journal menu
research papers
 | STRUCTURAL BIOLOGY |
accessThe impact of molecular variants, crystallization conditions and the on ligand–protein complexes: a case study on bacterial phosphotriesterase
aDepartment of Life Sciences Core Facilities, Weizmann Institute of Science, Rehovot, Israel, bDepartment of Biomolecular Sciences, Weizmann Institute of Science, Rehovot, Israel, cDepartment of Brain Sciences, Weizmann Institute of Science, Rehovot, Israel, and dDepartment of Chemical and Structural Biology, Weizmann Institute of Science, Rehovot, Israel
*Correspondence e-mail: orly.dym@weizmann.ac.il, joel.sussman@weizmann.ac.il
A bacterial phosphotriesterase was employed as an experimental paradigm to examine the effects of multiple factors, such as the molecular constructs, the ligands used during protein expression and purification, the crystallization conditions and the on the visualization of molecular complexes of ligands with a target enzyme. In this case, the ligands used were organophosphates that are fragments of the nerve agents and insecticides on which the enzyme acts as a bioscavenger. 12 crystal structures of various phosphotriesterase constructs obtained by directed evolution were analyzed, with resolutions of up to 1.38 Å. Both apo forms and holo forms, complexed with the organophosphate ligands, were studied. Crystals obtained from three different crystallization conditions, crystallized in four space groups, with and without N-terminal tags, were utilized to investigate the impact of these factors on visualizing the organophosphate complexes of the enzyme. The study revealed that the tags used for protein expression can lodge in the active site and hinder ligand binding. Furthermore, the in which the protein crystallizes can significantly impact the visualization of bound ligands. It was also observed that the crystallization precipitants can compete with, and even preclude, ligand binding, leading to false positives or to the incorrect identification of lead drug candidates. One of the co-crystallization conditions enabled the definition of the spaces that accommodate the substituents attached to the P atom of several products of organophosphate substrates after detachment of the The crystal structures of the complexes of phosphotriesterase with the organophosphate products reveal similar short interaction distances of the two partially charged O atoms of the P—O bonds with the exposed β-Zn2+ ion and the buried α-Zn2+ ion. This suggests that both Zn2+ ions have a role in stabilizing the transition state for substrate hydrolysis. Overall, this study provides valuable insights into the challenges and considerations involved in studying the crystal structures of ligand–protein complexes, highlighting the importance of careful experimental design and rigorous data analysis in ensuring the accuracy and reliability of the resulting phosphotriesterase–organophosphate structures.
Keywords: crystallization conditions; zinc ions; phosphotriesterases; organophosphates; space groups; bioscavengers; expression systems.
PDB references: A53_1, 8p7f; A53_2, 8p7h; A53_3, 8p7i; A53_4, 8p7k; A53_5, 8p7m; C23_1, 8p7n; C23_2, 8p7q; C23_3, 8p7r; C23_4, 8p7s; C23_5, 8p7t; C23M_1, 8p7u; C23M_2, 8p7v
1. Introduction
Structure-based drug design is a powerful tool in drug discovery, allowing researchers to predict how a potential drug will interact with its target protein. X-ray crystallography is widely used to obtain detailed structural information about the complex of a lead compound with the target protein, but also has certain limitations that may affect the accuracy and reliability of the structure determined. About 75% of the ∼175 000 macromolecular structures determined using X-ray crystallography in the Protein Data Bank (PDB; Berman et al., 2000![[Berman, H. M., Westbrook, J., Feng, Z., Gilliland, G., Bhat, T. N., Weissig, H., Shindyalov, I. N. & Bourne, P. E. (2000). Nucleic Acids Res. 28, 235-242.]](../../../../../../logos/arrows/d_arr.gif "Berman, H. M., Westbrook, J., Feng, Z., Gilliland, G., Bhat, T. N., Weissig, H., Shindyalov, I. N. & Bourne, P. E. (2000). Nucleic Acids Res. 28, 235-242."){kind=small}
; Burley et al., 2022; Sussman et al., 1998) contain at least one of nearly 5000 unique ligands. Some ligands may be present serendipitously, due to the purification or crystallization processes (Dym et al., 2016; Newman, 2004), while others are deliberately added to the sample to study protein function or in the context of structure-based drug design.
The experimental information that is utilized is the electron density of the protein and the ligand. The extent to which we can correctly interpret the electron density, and thus determine the positions of the atoms of both the protein and the ligand, determines the accuracy and reliability of the structure. Although stringent and strict validation tools are available for assessing the accuracy of a protein–ligand structure, in some of the structures deposited in the PDB overenthusiastic interpretation of the electron density of the ligand is encountered. This can lead to the placement of a ligand without supporting electron density. Another issue that may arise is incorrect assignment of the identity of the ligand (Dauter et al., 2014; Weichenberger et al., 2015). This can happen when the electron-density map is not clear enough to permit unambiguous interpretation. Finally, there may be cases in which the electron density cannot be fully accounted for, again leading to an uncertain or incorrect assignment. Thus, although crystallographic studies of protein–ligand complexes provide a valuable experimental tool in drug discovery, they require careful interpretation and validation of the experimental data. Here, we discuss some of the difficulties encountered in studying organophosphate–phosphotriesterase complexes and how these difficulties were overcome.
This study is based on our long-term interest in developing new therapies for treating organophosphate (OP) poisoning by insecticides and nerve agents. The enzyme phosphotriesterase (PTE) is capable of hydrolyzing such OP compounds. Accordingly, it has been characterized from a number of bacterial sources, including Brevundimonas diminuta (Bd; formerly Pseudomonas diminuta), Flavobacterium spp. and Agrobacterium radiobacter (Bigley et al., 2013; Cherny et al., 2013; Goldsmith et al., 2016; Harper et al., 1988; Horne et al., 2002; Masson & Rochu, 2009). The present study focuses on Bd_PTE (Holm & Sander, 1997), a dimeric metallohydrolase that has a (β/α)8 TIM-barrel fold with two Zn2+ ions embedded in the active site (Fig. 1). Carbon dioxide reacts with the side chain of Lys169 to form a carbamate within the active site, which serves as a bridging ligand to the α-Zn2+ and β-Zn2+ ions, both of which are required for full (Benning et al., 1995). The buried α-Zn2+ ion is ligated to the protein via direct interactions with His55, His57 and Asp301, and the exposed β-Zn2+ ion interacts with His201 and His230 (Fig. 1b).
![[Figure 1]](nj5323fig1thm.gif) | Figure 1 View of a typical PTE structure (PDB entry 1hzy; Benning et al., 2001 ). (a) The (β/α)8 TIM-barrel fold is shown as a cartoon, with helices in red, sheets in yellow, coils in green, the α-Zn2+ (buried) and β-Zn2+ (exposed) ions as magenta spheres and a single bridging water shown as a cyan sphere. The six residues that bind to the two Zn2+ ions are shown as stick representations, with C atoms colored yellow, N atoms blue and O atoms red. The N- and C-terminal residues are labeled N and C, respectively. (b) Close-up view of the active site of the apo PTE structure. The buried α-Zn2+ ion is directly bound to His55, His57 and Asp301, while the exposed β-Zn2+ ion is bound to His201 and His230. The carbamate bound to Lys169 interacts with both Zn2+ ions. Coloring is as in (a). |
There has been significant interest in utilizing enzymes that are capable of hydrolyzing OPs for the degradation and disposal of OP-based nerve agents and insecticides (Singh, 2009). Initial efforts to develop efficient enzyme variants for this purpose took human paraoxonase 1 (hPON1) as the starting point for multiple rounds of directed evolution, taken together with computational design and site-directed mutagenesis (Aharoni et al., 2005; Gupta et al., 2011; Worek et al., 2014).
However, more recently, bacterial phosphotriesterases (PTEs) have proved to have a greater potential and have been used extensively over the course of the past two decades to generate variants with high catalytic efficiency towards a broad range of OPs of interest (Bigley et al., 2015; Cherny et al., 2013; Goldsmith et al., 2016, 2017; Grimsley et al., 2005; Jackson, Weir et al., 2009; McLoughlin et al., 2005; Tsai et al., 2012; Yang et al., 2003, 2014). Thus, variants have been produced that can serve as effective prophylactic and therapeutic antidotes for the treatment of both V-type and G-type nerve-agent intoxication (Cherny et al., 2013; Despotović et al., 2019; Goldsmith et al., 2016, 2017). Recently, PTE isolated from a marine bacterium permitted growth on a pesticide analogue as the sole phosphorus source, suggesting that marine bacteria might serve for the remediation of diverse OP-containing pollutants (Despotović et al., 2022).
Since the evolved PTE scaffolds had been shown to act as highly effective bioscavengers both in vitro and in vivo, it was important to clarify the structural features of their active sites. To date, 75 crystal structures of apo PTE and OP–PTE complexes, which are displayed in Table 1, have provided valuable information for mapping their active sites. We thought it important to determine the crystal structures of our advanced PTE variants and, when possible, those of their complexes with OPs so as to better understand the structural features that are responsible for their enhanced effectiveness. This study focused on three PTE variants, A53, C23 and C23M (Fig. 2b), which correspond to G5-A53, G5-C23 and G5-B60, respectively, as described previously (Cherny et al., 2013). These variants showed enhanced efficiency in degrading and detoxifying OPs relative to wild-type (wt) PTE. The following features were investigated.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
![[Figure 2]](nj5323fig2thm.gif) | Figure 2 PTE variants. (a) Schematic presentation of the maltose-binding protein (MBP) fused before the factor Xa cleavage motif (IEGR) and the octapeptide spacer sequence, 26ISEFITNS33, followed by the mature PTE protein sequence starting with Gly34. Factor Xa cleaves after the arginine residue of the cleavage motif, leaving the 26ISEFITNS33 linker attached to the PTE, Gly34–Ser365. (b) Sequence alignment of wt PTE, PDB entry 1hzy, A53, C23, C23M, A53_T, C23_T and C23M_T. The last three bear the octapeptide tag. Secondary-structure elements of PDB entry 1hzy are labeled above the alignments: α-helices and 310-helices (shown with the symbol η) are indicated by coils and β-strands by arrows. Residues conserved in all variants are in red. Multiple sequence alignment was performed using MultAlin (Corpet, 1988 ) and the figure was created using ESPript (Robert & Gouet, 2014). |
(i) The effect of retaining a residual tag used for protein expression.
