The X‐ray crystal structure of PA3566 from <i>Pseudomonas aureginosa</i> at 1.8 Å resolution

The structure of the PA3566 protein from Pseudomonas aureginosa was determined at 1.78 Å resolution using the Se-Met multiwavelength anomalous diffraction (MAD) technique. The PA3566 gene encodes a small, conserved protein of unknown function, with a predicted molecular weight of 11.2 kDa (104 amino acids). The protein belongs to a family of conserved proteins of unknown function [COG 1359, Fig. 1(a), top six sequences], but has low sequence identity to any known structures. There is an annotational link to antibiotic biosynthesis monooxygenases (pfam 03992).1 Crystal structure of PA3566. (A) Sequence alignment of PA3566 with its closest homologs (based on a BLAST search) and structural homologs (based on a DALI11 search). Diagram showing the secondary structure elements in PA3566 superimposed on the sequence alignment. The top six sequences are the closest homologs based on sequence alignment, while the bottom two sequences are the two structural homologs. Black asterisks indicate the positions of conserved residues in hydrophobic pocket, blue triangles represent conserved residues in dimer interface and red circles are residues in active site of ActVA-Orf6. Aligned sequences from: PA3566 (Pseudomonas aureginosa), Pseudomonas fluorescens PfO-1, Exiguobacterium sp. 255-15, Burkholderia fungorum LB400, Geobacter metallireducens GS-15, Yersinia pestis. The two structural homologs are: protein yjcs from Bacillus subtilis (PDB code: 1Q8B) and ActaVA-Orf6 from Streptomyces coelicolor (PDB code: 1LQ9). The numbers after the sequences are the sequence identity compared to PA3566. (B) Stereoribbon diagram of Pseudomonas aureginosa PA3566 monomer, color coded from N-terminus (blue) to C-terminus (red). α-Helices (α1–α3), β-strands (β1–β4), and 310-helices (η1) are indicated and correspond to elements in (A). Alignment figures drawn using ESPript21 and other figures made using PYMOL22. (C) Stereoribbon diagram of a superposition of PA2566 dimer with ActVA-Orf6 from Streptomyces coelicolor (PDB code: 1LQ9; white). PA3566 is shown with the same color code as in (B). (D) Stereodiagram of the potential binding site of PA3566, based on structural alignment with ActVA-Orf6. The conserved residues and two water molecules are shown in ball and stick. The atoms are colored as follows: carbon (gray), nitrogen (blue), oxygen (red), selenium (purple). The PA3566 protein crystallizes as a trimer, although the functional unit is most likely a dimer, as are other members of this structural superfamily.2-4 A single monomer assumes a two-layer α + β sandwich fold (β1α1β2β3α2α3β4) shared by a large number of proteins (CATH number 3.30.70). Helices α2 and α3 form one long helix, with a 74° kink in the middle. The two βαβ motifs form a ferredoxin-like split βαβ-fold.5 The four β-strands form an antiparallel β-sheet with the β1 strand lying between the β3 and β4 strands [Fig. 1(b)] and the three α-helices cover one side of the molecule. The monomer is elongated with dimensions of 43 × 37 × 25 Å [Fig. 1(b)]. The 310 helix from residues 81–84 is not present in all monomers. The loop from 75–83 in monomer C is highly disordered, resulting in the disordering of part of a3 and the short 310 helix found in monomers A and B. This region contains none of the conserved residues [Fig. 1(a)], but is involved in the contacts between monomers A and B. This suggests that this loop is likely flexible in solution, but stabilized in one position through the crystal contacts. It is possible that the 310 helix is a crystallographic artefact caused by the contacts between the two monomers. There are four turns located in the protein, as defined by DSSP.6 The turns incorporating residues 14–17 and 35–38 are classical type II turns, while the turn incorporating residues 49–52 is a type I turn and the one at residues 78–81 is a type I′ turn, based on the turns defined by Hutchinson and Thornton.7 The monomers B and C form a likely biologically relevant dimer in the asymmetric unit [Fig. 1(c)], while monomer A forms a dimer with a related A subunit