|Home | About | Journals | Submit | Contact Us | Français|
Processing of the polyprotein of Turnip yellow mosaic virus is mediated by the protease PRO. PRO cleaves at two places, one of which is at the C-terminus of the PRO domain of another polyprotein molecule. In addition to this processing activity, PRO possesses an ubiquitin hydrolase (DUB) activity. The crystal structure of PRO has previously been reported in its polyprotein-processing mode with the C-terminus of one PRO inserted into the catalytic site of the next PRO, generating PRO polymers in the crystal packing of the trigonal space group. Here, two mutants designed to disrupt specific PRO–PRO interactions were generated, produced and purified. Crystalline plates were obtained by seeding and cross-seeding from initial ‘sea urchin’-like microcrystals of one mutant. The plates diffracted to beyond 2 Å resolution at a synchrotron source and complete data sets were collected for the two mutants. Data processing and analysis indicated that both mutant crystals belonged to the same monoclinic space group, with two molecules of PRO in the asymmetric unit.
Turnip yellow mosaic virus (TYMV) encodes a replication polyprotein of 206 kDa (206K) harbouring an RNA polymerase (POL) and an RNA helicase (HEL). TYMV replication depends on POL and HEL being successively cleaved from the C-terminus of 206K in an orchestrated sequence of events (Jakubiec et al., 2007 ). Proteolysis is mediated by a 17 kDa protease domain PRO encoded within 206K just upstream of HEL (Fig. 1 a). Thus, cleavage at the PRO–HEL junction entails specific recognition of one PRO by another PRO and the engagement of the C-terminus of the first PRO by the catalytic site of the other (trans cleavage; Fig. 1 a). Despite its small size, PRO is a multifunctional protein that also displays ubiquitin hydrolase (DUB) activity in vivo and in vitro (Chenon et al., 2012 ), a feature of PRO that allows regulation of the amount of POL in plant cells infected by TYMV (Camborde et al., 2010 ). Thus, PRO can specifically recognize several distinct substrates: both PRO–HEL and HEL–POL endopeptide cleavage sites, as well as POL–ubiquitin isopeptide bonds. We previously shed light on the self-recognition of PRO thanks to the atomic structure of PRO in a crystal form in which the C-terminus of one molecule inserts into the catalytic cleft of the next molecule (Lombardi et al., 2013 ). However, we could not obtain diffraction-quality crystals of disengaged PRO (Robin et al., 2012 ), i.e. PRO not engaged in this self-interaction, which would have allowed assessment of the conformational adjustments upon PRO self-recognition. Here, we report the production and crystallization of mutants of PRO designed to impair self-recognition (Fig. 1 b). The first mutant (PRO I847A) alters a hydrophobic patch prominently used by PRO to recognize its substrates (Lombardi et al., 2013 ). The second mutant is a five-residue C-terminal deletion (PRO ΔC5) that prevents engagement by the active site. PRO I847A should be severely impaired mostly in DUB activity, while PRO ΔC5 will not allow cleavage at the PRO–HEL junction. These mutants should thus be useful in disentangling the functional importance of the two distinct activities of PRO.
PRO I847A was generated by QuikChange site-directed mutagenesis (Agilent Technologies, Santa Clara, California, USA) according to the manufacturer’s instructions. The coding sequence of the PRO domain (residues 728–879 of 206K; Fig. 1 b) previously inserted into a modified pGEX plasmid was used as a template to generate the mutant. In this modified vector the protein sequences harbour a six-histidine tag at their N-terminus (Table 1 ). PRO ΔC5 lacking the last five residues was generated similarly by ShineGene Biotech (Shanghai, People’s Republic of China). Expression conditions in Escherichia coli and purification were as previously reported (Robin et al., 2012 ; Lombardi et al., 2013 ). The purified proteins were then concentrated to 12 mg ml−1 for PRO I847A and 17 mg ml−1 for PRO ΔC5, frozen and stored at 193 K. The concentrations were estimated from the absorbance at 280 nm assuming extinction coefficients of 9919 and 9975 M −1 cm−1 calculated from the sequences of PRO I847A and PRO ΔC5, respectively.
Screening for crystallization conditions was performed by robotics using commercially available kits (The Classics and PEGs Suites from Qiagen; 192 conditions in all). Experiments were set up with a Cartesian robot using the sitting-drop vapour-diffusion method. Equal volumes (100 nl) of protein solution and well solution were mixed. Fine screening around the crystallization conditions of PRO I847A was performed in larger drop volumes (1 µl protein solution and 1 µl crystallization reagent equilibrated against a 0.5 ml reservoir volume) using the hanging-drop vapour-diffusion setup. We obtained ‘sea urchin’-like microcrystals in a few days. In order to slow down crystal growth from the initial nuclei, we used the streak-seeding technique as described previously (Stura & Wilson, 1991 ). We also cross-seeded using the same technique from the PRO I847A microcrystals into hanging drops pre-equilibrated with PRO ΔC5 (Table 2 ).
