The Rel stringent factor from Thermus thermophilus: crystallization and X-ray analysis

Crystallization and structural analysis are reported of the Rel stringent factor from Thermus thermophilus in the resting state and bound to nucleotides.

Abstract

The stringent response, controlled by (p)ppGpp, enables bacteria to trigger a strong phenotypic resetting that is crucial to cope with adverse environmental changes and is required for stress survival and virulence. In the bacterial cell, (p)ppGpp levels are regulated by the concerted opposing activities of RSH (RelA/SpoT homologue) enzymes that can transfer a pyrophosphate group of ATP to the 3′ position of GDP (or GTP) or remove the 3′ pyrophosphate moiety from (p)ppGpp. Bifunctional Rel enzymes are notoriously difficult to crystallize owing to poor stability and a propensity for aggregation, usually leading to a loss of biological activity after purification. Here, the production, biochemical analysis and crystallization of the bifunctional catalytic region of the Rel stringent factor from Thermus thermophilus (Rel_Tt ^NTD) in the resting state and bound to nucleotides are described. Rel_Tt and Rel_Tt ^NTD are monomers in solution that are stabilized by the binding of Mn²⁺ and mellitic acid. Rel_Tt ^NTD crystallizes in space group P4₁22, with unit-cell parameters a = b = 88.4, c = 182.7 Å, at 4°C and in space group P4₁2₁2, with unit-cell parameters a = b = 105.7, c = 241.4 Å, at 20°C.

1. Introduction  

Nutrient starvation in bacteria is signalled by the alarmone guanosine polyphosphate [guanosine 3′,5′-bisdiphosphate and guanosine 5′-triphosphate-3′-diphosphate, collectively referred to as (p)ppGpp] synthesized by the transfer of pyrophosphate from ATP to the 3′ position of the ribose of GDP or GTP (Haseltine et al., 1972 ▸). An increased concentration of the alarmone drives the so-called stringent response, a crucial mechanism of bacterial adaptation to a rapidly and continuously changing environment (Laffler & Gallant, 1974 ▸). The stringent response affects transcription, translation and DNA replication (Durfee et al., 2008 ▸; Mechold et al., 2013 ▸). The effects of (p)ppGpp on gene expression are mediated via direct binding to its molecular targets, RNA polymerase, DnaG and translational GTPases (Ross et al., 2013 ▸; Zuo et al., 2013 ▸), or indirectly by the depletion of GTP from the available nucleotide pool (Geiger et al., 2014 ▸; Krásný & Gourse, 2004 ▸). Induction of the stringent response leads to growth arrest and is involved in virulence and biofilm formation (Hauryliuk et al., 2015 ▸; Liu et al., 2015 ▸).

The RSH (RelA/SpoT homologues) enzymes synthesize and degrade (p)ppGpp to control its concentration in the cell (Stent & Brenner, 1961 ▸). This protein superfamily is named after the Escherichia coli relA and spoT genes, which were discovered as a set of factors that control the synthesis of so-called stable RNA: rRNA and tRNA (Stent & Brenner, 1961 ▸). The RSH family consists of three subfamilies of enzymes that differ mainly in domain content and catalytic activity (Atkinson et al., 2011 ▸). The ribosome-associated long RSH subfamily includes the Rel enzymes (with bifunctional ppGpp synthetase and hydrolase activities), specialized ppGpp synthetases, such as RelA enzymes, that have lost their hydrolase activity, and the specialized SpoT ppGpp hydrolase enzymes that have retained only poor synthetase activity. While the vast majority of bacteria encode one long bifunctional Rel, a gene-duplication event in β- and γ-proteobacteria resulted in divergence into the two more specialized enzyme homologues RelA and SpoT (Atkinson et al., 2011 ▸). The N-terminal region of long RSHs contains the two catalytic domains [the (p)ppGpp hydrolase (HD) domain and the (p)ppGpp synthetase (SYN) domain] and the C-terminal half is involved in regulation of the enzyme [Fig. 1 ▸(a)]. In addition, the superfamily includes small alarmone synthetase (SAS) homologues that consist only of the ppGpp catalytic domain and small alarmone hydrolase (SAH) homologues containing only the hydrolase catalytic domain. Both SASs and SAHs are stripped of the conserved C-terminal regulatory domains.

Figure 1.

