Crystal structure of plant γ‐glutamyl peptidase 1: implications for sulfur metabolism and secondary metabolite biosynthesis
Kosei Sone, Takehiro Ito, Hibiki Sawada, Chihaya Yamada, Toma Kashima, Akimasa Miyanaga, Naoko Ohkama‐Ohtsu, Shinya Fushinobu
TL;DR
This study reveals the crystal structure of GGP1 in Arabidopsis, showing how it processes glutathione and its conjugates in both primary and secondary metabolism.
Contribution
The paper provides the first crystal structures of GGP1, including a covalent intermediate and inactive forms, revealing its catalytic mechanism and dual function.
Findings
GGP1's active site recognizes glutamate through extensive hydrogen bonds in the S1 subsite.
An open pocket in the S1′ subsite allows dual activity toward glutathione and its conjugates.
The disulfide-linked inactive form of GGP1 suggests a role in oxidative stress regulation.
Abstract
Gamma‐glutamyl peptidase 1 (GGP1) plays a dual role in primary and secondary sulfur metabolism in Arabidopsis thaliana. During glutathione (GSH) turnover, GGP1 hydrolyzes the isopeptide bond of GSH to degrade the tripeptide into glutamate and cysteinylglycine. During glucosinolate and camalexin biosynthesis, GGP1 processes GSH conjugates by hydrolyzing the same isopeptide bond of γ‐glutamate. In the present study, we determined the crystal structures of the following GGP1 forms: ligand‐free, glutamate complex, covalent γ‐glutamate intermediate, and disulfide‐linked S–S inactive forms. The intermediate structure, in which γ‐Glu is covalently linked to the catalytic nucleophile cysteine (C100), was trapped by mutating the catalytic histidine to asparagine (H192N). In the glutamate complex and γ‐glutamate intermediate structures, glutamate bound to the S1 subsite is extensively recognized…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8| Ligand‐free | Glu complex | H192N γ‐Glu intermediate | S–S inactive | |
|---|---|---|---|---|
| Data collectiona | ||||
| Beamline | SPring‐8 BL45XU | SPring‐8 BL45XU | KEK PF‐AR NE‐3A | SLS PSI X06SA |
| Wavelength (Å) | 1.000000 | 1.000000 | 1.000000 | 1.000030 |
| Resolution range (Å) | 46.67–1.87 (1.94–1.87) | 46.63–1.68 (1.78–1.68) | 46.10–1.59 (1.62–1.59) | 46.51–2.04 (2.10–2.04) |
| Space group |
|
|
|
|
| Unit cell (Å) |
|
|
|
|
| Total reflections | 632 775 (99972) | 845 503 (134536) | 467 799 (15838) | 241 573 (19347) |
| Unique reflections | 90 502 (14608) | 123 388 (19848) | 71 347 (2936) | 36 242 (2787) |
| Multiplicity | 7.0 (6.8) | 6.9 (6.8) | 6.6 (5.4) | 6.7 (6.9) |
| Completeness (%) | 100.0 (99.8) | 100.0 (100.0) | 97.0 (82.7) | 100.0 (99.9) |
| Mean | 8.16 (1.2) | 10.0 (1.1) | 16.2 (1.8) | 13.5 (3.4) |
|
| 44.9 (169.7) | 18.0 (186.7) | 6.7 (67.8) | 10.4 (61.6) |
| CC1/2 | 0.992 (0.544) | 0.999 (0.409) | 0.996 (0.791) | 0.998 (0.895) |
| Wilson B‐factor (Å2) | 27.50 | 27.77 | 13.89 | 22.91 |
| Refinement | ||||
| Resolution range (Å) | 46.72–1.90 | 46.67–1.74 | 46.15–1.59 | 46.55–2.04 |
|
| 0.170 | 0.155 | 0.166 | 0.177 |
|
| 0.219 | 0.195 | 0.200 | 0.233 |
| Number of atoms | ||||
| Amino acids | 4018 | 3982 | 4047 | 4019 |
| Ligands | 14 | 18 | 32 | 18 |
| Waters | 347 | 466 | 448 | 126 |
| RMS deviations | ||||
| Bonds lengths (Å) | 0.0149 | 0.0148 | 0.0102 | 0.0150 |
| Bond angles (°) | 2.441 | 2.340 | 1.843 | 2.554 |
| Ramachandran plot (%) | ||||
| Favored | 96.93 | 97.14 | 97.58 | 96.92 |
| Allowed | 3.07 | 2.86 | 2.42 | 3.08 |
| Outlier | 0.00 | 0.00 | 0.00 | 0.00 |
| Average B‐factor (Å2) | ||||
| Amino acids | 23.91 | 22.29 | 21.40 | 30.59 |
| Ligands | 30.54 | 41.66 | 44.68 | 38.34 |
| Waters | 30.12 | 30.45 | 30.88 | 27.88 |
| PDB ID |
|
|
|
|
| Protein | Organism | PDB ID (chain) | Z score | RMSD (Å) |
| %seq | References |
|---|---|---|---|---|---|---|---|
| Putative GATase |
|
| 24.9 | 2.7 | 220 | 20 | [ |
| Putative GATase |
|
| 24.6 | 2.3 | 220 | 16 | – |
| SMU.1228c |
|
| 23.5 | 2.4 | 214 | 17 | – |
| Glutaminase (MsGATase) |
|
| 22.7 | 2.6 | 211 | 20 | [ |
| HTS |
|
| 22.4 | 2.4 | 220 | 17 | [ |
