99.7 (100.0)	100.0 (100.0)	99.8 (100.0)	99.9 (100.0)	99.9 (100.0)	99.9 (100.0)	99.9 (100.0)
Mutliplicity	13.3 (12.4)	20.6 (21.1)	17.3 (12.4)	14.2 (13.9)	20.8 (21.0)	19.6 (19.3)	21.6 (21.7)	21.3 (19.4)	21.0 (18.4)	14.6 (15.0)
Rmerge†	0.087 (>1)	0.098 (>1)	0.104 (>1)	0.153 (>1)	0.138 (>1)	0.083 (>1)	0.101 (>1)	0.091 (>1)	0.098 (>1)	0.094 (0.847)
〈I/σ(I)〉	25.1 (2.14)	30.4 (2.19)	24.5 (2.13)	14.5 (2.06)	21.3 (2.16)	31.8 (2.01)	34.7 (2.78)	33.6 (2.20)	31.5 (1.77)	28.4 (3.49)
Phasing		SIRAS [Hg]			SIRAS [Hg]		MAD [Pt]
No. of identified heavy-atom sites		0.5‡			0.5‡		11
Phasing power
Iso (acentric/centric)		0.760/0.463			0.788/0.522		—	0.917/0.819	0.537/0.471
Ano		0.605			0.216		2.20	1.42	1.36
FOM (before DM/after DM)		0.260/0.325			0.231/0.680		0.460/0.811

^† R _merge = , where I_i(hkl) is the intensity of the ith measurement of reflection hkl and 〈I(hkl)〉 is the average value of I_i(hkl) for all i measurements.

^‡Special position.

In order to improve the diffraction resolution, methylation of the lysine residues of SerRS was examined. The complex with Lys-methylated SerRS was crystallized under the same conditions as used for SerRS without Lys methylation and somewhat larger crystals were obtained (Fig. 1 ▶ b). An ∼400 µm crystal diffracted to 6.1 Å resolution and belonged to space group P432, with unit-cell parameters a = b = c = 272 Å (Table 1 ▶). Although the crystallization efficiency was improved, the diffraction resolution was rather worse and was accompanied by a slight expansion of the unit-cell length.

3.2. Preliminary phasing of the wild-type SerRS–tRNA^Sec complex  

Point mutations targeting the protein surface that participates in crystal packing are expected to improve the diffraction resolution. To estimate the crystal-packing surface, various positions within SerRS and tRNA^Sec were investigated using the molecular-replacement method. M. kandleri SerRS consists of an N-terminal tRNA-binding domain (residues 1–167) and a C-terminal catalytic domain (residues 168–527). The position of the C-terminal domain was identified by using the corresponding domain structure from the structure of Methanosarcina barkeri SerRS (PDB entry 2cj9; Bilokapic et al., 2006 ▶) as the search model. The asymmetric unit contained one subunit of the SerRS homodimer, which was built by the crystallo­graphic symmetry. In contrast, the N-terminal domain of SerRS and tRNA^Sec were not found.

The SIRAS method was employed to estimate the positions of the N-terminal domain of SerRS and tRNA^Sec. A mercury-labelled SerRS–tRNA^Sec complex was obtained by cocrystallization in 5.0 mM HgCl₂ diluted with the reservoir solution (100 mM trisodium citrate–HCl buffer pH 5.6 containing 2.1 M ammonium sulfate and 200 mM potassium sodium tartrate). The mercury-labelled crystal (∼150 µm in size) was transferred into the cryoprotective solution and diffraction data were collected. The crystal diffracted to a resolution of 6.3 Å and belonged to space group I432, with unit-cell parameters a = b = c = 273 Å (Table 1 ▶). A 0.5 site (special position) for the Hg atom was identified based on peaks in the |F _{Hg-derivative}| − |F _native| map and the anomalous-difference Fourier map generated using the initial phase calculated by the molecular-replacement method. The experimental phase was calculated using the identified mercury site (Table 1 ▶). The |F _o| map generated by the experimental phase indicated the rough electron density of the components.

3.3. Improvement by point mutations (first trial)  

The |F _o| map indicated that the N-terminal domain of SerRS participates in protein–protein contacts in the crystal packing, in contrast to the C-terminal domain. To reinforce the crystal packing, Tyr residues were introduced into the contact surface. Four point mutations K100Y/R103Y/D126Y/D127Y (SerRS^Set1) and five point mutations R55Y/E58Y/E62Y/R116Y/D118Y (SerRS^Set2) were introduced into SerRS and crystallization of the SerRS–tRNA^Sec complexes was attempted. No crystals were obtained when the SerRS^Set1 mutant (with or without Lys methylation) was used. In contrast, the complex of the SerRS^Set2 mutant (SerRS^Set2–tRNA^Sec complex) crystallized, although crystal growth was quite slow. The growth rate was somewhat enhanced by using Lys-methylated SerRS^Set2 and a reservoir solution consisting of 100 mM trisodium citrate–HCl buffer pH 5.6 containing 1.6–1.7 M ammonium sulfate and 200 mM potassium sodium tartrate (Fig. 1 ▶ c). An ∼100 µm crystal of Lys-methylated SerRS^Set2–tRNA^Sec complex diffracted to 4.6 Å resolution and belonged to space group I432, with unit-cell parameters a = b = c = 274.4 Å (Table 1 ▶). The diffraction resolution was improved remarkably by introduction of the Tyr residues and was accompanied by expansion of the unit-cell length.