Biochemical and biophysical characterization of Rel_Tt and Rel_Tt ^NTD. (a) Domain composition of long RSH enzymes and structural model based on the structure of E. coli RelA in the ribosome-bound state (PDB entry 5l3p; Arenz et al., 2016 ▸). (b) Analytical SEC of Rel_Tt (dark blue) and Rel_Tt ^NTD (light blue) on a Superdex 200 Increase SEC column; the measurements were performed in 50 mM Tris pH 7.5, 500 mM NaCl, 500 mM KCl, 10 mM MgCl₂, 1 mM TCEP, 0.002% mellitic acid. Estimation of the molecular weights of the proteins based on the SEC profiles suggest that they are both monomers in solution under these conditions. (c) Stability of Rel_Tt ^NTD in the absence (orange) and presence (green) of mellitic acid monitored by thermal shift assay. (d) Stability of Rel_Tt ^NTD in the presence (blue) and absence (red) of Mn²⁺ monitored by thermal shift assay. Removal of the metal ion decreases the T _m of the enzyme by 3°C. (e) Binding of Mn²⁺ to metal-free Rel_Tt ^NTD monitored by isothermal titration calorimetry (K _d = 420 nM). The binding is largely enthalpically driven (ΔH = 3.5 kcal mol⁻¹), suggesting that the Mn²⁺ coordination involves structuring of the HD domain. (f) TLC showing the ppGpp hydrolase activity of Rel_Tt in the presence of ³²P-labelled ppGpp (lane 1 shows the migration of the control ³²ppGpp) either at 40°C (lane 2) or 4°C (lane 3). The migration of ppGpp and pyrophosphate (PP) is shown in red. (g) ppGpp synthetase activity of Rel_Tt in the presence of 100 µM cold ppGpp as activator, 1 mM APPNP or 1 mM ATP and 0.3 mM ³H-labelled GDP.

High-resolution structural information on long RSHs is limited to the structure of the N-terminal region of Streptococcus dysgalactiae subsp. equisimilis Rel (Rel_Seq ^NTD), which contains the two opposing catalytic domains of the enzyme (Hogg et al., 2004 ▸). Rel_Seq is a bona fide bifunctional enzyme, and interestingly the structure of Rel_Seq ^NTD contained two distinct conformations in the same crystal lattice that underscore the antagonistic functions of the enzyme. However, it failed to fully capture the regulatory mechanism, including the changes required for the incorporation of ATP into the active site of the synthetase domain (Hogg et al., 2004 ▸). Nevertheless, the conformational changes observed in the crystal structure suggested that a nucleotide-triggered swinging motion of the synthetase domain could be involved in the switch from one activity to the other. Further insights into the catalytic mechanism of RSHs came from the structures of the SASs RelP and RelQ that revealed the key catalytic residues of the synthetase domain and how ATP and GTP are engaged in the active site in a pre-catalytic state (Manav et al., 2018 ▸; Steinchen et al., 2015 ▸).

More nuanced regulatory mechanisms of RSH enzymes that control the enzymatic output by sensing the physiological state of the cell involve the interactions of cellular partners via the C-terminal regulatory domains (Gully et al., 2003 ▸; Ramagopal & Davis, 1974 ▸; Ronneau et al., 2019 ▸), including ‘starved’ ribosomal complexes containing deacylated tRNA in the A-site (Loveland et al., 2016 ▸; Brown et al., 2016 ▸; Arenz et al., 2016 ▸), the acyl-carrier protein (ACP; Gully et al., 2003 ▸) and the PtsN phosphorelay (Ronneau et al., 2019 ▸). In addition, (p)ppGpp itself controls E. coli RelA via positive allosteric feedback regulation (Shyp et al., 2012 ▸).

Here, we describe the purification and the structure determination by X-ray crystallography of three different crystal forms of the catalytic domain of a bifunctional Rel RSH enzyme from Thermus thermophilus.

2. Materials and methods  

2.1. Cloning  

The chemically synthesized rel gene from T. thermophilus (rel _Tt) was cleaved with the NdeI and XhoI restriction enzymes (NEB) and transferred to pET-21b, which was cut with the same restriction enzymes and dephosphorylated with alkaline phosphatase (NEB). The sequence of the resulting plasmid, pET-21b-HisTEV-Rel_Tt, was verified by sequencing (Eurofins Genomics). To acquire a plasmid for the overexpression of solely the catalytic (N-terminal) domain of the Rel_Tt protein (Rel_Tt ^NTD), the C-terminal part of the gene (from Val356) was removed by amplifying the plasmid with primers (Sigma–Aldrich) flanking the sequence to be deleted (Table 1 ▸) using Q5 High-Fidelity Polymerase (Sigma–Aldrich). The acquired PCR reaction mixtures were treated with DpnI (NEB) to remove the pET-21b-HisTEV-Rel_Tt plasmid used as the template (see Table 2 ▸ for the full sequence of each protein variant). Finally, the PCR products were purified using a PCR purification column (Sigma), phosphorylated with PNK (Sigma) and ligated. The ligation mixture was transferred into E. coli strain MC1061 by electroporation. The resulting plasmid pET-21b-HisTEV-Rel_Tt ^NTD (residues 1–355; Table 2 ▸) was then purified and its sequence was confirmed by standard sequencing methods (Eurofins Genomics).

Table 1. Sequences of the recombinant proteins used in this work.