| AnthS TrpG subunit (SsTrpG) |
|
| 21.9 | 2.3 | 189 | 22 | [ |
| HTS |
|
| 21.3 | 2.6 | 220 | 13 | [ |
| AnthS TrpG subunit (SmTrpG) |
|
| 21.2 | 2.3 | 184 | 17 | [ |
| Putative HTS |
|
| 21.1 | 2.5 | 219 | 16 | – |
| Enzyme | Relative activity (%) |
|
|---|---|---|
| WT | 100 ± 13.1 | 49.2 ± 0.3 |
| C100S | ND | 47.5 ± 0.4 |
| H192N | ND | 56.2 ± 0.8 |
| C154A | 92.5 ± 31.0 | 48.7 ± 0.3 |
| C154S | 36.5 ± 11.8 | 47.3 ± 0.6 |
| R206A | 5.6 ± 2.0 | 50.9 ± 1.7 |
| Protein | Helix | Antiparallel | Parallel | Turn | Others |
|---|---|---|---|---|---|
| Wild‐type | 31.6 | 16.0 | 5.5 | 8.6 | 38.3 |
| C100S | 26.1 | 17.6 | 8.9 | 9.8 | 37.7 |
| H192N | 24.5 | 11.7 | 8.7 | 13.0 | 42.1 |
| C154A | 23.8 | 13.8 | 4.9 | 15.2 | 42.3 |
| C154S | 20.6 | 12.3 | 12 | 16.1 | 39.1 |
| R206A | 16.1 | 18.5 | 11.6 | 13.8 | 40.1 |
- —Japan Society for the Promotion of Science10.13039/501100001691
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsNitrogen and Sulfur Effects on Brassica · Sulfur Compounds in Biology · Genomics, phytochemicals, and oxidative stress
Introduction
Glutathione (GSH, γ‐l‐Glu‐l‐Cys‐Gly) is a tripeptide containing an isopeptide bond between the γ‐carboxy group of Glu and Cys (Fig. 1A). GSH plays various essential roles in plants, one of which is the maintenance of cellular redox homeostasis [1, 2, 3]. Two GSH molecules form a disulfide bond at their Cys residues to yield glutathione disulfide (GSSG). Moreover, mixed disulfides are formed with proteins or other chemical compounds in addition to GSH, and the thiol group of GSH is possibly oxidized to sulfenic, sulfinic, or sulfonic acid. These oxidation reactions consume excess reactive oxygen species (ROS) [4]. GSSG is reduced by glutathione reductase with NADPH to form two GSH molecules [5], and this reversible redox reaction alleviates oxidative stress [6]. Furthermore, GSH serves as a major sulfur repository. Because the thiol group of free Cys has higher reactivity than that of GSH, elevated intracellular Cys levels are highly toxic to organisms [7]. Thus, in plants, GSH serves as a storage and transport form of sulfur instead of Cys [8, 9].
Substrates and the reaction of GGP1. (A) Chemical structures of substrates: glutathione (GSH), glutathione‐indole‐3‐acetonitrile (GS‐IAN), and S‐[(Z)‐phenylacetohydroximoyl]‐l‐glutathione (GS‐B). (B) The reaction and subsites. GSH and Cys‐Gly: ‐R = ‐H. (C) The catalytic mechanism of GGP1.
γ‐Glutamyl peptidase 1 (GGP1, EC3.4.19.16) is a cytosolic enzyme that hydrolyzes the isopeptide bond of GSH conjugates to Glu and Cys‐Gly conjugates (Fig. 1B) in the biosynthesis of glucosinolates and camalexin [10]. Among the five GGP genes in Arabidopsis thaliana (GGP1–5), GGP1 and GGP3 are highly expressed in all tissues and are involved in the biosynthesis of glucosinolates and camalexin through the processing of GSH conjugates, such as glutathione‐indole‐3‐acetonitrile (GS‐IAN, Fig. 1A) [11]. On the contrary, the expression of GGP2, GGP4, and GGP5 was very low, and their physiological function is unknown. We previously reported that GGP1, and possibly GGP3, is involved in GSH turnover [12]. GGP1 is expressed under both sulfur‐deficient and normal conditions and exhibits high GSH degradation activity to form Glu and Cys‐Gly (Fig. 1B), the latter of which is then utilized for sulfur and protein metabolism [13]. Accordingly, GGP1 has dual roles in primary and secondary metabolisms through the hydrolysis of the γ‐Glu‐Cys isopeptide bond in both GSH and GSH conjugates. GGP1 is a member of the cysteine peptidase family C26 (γ‐glutamyl hydrolase family) in the MEROPS database [14], which classifies peptidases. In this study, we report the crystal structures of four different forms of GGP1 in the active site: ligand‐free, Glu complex, covalently linked γ‐Glu complex, and an inactive state in which the catalytic cysteine formed a disulfide bond. The crystal structures provided structural insights into the dual activity of GGP1 and its possible regulation by redox reactions.