(ii) Three different crystallization conditions for A53, C23 and C23M were used to optimize the formation of PTE–OP complexes.
(iii) Four different space groups allowed the investigation of the impact of crystal packing on the formation of PTE–OP complexes.
(iv) The use of four different OP ligands permitted the identification of ligand-specific interactions and differences in these interactions between the three PTE variants.
2. Materials and methods
2.1. Expression and purification of PTE variants
PTE variants were expressed either without a fusion protein or tag, or as a fusion protein with maltose-binding protein (MBP). In the latter case, MBP was fused to PTE via a factor Xa cleavage motif sequence: IEGR (Cherny et al., 2013) and a spacer sequence of eight amino-acid residues: ISEFITNS (see the schematic representations in Fig. 2a). The purpose of using the MBP fusion partner was to increase the expression levels and solubility of the PTE variants. However, the MBP fusion partner was removed by digestion with factor Xa prior to crystallization trials.
2.1.1. Expression of tag-free PTE variants
The PTE constructs A53, C23 and C23M (Cherny et al., 2013) were generated based on a modified protocol (Tokuriki et al., 2012; Fig. 2b). These variants (residues Gly34–Ser365), devoid of any tag, were cloned into pET-21. They were grown in a 5 l culture of Escherichia coli BL21 cells at 30°C following induction with 0.5 mM isopropyl β-D-1-thiogalactopyranoside (IPTG). The bacterial pellet was resuspended in 20 mM HEPES pH 7.5 supplemented with 0.1 mM ZnCl2. The cells were sonicated for 2 min, employing pulses of 30 s on and 30 s off, at 35% amplitude in a Vibracell apparatus (Sonics & Materials Inc., Newtown, Connecticut, USA) at 4°C. The clarified lysate was loaded onto a HiTrap DEAE FF 5 ml column (GE Healthcare) pre-equilibrated with 0.1 mM ZnCl2, 20 mM HEPES pH 7.5. PTE-containing fractions were eluted with a 0–1 M NaCl gradient in the same buffer. The peak fractions containing PTE were dialyzed against 0.1 mM ZnCl2, 50 mM MES pH 6.0 and loaded onto a Tricorn Mono S 10/100 GL cation-exchange column (GE Healthcare) equilibrated with the same buffer. PTE was again eluted with a 0–1 M NaCl gradient in the same buffer. Fractions containing PTE were then concentrated to 5 ml and loaded onto a gel-filtration column (HiLoad Superdex 75 16/60) equilibrated with 50 mM NaCl, 25 mM HEPES pH 8.0. The pooled fractions were applied onto a Tricorn Q 10/100 GL anion-exchange column equilibrated with 50 mM Tris pH 8.0 and eluted with a 0–1 M NaCl gradient in the same buffer. The final protein solution was concentrated to 13–17 mg ml−1 for crystallization screening.
2.1.2. Expression and purification of MBP-PTE fusion proteins
To increase the expression levels and solubility of PTE, maltose-binding protein (MBP) was introduced into the three PTE variants as a fusion partner, together with a factor Xa cleavage motif squence: IEGR (Zhao et al., 2013), followed by an octapeptide spacer (ISEFITNS) preceding PTE (Fig. 2a). These fusion proteins were expressed and purified as described previously (Cherny et al., 2013). Briefly, E. coli BL21 cells were transformed with a PTE plasmid bearing the desired mutation. The E. coli BL21 cells were grown overnight at 37°C in lysogeny broth (LB; Bertani, 1951) supplemented with ampicillin (100 µg ml−1). They were then subcultured [1%(v/v) inoculum] into LB supplemented with 0.2 mM ZnCl2 and ampicillin (100 µg ml−1) and allowed to grow at 37°C until OD600 nm reached 0.6–0.8, followed by induction with 0.4 mM IPTG and further growth for 14–18 h at 20°C. The cell pellet was suspended in buffer A (0.1 mM ZnCl2, 10 mM NaHCO3, 100 mM NaCl, 100 mM Tris pH 8.0) containing 0.4 mg ml−1 lysozyme, 50 units of Benzonase and 1 mM phenylmethylsulfonyl fluoride, and sonicated as described for the unpacked protein. A clear cell lysate was obtained by centrifugation at 7500 rev min−1 for 30 min. Amylose beads equilibrated with buffer A were packed into a 10 ml column. The clarified lysate was loaded onto the column by gravitation and washed with buffer A. The MBP-PTE fusion variants were eluted with 10 mM maltose in buffer A. The eluted fractions were analyzed by 12% SDS–PAGE. Fractions containing MBP-PTE were pooled and dialyzed overnight against buffer A. The protein concentration was measured at 280 nm using a molar extinction coefficient for MBP-PTE of 95 800 cm−1 M−1.
For crystallization trials, the A53, C23 and C23M variants were expressed and purified from 2.5–5 l cultures as above. The suspended pellet was disrupted using a cell disruptor (Constant Systems, Low March, United Kingdom), clarified by centrifugation and applied onto a column containing amylose beads. MBP-PTE variants were eluted with maltose, and MBP was cleaved by incubation with 25 µg factor Xa per 800 ml culture (Zhao et al., 2013) for 40 h at 4°C. Post-cleavage, the solution was passed four times over a column packed with fresh amylose beads to retain the cleaved MBP and any residual uncleaved MBP-PTE, while tagless PTE appeared in the flowthrough. In some cases where PTE was still contaminated with residual MBP, an anion-exchange column was used to separate the tagless PTE from residual MBP. In these cases, the protein solution was dialyzed against 0.1 mM ZnCl2, 10% glycerol, 200 mM NaCl, 20 mM Tris pH 8.0 and loaded onto a Tricorn Q10/100 column (GE Healthcare). Under these conditions, pure PTE was obtained in the flowthrough of the column, while MBP was retained. The tagless PTE was dialyzed against buffer A and concentrated to ∼11 mg ml−1.
2.2. Synthesis of OP ligands
2.2.1. Methylphosphonic acid
Methylphosphonic acid (Fig. 3a) was obtained by the dropwise addition of 266 mg methylphosphonyl dichloride in 3 ml acetone to 5 ml H2O at 4°C. After three days at room temperature (RT) the mixture of water and acetone was removed under vacuum to yield a viscous oil, which then solidified. Only one kind of 31P-NMR signal was observed, with a (δ) of 34.0 p.p.m. in D2O. This signal corresponds to the resonance of the P atom in methylphosphonic acid (Fig. 3a).
![[Figure 3]](nj5323fig3thm.gif){kind=small} | Figure 3 Methylphosphonates that were crystallized with the PTE variants. (a) Methylphosphonic acid, (b) O-ethyl methylphosphonic acid, (c) O-ethyl-O-(N,N-diisopropylaminoethyl) methylphosphonate (the oxo analogue of VX), (d) O-isopropyl methylphosphonic acid. In all four structures the CH3 moiety is pointing away from the viewing plane and thus is not seen. The OPs are shown as stick figures with C atoms colored yellow, N atoms blue, O atoms red and P atoms orange. |
2.2.2. O-Ethyl methylphosphonic acid
A solution of 100 mg O-ethyl methylphosphonyl chloride (Levy & Ashani, 1986) in 5 ml acetone was added dropwise to 10 ml H2O at 4°C and stirred for 30 min. The mixture of water and acetone was removed under vacuum to yield a viscous oil. The 31P-NMR δ was 35.5 p.p.m. in D2O. This signal corresponds to the resonance of the P atom in O-ethyl methylphosphonic acid (Fig. 3b).
2.2.3. O-Ethyl-O-(N,N-diisopropylaminoethyl) methylphosphonate
O-Ethyl methylphosphonyl chloride (7.7 g in 25 ml dry ether) was added dropwise to a cold (ice-bath) solution of equimolar amounts of 5 g triethylamine and 7.5 g N,N-diisopropylaminoethanol in 30 ml dry ether to obtain O-ethyl-O-(N,N-diisopropylaminoethyl) methylphosphonate (Fig. 3c). After stirring for 48 h at RT, the precipitated solid was filtered off and the ether was washed several times with cold distilled water. The ether was dried over anhydrous sodium sulfate, filtered and removed under vacuum to leave behind 8.5 g of a pale yellow viscous oil. A 31P-NMR signal was observed at a δ of 33.8 p.p.m. in CDCl3. This signal corresponds to the resonance of the P atom in O-ethyl-O-(N,N-diisopropylaminoethyl) methylphosphonate (Fig. 3c). The chemical shifts and signal patterns of the 1H-NMR spectrum (300 MHz, CDCl3) are consistent with the structure of the purified ligand. δ: 4.08 (m, 10 lines, 2H), 2.85 (m, 10 lines, 2H), 2.97 (septate, 2H), 2.65 (triplet, 3H), 1.45 (doublet, 3H), 1.30 (triplet, 3H), 0.96 (doublet, 12H).