Name	Sequence
His-TEV-RelTt	MGSSHHHHHHSSGENLYFQGASVGADLGLWNRLEPALAYLAPEERAKVREAYRFAEEAHRGQLRRSGEPYITHPVAVAEILAGLQMDADTVAAGLLHDTLEDCGVAPEELERRFGPTVRRIVEGETKVSKLYKLANLEGEERRAEDLRQMFIAMAEDVRIIIVKLADRLHNLRTLEHMPPEKQKRIAQETLEIYAPLAHRLGMGQLKWELEDLSFRYLHPEAFASLSARIQATQEARERLIQKAIHLLQETLARDELLQSQLQGFEVTGRPKHLYSIWKKMEREGKTLEQIYDLLAVRVILDPKPAPTRESQALREKQVCYHVLGLVHALWQPIPGRVKDYIAVPKPNGYQSLHTTVIALEGLPLEVQIRTREMHRVAEYGIAAHWLYKEGLTDPEEVKRRVSWLKSIQEWQKEFSSSREFVEAVTKDLLGGRVFVFTPKGRIINLPKGATPVDFAYHIHTEVGHHMVGAKVNGRIVPLSYELQNGEIVEILTSKNAHPSKGWLEFAKTRSAKSKIRQYFRAQERQETLEKGQHLLERYLKRRGLPKPTDSQLEEAAKRLSLPPSPEELYLALALNRLTPRQVAEKLYPKALLKPEKPKPQARNEWGIRLEQDLQAPIRLASCCEPMKGDPILGFVTRGRGVTVHRADCPNLRRLLQGPEADRVIGAYWEGVGGKVVVLEVLAQDRAGLLRDVMQVVAEAGKSALGSETRVLGPLARIRLRLAVQDGEEERILQAVQKVSGVKEARWA
His-TEV-RelTt NTD	MGSSHHHHHHSSGENLYFQGASVGADLGLWNRLEPALAYLAPEERAKVREAYRFAEEAHRGQLRRSGEPYITHPVAVAEILAGLQMDADTVAAGLLHDTLEDCGVAPEELERRFGPTVRRIVEGETKVSKLYKLANLEGEERRAEDLRQMFIAMAEDVRIIIVKLADRLHNLRTLEHMPPEKQKRIAQETLEIYAPLAHRLGMGQLKWELEDLSFRYLHPEAFASLSARIQATQEARERLIQKAIHLLQETLARDELLQSQLQGFEVTGRPKHLYSIWKKMEREGKTLEQIYDLLAVRVILDPKPAPTRESQALREKQVCYHVLGLVHALWQPIPGRVKDYIAVPKPNGYQSLHTTVIALEGLPLEVQIRTREMHR

Table 2. Crystallization conditions observed for Rel_Tt ^NTD .

Conditions shown in bold yielded the crystals that were used for data collection.

Enzyme	Condition	Temperature (K)	Cryoprotectant
RelTt NTD	Morpheus II B11	20	—
RelTt NTD	Morpheus II C10	20	—
RelTt NTD	Morpheus II C11	20	—
RelTt NTD	Morpheus II E10	20	—
RelTt NTD	Morpheus II F10	20	—
RelTt NTD	Morpheus II G6	20	—
RelTt NTD	Morpheus II G10	20	—
RelTt NTD	Morpheus II H11	20	—
RelTt NTD	PACT premier E4	20	PEG 400
RelTtNTD	PACT premier G2	20	PEG 400
RelTt NTD	PACT premier G3	20	25% MPD, 20% glycerol, 1 M TMAO
RelTt NTD	PACT premier H4	20	20% glycerol
RelTtNTD + APPNP	PACT premier D12	20	3 M TMAO
RelTt NTD + APPNP	PACT premier G5	20	3 M TMAO, 20% glycerol, PEG 400
RelTt NTD + APPNP	PACT premier G10	20	1.5 M TMAO, 20% glycerol, 50 mM APPNP, 50 mM GDP, PEG 400
RelTt NTD + GDP	Morpheus II G5	20	—
RelTt NTD + GDP	Morpheus II G6	20	—
RelTtNTD	Morpheus II G8	4	—

2.2. Protein expression and preparation for purification  

Rel_Tt or its catalytic domain Rel_Tt ^NTD with an N-terminal His₆ tag and a TEV protease recognition site were expressed from the plasmids pET-21b-HisTEV-Rel_Tt and pET-21b-HisTEV-Rel_Tt ^NTD, respectively. The plasmids were transferred into E. coli BL21(DE3) cells by electroporation. The cells carrying the plasmid were grown to an OD₆₀₀ of 0.6 at 37°C in LB supplemented with ampicillin (100 mg l⁻¹) prior to induction of protein expression by the addition of IPTG (final concentration of 0.5 mM) at 28°C. Expression took place for 5 h for Rel_Tt ^NTD or overnight for Rel_Tt. Next, the cells were collected and resuspended in resuspension buffer (50 mM Tris–HCl pH 8, 1.5 M KCl, 2 mM MgCl₂, 1 mM TCEP) supplemented with cOmplete protease-inhibitor cocktail (Roche). If not used immediately for downstream applications, the cells were flash-frozen in liquid nitrogen and stored at −80°C.