Results and discussion
Overall structure
C‐terminally His_6_‐tagged recombinant protein of GGP1 was expressed in Escherichia coli and purified as a single band on SDS‐PAGE. The crystal structure of ligand‐free GGP1 was determined at 1.9 Å resolution (Table 1). The crystal belonged to space group P2_1_2_1_2_1_, and the asymmetric unit contained two noncrystallographic symmetry‐related GGP1 molecules. Molecular interface analysis using the PISA server [15] suggested that the protein is a monomer in solution. The final model contained the protein Residues 4–246 in chain A and 4–250 in chain B. The overall structure of GGP1 comprises seven α‐helices, two 3_10_ helices, and eleven β‐strands, adopting a three‐layer α/β/α sandwich fold (Fig. 2). The central β‐sheet consists of six parallel strands (β2, β1, β3, β4, β11, and β9) and an antiparallel strand (β10). This β‐sheet is sandwiched between three α‐helices (α1, α2, and α7) and two β‐strands (β6 and β7) on one side and two α‐helices (α3 and α4), two 3_10_ helices (η1 and η2), and two β‐strands (β5 and β8) on the other side. A two‐helix bundle (α5 and α6) is extended from the core α/β/α fold. The catalytic triad of GGP1 comprises C100, H192, and E194. These residues are completely conserved in the C26 family and play essential roles in catalysis [16, 17]. Therefore, the reaction mechanism of GGP1 was presumed to follow that of general cysteine peptidases (Fig. 1C). C100 is located in the loop between β4 and α4, while H192 and E194 are in the loop between β11 and α5 (Fig. 2). In the crystal structure of ligand‐free GGP1, the distance between H192 and E194 is 2.7 Å, where a hydrogen bond is formed, and the distance between C100 and H192 is 3.6 Å (Fig. 3A). Although this distance is within the acceptable range for a hydrogen bond between the sulfur and nitrogen atoms, the relative orientation of the C100 side chain was not compatible with that of the hydrogen bond. This structural arrangement of the catalytic triad is common among C26 family peptidases [17].
Overall structure of GGP1 in ligand‐free form. The side chains of the catalytic triad (C100, H192, and E194) and mutated residues in this study (C154 and R206) are shown as sticks. Helices, β‐sheets, and loops are shown in cyan, yellow, and orange, respectively. The image was prepared using Chimera.
Active site of crystal structures of GGP1. (A) Ligand‐free form, (B) Glu complex structure, (C) γ‐Glu intermediate structure, and (D) S–S inactive structure. In (A), (B), and (C), polder maps at 6σ are shown. In (D), a polder map at 3.5σ is shown to indicate the two different conformations of C100 and C154. Hydrogen bonds with the ligand and within the catalytic triad are shown as blue and black dotted lines, respectively. (E) The distance between the putative catalytic water (red sphere) in the H192N γ‐Glu intermediate structure (pink) and superimposed H192 side chain in the ligand‐free structure (white sticks). (F) Superimposition of the catalytic triad in the Glu complex (protein in cyan and Glu in orange) and H192N γ‐Glu intermediate (pink) structures. The image was prepared using Chimera.
Ligand complex structures
The GGP1–Glu complex structure was determined at a resolution of 1.74 Å using a co‐crystallization method (Table 1). The Glu complex crystal was isomorphous with the ligand‐free crystal, and their main chain structures were similar. Their root mean square deviations (RMSD) for the Cα atoms were 0.24 Å. The electron density of Glu was clearly observed in the active site, and the bound Glu forms numerous hydrophilic interactions with the protein (Fig. 3B). The γ‐carboxy group of Glu is located near the nucleophilic C100 residue, representing the product complex after hydrolysis of the thioester intermediate in the reaction cycle (Fig. 1C). According to the subsite nomenclature of peptidases [18], γ‐Glu, Cys, and Gly of GSH correspond to P1, P1′, and P2′, and bind to the S1, S1′, and S2′ subsites, respectively (Fig. 1B). Therefore, this Glu molecule (P1) was bound to the S1 subsite of GGP1. The α‐carboxy group of Glu forms hydrogen bonds with the main chain amides of D157 and Q156 and the side chain amide of Q104, whereas the amino group forms bifurcated hydrogen bonds with the main chain carbonyl groups of S68 and H70 (Fig. 3B). The γ‐carboxy group is hydrogen‐bonded with the main chain amides of S68 and F101 and the side chain of H192.