2.2.4. O-Isopropyl methylphosphonic acid
O-Isopropyl methylphosphonyl chloride was synthesized by reacting O,O-diisopropyl methylphosphonate with oxalyl chloride following the method described above for O-ethyl methylphosphonyl chloride. Thus, 7.8 g O-isopropyl methylphosphonyl chloride was added dropwise, while stirring, to 60 ml water precooled in an ice bath. After 60 min, water was removed under vacuum and the residual oil was dissolved in 100 ml benzene. The organic solution was dried over anhydrous sodium sulfate, filtered and evaporated under vacuum, leaving behind ∼5 g of a colorless oil. Only one kind of 31P-NMR signal was observed, with a δ of 36.0 p.p.m. in D2O. This signal corresponds to the resonance of the P atom in O-isopropyl methylphosphonic acid (Fig. 3d).
2.2.5. Determination of the dissociation constants of the complexes between wt PTE and OP product ligands
In order to determine the association strength between PTE and the two acid products, methylphosphonic acid (Fig. 3a) and O-ethyl methylphosphonic acid (Fig. 3b), the dissociation constants of the corresponding complexes were evaluated by use of the secondary plots of Lineweaver–Burk slopes (Radić et al., 1993). Increasing amounts of the OP acid were added to 1 mM paraoxon in 50 mM NaCl, 50 mM Tris pH 8.0 containing 0.1% Tergitol. The initial velocity of hydrolysis of paraoxon by wt PTE was recorded over the first 2 min by monitoring the release of p-nitrophenol at 400 nm.
2.3. Crystallization and of PTE variants
Crystals of the three PTE variants, A53, C23 and C23M, were grown at 19°C by the hanging-drop vapor-diffusion technique using a Mosquito robot (SPT Labtech, Cambridge, Massachusetts, USA). Crystals of the PTE variants were obtained under three different crystallization conditions: (i) polyethylene glycol (PEG) 6000 and 2-methyl-2,4-pentanediol (MPD), (ii) glycerol and ammonium sulfate (AS) and (iii) polyacrylic acid (PAA). The crystals grew in four different space groups: P21212 (PEG 6000 and MPD, A53_1), P212121 (PEG 6000 and MPD, C23_1), P21 (PEG 6000 and MPD, A53_3; PAA, A53_4) and P43212 (AS and glycerol, A53_2, A53_5, C23_2, C23_3, C23_4, C23_5, C23M_1 and C23M_2). With one exception, C23_1, the crystals diffracted to high resolutions in the range 1.38–2.00 Å. Crystallization optimization experiments were set up manually using the hanging-drop vapor-diffusion technique. Four methylphosphonate ligands were co-crystallized with the A53, C23 and C23M variants (Fig. 3). 12 crystal structures were analyzed, including the apo A53, C23 and C23M variants, as well as several complexes with the phosphonate ligands.
X-ray data collection was performed under cryo-conditions (100 K) either in-house on a Rigaku RU-H3R instrument or on beamlines ID14-4 and ID23-1 at the ESRF, Grenoble, France (Table 2). All diffraction images were indexed and integrated using MOSFLM (Leslie & Powell, 2007) and the integrated reflections were scaled using SCALA (Evans, 2006). Structure-factor amplitudes were calculated using TRUNCATE (French & Wilson, 1978) from the CCP4 suite (Agirre et al., 2023). The structure of apo A53 (A53_1) was solved by with Phaser (McCoy et al., 2007) using the homologous refined structure of Bd_PTE (PDB entry 1hzy; Benning et al., 2001) as a starting model. All steps of atomic were performed with REFMAC (Murshudov et al., 2011) and Phenix (Liebschner et al., 2019). The models were built into the electron-density map using Coot (Emsley et al., 2010). Following this, the refined A53_1 structure was used as a model to determine the structures of all of the other PTE variants (Table 2). All models were optimized using PDB-REDO (Joosten et al., 2014) and were evaluated with MolProbity (Chen et al., 2010). Details of the are presented in Table 2. The coordinates of the A53_1, A53_2, A53_3, A53_4, A53_5, C23_1, C23_2, C23_3, C23_4, C23_5, C23M_1 and C23M_2 structures have been deposited in the Protein Data Bank under PDB codes 8p7f, 8p7h, 8p7i, 8p7k, 8p7m, 8p7n, 8p7q, 8p7r, 8p7s, 8p7t, 8p7u and 8p7v, respectively.
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
3. Results
3.1. Correlation between the presence of Zn2+ and the absence of tags: A53_1 and C23_1
The A53 variant, which is devoid of any tags, was crystallized in the presence of 12% PEG 6000, 5% MPD, 0.1 M HEPES pH 7.5 and the crystals diffracted to 2.0 Å resolution (A53_1 in Table 2). The (β/α)8 TIM-barrel fold of A53_1 is very similar to that of wt Bd_PTE (PDB entry 1hzy; Benning et al., 2001). However, A53_1 is a monomer, with only one Zn2+ ion in the active site. PTEs require two divalent metal ions in their active site, which contribute to both their activity and their stability (Benning et al., 1995). The removal of either of the two metal ions results in a loss of enzymatic activity. The exposed α-Zn2+ ion in the PTE structures, as seen, for example, in PDB entry 1hzy (Benning et al., 2001), is ligated by His55, His57 and Asp301, and the buried β-Zn2+ ion is ligated by His201 and His230, while the carbamate bound to Lys169 interacts with both Zn2+ ions (Fig. 1b). These findings are consistent with the general observation in metalloenzymes that a hydrophilic active site is embedded in a hydrophobic core (Yamashita et al., 1990).
A53_1 and PDB entry 1pta (Benning et al., 1994) crystallize in the same P21212, with virtually identical unit cells: a = 79.7, b = 93.7, c = 44.6 Å and a = 80.2, b = 93.7, c = 45.0 Å, respectively. It is interesting that the structure with PDB code 1pta does not contain any Zn2+ ions in its active site (Fig. 4a), despite being structurally very similar to the A53_1 variant (r.m.s.d. of 0.498 Å), which contains one Zn2+ ion, corresponding to the buried α-Zn2+ ion, in its active site (Fig. 4b). The putative active sites of both the PDB entry 1pta and A53_1 structures differ significantly from the canonical active site containing two Zn2+ ions, as is the case for other PTE structures, such as PDB entry 1hzy (Fig. 4c).
![[Figure 4]](nj5323fig4thm.gif){kind=small} | Figure 4 Comparison of the active-site regions of PTEs containing different numbers of Zn2+ ions. (a) PDB entry 1pta, which is devoid of Zn2+ ions. (b) A53_1, containing one Zn2+ ion. (c) PDB entry 1hzy, containing two Zn2+ ions. Zn2+ ions are shown in magenta and the bridging water is in cyan. |
It has been reported that some reagents used in crystallization processes can act as chelators, creating coordinate bonds to the Zn2+ ions in solution and thus decreasing their with Tris being one such reagent (Fischer et al., 1979). In the present study, Tris was used at high concentrations, >50 mM, both in the purification of the PTE variants and in some of the crystallization trials. Tris can chelate metal ions, especially Zn2+, via its amine N atom (Handing et al., 2018). Its presence in the purification buffer thus decreases the of free Zn2+, making it difficult to incorporate two Zn2+ ions into the active sites of the PTE variants. Since the presence of both Zn2+ ions and the correct orientation of the residues in the active site are crucial for maintaining the of PTE, it is reasonable to assume that the A53_1 monomer is catalytically inactive. To avoid forming inactive PTE molecules containing less than two Zn2+ ions in the active site, it was crucial to add ZnCl2 during protein expression, purification and crystallization. All of the structures listed in Table 2, except for A53_1 and C23_1, were obtained using preparations in which the purification media were supplemented with ZnCl2, and indeed display molecular dimers with two Zn2+ ions in the active site of each monomer.
Comparison of PDB entry 1hzy, containing two Zn2+ ions, with A53_1, which contains only one, reveals an r.m.s.d. of 0.65 Å on Cα atoms. Inspection of the active site of A53_1 shows that the α-Zn2+ ion is ∼2.5 Å away from the corresponding ion in PDB entry 1hzy (Figs. 4b and 4c, respectively). Moreover, there are noticeable conformational deviations in key active-site residues (Fig. 4). These deviations can be attributed to the absence of the β-Zn2+ ion, which would have been expected to coordinate to the side chains of His201 and His230 (Figs. 1b and 4c). Interestingly, the structure of A53_1 more closely resembles that of PDB entry 1pta, which lacks any Zn2+ ions (Figs. 4b and 4a, respectively).
While almost all of the PTEs in the PDB are seen to crystallize with molecular dimers in the A53_1 displays only one molecule in the A dimer is formed by applying the crystallographic twofold axis in P21212. However, this dimer interface utilizes different residues and is rather loose compared with the canonical noncrystallographic dimer seen in other PTEs, suggesting that it is not a physiological dimer. The two segments, residues 60–79 and 301–313, which are involved in the canonical PTE dimer interface show pronounced conformational differences in the A53_1 structure cartoon shown in Fig. 5. Furthermore, three regions, 203–209, 254–275 and 314–320, are disordered and thus are not visible in the electron-density map (Fig. 5). Residues 314–320 are close to residues 301–313, which are involved in the canonical PTE dimer interface. Since some of the residues in the disordered regions are in the vicinity of the active site, it is plausible that their disorder, and the significant conformational changes observed in residues 60–79 and 301–313, as well as the presence of only one Zn2+ ion, preclude the dimerization of the A53_1 structure and eliminate Indeed, A53_1 occurs as a monomer in solution, as demonstrated by gel filtration (not shown).