through a crystallographic twofold axis. Gel filtration results confirmed that the PA3566 protein is a dimer in solution (see Methods). The dimer buries 1146.2 Å2 surface area of each monomer and includes 17 hydrogen bonds and two salt bridges.8 The dimer is predominantly stabilized by contacts along the edge of the β-sheets (contacts with all strands, but longest stretches with β2 and β4), plus contacts between the C-terminus and helix α1. The highly conserved residues Arg33 and His44 are buried in the dimer interface. The conserved His44 is noteworthy because previous work shows that this residue is critical for proper folding and thus activity of structurally similar proteins.4, 9 Arg33 forms a water-mediated hydrogen bond network across the dimer interface, composed of Arg33, two waters, and the carbonyl of Leu96 from the other monomer. Arg33 is also involved in crystal contacts through bridging water molecules. Phe53, an absolutely conserved residue, forms part of a small hydrophobic pocket with the highly conserved residues Ile11, Leu21, and Leu43. This pocket docks the conserved hydrophobic residue Leu96 (also found as Ile or Val) [Fig. 1(a)] of the other monomer, burying 137 Å3 of the residue and accounting for 12% of the interface. Salt bridges are formed between Arg47 and Asp62′ and Asp42 and Arg95′, although neither of these salt bridges appears to be highly conserved across this family of proteins, as all four residues are highly variable [Fig. 1(a)]. A search of the Protein Data Bank10 using the Dali server11 yields eight structures with a Z-score over 7, and two of these with a Z-score over 8. The best matches (with Z-scores of 11.5 and 10.3, respectively) are protein yjcs (PDB code: 1Q8B), of unknown function from Bacillus subtilis and ActVA-Orf6 (PDB code: 1LQ9), a monooxygenase involved in actinorhodin biosynthesis4 from Streptomyces coelicolor [Fig. 1(c)]. The sequence identity between PA3566 and each of these proteins is less than 15% [Fig. 1(a), bottom 2 sequences]. The DALI search shows that all of the similar structures have less than 15% identity with PA3566, suggesting that this conserved fold is evolutionarily convergent. The dimer assembly of ActVA-Orf has similarities and differences with PA3566. Both dimers are stabilized by many hydrophobic contacts between β-strands, and the dimer interfaces both contain the absolutely conserved histidine residue (His44) described above. The major difference is the C-terminal β-strand of ActVA-Orf6 that is swapped into the other monomer [Fig. 1(c)]. This strand contributes a significant portion of the buried surface area for ActVA-Orf6 (approx. 66% of the buried surface area), and is apparently unique to that class of proteins.4 In PA3566, this swapped β-strand is still present, but is much shorter and contributes significantly less to the buried surface area (42%).8 Although only one residue (Trp66 in ActVA, Trp59 in PA3566) in the active site of ActVA-Orf6 is conserved with PA3566, structural alignment of ActVA-Orf6 in complex with substrates and PA3566 reveals the presence of a conserved hydrophobic pocket in PA3566 at approximately the same location as the substrate binding site in ActVA-Orf6. In PA3566, the volume of this pocket is 424.6 Å3 (using VOIDOO12) and is made up of conserved residues Ile8, Ala9, Ile11, Leu25, Tyr41, Glu57, Trp59, Leu65, His68, Phe74, plus the semiconserved (hydrophobic nature retained) residues Leu7 and Met55 [Fig. 1(a) and (d)]. In the structure of PA3566, this pocket contains two water molecules that are hydrogen bonded to each other and Tyr41/His68 and Glu57, respectively. The second water molecule is also bonded to Glu82 from the third crystallographic monomer (i.e., no biological relevance). The gates described for ActVA-Orf64 are not conserved in this protein and in the case of the O2/H2O gate, the location is actually blocked by bulky hydrophobic residues. Without any further information regarding the substrate or catalytic reaction, it is impossible to comment further about the role of conserved residues. It does appear that the reaction will involve a predominantly hydrophobic