LachrymatoryFactorSynthase
Lachrymatory Factor Synthase (PBD ID: 5VGS) from Allium cepa
Created by: Inaara Malick

     Lachrymatory factor (LF) ((Z)-propanethial S-oxide) causes eye irritation associated with onion chopping. This is part of the onion's chemical warfare against microbes and animals. LF is produced in a reaction catalyzed by lachrymatory factor synthase (LFS) (PBD ID: 5VGS) from Allium cepa. The crystal structure of LFS was found in apo-form and in complex with a substrate analogue, crotyl alcohol. Since the enzyme closely resembles the helix-grip fold characteristic for plant representatives of the START (star-related lipid transfer) domain-containing protein superfamily, LFS was compared to the abscisic acid receptor PYL10 (PBD ID: 3RT2), a representative of the START protein superfamily. From this comparison, structural adaptations underlying the catalytic activity of LFS and the structure of the active site were determined. Based on this information and the orientation of the ligand, a mechanism of catalysis that involves sequential proton transfer accompanied by formation of a carbanion intermediate was delineated.

     Using the sitting drop vapor diffusion method, LFS crystals were grown. At concentrations between 6 and 10 mg mL-1, LFS was mixed in a 1:1 ratio with 0.1 M sodium acetate at a pH of 4.5 containing PEG 3350 at concentrations between 25% and 32% (w/v). Rod-shaped crystals 20 µm x 500 µm formed within one day at 25?C. These crystals were cryo-protected in 0.1 M sodium acetate at a pH of 4.5 and 30% PEG 3350 (w/v) with 10% PEG 2000 and 10% glycerol (v/v) before snap freezing in liquid nitrogen for X-ray diffraction data collection. For cocrystallization of LFS with crotyl alcohol (Sigma-Aldrich), the protein was preincubated with 1 mM of the ligand for 30 minutes on ice before setting up crystal drops. The crystals were grown and cryo-protected in the same manner as previously stated (1).

     From Expasy, it was determined that lachrymatory factor synthase has a molecular weight of 17559.98 Da and an isoelectric point of 4.99 (2). LFS has only one subunit, or one unique chain. Thus, the function of the subunit is the same as the function of LFS as a whole (3). A database survey failed to identify any known functional domains exhibiting high sequence homology with onion LFS. By sequence comparison alone, no catalytic domain for the LFS activity could be determined because conserved amino acids were evenly distributed over the entire length of the LFS sequences (4). LFS contains 157 residues. However, the protein was not completely crystallized. From length 0-25, there is a section that was not crystallized. Additionally, part of the P59082 residue, an unstable fragment of LFS, and the propeptide residue were not crystallized (3).

     LFS's primary structure consists of 157 residues. The secondary structure contains two alpha helices and two 3/10-helices for a total of 4 helices or 40 residues (25% of the total molecule). It also contains ten beta sheets or 69 residues (43% of the total molecule). There are nine bends, six turns, and one beta bridge. Therefore, approximately 32% of the molecule is random coils. LFS contains both polar and nonpolar residues and both acidic and basic residues in no particular order. Since LFS only has one subunit, it is considered an asymmetric monomer. Thus, there is no quaternary structure (3). The structure of LFS displays a compact fold composed of a seven-stranded antiparallel Β-sheet (strands Β1-Β7), which enfolds a long C-terminal α-helix. Two other short α-helices (α1 and α2) located between Β1 and Β2 complete the structure. The regularity of the extended Β-sheet is perturbed by Β-bulging at Ser-43 to Val-44, Val-79 to Ala-80, and Thr-93 to Glu-94, which results in the curved shape of the entire Β-sheet.

     LFS catalyzes a reaction in which LF is produced. However, the enzymatic mechanism by which LFS converts (E)-1-propenesulfenic acid into LF is currently unknown. The conversion of (E)-1-propenesulfenic acid into its corresponding thioaldehyde S-oxide signifies a constitutive isomerization reaction that requires shuffling the position of a double bond in the alkene chain without changing the chemical formula of the reactants. This kind of enzymatic reaction usually involves protonation/deprotonation steps followed by double bond rearrangement. However, mechanistic studies that could provide unambiguous information regarding LF formation have been hindered by numerous challenges. To overcome these limitations, an alternative method was used to determine the mechanism of the reaction. A mechanistic framework for the production of LF based on high-resolution crystallographic structures of LFS from A. cepa in apo-form as well as in complex with a substrate analogue, crotyl alcohol ((2E)-but-2-en-1-ol). The active site architecture indicates that two solvent-inaccessible polar amino acids, Glu-88 and Arg-71, are in close proximity to the substrate molecule. Because the gamma carboxyl group of Glu-88 is located within hydrogen-bonding distance from sulfenic acid, it is proposed that the carboxylate oxygen of this side chain is required to polarize the substrate by abstracting an acidic proton through a general base mechanism. Under physiological conditions, glutamic acid carboxylate exists in a resonance form with the negative charge shared between the two oxygens. Interaction with Nε of the neighboring Arg-71 via hydrogen bonding stabilizes a negative charge on one of the oxygens, effectively lowering the pKa of Glu-88 and preventing it from becoming protonated. The presence of Arg-71 is necessary for the catalytic function of Glu-88. The side chain of Glu-88 is inaccessible to the solvent; this nonpolar environment increases the pKa value for the carboxylate to 8.0. The side chain hydrogen bond and Coulombic interactions with Arg-71 lower the pKa by 0.7 and 1.7 units, respectively, to a final value of 5.6. The pKa values of sulfenic acids are relatively high, ranging from 7.5 for sterically hindered 1-anthraquinone to 10.5 for 2-methyl-2-propenesulfenic acids. Because of this moderate acidity, deprotonation of the substrate solely depends on the proximity to the carboxylate anion of glutamic acid which acts as a base, for which the negative charge is stabilized via hydrogen bonding with Arg-71. Another factor that might contribute to the efficient deprotonation of the substrate is formation of a hydrogen bond between O1 and Nη2 of Arg-71and hydroxyl group of Tyr-102. This interaction increases the acidity of the sulfenic acid proton and lowers the energy barrier required for substrate deprotonation. As a consequence of proton extraction, a double bond forms between the O1 and S2 atoms that rapidly rearranges to form S-oxide. This reorganization causes a single proton deficiency at C4 that results in a localized carbanion reaction intermediate. In the secluded environment of the active site, the hydroxyl group of Tyr-102 could serve as a proton donor, forming the final product (LF) of the enzymatic reaction with a newly acquired hydrogen atom bond at C4. Because of the high pKa value of the tyrosine side chain, deprotonation of Tyr-102 could occur through the transfer of a hydrogen from the carboxyl group of Glu-88, restoring the protonation state of the active site prior to the enzymatic reaction. An alternative scenario could include donation of a proton abstracted from the substrate by Glu-88 back to the catalytic intermediate yielding the final product (LF) (1).