A previous study showed that a His‐to‐Asn mutant of the catalytic triad of the small (glutaminase) subunit of carbamoyl phosphate synthetase trapped a covalent intermediate structure with nucleophilic Cys [19]. Therefore, we co‐crystallized the H192N mutant of GGP1 with GSH and successfully obtained a high‐resolution (1.59 Å) intermediate structure in which C100 and γ‐carboxy of Glu formed a covalent thioester bond (Table 1, Fig. 3C). The electron density of the thioester bond was evident, and this structure mimicked the covalent intermediate state of cysteine peptidase (Fig. 1C, State 4). The Cα RMSD between the γ‐Glu intermediate and ligand‐free structures is 0.32 Å. H192 in the catalytic triad facilitates the deprotonation of nucleophilic water during the thioester hydrolysis step (Fig. 1C, State 5) as well as the nucleophilic attack of C100 (State 2). The H192N mutation blocked thioester release and allowed the trapping of the covalent intermediate structure. Therefore, H192 plays a more crucial role in hydrolysis (steps in States 5–6–1) than in the thioester intermediate formation (steps in States 2–3–4). Although the positions of the catalytic triad in the covalent complex structure are almost identical to those in the Glu complex structure, the thioester bond formation induced a slight movement of the γ‐carboxyl group of Glu toward C100 (Fig. 3F). Interestingly, a prominent electron density peak for a water molecule was observed near the thioester carbon atom (Fig. 3C). This water is at a distance of 3.3 Å from the Cδ atom of Glu and forms hydrogen bonds with the N192 side chain and the main chain carbonyl of C154. When the His side chain is modeled in the γ‐Glu intermediate, the distance between the water and the side chain nitrogen atom is 1.7 Å (Fig. 3E). We hypothesize that this water represents the position of nucleophilic water.
In the γ‐Glu intermediate, the oxygen atom of the thioester bond forms bifurcated hydrogen bonds with the main chain amides of S68 and F101 (Fig. 3C). Combined with the observation of the Glu complex structure (Fig. 3B), it is strongly suggested that the main chain amides of S68 and F101 form an oxyanion hole of GGP1. The oxyanion hole is crucial for the catalytic function of serine and cysteine hydrolases and plays a pivotal role in stabilizing the tetrahedral intermediate during the reaction [20]. The oxyanion hole structure of GGP1 formed solely by the main chain amides is conserved in the C26 family [21, 22, 23]. In contrast, a glutamine side chain is involved in the oxyanion hole formation in the C12 family, including ubiquitin C‐terminal hydrolases, and in other cysteine peptidases [24].
Comparison with structural homologs
Structural homologs of GGP1 were searched using the Dali server [25]. Table 2 lists the top nine structures with the highest Z‐scores. Three of these structures (PDB IDs: 3M3P, 3L7N, and 2H2W) remain unpublished, and one is a putative class I glutamine amidotransferase (GATase) from Thermotoga maritima (PDB ID: 1O1Y) [26]. The structural homologs of GGP1 with known enzyme activities include the glutaminase subunit of GATase from Mycolicibacterium smegmatis (MsGATase) [27], the TrpG (glutaminase) subunit of anthranilate synthase (EC4.1.3.27) from Saccharolobus solfataricus (SsTrpG) and Serratia marcescens (SmTrpG) [29, 31], and homoserine O‐succinyltransferase (E.C 2.3.1.46) from B. cereus and E. coli [28, 30]. Putative T. maritima GATase, SsTrpG, and SmTrpG form part of the Class I glutamine amidotransferases family, according to the Structural Classification Of Proteins‐extended (SCOPe) classification [32]. MsGATase and the two TrpG proteins exhibit glutaminase activity that hydrolyzes Gln to Glu and NH_3_ (EC3.5.1.35), and their structures have folds similar to that of GGP1 (Fig. 4A).
Structural comparison with GGP1 homologs. (A) Overall structures of GGP1 (cyan), glutaminase subunit of glutamine amidotransferase from M. smegmatis (MsGATase, green), TrpG (glutaminase) subunit of anthranilate synthase from S. solfataricus (SsTrpG, yellow), and TrpG from S. marcescens (SmTrpG, purple). (B) Superimposition of the Glu complex structure of GGP1 (protein in cyan and Glu in orange) and MsGATase (PDB ID: 7D54, chain B in green) complexed with Gln (magenta). (C) Superimposition of the γ‐Glu intermediate structure of GGP1 (protein in pink and Glu moiety in orange) and SmTrpG (protein in purple and Glu moiety in magenta). The image was prepared using Chimera.