The cell extracts were prepared using a cell cracker equilibrated with lysis buffer (50 mM Tris–HCl pH 8, 500 mM NaCl, 500 mM KCl, 1 mM TCEP) supplemented with cOmplete protease-inhibitor cocktail (Roche). The pellet was separated from the supernatant by centrifugation for 30 min at 25 000g.

2.3. Protein purification and preparation for crystallization  

The supernatant was loaded onto a gravity-flow TALON column previously equilibrated with buffer A (25 mM Tris pH 8, 500 mM NaCl, 500 mM KCl, 10 mM MgCl₂, 0.002% mellitic acid). The column was washed with seven column volumes of buffer A and the bound protein was eluted stepwise with buffers B1–B3 (buffer A containing 50, 125 and 500 mM imidazole, respectively). The elution fractions containing His-Rel_Tt ^NTD were immediately concentrated using spin filters (Amicon) and loaded onto a GE Healthcare Superdex 200 16/60 gel-filtration column equilibrated with GF buffer (50 mM Tris pH 8, 500 mM NaCl, 500 mM KCl, 10 mM MgCl₂, 0.002% mellitic acid). The purity of the protein fractions was assessed by SDS–PAGE. The pure protein was immediately used for further applications or stored in 35% glycerol at −20°C for up to one week.

The His tag was cleaved off from the His-TEV-Rel_Tt ^NTD protein by adding TEV protease in a 1:100 molar ratio and incubating overnight at room temperature. The quality of the protein samples post-cleavage was then analysed by SDS–PAGE. Complete removal of the tag was confirmed by Western blotting using an anti-poly-His antibody (Sigma; catalogue No. H1029). To separate Rel_Tt ^NTD from the His-tagged protease and free poly-His tag, the mixture was passed through a TALON column. The cleaved protein was subsequently concentrated to 10–12 mg ml⁻¹ for crystallization.

2.4. Preparation of metal-free Rel_Tt ^NTD  

Metal-free Rel_Tt ^NTD (10 µM) was prepared as described for other Mn²⁺-binding proteins (Garcia-Pino et al., 2006 ▸) with minor modifications. Briefly, the enzyme was dialysed at 8°C into 50 mM HEPES pH 7.5, 500 mM KCl, 500 mM NaCl, 1 mM TCEP, 0.002% mellitic acid, 100 mM EDTA to remove the Mn²⁺ ion bound to Rel_Tt ^NTD. For the dialysis we used 12–14 kDa cutoff dialysis membranes (SpectrumLabs) and incubated the sample for 1 h, after which the buffer was exchanged and dialysis was continued overnight. After dialysis, the buffer was exchanged to the isothermal titration calorimetry buffer (50 mM HEPES pH 7.5, 500 mM KCl, 500 mM NaCl, 1 mM TCEP, 0.002% mellitic acid) by size-exclusion chromatography on a Superdex 200 column (GE Healthcare).

2.5. Isothermal titration calorimetry  

For the ITC titrations, the Mn²⁺-free Rel_Tt ^NTD was concentrated to around 50 µM using ultrafiltration units (Amicon). The MnCl₂ (Merck) used to titrate Rel_Tt ^NTD was dissolved in a matching buffer to a concentration of 400–1000 µM. The titrations were performed on an affinity ITC instrument (TA Instruments) at 10°C by injecting 2 µl of the MnCl₂ solution for each titration point at 200 s intervals with the stirring rate set to 75 rev min⁻¹. Data analysis and modelling were performed as described previously (Garcia-Pino et al., 2016 ▸; Talavera et al., 2018 ▸) with NanoAnalyze (TA Instruments) and MicroCal Origin 7.0.

2.6. ppGpp synthesis and hydrolysis assays  

H³-ppGpp synthesis assays were performed as described for E. coli RelA (Kudrin et al., 2018 ▸) with minor modifications. The reaction mixtures typically contained 120 nM T. thermophilus 70S IC (MV), 30 nM Rel_Tt, guanosine nucleotide substrate (300 µM H³-GDP, PerkinElmer), ppGpp 100 µM and 2 µM E. coli tRNA^Val, all in HEPES:Polymix buffer at a final Mg²⁺ concentration of 5 mM. After pre-incubation at 40°C for 3 min, the reaction was started by the addition of preheated ATP or APPNP {adenosine 5′-[(β,γ)-imido]triphosphate} to a final concentration of 1 mM and quenched with 4 µl 70% formic acid supplemented with a cold nucleotide standard (4 mM GDP) for UV-shadowing. Individual quenched time points were spotted onto PEI-TLC plates (Macherey-Nagel) and the nucleotides were resolved in 0.5 M KH₂PO₄ pH 3.5 buffer. ³H radioactivity was quantified by scintillation counting in an Optisafe-3 (Fisher Scientific) scintillation cocktail.