![[Figure 5]](nj5323fig5thm.gif) | Figure 5 Cartoon tube diagrams of the backbones of the monomer structures of apo A53_2 (beige) and apo A53_1 (green). A region displaying sizeable conformational differences (residues 60–79) is circled by a black dashed line. This region participates in dimer formation in the A53_2, A53_3 and A53_4 structures, which are very similar and differ significantly from that of A53_1. The missing regions in A53_1, i.e. residues 203–210, 254–274 and 314–320, are colored red and circled by red dashed lines. The two Zn2+ ions in A53_2 are shown as magenta spheres and the residues which bind them are displayed as sticks. |
The tag-free C23 variant was expressed without the MBP fusion tag and purified by conventional purification techniques. Crystals of C23_1 obtained from 10% PEG 6000, 15% MPD, 2% PEG 400, 0.1 M HEPES pH 7.5 diffracted to 3.2 Å resolution (Table 2). Since the crystals diffracted so poorly, it was deemed to be unsuitable for use in OP co-crystallization experiments.
3.2. The structures of tagged constructs of A53, C23 and C23M
The addition of ZnCl2 during expression, purification and crystallization resulted in the presence of two Zn2+ ions in the active sites of A53_2, A53_3 and A53_4, all three of which crystallized under different crystallization conditions. The coordination geometry of the two Zn2+ ions in the active site of these three structures is similar to that observed in other PTE structures, such as PDB entry 1hzy. Interestingly, the of A53_2 has a monomer in the and the dimer is formed through a crystallographic twofold axis in P43212, as observed in the structures of several other PTEs: A53_5, C23_2, C23_3, C23_4, C23_5, C23M_1 and C23M_2. However, A53_3 and A53_4 crystallize in P21 with a dimer in the (Table 2).
3.3. Impact of a residual tag on the apo structures of the A53, C23 and C23M variants
To facilitate purification, the A53, C23 and C23M variants were expressed as constructs with an N-terminal MBP tag (Fig. 2a). Factor Xa cleavage resulted in proteins with an octapeptide linker (ISEFITNS) at the N-terminus followed by the mature PTE protein sequence, starting at Gly34, for constructs A53_T, C23_T and C23M_T (Fig. 2b).
Apo structures of the octapeptide-tagged A53 variant were obtained using three different crystallization precipitants: AS with glycerol, PEG 6000 with MPD, and PAA, which resulted in crystal structures A53_2, A53_3 and A53_4, respectively. Crystals of apo C23 were obtained using three precipitants: PEG 6000 with MPD (C23_1), PAA, and AS with glycerol (data not shown).
C23_1 crystallized in P212121, with two dimers in the The crystals of tagged C23 (C23_2, C23_3, C23_4 and C23_5) and C23M (C23M_1 and C23M_2) grew from AS and glycerol, with one monomer in P43212, such that the canonical dimer is generated by crystallographic symmetry.
In the crystal structures of A53_3 and A53_4, electron density corresponding to residues from the octapeptide spacer 26ISEFITNS33 was unexpectedly observed (A53_T in Fig. 2b). Thus, the residual octapeptide of subunit B was found to penetrate into the active site of the symmetry-related subunit A, such that in A53_3 the distance between the active site of the A subunit and the N-terminus (Ile26) of the octapeptide of subunit B was approximately 4.5 Å (Fig. 6). The residues from the octapeptide of subunit B made contact with Trp131, Glu132, Gln173, Phe203, Ala270, Phe306, Ser308 and Tyr309 in the symmetry-related subunit A. Interestingly, although the crystals of A53_3 and A53_4 were obtained from different crystallization conditions (PEG 6000 with MPD and PAA, respectively), both crystallized in P21 and both retained the residual octapeptide. Notably, A53 crystallized in the presence of PEG 8000 in P21 also contains the residual octapeptide in the active site (data not shown). Since the octapeptide was only seen in crystals in P21, it is reasonable to assume that penetration of the residual octapeptide into the active site is space-group dependent. The presence of the octapeptide in the active site might be expected to interfere with the binding of OPs. This indeed explains why no OPs were found in crystals of A53_3 and A53_4 (Fig. 6). The octapeptide behaves like a peptide-inhibitor mimic and thus may help to define the mode of binding of a substrate with a long side chain.
![[Figure 6]](nj5323fig6thm.gif) | Figure 6 Ribbon representation of the A53_3 variant. It shows the octapeptide tag on subunit B penetrating the active-site region of the symmetry-related subunit A. The tag is shown in magenta. The active-site residues of the symmetrically related chain A are shown in yellow, with those residues within 5 Å of the tag shown in cyan. The green electron density corresponds to an omit map with the octapeptide omitted (contoured at 3σ). The cyclic compound X3E, which was presumably carried over from the protein expression and purification process, is seen in the active site, and the black electron density corresponds to a 2Fo − Fc map (contoured at 1σ). The two Zn2+ ions are shown as magenta spheres. |
3.4. PTE crystal structures obtained using polyacrylic acid as the precipitant
PTE variants were also crystallized using polyacrylic acid (PAA), including A53 (A53_4; Table 2) and C23 (data not shown). In these cases, no water molecules are directly bound to the Zn2+ ions, since the PAA monomer, i.e. acrylic acid (AA), is detected in the active site. The bound AA acts as a ligand to bridge the two Zn2+ metal ions, mimicking the tetrahedral intermediate formed during the hydrolysis of carboxylate and OPs (Fig. 7). The presence of crystallization reagents in the active site of PTE is not unusual, since similar observations have also been made in other PTE structures (Table 2). Thus, cacodylate was observed in many PTE structures, with its two O atoms bound to the two Zn2+ ions (for example, PDB entries 4xd3, 4xd4, 4xd5, 4xd6, 4xaf, 4xag, 4xay, 6gbl and others; Campbell et al., 2016). In the structure of organophosphorus acid anhydrolase (OPAA; PDB entry 4zwo), glycolic acid, which resembles acrylic acid, was observed in the active site, with its two O atoms bound to the two Mn2+ ions (Daczkowski et al., 2015). Many studies have been published on nasal drug delivery making use of PAA derivatives, and its bioadhesive properties are well recognized (Arkaban et al., 2022; Sabale et al., 2020). Indeed, its strong bioadhesive properties and its high capacity to bind to proteins (Dai et al., 2006) can explain its presence within the active site of the PTE variants. Furthermore, in addition to AA, the A53_4 active site also contains the octapeptide, which is exclusively observed in P21 (as described in Section 3.3). This suggests that the crystallization reagent PAA initially binds to the active site of PTE in solution. Subsequently, the octapeptide spacer 26ISEFITNS33 penetrates the active site during crystal formation in P21. Since no electron density for the co-crystallized OP ligands was observed in any PTE variants crystallized from PAA, the latter may have a higher affinity for the Zn2+ ions than the OPs.
![[Figure 7]](nj5323fig7thm.gif) | Figure 7 Active-site region of A53_4. Acrylic acid (AA; blue sticks) is clearly seen bound at the active site. The green electron density corresponds to an omit map with AA and the two Zn2+ ions omitted (contoured at 3σ). The black electron density corresponds to a 2Fo − Fc map (contoured at 1σ). The two Zn2+ ions are shown as magenta spheres. |
3.5. Identifying the metal ion within the active site of PTE
Identifying the intrinsically bound metal ion(s) in a protein–metal complex structure is crucial to ensure that they are consistent with the solutions employed in the expression, purification and crystallization steps (Zheng et al., 2008). Unanticipated metal ions can potentially replace the expected ones, resulting in incorrect or misleading data. In the case of the PTE variants, it was essential to confirm the identity of the metal ions in the active site. Accordingly, the X-ray data for C23M_1 crystallized from AS and glycerol were collected at the zinc absorption-edge wavelength (λ = 1.2724 Å) on beamline ID23-1 at the ESRF from a crystal that diffracted to 1.38 Å resolution (C23M_1 in Table 2; Fig. 8a), showing a peak corresponding to the zinc The anomalous omit electron-density maps of the active-site region show unequivocally that the two metal ions within the active site are indeed both Zn2+ ions, thus confirming their expected identity (Figs. 8b and 9).
![[Figure 8]](nj5323fig8thm.gif) | Figure 8 (a) Scan of the C23M_1 crystal at a range of energies showing a peak corresponding to the zinc (top). Scattering factors (f′ and f′′) are plotted as a function of energy (bottom). (b) An anomalous omit electron-density map of the active-site region of C23M_1, contoured at 6σ, is shown in black. The two Zn2+ ions are shown as magenta spheres and were omitted in calculating the electron density. |
![[Figure 9]](nj5323fig9thm.gif) | Figure 9 Electron-density omit map of the active-site region of C23M_1. The two Zn2+ ions and the electron density of an unidentified six-membered ring ligand, labeled X3T, were omitted in calculating the electron densities. The 2Fo − Fc omit map, contoured at 1σ, is shown in black. The Fo − Fc omit map, contoured at 3σ, is shown in green. The two Zn2+ ions are shown as magenta spheres. |
It can be seen that there is well defined electron density in the active site of C23M_1 corresponding to an unidentified six-membered ring ligand (X3T). This ligand will be discussed in more detail in the following section.
3.6. Cyclic compounds in PTE active sites
As already noted, it is not uncommon for the crystallization precipitants and compounds used in protein expression and purification to be observed within the active sites of protein structures (Dym et al., 2016). PEG and MPD were used as precipitants for the crystallization of A53_1, A53_3 and C23_1 (Table 2). As described above, A53_1 is presumably an inactive monomer with one Zn2+ ion in its active site and C23_1 diffracted to low resolution. A53_3 crystallized in P21, with the residual octapeptide pointing into the active site. Additionally, an unidentified six-membered ring ligand (X3E) was observed in the active site with two O atoms, likely acting as a chelator that coordinates the two Zn2+ ions present in the active site of PTE (Fig. 6). For A53, which was crystallized from PEG 6000 and MPD, no electron density for the co-crystallized OP ligands was detected, as was the case for A53 crystallized from PAA. Interestingly, both A53_3 and A53_4, which contain X3E and AA, respectively, in their active sites, crystallized in P21, with the residual octapeptide pointing into their active site.