The crystal structure of MsGATase complexed with Gln was obtained using a co‐crystallization method [27]. Figure 4B shows a superimposition of the Glu complex structure of GGP1 and the Gln complex structure of MsGATase (chain B). Although chain A of MsGATase bound to Gln in a catalytically nonproductive reverse orientation (not shown), Gln (magenta in Fig. 4B) in chain B (green in Fig. 4B) was bound in an apparently correct manner, placing the side chain amide nitrogen near the catalytic Cys. Although it is unclear why the Gln substrate was not hydrolyzed during the crystal growth of MsGATase, the catalytic residue (Cys99) was oxidized to cysteine sulfonate, and the bound ligand was assigned as a Gln molecule (not Glu). The superimposition of GGP1 and MsGATase indicated that their ligand‐binding modes were notably different (Fig. 4B). The α‐carboxy and amino groups of the Glu molecule in the S1 subsite of GGP1 are recognized by multiple residues, whereas those groups of Gln in MsGATase have no evident interactions with the protein. Considering that the residues recognizing the main chain groups of Glu in GGP1 (S68, H70, Q104, Q156, and D157) are highly conserved in MsGATase, they may share the same subsite S1 architecture, and the binding mode of Gln in chain B of MsGATase is probably incorrect.
Spraggon et al. successfully obtained a glutamyl thioester intermediate of SmTrpG with no mutations in the catalytic triad [31]. When the γ‐Glu intermediate structure of GGP1 was overlaid on the γ‐Glu intermediate structure of SmTrpG (Fig. 4C), the positions of the catalytic triad and Glu were nearly identical, indicating that the substrate recognition at the S1 subsite is conserved in these enzymes. The Class I glutamine amidotransferases family in SCOPe and the C26 γ‐glutamyl hydrolase family in MEROPS generally have specificity for the isopeptide bond cleavage of P1 Glu, and this specificity is common to glutaminase that hydrolyzes the side chain amide nitrogen of Gln. Therefore, we concluded that the S1 subsite recognition architecture is basically conserved in these enzyme families.
S–S inactive structure
Although analysis of the Glu complex and γ‐Glu intermediate structures helped elucidate the S1 subsite (γ‐Glu) recognition architecture of GGP1, the substrate recognition at subsites S1′ and S2′ for the Cys‐Gly dipeptide remained unknown. We soaked wild‐type GGP1 crystals in a solution containing the Cys‐Gly dipeptide and collected a diffraction dataset at 2.04 Å resolution (Table 1). Unexpectedly, the electron density map showed that the S1′ and S2′ subsites were not occupied, and C100 partially formed a disulfide bond with C154 in the next β‐strand (β7) (Fig. 3D). This is an inactive state because C100 is the catalytic nucleophile residue. We also observed a disulfide‐free state in the crystal, and the occupancies of the S–S inactive and disulfide‐free states were refined to 0.7 and 0.3, respectively, based on the electron density map of K153‐C154‐H155‐Q156 (Fig. S1). The ligand‐free and S–S inactive structures are nearly the same except for the disulfide bond in the active site; the Cα RMSD between them is 0.32 Å. Considering that the Cys‐Gly dipeptide is a pro‐oxidant that acts as an iron reductant as well as a lipid peroxidation inducer [33], soaking crystals in the dipeptide‐containing solution may have prompted the oxidization of two neighboring cysteine residues. In plant cells, GSH functions as an antioxidant by consuming excess ROS in the cytosol through reversible redox reactions between GSH and GSSG [1]. We hypothesized that the disulfide bond involving the catalytic C100 residue is formed under oxidative stress and that the loss of GGP1 function prevents the degradation of the antioxidant GSH. The partial amino acid sequence alignment of 30 plant GGP1 homologs with the lowest protein BLAST e‐values (<1 × 10^−162^) indicates that C154 is not conserved among plants (Fig. 5). In contrast, GGP2‐GGP5 of A. thaliana completely conserve C154. Several response pathways for oxidative stress in Arabidopsis have been reported [11, 34, 35, 36], and numerous factors are involved in the regulation of GSH supply under oxidative stress. The GGP1 inactivation by disulfide bond formation may be one such factor in the regulation of oxidative stress in Arabidopsis.