For ³²ppGpp hydrolysis assays, ³²P-labelled ppGpp was synthesized by incubating Rel from Chlorobaculum tepidum with 1 mM GDP and 3 nM ATP-[γ-³²P] (PerkinElmer) for 90 min. Hydrolysis assays were performed in 20 µl reactions with 1 µM Rel_Tt and 1.5 nM ³²ppGpp in reaction buffer: 1 mM MgCl₂, 1 mM TCEP, 10 mM Tris pH 7.4, 50 mM NaCl. The reaction mixtures were incubated at 4 or 40°C for 20 min and quenched by the addition of 2 µl formic acid (Sigma–Aldrich). 2 µl of each mixture was spotted onto a PEI-Cellulose F TLC plate (Merck Millipore) and resolved in 1 M KH₂PO₄ at pH 3. The dried TLC plates were exposed to a BAS storage phosphor screen (GE Healthcare) for 1 h. The storage screen was scanned with an Amersham Typhoon Phosphorimager (GE Healthcare).

2.7. Thermal shift assay  

Thermal shift assays were performed on 25 µl mixtures that consisted of 2.5 µl Rel_Tt ^NTD (100 µM) and 2.5 µl of a 100× dilution (in water) of SYPRO Orange 5000× concentrate (ThermoFisher). The final volume was achieved with 16 µl TF buffer (50 mM HEPES pH 7.5, 500 mM KCl, 500 mM NaCl, 1 mM TCEP, 0.002% mellitic acid) and 4 µl 100 mM EDTA when testing the stabilizing effect of Mn²⁺ on Rel_Tt ^NTD and with 20 µl TF buffer with (0.002%) or without mellitic acid to test the stabilizing effect of mellitic acid. The experiment was set up in 96-well real-time PCR plates and the fluorescence of the solutions was measured at 490 nm excitation and 550 nm emission in real-time PCR instruments (Bio-Rad C1000 Thermal cycler for EDTA experiments and Applied Biosystems StepOnePlus for the mellitic acid experiment), increasing the temperature of the wells by 1°C min⁻¹ from 15 to 95°C.

2.8. Crystallization  

Crystallization conditions were screened at 20 and 4°C by the sitting-drop vapour-diffusion method. The drops were set up in Swissci (MRC) 96-well 2-drop UVP sitting-drop plates using a Mosquito robotic system (TTP Labtech). Drops consisting of 0.1 µl protein solution and 0.1 µl precipitant solution were equilibrated against 80 µl precipitant solution in the reservoir. Crystallization conditions were tested with several commercially available screens: Crystal Screen, Crystal Screen 2 (Hampton Research), Helix, ProPlex, PACT premier, JCSG-plus, The LMB Crystallization Screen and Morpheus II (Molecular Dimensions). The concentrations of the protein solutions were determined from the absorbance at 280 nm and were corrected by theoretical extinction coefficients calculated using ProtParam (Gasteiger et al., 2003 ▸).

For the co-crystallization of Rel_Tt ^NTD with nucleotides, 12 mg ml⁻¹ Rel_Tt ^NTD was mixed with nucleotides at 50 mM or 100 mM and incubated for 10 min at room temperature. The commercially available screens Morpheus II, PACT premier and SG1 (Molecular Dimensions) were used to screen for crystallization conditions using a Mosquito HTS robot.

2.9. Data collection and processing  

Prior to data collection, all crystals were vitrified in liquid N₂ after cryoprotection (see Table 2 ▸ for details of the cryoprotective solutions that were used). In the case of co-crystals with nucleotides, the cryoprotective solution included the nucleotides at 20 mM. X-ray diffraction data were collected on the PROXIMA 1 (PX1) and PROXIMA 2A (PX2A) beamlines at the SOLEIL Synchrotron, Gif-sur-Yvette, Paris, France. All data were indexed, integrated with XDS and scaled with XSCALE (Kabsch, 2010 ▸) or AIMLESS (Evans, 2006 ▸). Data quality and twinning were assessed with phenix.xtriage (Afonine et al., 2012 ▸), and POINTLESS (Evans, 2006 ▸) was used for determination of the space group.

2.10. Structure solution  

Data were collected from vitrified crystals of Rel_Tt ^NTD at 100 K. MATTHEWS_COEF was used to calculate Matthews coefficients for cell-content analysis (Winn et al., 2011 ▸). Self-rotation functions were calculated with MOLREP (Murshudov et al., 2011 ▸) and molecular replacement was performed with Phaser (McCoy et al., 2007 ▸) using the coordinates of Rel_Seq ^NTD (PDB entry 1vj7) as an initial search model (Hogg et al., 2004 ▸). Automated model building of the data set collected at 4°C was performed with ARP/wARP (Langer et al., 2008 ▸).