3.4. Improvement by point mutations (second trial)  

An ∼100 µm crystal of Lys-methylated SerRS^Set2–tRNA^Sec complex was soaked in 10 mM HgCl₂ and diffraction data were collected at the peak X-ray wavelength of mercury. The crystal diffracted to 4.7 Å resolution and belonged to space group I432, with unit-cell parameters a = b = c = 274.7 Å (Table 1 ▶). SIRAS phasing was performed using diffraction data from the native and mercury-labelled crystals of the Lys-methylated SerRS^Set2–tRNA^Sec complex. Based on the updated |F _o| map, Tyr and Arg residues were introduced into the SerRS surface involved in the protein–RNA contact of the crystal packing. The six-site point mutation E189R/D193R/E379Y/E383R/E497Y/E499Y was introduced into SerRS^Set2 to create SerRS with 11 point-mutation sites (SerRS^Set3).

The Lys-methylated SerRS^Set3–tRNA^Sec complex was crystallized using a reservoir solution consisting of 100 mM trisodium citrate–HCl buffer pH 5.6 containing 1.3 M ammonium sulfate, 210 mM potassium sodium tartrate and 50 mM imidazole. However, the growth rate was extremely slow and the crystals were too small. In order to obtain larger crystals, the volume of the crystallization drop was increased (from 1.5 to 13.2 µl) and the number of crystal nuclei was controlled by seeding 20–50 µm crystals. The crystals grew to 500–800 µm in 6–8 weeks (Fig. 1 ▶ d). They diffracted to ∼4.0 Å resolution and belonged to space group I432, with unit-cell parameters a = b = c = 274 Å (Table 1 ▶). Crystallization of the Lys-methylated SerRS^Set3–tRNA^Sec complex using SeMet-substituted SerRS^Set3 was also performed in the same manner as for the native protein complex (Fig. 1 ▶ e). An ∼500 µm crystal of the complex diffracted to ∼4.0 Å resolution and belonged to space group I432, with unit-cell parameters a = b = c = 274 Å. The improvement in the diffraction resolution was likely to be a consequence of the larger crystal sizes; the contributions of the additional point mutations were unclear.

3.5. Improvement by soaking in a heavy-atom reagent  

To reinforce the phase information, heavy-atom labelling was examined by soaking crystals of the Lys-methylated SerRS^Set3–tRNA^Sec complex in 20 mM HgCl₂, K₂PtCl₄ or K₂Pt(CN)₄. The HgCl₂-soaked and the K₂PtCl₄-soaked crystals diffracted to 4.0 and 5.5 Å resolution, respectively. The heavy-atom labelling was monitored by XAFS measurements. In contrast, K₂Pt(CN)₄-soaked crystals diffracted to ∼3.2 Å resolution and belonged to space group I432, with unit-cell parameters a = b = c = 273 Å. Small cracks appeared in the crystals when they were transferred into the K₂Pt(CN)₄ solution, but the cracks became invisible during incubation. Therefore, K₂Pt(CN)₄ soaking may trigger slight slippage at the packing interface which results in an unexpected improvement of the diffraction resolution.

Diffraction data were collected from the K₂Pt(CN)₄-soaked crystal at three wavelengths (peak, edge and remote). 11 Pt-atom sites were identified based on the anomalous difference Fourier map generated using phases calculated from the preliminary models of SerRS and tRNA^Sec. The refined experimental phase was calculated using the MAD method (Table 1 ▶). In contrast, the highest resolution data (3.10 Å) were obtained from an SeMet-substituted crystal soaked with K₂Pt(CN)₄ (Table 1 ▶).

4. Discussion  

M. kandleri SerRS contains 24 Lys residues (4.55%) and its theor­etical isoelectric point is 4.9. The Lys methylation examined in this study did not alter the crystallization conditions and had either a neutral or a slightly negative effect on the diffraction resolution. However, it exerted a positive effect on crystal formation and growth.

Introducing point mutations in the protein–protein interface in the crystal packing had a negative effect on crystal formation and growth, and no crystals were obtained when SerRS^Set1 was used. However, a dramatic improvement in the diffraction resolution was observed using SerRS^Set2. This indicated that the introduction of Tyr residues into the protein–protein interface could improve the diffraction resolution. In contrast, point mutations targeted to the protein–RNA interface had little effect on the resolution. Heavy-atom labelling can also affect the diffraction resolution. K₂Pt(CN)₄ soaking exhibited a positive effect, while neutral and negative effects were observed on HgCl₂ and K₂PtCl₄ soaking, respectively.