The apo structures of A53_2 and C23M_1, which were crystallized from AS and glycerol, also showed the presence of six-membered ring ligands: X3B (Fig. 10) and X3T (Fig. 9), respectively. These ligands, similarly to X3E, act as chelators that coordinate the Zn2+ ions in the active site. It is therefore very likely that they were carried over from the protein expression and purification process. It is worth mentioning that in some published PTE structures, i.e. PDB entry 3a4j (Jackson, Foo et al., 2009), water molecules were assigned within the active site. However, they also contain electron density in the active site that might correspond to unidentified six-membered rings similar to those that we observed.
![[Figure 10]](nj5323fig10thm.gif) | Figure 10 Electron-density omit map of the active-site region of A53_2. The two Zn2+ ions and the electron density of an unidentified six-membered ring, labeled X3B, were omitted in calculating the electron densities. The 2Fo − Fc omit map, contoured at 1σ, is shown in black. The Fo − Fc omit map, contoured at 3σ, is shown in green. The two Zn2+ ions are shown as magenta spheres. |
Notably, when AS and glycerol are used as precipitants, the binding of OP ligands displaces the cyclic compounds X3B or X3T initially observed in the apo structures. This displacement and binding of the OPs will be discussed in more detail below.
3.7. Co-crystallization and soaking of OPs into PTEs
Soaking ligands into crystals is a common approach to obtain the structures of protein–ligand complexes. However, it requires the careful consideration of several factors, including the requirement to dissolve the ligand either in the crystallization precipitant or in a solvent that will not destroy the protein crystal, the anticipated soaking time and the choice of an effective ligand concentration. One significant limitation of soaking is the requirement for a crystal form with an accessible ligand-binding site or with a bound ligand that can easily be replaced by the ligand of interest. In the case of the PTE variants, attempts to soak the products of OP substrates into existing crystals of the native enzymes were unsuccessful because the active site was already occupied by ligands, such as the octapeptide spacer (Fig. 6), AA (Fig. 7) or the unidentified six-membered rings shown in Figs. 9 and 10, coordinated to the α-Zn2+ and β-Zn2+ ions, thus making it difficult for the soaked ligands to replace them.
Failure of the employed OP ligands to displace the ligands in the active site could be due to a lack of flexibility of the crystalline protein, to crystal-packing constraints or to the fact that the products may be bound more tightly than the incoming OP substrate. These limitations, taken together, contribute to the challenges faced when attempting to soak OP ligands into existing crystals of PTE variants. One way to overcome these challenges is to employ co-crystallization, which can facilitate ligand exchange within the active site in the absence of the constraints imposed by a pre-existing crystal structure.
Co-crystallization is the method of choice when the ligands are insoluble in the crystallization precipitant, the crystal-packing constraints preclude soaking, ligand binding is associated with conformational changes or the active site is occupied by a ligand that cannot be displaced by the ligand of interest. Indeed, in the PTE variants studied, OP hydrolysis products could only be observed in the active site when co-crystallization was the method adopted, and only in crystals obtained from AS and glycerol in P43212. The electron density in the active sites of the A53_5, C23_3 and C23M_2 structures co-crystallized with methylphosphonic acid (Fig. 11a) clearly indicated that the ligand replaced the six-membered ring present in the apo structures (X3B in A53_2 and X3T in C23M_1), and two of the three O atoms of the OPs are in close contact with the two Zn2+ ions at similar interatomic distances of 1.9–2.0 Å. It is somewhat puzzling that in the complexes of methylphosphonic acid with the three PTE variants, A53, C23 and C23M, the methyl group of CH3P unequivocally projects in an identical direction, which clearly differs from that observed in the other OP conjugates (Figs. 11b, 11c and 11d). In the case of co-crystallization with authentic O-ethyl methylphosphonic acid (C23_5; Fig. 11), the electron density observed could account only for the first C atom of the OCH2CH3 substituent. Since the second C atom would not make contact with any other it is likely to be disordered. In the case of the C23_4 structure co-crystallized with O-ethyl-O-(N,N-diisopropylaminoethyl) methylphosphonate (Fig. 3c), the electron density is consistent with its hydrolysis product, O-ethyl methylphosphonic acid (Fig. 11b). This suggested unexpected hydrolysis of O-ethyl-O-(N,N-diisopropylaminoethyl) methylphosphonate, either in the stored stock or in the crystallization medium. The rapid hydrolysis of this compound will be described elsewhere. The presence of a bound disordered moiety of detached N,N-diisopropylaminoethanol cannot be excluded.
![[Figure 11]](nj5323fig11thm.gif) | Figure 11 Electron-density omit maps (Fo − Fc) of the active-site regions of PTEs. The two Zn2+ ions and the OPs were omitted from the calculations. (a) Methylphosphonate was observed in the C23M_2 structure. (b) O-Ethyl methylphosphonic acid was observed in C23_5. However, the electron density could only account for the first C atom of the OCH2CH3 substituent. (c) O-Ethyl methylphosphonic acid, the hydrolysis product of O-ethyl-O-(N,N-diisopropylaminoethyl) methylphosphonate, was observed in the C23_4 structure. (d) O-Isopropyl methylphosphonic acid was observed in C23_2. The interatomic distances observed between the O atoms of the P—O bond in all OP acid products are 1.9–2.0 Å. The Fo − Fc omit map, contoured at 3σ, is shown in green. The six residues that bind to the two Zn2+ ions are shown as stick representations, with C atoms colored yellow, N atoms blue, O atoms red and P atoms orange. |
The dissociation constant generated in this study for the reversible complex between wt PTE and authentic O-ethyl methylphosphonic acid (Ki = 4.3 mM) is eightfold greater than that obtained for methylphosphonic acid (Ki = 0.54 mM). Both are poor inhibitors of PTE when paraoxon is the substrate. The enhanced affinity of methylphosphonic acid is attributed to the greater density of negative charge due to the third P—O bond. This confirms the contribution of the negative charge density to association with the PTE binuclear center. The O-isopropyl methylphosphonic acid in C23_2 (Fig. 11d) is oriented similarly to O-ethyl methylphosphonic acid in the active sites of C23_5 and C23_4 (Figs. 11b and 11c, respectively). Thus, based on visualization of the projections of different product ligands, in C23_2, C23_4 and C23_5 the following residues accommodate the P–O–ethyl and P–O–isopropyl moieties following detachment of the His57, Gly60, Ile106, Trp131, Asp301, Leu303, Phe306 and Ser308. The CH3 group of the CH3P moiety in these OP ligands is projected into a space defined by His230, His257, Leu271, Asp301 and Phe306, an observation consistent with the previous report using the nonhydrolysable ligand O,O-diisopropyl methylphosphonate (Benning et al., 2000). However, in the case of the three crystal structures of methylphosphonic acid (A53-5, C23-3 and C23M-2) the same methyl group is oriented in a different direction, into a space defined by His57, Ile106, Trp131, Asp301 and Leu303. The multiple orientations observed in the crystal structures of a variety of substituents attached to the P atom of either nonhydrolysable OPs or acid products of OP substrates can explain the promiscuity of PTEs.
Earlier studies suggested that the phosphonyl O atoms of O,O-diethyl 4-methylbenzylphosphonate (PDB entry 1dpm; Table 1; Vanhooke et al., 1996) and diisopropyl methylphosphonate (PDB entry 1ez2; Table 1; Benning et al., 2000), in their wt PTE complexes, interacted with the more exposed β-Zn2+ ion at distances of 3.5 and 2.5 Å, respectively, thus assigning it as the catalytic Zn2+. Notably, in both structures the phosphonyl O atoms were observed to be at distances of 4.7 and 5.0 Å, respectively, from the buried α-Zn2+ ion. The experimentally determined 3D structures of the OP acid products observed in the present study are consistent with the reported 3D structures of the complexes of PTEs with two other acid products: O,O-diethylphosphoric acid (DEP; PDB entry 3cak; Table 1; Kim et al., 2008) and ethyl-4-methylbenzylphosphonate (mEBP; PDB entry 7p85; Table 1; Job et al., 2023). The P—O O atoms of PTE–DEP and PTE–mEBP are placed symmetrically at 2.0–2.2 Å away from the two Zn2+ ions, whereas in the case of the complex of the triester substrate analogue (C2H5O)2P(O)CH2phenyl-pCH3 the P=O O atom is 3.5 and 4.7 Å from the exposed β-Zn2+ and the buried α-Zn2+, respectively (Vanhooke et al., 1996). These observations are consistent with the flexibility of the PTE active site, which results in a broad specificity.
The short distance between the P—O O atom of the acid products and the buried α-Zn2+ in the C23_2, C23_4 and C23_5 structures, together with the similar short distances of the P—O O atoms observed for O,O-diethylphosphoric acid (Kim et al., 2008) and O-ethyl-4-methylbenzylphosphonic acid (Job et al., 2023), suggests the involvement of both Zn2+ ions in catalysis.
4. Discussion
This study describes the challenges encountered in obtaining crystal structures of complexes of OP ligands with the enzyme PTE. Crystallization solutions contain a spectrum of chemicals that act as protein precipitants, buffers and/or reagents to increase protein stability. The effective metal concentration in solution may be substantially decreased by the formation of metal complexes with some of these chemicals. Some crystallization reagents or compounds used in the expression and purification processes may also act as chelators that coordinate metal ions. Therefore, it is essential to carefully consider the choice and concentration of reagents used in the pipeline from expression to crystallization so as to ensure that they do not perturb the functional state of the target protein (Newman, 2004).