3. Results and discussion  

3.1. Purification of Rel_Tt and Rel_Tt ^NTD  

For the purification of Rel_Tt, we used a gravity-flow column filled with 1 ml TALON resin followed by size-exclusion chromatography (SEC), as shown in Fig. 1 ▸(b). Rel_Tt elutes at approximately 14.0 ml from a Superdex Increase 200 13/30 column (GE Healthcare), corresponding to a size of approximately 95 kDa, which is of the order of the molecular weight expected for a Rel_Tt monomer. For biochemical assays, the enzyme underwent an additional desalting step using a Hi-Prep 26/10 desalting column to remove potential nonspecific RNA contaminants, followed by another SEC step, and only samples with a ratio of absorption at 260 nm/280 nm of between 0.6 and 0.7 were used in further experiments. After purification by SEC, both versions of the protein were concentrated to 10–12 mg ml⁻¹ and used to screen for crystallization conditions.

Long Rel enzymes are notoriously difficult to study because of their poor stability and their tendency to oligomerize. Therefore, we also focused on the crystallization of the N-terminal catalytic domains of Rel_Tt (Rel_Tt ^NTD) to study the allosteric control of the different substrates of the enzyme on its conformational landscape. Rel_Tt ^NTD included residues 1–355 preceded by an N-terminal His tag and a TEV cleavage sequence. Rel_Tt ^NTD was purified using the same strategy as described for Rel_Tt. Analytical SEC performed on a Superdex Increase 200 13/30 column (GE Healthcare) after the initial affinity chromatography step showed that Rel_Tt ^NTD elutes at approximately 16.0 ml, which is consistent with a monomeric species [Fig. 1 ▸(b)]. After removing the His tag and additional SEC, Rel_Tt ^NTD was concentrated for crystallization.

3.2. Rel_Tt is stabilized by mellitic acid  

To further stabilize Rel_Tt, we used the Silver Bullets screen from Hampton Research to screen for compounds that would increase the melting temperature of Rel_Tt as monitored by a thermal shift assay. The screen contains a varied collection of metal ions and organic compounds that could bind and stabilize the enzyme and increase its chances of crystallization. This screening highlighted mellitic acid as a compound that increases the T _m of Rel_Tt ^NTD by between 1.5 and 2°C at concentrations of between 0.002% and 0.02% [Fig. 1 ▸(c)]. Therefore, we decided to include mellitic acid at 0.002% in all of the purification steps of Rel_Tt and Rel_Tt ^NTD prior to crystallization. This allowed us to concentrate the enzyme to above 10 mg ml⁻¹.

3.3. Biochemical and biophysical characterization of Rel_Tt  

The N-terminal region of long Rel enzymes harbours two catalytic domains with opposing activities. In Rel_Tt ^NTD the HD domain (residues 1–158) has a pyrophospho-hydrolase activity and is involved in the cellular degradation of (p)ppGpp, whereas the SYN domain (residues 198–355) binds GTP/GDP and ATP to synthesize (p)ppGpp. It has long been speculated that the nucleotide substrates of the enzyme dictate the conformational arrangement of the domains in the NTD region based on the structures of Rel_Seq ^NTD. However, these structures failed to capture the full regulatory mechanism that involves allosteric triggering of metal ion-catalysed hydrolysis and a strong protein-dynamics component.

Therefore, we decided to characterize the interaction of Rel_Tt ^NTD with Mn²⁺ (the metal ion bound in the HD domain) and its substrates, as well as the effect of temperature on catalysis, to guide the crystallization process. The presence of Mn²⁺ seems to stabilize the enzyme based on a decrease in the T _m of 3°C observed after incubating the enzyme with EDTA to remove the Mn²⁺ ion [Fig. 1 ▸(d)]. Isothermal titration calorimetry (ITC) shows that the enzyme binds Mn²⁺ with an affinity of 420 nM [Fig. 1 ▸(e)]. The observed K _d is lower than expected based on earlier observations for other Rel enzymes (Avarbock et al., 2005 ▸). However, it is in agreement with the cellular concentration of Mn²⁺ in bacterial cells which is in the µM range (Foster et al., 2014 ▸). In addition, temperature also has a strong effect on catalysis, precluding ppGpp hydrolysis at 4°C [Fig. 1 ▸(f)].