Partial amino acid sequence alignment of A. thaliana GGP1–GGP5 and plant GGP1 homologs around C154 and R206. Strictly conserved residues are shown as red characters, and partially conserved residues are boxed in blue frames. The image was prepared using ESPript. Database accession numbers of the sequences are as follows: Arabidopsis thaliana GGP1, NP_194782.1; Arabidopsis thaliana GGP2, NP_194783.1; Arabidopsis thaliana GGP3, NP_194784.1; Arabidopsis thaliana GGP4, NP_179974.2; Arabidopsis thaliana GGP5, NP_179975.1; Arabidopsis arenosa, CAE6005314.1; Arabidopsis suecica, KAG7641848.1; Arabidopsis thaliana x Arabidopsis arenosa, KAG7540889.1; Arabidopsis lyrata subsp. lyrata, XP_002869373.1; Eutrema salsugineum, XP_006412693.1; Isatis tinctoria, QWJ73366.1; Sinapis alba, KAF8084786.1; Capsella rubella, XP_006284418.1; Arabis nemorensis, VVB10962.1; Thlaspi arvense, CAH2077453.1_1; Brassica rapa, XP_009127746.1; Brassica napus, XP_013689141.2; Eruca vesicaria subsp. sativa; Raphanus sativus, XP_018444830.1; Brassica rapa subsp. trilocularis, KAG5413098.1; Brassica carinata, KAG2243853.1; Camelina sativa, XP_010438166.1; Hirschfeldia incana, KAJ0262993.1; Brassica oleracea var. oleracea, XP_013598712.1.
Substrate modeling and implications for dual physiological roles
A notable physiological feature of GGP1 is its dual role in primary (GSH degradation) and secondary (processing GSH conjugates) sulfur metabolism [12]. In the biosynthetic pathways of glucosinolates and camalexin, GGP1 cleaves the γ‐Glu isopeptide bond of GSH conjugates or GS‐IAN by accommodating a large thiol‐linked substituent at P1′ Cys [11]. In our previous study, to examine the ability of GGP1 to process GSH and its conjugates, we modeled GSH in a predicted structure of GGP1 using AlphaFold2 [12]. In the present study, we predicted the complex structure with a model GSH conjugate substrate, S‐[(Z)‐phenylacetohydroximoyl]‐l‐glutathione (GS‐B, Fig. 1A), using AlphaFold3 (Fig. 6A) [37]. We previously suggested that R206 may form a salt bridge with the carboxylate group of P2′ Gly [12]. However, the R206 side chain was located ~16 Å away from the active site in all crystal structures (Fig. 6C). In the predicted structure (Fig. 6A), R206 is moved to the vicinity of the carboxylate of P2′ Gly (3.9 Å), suggesting that it is involved in substrate recognition. A surface representation of the predicted structure (Fig. 6B) shows that the phenyl group of GS‐B is accommodated in an open pocket. The pocket consists of C154, Q156, and Y212, and these residues may affect the substrate specificity for GSH conjugates. GGP1 has a higher affinity for GS‐B than for GSH [10, 12]. This can be explained by a possible π‐π stacking between Y212 and the phenyl group of GS‐B.
An AlphaFold3‐predicted structure of GGP1 complexed with GS‐B and superimposition with the crystal structures. (A) The active site interactions of predicted GGP1 structure (green) with GS‐B (purple). The distance between the R206 side chain and the carboxy group of P2’ Gly is shown as a black dotted line. (B) The molecular surface of the active site of the predicted GGP structure (green) with GS‐B (purple). The side chains of the catalytic triad and pocket‐forming residues for the substituents of GSH conjugates (C154, Q156, and Y212) are shown. (C) Comparison around R206 in the crystal and predicted structures. Ligand‐free (gray), Glu complex (protein in cyan and Glu in orange), H192N γ‐Glu intermediate (pink and orange), S–S inactive (blue), and AlphaFold3‐predicted GS‐B complex (green and purple) structures are shown. Distance between R206 and γ‐carboxyl in the Glu complex structure is shown. The image was prepared using Chimera.
Mutational analysis
To verify the substrate interactions predicted by the substrate modeling, a site‐directed mutational analysis of GGP1 was conducted. First, we optimized the activity assay condition. The K m values of recombinant GGP1 toward GSH and a GSH conjugate (GS‐B) were reported to be 5.0 mm and 37 μm, respectively [10, 12]. However, we found that under our previous assay condition for GSH (50 mm Tris–HCl buffer adjusted to pH 8.0), the pH in the reaction buffer decreased significantly as the GSH concentration increased (≥10 mm). In this study, we reexamined the buffer system to ensure that the pH did not change at high substrate concentrations. Moreover, in our previous study, we conducted the activity measurements at 37 °C [12], which is not the physiological temperature of the plant. The pH‐activity profile showed that the optimum pH of GGP1 was 7.5–8.0 at both 25 °C and 37 °C, while there was a difference in the profile (Fig. S2). This result suggested that the enzyme was partially unstable at 37 °C, especially at pH below 7.0. Therefore, we used 200 mm HEPES‐NaOH buffer (pH 7.5), and the initial velocity was measured at 25 °C toward various different substrate concentrations. Unexpectedly, the activity did not saturate at 15 mm GSH (Fig. 7). Although we could not determine the K m value of GGP1 toward GSH, the k cat/K m value calculated from the slope of the plot was 19.3 s^−1^ m ^−1^. Considering that the cytosolic GSH concentration in Arabidopsis is 1–5 mm [2], the K m value of GGP1 for GSH is far higher than its concentration in the cytosol.