Zn2+ ions play a crucial role in the activity of more than 300 enzymes, including PTE (McCall et al., 2000). In initial attempts to obtain PTE crystals suitable for soaking with OPs, we obtained crystals of the A53 variant, which proved to contain an inactive monomer with only one Zn2+ ion, corresponding to the α-Zn2+ ion, in its active site. In contrast, the active form of the enzyme is a dimer, containing buried α-Zn2+ and exposed β-Zn2+ ions in each subunit, which are essential both for stabilizing the structure of the enzyme and for facilitating its (Holden & Raushel, 2021). That the two metal ions seen in the crystal structures studied were indeed Zn2+ ions was confirmed by the anomalous difference maps of X-ray data collected at the zinc To avoid forming inactive PTE molecules containing fewer than two Zn2+ ions in the active site, it was crucial to add ZnCl2 during protein expression, purification and crystallization.
To increase the expression levels and solubility of PTE, maltose-binding protein (MBP) was introduced as a fusion partner and was removed by digestion with factor Xa prior to crystallization trials. Unexpectedly, electron density was observed that could be ascribed to penetration of the PTE active site by the residual octapeptide spacer. The distance between the active site and the N-terminus of the octapeptide was approximately 4.5 Å. Notably, the residual octapeptide was observed in PTE crystals obtained from several crystallization conditions, all of which crystallized exclusively in P21. The presence of the octapeptide in the active site might be expected to interfere with the binding of OPs. Thus, it is clear that the in which a protein crystallizes can affect the conformation of the protein and the binding of ligands. This is because the determines the crystal packing, which in turn affects the local environment around the protein molecules. It is worth noting that the in which a protein crystallizes is not typically taken into account in the computer-aided drug-design methods used to identify promising drug candidates, which rely on the protein structure alone to predict binding. Furthermore, our results highlight the importance of removing residual tags used to increase expression and purification levels before attempting to study the 3D structures of complexes and conjugates of proteins. In fact, when one overlays the recently determined structure of PTE complexed with ethyl-4-methylbenzylphosphonate (mEBP; PDB entry 7p85; Table 1; Job et al., 2023) on that of A53_3, the benzyl group of mEBP is seen to be very near the side chain of Ile26 of the octapeptide tag. This is strong supporting evidence that the octapeptide binds in a position similar to that of the large of the OP.
In the present study, it was observed that some of the crystallization precipitants and compounds used in the expression and purification processes could lodge within the active site of PTE. Specifically, the AA monomer of the PAA used as a precipitant was detected in the active site of the PTE variants, and well defined electron density corresponding to unidentified six-membered ring compounds was observed in the PTE active site when either PEG 6000 together with MPD, or AS together with glycerol, were employed in crystallization trials. Soaking OPs into crystals of the apo PTE variants grown from the three different crystallization conditions failed to replace either the six-membered ring compounds or the AA observed in the active site. We overcame this limitation by adapting co-crystallization protocols. We thus co-crystallized the A53, C23 or C23M variants in the presence of OPs under the same conditions used to crystallize the apo forms. However, in the co-crystallization experiments where either PEG 6000 together with MPD or PAA were used as crystallization conditions, the six-membered ring (X3E) and AA, respectively, also lodged within the active site of PTE and impeded OP binding. In the crystals obtained in P21 the active site also contains the residual octapeptide tag used for expression, which is only observed in this This octapeptide behaves like a peptide-inhibitor mimic and thus may help to define the mode of binding of substrates with an aliphatic leaving group.
Only when the PTE variants and the OPs are co-crystallized from AS with glycerol in P43212 can the OP ligands displace the cyclic compounds X3B or X3T that were initially observed in the apo structures. Thus, electron density corresponding to methylphosphonate was observed in the active sites of the three PTE variants A53, C23 and C23M (see Table 2). The electron-density map indicates that the methyl group (CH3) of the methylphosphonate moiety unequivocally projects in an identical direction across all three variants. The O-ethyl methylphosphonate and O-isopropyl methylphosphonate moieties were observed in the active sites of the C23 structures (see Table 2). Electron density corresponding to O-ethyl methylphosphonate, the less expected hydrolysis product of O-ethyl-O-(N,N-diisopropylaminoethyl) methylphosphonate, was observed in the active site of the C23 structure, implying rapid hydrolysis of this ligand in aqueous solutions. The isopropyl group of O-isopropyl methylphosphonate and the ethyl group of O-ethyl methylphosphonate project in the same direction, which clearly differs from that observed for the methyl group of methylphosphonate. The results obtained in the co-crystallization experiments in AS with glycerol in P43212 defined an O-alkyl binding pocket.
The accepted catalytic mechanism for the hydrolysis of OP triesters by PTEs is based on data that are consistent with an SN2-type nucleophilic displacement of the by direct attack at the P atom (Bigley & Raushel, 2013; Koca et al., 2001). The is believed to be a water molecule (or hydroxide ion) that is clearly seen as well defined electron density bridging the two Zn2+ ions in a high-resolution PTE structure (PDB entry 2ob3). Thus, in all likelihood, the first step is the formation of a pentacoordinated transition state (TS) that develops two partially charged P—O bonds, as shown below schematically for the hydrolysis of an O-alkyl methylphosphonate with a suitable (LG) (Fig. 12).
![[Figure 12]](nj5323fig12thm.gif){kind=small} | Figure 12 Schematic depiction of the putative TS for the hydrolysis of an O-alkyl methylphosphonate. The is a hydroxide ion that originates from a water molecule that bridges the two Zn2+ ions. LG represents the The geometries displayed are for the purposes of illustration and discussion. |
The findings presented in this study may help to rationalize the role of the active-site metal center in PTEs. The OP products observed in the C23 crystal structures (Fig. 11) are in their ground state. Yet, they mimic the putative charged TS for the hydrolysis of O-alkyl methylphosphonates by offering two P—O bonds that ligate the two Zn2+ ions in a bidentate mode at interatomic distances of 1.9–2.0 Å (Fig. 11). Together with the similar short distances (2.0–2.2 Å) reported for PTE complexes with acidic OP products such as O,O-diethylphosphoric acid (Kim et al., 2008) and O-ethyl-4-benzylphosphonic acid (Job et al., 2023), it is suggested that, regardless of the size of the substituent, the enzyme is sufficiently flexible to utilize the stabilization machinery offered by the two Zn2+ ions. Accordingly, the buried and exposed Zn2+ ions are proposed to stabilize the developing high-energy charged TS, thereby lowering its energy content, with concomitant acceleration of the reaction relative to water alone. A similar possibility was suggested in the case of a Zn2+-containing carbonic anhydrase in which a negatively charged transition state was envisaged (Christianson & Cox, 1999). The bimetal active-site centers formed by the Zn2+ ions seem to play a major role not only in polarizing the P=O bond and making the P atom more susceptible to nucleophilic attack, but also by stabilizing the TS through an electrostatic contribution, which is consistent with its key role in enzymatic catalysis (Warshel et al., 2006). The rapid departure of the acid OP product guarantees the high turnover of OP substrates that in the case of paraoxon, for example, approach of the reaction rate. Indeed, despite the the tight coordination of the acid OP product in the solid state, the dissociation constant Ki of 4.3 mM observed for O-ethyl methylphosphonic acid reveals poor affinity for PTE in aqueous solution. Similarly, analysis of the of the complex of the P=S-containing product of dimethoate hydrolysis, dimethylthiophosphate, suggested that the ∼1000-fold slower turnover of the phosphorothiolates by phosphotriesters when compared with the P=O homologues may be attributed to slow departure of the bound product seen in the bimetal catalytic center (Jackson et al., 2005). However, further experiments are required to substantiate this contention.
Our study highlights the significant influence of several factors on the successful crystallization of protein–ligand complexes. These factors include the molecular constructs used, the presence of residual tags, the choice of space groups, the compounds employed for crystallization and the compounds carried over from protein expression and purification processes. In the case of PTE, we observed that these compounds could occupy the active site, effectively competing with, or even preventing, the binding of OPs, which are of particular interest for drug development. This can potentially lead to the misidentification of lead drug candidates. By considering and by constructively addressing these factors, one can enhance the likelihood of obtaining crystals of protein–ligand complexes that faithfully represent the desired interaction, allowing more accurate and reliable characterization of ligand binding and aiding in the rational design of potential therapeutic agents.
Supporting information
PDB references: A53_1, 8p7f; A53_2, 8p7h; A53_3, 8p7i; A53_4, 8p7k; A53_5, 8p7m; C23_1, 8p7n; C23_2, 8p7q; C23_3, 8p7r; C23_4, 8p7s; C23_5, 8p7t; C23M_1, 8p7u; C23M_2, 8p7v
Acknowledgements
This study benefited from access to the Structural Proteomics Unit at the Weizmann Institute of Science, Israel, an Instruct-ERIC Center.
Funding information
Financial support was provided by the Center for Scientific Excellence at the Weizmann Institute of Science, Israel.