The synthesis of ppGpp involves the pyrophosphorolysis of the α–β phosphoric anhydride bond of ATP. Therefore, we reasoned that substituting the O atom of the adjacent β–γ phosphoric anhydride bond could impact the catalytic reaction, facilitating structural studies of the catalytic mechanism. Indeed, using APPNP (an analogue of ATP in which the β–γ oxygen is substituted by an NH group) instead of ATP slows the reaction by around 50-fold [Fig. 1 ▸(g)]. Based on these results, we decided to pursue the crystallization of Rel_Tt ^NTD at low temperatures and in the presence of APPNP (besides the obvious natural substrates and real nonhydrolysable substrates such as APCPP) to increase the chance of obtaining catalytically meaningful conformations.

3.4. Preliminary X-ray analysis of the Rel_Tt ^NTD crystals  

Rel_Tt ^NTD crystallized in various conditions, resulting in crystals with various morphologies and size distributions [Figs. 2 ▸(a)–2 ▸(c)]. These crystals diffracted to around 10 Å resolution on average; however, a series of changes of the sodium salt used during crystallization regularly improved the resolution of the diffraction pattern to around 2.8 Å. The addition of nucleotides to Rel_Tt ^NTD also resulted in new crystallization conditions, with hits obtained using different nucleotide combinations, and nucleotides were also included in the cryoprotective solutions to soak the crystals of Rel_Tt ^NTD prior to vitrification in liquid N₂ (Table 2 ▸). Temperature was also a crucial factor in improving the quality of the diffraction pattern. The best diffracting crystals in terms of resolution and lower anisotropic features were obtained with crystals grown at 4°C.

Figure 2.

Typical crystals of Rel_Tt ^NTD grown under different conditions. Crystals of Rel_Tt ^NTD in the presence of APPNP at 20°C (a) and of Rel_Tt ^NTD at 20°C (b) and at 4°C (c).

These crystals of Rel_Tt ^NTD obtained at 4°C were around 70 µm along the largest direction on average and grew after two weeks of incubation. These crystals diffracted to 2.7 Å resolution (Table 3 ▸) and belonged to space group P4₁22 (or P4₃22). Based on the unit-cell content analysis, there could be one or two molecules in the asymmetric unit (with a solvent content of 70% or 40% based on Matthews coefficients of 4.1 and 2.0 ų Da⁻¹, respectively). However, analysis of the self-rotation function is consistent with only one molecule in the asymmetric unit, with only trivial peaks observed in the κ = 180° section of the self-rotation function corresponding to the crystallographic twofold [Fig. 3 ▸(a)]. An initial search using each of the two conformations of Rel_Seq ^NTD (PDB entry 1vj7) failed to deliver a conclusive molecular-replacement solution. Rel_Tt ^NTD and Rel_Seq ^NTD have an overall sequence identity of 40%, with the HD domain having a local sequence identity of 47% and the SYN domain having an identity of around 40%. Considering that the enzyme is also notoriously dynamic, we decided to search using ensembles generated from each individual Rel_Seq ^NTD domain. Using this approach, we obtained a molecular-replacement solution with Phaser in space group P4₁22 consistent with one molecule in the asymmetric unit, with a final TFZ of 30.1, PAK = 1 and LLG = 650. This solution is significantly better than that in the alternate space group P4₃22, with a TFZ of 12.3, PAK = 4 and LLG = 145, which confirmed P4₁22 as the correct space group. We then used ARP/wARP for automated model building starting from this molecular-replacement solution and were able to reconstruct 80% of the backbone [Fig. 3 ▸(b)].

Table 3. Data collection and processing.

The CC_1/2 criterion was used to determine the resolution range. Values in parentheses are for the outer shell.

	RelTt NTD (4°C)	RelTt NTD (20°C)	RelTt NTD–GDP–APPNP (20°C)
Diffraction source	PX1, SOLEIL	PX1, SOLEIL	PX2A, SOLEIL
Wavelength (Å)	1.008	0.9786	0.9801
Temperature (K)	100	100	100
Detector	PILATUS 6M	PILATUS 6M	EIGER
Crystal-to-detector distance (mm)	582.01	623.60	342.51
Rotation range per image (°)	0.1	0.1	0.1
Total rotation range (°)	190	200	360
Exposure time per image (s)	0.1	0.1	0.1
Space group	P4122	P41212	C2
a, b, c (Å)	88.4, 88.4, 182.7	105.7, 105.7, 241.4	120.8, 50.3, 86.1
α, β, γ (°)	90.0, 90.0, 90.0	90.0, 90.0, 90.0	90.0, 111.0, 90.0
Mosaicity (°)	0.07	0.07	0.21
Resolution range (Å)	51.60–2.75 (2.85–2.75)	240.00–2.90 (2.94–2.90)	56.36–2.95 (3.06–2.95)
Total No. of reflections	156684 (2496)	297021 (487)	67782 (2399)
No. of unique reflections	12385 (223)	21377 (50)	6572 (255)
Completeness after anisotropic correction (%)	94.7 (97.9)	93.7 (90.9)	91.9 (94.8)
Multiplicity	12.7 (11.2)	13.9 (9.7)	10.3 (9.4)
〈I/σ(I)〉	10.2 (1.2)	14.4 (1.2)	13.8 (2.2)
CC1/2	1.00 (0.61)	1.00 (0.45)	1.00 (0.87)
R r.i.m.	0.21 (1.69)	0.12 (1.70)	0.12 (1.1)
R p.i.m.	0.06 (0.52)	0.05 (0.75)	0.04 (0.36)
Overall B factor from Wilson plot (Å2)	47.95	105.42	79.99

Figure 3.