     The purpose of PSI-BLAST is to find proteins with similar primary structure (subjects) to a certain protein (query). PSI-blast assigns an E value to subjects that have sequence homology to the query. The E value is determined by looking at total sequence homology and by assigning gaps, which are a single amino acid or group of amino acids that exist in the subject's sequence but not the query's. Sequence homology continues after the gap. Total sequence homology decreases the E value, while gaps and mutation events increase the E value. An E value of less than 0.05 is considered significant for proteins (5). The Dali Server is used for finding proteins with tertiary structure similarities to a query. This server uses a sum-of-pairs method, which produces a measure of similarity by comparing intramolecular distances. Similarity is measured by evaluating the Z-score. Structures that have significant similarities have a Z-score above 2, which means the protein has similar folds (6). When compared to LFS, abscisic acid receptor, PYL10, has an E value of 0.025 and a Z-score of 18.9 (5, 6).

     A dominant feature of PYL abscisic acid receptors is a large internal hydrophobic cavity of elongated shape. It is formed by the inside surface of a curved Β-sheet and three α-helices. The total volume of this pocket exceeds 500 Å in the apo-forms of PYL proteins in order to accommodate relatively large ligands, such as abscisic acid (molecular volume of 254.1 Å). The opening to this binding site is surrounded by two Β-loops (Β3--Β4 and Β5--Β6) and a linker between Β7 and α3. They form a gate for lipid-binding proteins. In comparison to abscisic acid receptors, the size of the intramolecular cavity in the LFS structure is only 216 Å. The structural factor that contributes to the reduced size of this pocket is the extended Β3--Β4 loop. Although spatial positions of the secondary elements are comparable when looking at the examined structures, part of Β3 and Β4, as well as a loop that connects these two Β-strands in LFS, revealed major conformational differences that have functional consequences for LFS. The main chain of this region appears shifted inward toward the core of the protein and α3. Consequently, side chains of two large hydrophobic amino acids (Met-77 and Phe-84), which are part of the Β3--Β4 loop, protrude into the space that corresponds to the abscisic acid binding site in PYL proteins. In addition, two tryptophan residues, Trp-133 and Trp-155, limit the volume of the internal cavity in LFS. Reduced size of the binding pocket in LFS is accompanied by substitution of key residues involved in interaction with abscisic acid in PYL proteins. The carboxyl of abscisic acid forms a salt bridge with the ε-amino group of lysine and a water-mediated hydrogen bond network with several side chains of polar residues, including two glutamic acids, an asparagine, and a serine. In LFS, the lysine residue, as well as the cluster of polar amino acids, is replaced by hydrophobic methionine, phenylalanine, leucine, and tryptophan residues. Thus, the reduced size of the binding pocket, its much higher hydrophobicity, and the absence of key residues involved in hydrogen bonding with the carboxylic acid group of the ligand in PYL proteins exemplify the primary adaptations of LFS for specific interaction with a relatively small 1-propenesulfenic acid molecule (molecular volume of 82.8 Å). As demonstrated by comparing LFS to PYL proteins, small changes in the geometry of the binding pocket, accompanied by the introduction of polar amino acids in the active site, can have major functional consequences.

     The production of LF, which causes one to cry when cutting onions, depends on the activity of a specific enzyme, which was discovered to be LFS. Silencing LFS led to a generation of a tearless onion that did not exert eye irritation upon disruption of the onion tissue. Changes in the relative concentration of alliinase and LFS has had a profound effect on the organosulfur metabolite profile in genetically modified plants (1). However, more research is needed to determine the consequences of silencing LFS on the health and flavor attributes of the onion (7). For example, a preliminary rat feeding trail indicated that tearless onions may play a key role in reducing weight gain. Additionally, more research could be done to compare tearless onion extract and normal onion extract and their effects on collagen-induced in vitro platelet aggregation (8).