S‐v plot of GGP1 activity toward GSH. The activity was measured in 200 mm HEPES‐NaOH (pH 7.5) and 200 mm NaCl at 25 °C. Error bars represent standard deviations (n = 3).
Single‐site substitution mutants (C100S, H192N, C154S, C154A, and R206A) were constructed for the analysis. All the mutants were expressed in E. coli and the activity of the purified enzyme was measured (Table 3). The activities of the two catalytic triad mutants, C100S and H192N, were not detected. The significant activity decrease of R206A indicates that R206 recognizes P2′ Gly and is essential for GSH degradation, whereas its side chain was flipped away from the active site in the crystal structures without Cys‐Gly (Fig. 6C). We hypothesize an induced‐fit motion of R206, in which the side chain moves toward P2′ Gly on the binding of the substrate as seen in the predicted structure. The importance of R206 is supported by the high degree of conservation among plant GGP1 homologs and GGP2‐GGP5 of A. thaliana (Fig. 5). C154 formed a disulfide bond with C100 in the S–S inactive structure (Fig. 3D). C154A retained 92.5% of WT enzyme activity, whereas the activity of C154S was reduced to 36.5% (Table 3). Because the C154 side chain is located near the P1′ Cys side chain in the predicted structure (Fig. 6A), the residue substitution at this position may have affected the substrate binding ability.
Table 3: Activity and melting temperature of the wild‐type (WT) GGP1 and mutants. ND, not detected (<1%). Activity toward 5 mm GSH at pH 7.5 and 25 °C was measured. All activity measurement experiments were performed in triplicate (n = 3) and expressed as mean ± standard error. The melting temperatures were measured by CD at 222 nm.
In the circular dichroism (CD) spectra of the mutants, C100S and H192N retained a negative peak at 210–230 nm, whereas the peaks of the other mutants decreased (Fig. 8). For the negative signal around 220 nm, which indicated helix content, the mutants showed a decrease in C154A, C154S, and R206A. Secondary structure content analysis showed that the mutants had decreased helical content and increased turn content (Table 4). The decrease in helix content was particularly noticeable in C154S and R206A, suggesting that C154 and R206 also contributed to helical structure stabilization of the protein. To assess stabilities of each mutant, we measured decreases of CD signal at 222 nm with increasing temperature, and calculated melting temperature (T m) (Fig. S3). WT showed T m = 49.2 °C, and H192N and R206A mutants unexpectedly exhibited increased heat stability (Table 3). Three cysteine mutants (C100S, C154S, and C154A) showed reduced stability. Among them, C154S mutant showed largest destabilization, the T m changes from WT being 1.9 °C. This result implicates that the larger activity decrease of C154S than C154A was at least partially due to protein instability.
Circular dichroism (CD) spectra of wild‐type (WT) GGP1 and mutants. The spectra were measured for protein solutions (0.04 mg·mL−1) in 10 mm NaHPO4 (pH 7.5) using a 1‐cm path length cuvette at room temperature. The averages of the four scans are presented as the molar ellipticity per residue.
Conclusions
In this study, we determined the high‐resolution crystal structures of GGP1 from A. thaliana, which is classified in the C26 γ‐glutamyl hydrolase family of the MEROPS database. To the best of our knowledge, this is the first three‐dimensional structure of a plant enzyme belonging to the C26 family to be determined. The Glu complex and γ‐Glu intermediate structures indicate that GGP1 shares a common catalytic mechanism for cysteine hydrolases that involves nucleophilic attack by the thiol side chain of cysteine, formation of an oxyanion hole by the main chain amides, and the presence of nucleophilic water near the catalytic histidine. Moreover, the elaborate P1 γ‐Glu recognition at the S1 subsite, which may be common in γ‐glutamyl hydrolase family enzymes and glutaminases, was elucidated. The predicted structure complexed with GS‐B illustrated the interactions with the P1′–P2′ Cys conjugate‐Gly moiety. The S–S inactive state structure, in which the catalytic C100 residue and C154 formed a disulfide bond, showed a possible oxidated inactive state in Arabidopsis species because C154 is not conserved in other plants. The structural basis for the dual physiological roles of GGP1 in primary and secondary metabolism was elucidated from the predicted structure with GS‐B in the active site. The thiol side chain of P1′ Cys was shown to be in an open pocket that accommodates a large substituent group, thereby explaining its activity toward GSH and its conjugates. Based on the structural insights revealed in this study, the substrate specificity and catalytic efficiency of GGP1 can be modified for a particular GSH conjugate by mutating the pocket‐forming residues. Because specific glucosinolates have been reported to exhibit chemopreventive actions that decrease cancer risk [38, 39], engineering GGP1 may lead to the establishment of a production platform for glucosinolates in a controlled manner.