substrate. It seems likely that the reaction catalyzed will be significantly different than the reaction catalyzed by the structurally related protein ActVA-Orf6. Further biochemical characterization is required to define the role of this protein. The PA3566 gene was subcloned, expressed, and its product purified and screened for crystallization as described previously.13 Crystals for X-ray diffractions data collection were obtained from hanging drop vapour diffusion conditions containing 2 μL of Se-Met derivative of PA3566 plus 2 μL of well buffer, containing 2 M ammonium sulphate, 0.1 M sodium citrate (pH 5.8). The crystallization was carried out at 21°C. The crystals were flash frozen with 25% glycerol in crystallization buffer. Size-exclusion chromatography, using a Superdex-75 column (10 × 300 mm) preequilibrated with 10 mM HEPES pH 7.5, 0.5 M NaCl was performed to determine the solution state of PA3566. The column was calibrated with Blue Dextran (2000 Da), cytochrome c (12.4 kDa), carbonic anhydrase (29 kDa), bovine serum albumin (66 kDa), alcohol dehydrogenase (150 kDa), and β-amylase (200 kDa). A 25-μL sample of PA3566 protein at a concentration of 2 mg/mL alone or premixed with the standard proteins was centrifuged at 14,000 rpm for 10 min before being injected into the column through a 20-μL injection loop. Elution was carried out at a flow rate of 1 mL/min at 20°C. The eluted proteins were detected by measuring the absorbance at 280 nm. PA3566 eluted with an apparent molecular weight of 20.3 kDa. Diffraction data (Table I) were collected at beamline 19ID of APS, Argonne National Laboratory following the approach described earlier.14 The three-wavelength inversed beam MAD adapt up to 1.7 Å were collected form one Se-Met labeled protein crystal at 100 K with 3-s exposure/0.5°/frame using a 180 mm crystal-to-detector distance. The total oscillation range was 180°. All data were processed and scaled with HKL2000.15 The space group was determined to be P21212 with unit cell dimensions of a = 104.56, b =38.40, c =63.72, α = β = δ = 90. Calculation of the Matthews volume16 suggested that the unit cell contained 2 monomers (Vm = 2.9 Å3Da−1, solvent content = 56.8%), assuming a molecular weight of 11.15 kDa. The structure was determined by MAD phasing; SHELXE17 was used for the initial phase determination. Four Se sites were located, consistent with two monomers in the unit cell. MLPHARE was used to improve the phases and these were then input into RESOLVE for initial model building. The initial model consisted of a 17 residue stretch that could be placed with side chains and 142 residues without side chains in 20 fragments. Further model building was carried out using ARP 6.1,18 with a total of 236 residues being built automatically. This showed that the unit cell actually contained a third monomer (Vm = 1.9 Å3Da−1 and solvent content = 35.2% for three monomers in a.s.u.). The remaining residues were built by superimposing the three monomers to fill in the missing stretches and manual rebuilding using “O.”19 All programs not individually referenced are part of the CCP4 package.20 The remote wavelength data was chosen for structure refinement through a trial- and-error approach to determine the best statistics at the initial stages of refinement. Refinement was carried out using REFMAC5, using TLS and anisotropic B-factor refinement. The final R-factor was 0.161 and the free R was 0.223 (Table II). The N-termini were disordered in the three monomers, the first residue of Chain A (MSE 1) had an average B-factor of 74.7 compared to the average B-factor of 45.2 while for Chain B and Chain C it was possible to model in the first residue and also one (for B) or two (for C) amino acids (GH) belonging to the TEV protease cleavage site. Atomic coordinates have been deposited in the Protein Data Bank (PDB) with the PDB accession code: 1X7V. We wish to thank all members of the Structural Biology Center at Argonne National Laboratories for their help in data collection and structure solution. For Dr. C.J. Sanders (1935–2005): “The stories go on.”

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.