Crystals of PRO I847A and PRO ΔC5 were transferred into cryoprotected solutions following the protocols and using the cryosolutions of Vera & Stura (2014 ). After a 1 min soak, crystals were flash-cooled by plunging into liquid nitrogen. The plate-shaped crystals diffracted to beyond 2 Å resolution on the PROXIMA2 microfocus beamline at the SOLEIL synchrotron. The programs XDS and XSCALE (Kabsch, 2010 ) were used for data processing.
PRO mutants were produced and purified by nickel-affinity and size-exclusion chromatography essentially as for the wild-type PRO (Robin et al., 2012 ). The similarity extended to a small but persistent contamination by the bacterial ribosomal protein S15 that co-purified with the mutants and, as for the wild type, was only partially removed by the gel-filtration polishing step (Fig. 1 c).
Initial screening of crystallization conditions yielded polycrystalline ‘sea urchins’ within a few days for PRO I847A with a well solution consisting of 0.1 M sodium acetate pH 4.7, 0.2 M ammonium acetate, 30%(w/v) PEG 4000. We easily reproduced the sea urchin-like microcrystals manually (Fig. 2 a), but could not obtain single crystals in a first round of optimization. Seeding is often a good way to circumvent this difficulty by bypassing the nucleation step and thus preventing multiple or secondary nucleation. Indeed, streak-seeding proved to be successful at the first attempt and yielded small but single-looking plate-like crystals that became visible overnight and grew to maximum size within a week (Fig. 2 b, Table 2 ). We therefore tried cross-seeding from PRO I847A crystals, and in this way also obtained plate-shaped crystals of PRO ΔC5 using essentially the same crystallization conditions (Table 2 ) and in the same time frame (Fig. 2 c).
The plate-shaped crystals were cryoprotected using cryosolutions consisting of the crystal-growth solution supplemented with mixtures of cryoprotectants (Vera & Stura, 2014 ). Specifically, cryomixes C5 (final concentrations 5% diethylene glycol, 10% ethylene glycol, 5% MPD, 5% 1,2-propanediol, 5% glycerol, 5 mM NDSB-201) and C6 (5% ethylene glycol, 10% MPD, 5% 1,2-propanediol, 5% DMSO, 5% glycerol) proved especially effective at cryoprotecting the PRO I847A and PRO ΔC5 crystals. The plate-shaped crystals were tested for diffraction, and diffraction data were collected on the PROXIMA2 microfocus beamline of the SOLEIL synchrotron. Complete data sets were collected to 1.9 Å resolution and processed from a single plate-shaped crystal each of PRO I847A (Fig. 3 ) and PRO ΔC5 (Table 3 ).
The crystal plates obtained for both mutants belong to the same monoclinic space group. The space group and unit cell indicate the presence of two molecules per asymmetric unit with a solvent content of 45% and a Matthews coefficient of 2.3 Å3 Da−1. Indeed, preliminary molecular replacement using Phaser (McCoy et al., 2007 ) with the wild-type PRO structure (chain A of PDB entry 4a5u) trimmed of its last five residues unambiguously confirmed this. S15 is not present in the crystals as it was for the wild type (Robin et al., 2012 ). Crystal packing shows that neither PRO molecule in the asymmetric unit has its C-terminus oriented towards the catalytic cleft of another molecule (not shown).
In our previous work, a crystal form with a single molecule of PRO in the asymmetric unit yielded insights into self-recognition by PRO (Lombardi et al., 2013 ). These new crystal structures will provide views of PRO in environments that are not constrained by self-recognition. In turn, these may provide further hints as to how PRO switches between its protease and ubiquitin hydrolase substrates.
We thank Sonia Fieulaine and Isabelle Jupin for critical reading of the manuscript. This work was funded by the French Agence Nationale de la Recherche (ANR) under grant ANR-11-BSV8-011 ‘Ubi-or-not-ubi’. We acknowledge the core I2BC protein crystallization facility (IBBMC, Paris Sud) and its support by the French Infrastructure for Integrated Structural Biology (FRISBI) ANR-10-INSB-05-01. We thank Synchrotron SOLEIL for the allocation of beamtime and Martin Savko of the PROXIMA2 beamline for assistance in data collection.