References
Agirre, J., Atanasova, M., Bagdonas, H., Ballard, C. B., Baslé, A., Beilsten-Edmands, J., Borges, R. J., Brown, D. G., Burgos-Mármol, J. J., Berrisford, J. M., Bond, P. S., Caballero, I., Catapano, L., Chojnowski, G., Cook, A. G., Cowtan, K. D., Croll, T. I., Debreczeni, J. É., Devenish, N. E., Dodson, E. J., Drevon, T. R., Emsley, P., Evans, G., Evans, P. R., Fando, M., Foadi, J., Fuentes-Montero, L., Garman, E. F., Gerstel, M., Gildea, R. J., Hatti, K., Hekkelman, M. L., Heuser, P., Hoh, S. W., Hough, M. A., Jenkins, H. T., Jiménez, E., Joosten, R. P., Keegan, R. M., Keep, N., Krissinel, E. B., Kolenko, P., Kovalevskiy, O., Lamzin, V. S., Lawson, D. M., Lebedev, A. A., Leslie, A. G. W., Lohkamp, B., Long, F., Malý, M., McCoy, A. J., McNicholas, S. J., Medina, A., Millán, C., Murray, J. W., Murshudov, G. N., Nicholls, R. A., Noble, M. E. M., Oeffner, R., Pannu, N. S., Parkhurst, J. M., Pearce, N., Pereira, J., Perrakis, A., Powell, H. R., Read, R. J., Rigden, D. J., Rochira, W., Sammito, M., Sánchez Rodríguez, F., Sheldrick, G. M., Shelley, K. L., Simkovic, F., Simpkin, A. J., Skubak, P., Sobolev, E., Steiner, R. A., Stevenson, K., Tews, I., Thomas, J. M. H., Thorn, A., Valls, J. T., Uski, V., Usón, I., Vagin, A., Velankar, S., Vollmar, M., Walden, H., Waterman, D., Wilson, K. S., Winn, M. D., Winter, G., Wojdyr, M. & Yamashita, K. (2023). Acta Cryst. D79, 449–461. Web of Science CrossRef IUCr Journals Google Scholar
Aharoni, A., Amitai, G., Bernath, K., Magdassi, S. & Tawfik, D. S. (2005). Chem. Biol. 12, 1281–1289. CrossRef PubMed CAS Google Scholar
Arkaban, H., Barani, M., Akbarizadeh, M. R., Pal Singh Chauhan, N., Jadoun, S., Dehghani Soltani, M. & Zarrintaj, P. (2022). Polymers, 14, 1259. CrossRef PubMed Google Scholar
Benning, M. M., Hong, S. B., Raushel, F. M. & Holden, H. M. (2000). J. Biol. Chem. 275, 30556–30560. CrossRef PubMed CAS Google Scholar
Benning, M. M., Kuo, J. M., Raushel, F. M. & Holden, H. M. (1994). Biochemistry, 33, 15001–15007. CrossRef CAS PubMed Web of Science Google Scholar
Benning, M. M., Kuo, J. M., Raushel, F. M. & Holden, H. M. (1995). Biochemistry, 34, 7973–7978. CrossRef CAS PubMed Web of Science Google Scholar
Benning, M. M., Shim, H., Raushel, F. M. & Holden, H. M. (2001). Biochemistry, 40, 2712–2722. Web of Science CrossRef PubMed CAS Google Scholar
Berman, H. M., Westbrook, J., Feng, Z., Gilliland, G., Bhat, T. N., Weissig, H., Shindyalov, I. N. & Bourne, P. E. (2000). Nucleic Acids Res. 28, 235–242. Web of Science CrossRef PubMed CAS Google Scholar
Bertani, G. (1951). J. Bacteriol. 62, 293–300. CrossRef PubMed CAS Web of Science Google Scholar
Bigley, A. N., Mabanglo, M. F., Harvey, S. P. & Raushel, F. M. (2015). Biochemistry, 54, 5502–5512. CrossRef CAS PubMed Google Scholar
Bigley, A. N. & Raushel, F. M. (2013). Biochim. Biophys. Acta, 1834, 443–453. CrossRef CAS PubMed Google Scholar
Bigley, A. N., Xu, C., Henderson, T. J., Harvey, S. P. & Raushel, F. M. (2013). J. Am. Chem. Soc. 135, 10426–10432. CrossRef CAS PubMed Google Scholar
Burley, S. K., Berman, H. M., Duarte, J. M., Feng, Z., Flatt, J. W., Hudson, B. P., Lowe, R., Peisach, E., Piehl, D. W., Rose, Y., Sali, A., Sekharan, M., Shao, C., Vallat, B., Voigt, M., Westbrook, J. D., Young, J. Y. & Zardecki, C. (2022). Biomolecules, 12, 1425. CrossRef PubMed Google Scholar
Campbell, E., Kaltenbach, M., Correy, G. J., Carr, P. D., Porebski, B. T., Livingstone, E. K., Afriat-Jurnou, L., Buckle, A. M., Weik, M., Hollfelder, F., Tokuriki, N. & Jackson, C. J. (2016). Nat. Chem. Biol. 12, 944–950. Web of Science CrossRef CAS PubMed Google Scholar
Campbell, E. C., Grant, J., Wang, Y., Sandhu, M., Williams, R. J., Nisbet, D. R., Perriman, A. W., Lupton, D. W. & Jackson, C. J. (2018). Adv. Biosys. 2, 1700240. CrossRef Google Scholar
Chen, V. B., Arendall, W. B., Headd, J. J., Keedy, D. A., Immormino, R. M., Kapral, G. J., Murray, L. W., Richardson, J. S. & Richardson, D. C. (2010). Acta Cryst. D66, 12–21. Web of Science CrossRef CAS IUCr Journals Google Scholar
Cherny, I., Greisen, P. Jr, Ashani, Y., Khare, S. D., Oberdorfer, G., Leader, H., Baker, D. & Tawfik, D. S. (2013). ACS Chem. Biol. 8, 2394–2403. CrossRef CAS PubMed Google Scholar
Christianson, D. W. & Cox, J. D. (1999). Annu. Rev. Biochem. 68, 33–57. Web of Science CrossRef PubMed CAS Google Scholar
Corpet, F. (1988). Nucleic Acids Res. 16, 10881–10890. CrossRef CAS PubMed Web of Science Google Scholar
Daczkowski, C. M., Pegan, S. D. & Harvey, S. P. (2015). Biochemistry, 54, 6423–6433. CrossRef CAS PubMed Google Scholar
Dai, J., Bao, Z., Sun, L., Hong, S. U., Baker, G. L. & Bruening, M. L. (2006). Langmuir, 22, 4274–4281. CrossRef PubMed CAS Google Scholar
Dauter, Z., Wlodawer, A., Minor, W., Jaskolski, M. & Rupp, B. (2014). IUCrJ, 1, 179–193. Web of Science CrossRef CAS PubMed IUCr Journals Google Scholar
Despotović, D., Aharon, E., Dubovetskyi, A., Leader, H., Ashani, Y. & Tawfik, D. S. (2019). Protein Eng. Des. Sel. 32, 169–174. PubMed Google Scholar
Despotović, D., Aharon, E., Trofimyuk, O., Dubovetskyi, A., Cherukuri, K. P., Ashani, Y., Eliason, O., Sperfeld, M., Leader, H., Castelli, A., Fumagalli, L., Savidor, A., Levin, Y., Longo, L. M., Segev, E. & Tawfik, D. S. (2022). Proc. Natl Acad. Sci. USA, 119, e2203604119. PubMed Google Scholar
Dym, O., Song, W., Felder, C., Roth, E., Shnyrov, V., Ashani, Y., Xu, Y., Joosten, R. P., Weiner, L., Sussman, J. L. & Silman, I. (2016). Protein Sci. 25, 1096–1114. Web of Science CrossRef CAS PubMed Google Scholar
Ely, F., Hadler, K. S., Gahan, L. R., Guddat, L. W., Ollis, D. L. & Schenk, G. (2010). Biochem. J. 432, 565–573. CrossRef CAS PubMed Google Scholar
Ely, F., Pedroso, M. M., Gahan, L. R., Ollis, D. L., Guddat, L. W. & Schenk, G. (2012). J. Inorg. Biochem. 106, 19–22. CrossRef CAS PubMed Google Scholar
Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. (2010). Acta Cryst. D66, 486–501. Web of Science CrossRef CAS IUCr Journals Google Scholar
Evans, P. (2006). Acta Cryst. D62, 72–82. Web of Science CrossRef CAS IUCr Journals Google Scholar
Fischer, B. E., Häring, U. K., Tribolet, R. & Sigel, H. (1979). Eur. J. Biochem. 94, 523–530. CrossRef CAS PubMed Web of Science Google Scholar
French, S. & Wilson, K. (1978). Acta Cryst. A34, 517–525. CrossRef CAS IUCr Journals Web of Science Google Scholar
Goldsmith, M., Aggarwal, N., Ashani, Y., Jubran, H., Greisen, P. J., Ovchinnikov, S., Leader, H., Baker, D., Sussman, J. L., Goldenzweig, A., Fleishman, S. J. & Tawfik, D. S. (2017). Protein Eng. Des. Sel. 30, 333–345. CrossRef CAS PubMed Google Scholar
Goldsmith, M., Eckstein, S., Ashani, Y., Greisen, P. Jr, Leader, H., Sussman, J. L., Aggarwal, N., Ovchinnikov, S., Tawfik, D. S., Baker, D., Thiermann, H. & Worek, F. (2016). Arch. Toxicol. 90, 2711–2724. CrossRef CAS PubMed Google Scholar
Grimsley, J. K., Calamini, B., Wild, J. R. & Mesecar, A. D. (2005). Arch. Biochem. Biophys. 442, 169–179. CrossRef PubMed CAS Google Scholar
Gupta, R. D., Goldsmith, M., Ashani, Y., Simo, Y., Mullokandov, G., Bar, H., Ben-David, M., Leader, H., Margalit, R., Silman, I., Sussman, J. L. & Tawfik, D. S. (2011). Nat. Chem. Biol. 7, 120–125. CrossRef CAS PubMed Google Scholar