Preliminary X-ray analysis of the Rel_Tt ^NTD crystals. (a) Self-rotation function at the κ = 180° section of the data collected from the Rel_Tt ^NTD crystals grown at 4°C, consistent with only one molecule in the asymmetric unit. (b) Initial 2mF _o − DF _c map (σ = 1.5) for the Rel_Tt ^NTD crystals grown at 4°C after automated model building with ARP/wARP. The autotraced C^α backbone is shown in red. Self-rotation function at the κ = 180° section (c) and κ = 120° section (d) of the data collected from the Rel_Tt ^NTD crystals grown at 20 °C.

In addition, we grew crystals of Rel_Tt ^NTD at 20°C that were around 100 µm in size on average and appeared within a week of incubation. These crystals diffracted to 2.8 Å resolution on average (Table 3 ▸) and belonged to space group P4₁2₁2 (or P4₃2₁2). However, these crystals were very anisotropic, with the magnitude of the anisotropy expressed as a function of the resolution along the h, k and l axes estimated as 2.5, 3.1 and 4.5 Å, respectively. Analysis of the unit-cell content of these crystals predicted that they contain three molecules of Rel_Tt ^NTD (52% solvent content and a Matthews coefficient of 2.65 ų Da⁻¹). The self-rotation function of these Rel_Tt ^NTD crystals had strong features at χ values of 180° and 120° consistent with three molecules in the asymmetric unit, in agreement with the cell-content analysis [Figs. 3 ▸(c) and 3 ▸(d)]. Indeed, using the individual HD and SYN domains of Rel_Seq ^NTD (PDB entry 1vj7) as a search model, Phaser found the equivalent of three molecules in the asymmetric unit in space group P4₁2₁2, with a final TFZ of 17.2, PAK = 2 and LLG = 1151. This solution was of considerably higher quality than that of the alternate space group P4₃2₁2 (TFZ of 10.9, PAK = 6 and LLG = 28) to conclusively assign P4₁2₁2 as the correct space group.

We also took advantage of the aforementioned slower reaction of Rel_Tt ^NTD involving APPNP for crystallization. This resulted in crystals of a C2 form of Rel_Tt ^NTD that were around 200 µm in size on average and typically appeared within a week after setting up the drops. These crystals were typically very anisotropic, diffracting on average to a resolution of around 2.9 Å in the best direction (Table 3 ▸), with the anisotropy expressed as a function of the resolution along the h, k and l axes estimated as 3.4, 3.4 and 2.9 Å, respectively. For data collection the cryoprotective solution was supplemented with 50 mM APPNP and 50 mM GDP and was incubated for 5–10 min before vitrification. Analysis of the unit-cell content suggested that there was only one molecule in the asymmetric unit based on a Matthews coefficient of 2.77 ų Da⁻¹, which corresponds to a solvent content of 55.7%. Molecular replace­ment with Phaser, using ensembles of HD and SYN domains generated from the structure of Rel_Seq ^NTD as search models, found a solution in space group C2 that was consistent with one molecule in the asymmetric unit, with a final TFZ of 26.7, PAK = 0 and LLG = 1185. Structure refinement and completion is ongoing in all cases.

4. Conclusions  

Long Rel stringent factors are the most broadly distributed class of RSHs. This family of enzymes regulate the synthesis and hydrolysis of the bacterial alarmone (p)ppGpp, which is the key factor in the stringent response. The structural basis of the regulation of Rel is not completely understood, in particular the role of nucleotides in the allosteric triggering of the opposing catalytic activities. Because of its intrinsic stability, we used T. thermophilus Rel as a model system to study this nucleotide-based allosteric control and obtained crystals of the catalytic region of the enzyme in different conditions and in the presence and absence of nucleotides. The corresponding crystal structures are likely to enhance our understanding of the function and regulation of the catalytic activity of Rel enzymes by their nucleotide substrates.

Funding Statement

This work was funded by Fonds De La Recherche Scientifique - FNRS grants FNRS-EQP U.N043.17F,, FRFS-WELBIO CR-2017S-03, FNRS-PDR T.0066.18, and CR/DM-392. Université Libre de Bruxelles grant . Fonds Jean Brachet grant . Fonds Alice et David van Buuren grant . Umeå Centre for Microbial Research grant . Molecular Infection Medicine Sweden grant . Vetenskapsrådet grant .