Handing, K. B., Niedzialkowska, E., Shabalin, I. G., Kuhn, M. L., Zheng, H. & Minor, W. (2018). Nat. Protoc. 13, 1062–1090. Web of Science CrossRef CAS PubMed Google Scholar
Harper, L. L., McDaniel, C. S., Miller, C. E. & Wild, J. R. (1988). Appl. Environ. Microbiol. 54, 2586–2589. CrossRef CAS PubMed Web of Science Google Scholar
Hill, C. M., Li, W. S., Thoden, J. B., Holden, H. M. & Raushel, F. M. (2003). J. Am. Chem. Soc. 125, 8990–8991. CrossRef PubMed CAS Google Scholar
Holden, H. M. & Raushel, F. M. (2021). Biochemistry, 60, 3413–3415. CrossRef CAS PubMed Google Scholar
Holm, L. & Sander, C. (1997). Proteins, 28, 72–82. CrossRef CAS PubMed Google Scholar
Horne, I., Sutherland, T. D., Harcourt, R. L., Russell, R. J. & Oakeshott, J. G. (2002). Appl. Environ. Microbiol. 68, 3371–3376. CrossRef PubMed CAS Google Scholar
Jackson, C., Kim, H.-K., Carr, P. D., Liu, J.-W. & Ollis, D. L. (2005). Biochim. Biophys. Acta, 1752, 56–64. Web of Science CrossRef PubMed CAS Google Scholar
Jackson, C. J., Foo, J.-L., Kim, H.-K., Carr, P. D., Liu, J.-W., Salem, G. & Ollis, D. L. (2008). J. Mol. Biol. 375, 1189–1196. CrossRef PubMed CAS Google Scholar
Jackson, C. J., Foo, J.-L., Tokuriki, N., Afriat, L., Carr, P. D., Kim, H.-K., Schenk, G., Tawfik, D. S. & Ollis, D. L. (2009). Proc. Natl Acad. Sci. USA, 106, 21631–21636. Web of Science CrossRef PubMed CAS Google Scholar
Jackson, C. J., Weir, K., Herlt, A., Khurana, J., Sutherland, T. D., Horne, I., Easton, C., Russell, R. J., Scott, C. & Oakeshott, J. G. (2009). Appl. Environ. Microbiol. 75, 5153–5156. CrossRef PubMed CAS Google Scholar
Job, L., Köhler, A., Eichinger, A., Testanera, M., Escher, B., Worek, F. & Skerra, A. (2023). Biochemistry, 62, 942–955. CrossRef CAS PubMed Google Scholar
Joosten, R. P., Long, F., Murshudov, G. N. & Perrakis, A. (2014). IUCrJ, 1, 213–220. Web of Science CrossRef CAS PubMed IUCr Journals Google Scholar
Khersonsky, O., Lipsh, R., Avizemer, Z., Ashani, Y., Goldsmith, M., Leader, H., Dym, O., Rogotner, S., Trudeau, D. L., Prilusky, J., Amengual-Rigo, P., Guallar, V., Tawfik, D. S. & Fleishman, S. J. (2018). Mol. Cell, 72, 178–186. CrossRef CAS PubMed Google Scholar
Kim, J., Tsai, P. C., Chen, S. L., Himo, F., Almo, S. C. & Raushel, F. M. (2008). Biochemistry, 47, 9497–9504. CrossRef PubMed CAS Google Scholar
Koca, J., Zhan, C. G., Rittenhouse, R. C. & Ornstein, R. L. (2001). J. Am. Chem. Soc. 123, 817–826. PubMed CAS Google Scholar
Leslie, A. & Powell, H. (2007). Evolving Methods for Macromolecular Crystallography, edited by R. J. Read & J. Sussman, pp. 41–51. Dordrecht: Springer. Google Scholar
Levy, D. & Ashani, Y. (1986). Biochem. Pharmacol. 35, 1079–1085. CrossRef CAS PubMed Google Scholar
Liebschner, D., Afonine, P. V., Baker, M. L., Bunkóczi, G., Chen, V. B., Croll, T. I., Hintze, B., Hung, L.-W., Jain, S., McCoy, A. J., Moriarty, N. W., Oeffner, R. D., Poon, B. K., Prisant, M. G., Read, R. J., Richardson, J. S., Richardson, D. C., Sammito, M. D., Sobolev, O. V., Stockwell, D. H., Terwilliger, T. C., Urzhumtsev, A. G., Videau, L. L., Williams, C. J. & Adams, P. D. (2019). Acta Cryst. D75, 861–877. Web of Science CrossRef IUCr Journals Google Scholar
Masson, P. & Rochu, D. (2009). Handbook of Toxicology of Chemical Warfare Agents, edited by R. C. Gupta, pp. 1053–1065. San Diego: Academic Press. Google Scholar
McCall, K. A., Huang, C. & Fierke, C. A. (2000). J. Nutr. 130, 1437S–1446S. Web of Science CrossRef PubMed CAS Google Scholar
McCoy, A. J., Grosse-Kunstleve, R. W., Adams, P. D., Winn, M. D., Storoni, L. C. & Read, R. J. (2007). J. Appl. Cryst. 40, 658–674. Web of Science CrossRef CAS IUCr Journals Google Scholar
McLoughlin, S. Y., Jackson, C., Liu, J. W. & Ollis, D. (2005). Protein Expr. Purif. 41, 433–440. CrossRef PubMed CAS Google Scholar
Murshudov, G. N., Skubák, P., Lebedev, A. A., Pannu, N. S., Steiner, R. A., Nicholls, R. A., Winn, M. D., Long, F. & Vagin, A. A. (2011). Acta Cryst. D67, 355–367. Web of Science CrossRef CAS IUCr Journals Google Scholar
Naqvi, T., Warden, A. C., French, N., Sugrue, E., Carr, P. D., Jackson, C. J. & Scott, C. (2014). PLoS One, 9, e94177. CrossRef PubMed Google Scholar
Newman, J. (2004). Acta Cryst. D60, 610–612. Web of Science CrossRef CAS IUCr Journals Google Scholar
Radić, Z., Pickering, N. A., Vellom, D. C., Camp, S. & Taylor, P. (1993). Biochemistry, 32, 12074–12084. PubMed Google Scholar
Reeves, T. E., Wales, M. E., Grimsley, J. K., Li, P., Cerasoli, D. M. & Wild, J. R. (2008). Protein Eng. Des. Sel. 21, 405–412. CrossRef PubMed CAS Google Scholar
Robert, X. & Gouet, P. (2014). Nucleic Acids Res. 42, W320–W324. Web of Science CrossRef CAS PubMed Google Scholar
Sabale, A., Kulkarni, A. & Sabale, A. (2020). J. Drug. Deliv. Ther. 10, 183–197. CrossRef Google Scholar
Selleck, C., Guddat, L. W., Ollis, D. L., Schenk, G. & Pedroso, M. M. (2017). J. Inorg. Biochem. 177, 287–290. CrossRef CAS PubMed Google Scholar
Singh, B. K. (2009). Nat. Rev. Microbiol. 7, 156–164. CrossRef PubMed CAS Google Scholar
Sussman, J. L., Lin, D., Jiang, J., Manning, N. O., Prilusky, J., Ritter, O. & Abola, E. E. (1998). Acta Cryst. D54, 1078–1084. Web of Science CrossRef CAS IUCr Journals Google Scholar
Tokuriki, N., Jackson, C. J., Afriat-Jurnou, L., Wyganowski, K. T., Tang, R. & Tawfik, D. S. (2012). Nat. Commun. 3, 1257. CrossRef PubMed Google Scholar
Tsai, P. C., Fox, N., Bigley, A. N., Harvey, S. P., Barondeau, D. P. & Raushel, F. M. (2012). Biochemistry, 51, 6463–6475. CrossRef CAS PubMed Google Scholar
Vanhooke, J. L., Benning, M. M., Raushel, F. M. & Holden, H. M. (1996). Biochemistry, 35, 6020–6025. CrossRef CAS PubMed Web of Science Google Scholar
Warshel, A., Sharma, P. K., Kato, M., Xiang, Y., Liu, H. & Olsson, M. H. (2006). Chem. Rev. 106, 3210–3235. CrossRef PubMed CAS Google Scholar
Weichenberger, C. X., Afonine, P. V., Kantardjieff, K. & Rupp, B. (2015). Acta Cryst. D71, 1023–1038. Web of Science CrossRef IUCr Journals Google Scholar
Worek, F., Seeger, T., Goldsmith, M., Ashani, Y., Leader, H., Sussman, J. L., Tawfik, D., Thiermann, H. & Wille, T. (2014). Arch. Toxicol. 88, 1257–1266. CrossRef CAS PubMed Google Scholar
Yamashita, M. M., Wesson, L., Eisenman, G. & Eisenberg, D. (1990). Proc. Natl Acad. Sci. USA, 87, 5648–5652. CrossRef CAS PubMed Web of Science Google Scholar
Yang, C.-Y., Renfrew, P. D., Olsen, A. J., Zhang, M., Yuvienco, C., Bonneau, R. & Montclare, J. K. (2014). ChemBioChem, 15, 1761–1764. CrossRef CAS PubMed Google Scholar
Yang, H., Carr, P. D., McLoughlin, S. Y., Liu, J. W., Horne, I., Qiu, X., Jeffries, C. M., Russell, R. J., Oakeshott, J. G. & Ollis, D. L. (2003). Protein Eng. 16, 135–145. Web of Science CrossRef PubMed CAS Google Scholar
Zhao, X., Li, G. & Liang, S. (2013). J. Anal. Methods. Chem. 2013, 581093. CrossRef PubMed Google Scholar
Zheng, H., Chruszcz, M., Lasota, P., Lebioda, L. & Minor, W. (2008). J. Inorg. Biochem. 102, 1765–1776. Web of Science CrossRef PubMed CAS Google Scholar
This is an open-access article distributed under the terms of the Creative Commons Attribution (CC-BY) Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original authors and source are cited.
| STRUCTURAL BIOLOGY |
access