Materials and methods
Recombinant protein production and purification
From the previously constructed expression plasmid [12], an unnecessary 18‐amino acid sequence (LESTSLYKKAGSAAAPFT) derived from pDEST17 and the Gateway entry vector pENTR/D‐TOPO (Invitrogen, Carlsbad, CA, USA) was removed using PCR with the KOD One PCR Master Mix (TOYOBO Co., Ltd., Osaka, Japan) and the primers listed in Table S1. The resulting plasmid contained a short His_6_‐tag (MSYYHHHHHH) at the N terminus of GGP1. E. coli BL21 (DE3) cells harboring the expression plasmid were cultured at 37 °C in lysogeny broth medium (1% tryptone, 0.5% yeast extract, and 1% NaCl) with 34 μg·mL^−1^ ampicillin until the OD_600_ reached 0.5. Subsequently, overexpression was induced by adding 0.1 mm isopropyl β‐d‐thiogalactopyranoside and continued for 20 h at 16 °C. The cells were harvested through centrifugation at 8000 ** g ** and resuspended in 50 mm HEPES‐NaOH (pH 7.5) and 200 mm NaCl. The cells were disrupted using an ultrasonic homogenizer (Branson Sonifier250D; Branson Ultrasonics Division of Emerson Japan, Kanagawa, Japan), and the supernatant was collected through centrifugation at 15 000 ** g ** and filtered using Minisart Hydrophilic 0.45‐μm filters (Sartorius Stedim Biotech, Göttingen, Germany). The crude enzyme solution was purified using Ni‐affinity chromatography (cOmplete His‐Tag; Roche Diagnostic GmbH, Mannheim, Germany). Solutions of 5 and 250 mm imidazole in 50 mm HEPES‐NaOH (pH 7.5) and 200 mm NaCl were used as wash and elution buffers, respectively. The eluted fraction was concentrated using an ultrafiltration centrifugal membrane unit (Amicon Ultra 15 MWCO 10 kDa; Millipore, Billerica, MA, USA) and loaded onto a gel filtration chromatography column (HiLoad 16/60 Superdex 200 pg; Cytiva, Marlborough, MA, USA) equilibrated with 20 mm HEPES‐NaOH (pH 7.5) and 200 mm NaCl. Protein concentrations were determined using the bicinchoninic acid (BCA) Protein Assay Kit (Thermo Fisher Scientific, Waltham, MA, USA) or absorbance at 280 nm. The theoretical extinction coefficient was calculated from the amino acid sequence using the ProtParam tool in ExPASy (https://web.expasy.org/protparam/). The purity of the enzymes was confirmed at each purification step using SDS‐PAGE.
Crystallography and structure prediction
Solutions of GGP1 were desalted before crystallization. Crystals were grown at 4 °C using the sitting drop vapor diffusion method. A protein solution (0.5–0.6 μL) containing 4–7 mg·mL^−1^ protein was mixed with a reservoir solution (0.4–0.5 μL) containing 0.1 m Na‐acetate (pH 4.0–4.75), 1.20–1.32 m K_2_HPO_4_, and 0.80–0.88 m NaH_2_PO_4_. Crystals of the Glu complex were prepared through co‐crystallization with 7.5 mm Glu. For the γ‐Glu intermediate, 5.0 mm GSH was co‐crystallized. To prepare S–S inactive crystals, GGP1 crystals were soaked in 100 mm Cys‐Gly dipeptide for 5 s. The crystals were cryoprotected in a reservoir solution supplemented with 20% glycerol or 20% PEG200 and flash‐frozen in liquid nitrogen. X‐ray diffraction data were collected at the beamlines of the Photon Factory of the High Energy Accelerator Research Organization (KEK, Tsukuba, Japan), the Swiss Light Source (SLS) of the Paul Scherrer Institut (PSI, Villigen, Switzerland), and SPring‐8 (Hyogo, Japan). For the diffraction datasets collected at SPring‐8, initial data processing was performed using KAMO [40] based on XDS [41] and DIALS [42]. Only XDS [41] was used for the other datasets. Data reduction to proper resolution was performed using the Aimless software [43]. The initial structures were solved using the molecular replacement method implemented in PHASER [44] using a predicted structure provided by ColabFold [45]. Manual model building and refinement were performed using Coot [46] and Refmac5 [47]. Polder maps were prepared using the PHENIX software [48]. Molecular graphic images were prepared using Chimera [49]. Amino acid sequence alignment was performed using Clustal Omega [50] and ESPript [51]. Protein structure prediction was performed using the local AlphaFold3 software [37] installed on a PC with the source code (https://github.com/google‐deepmind/alphafold3) and the model parameters distributed by Google DeepMind.
Construction of mutants and enzyme assay
The KOD One PCR Master Mix (TOYOBO Co., Ltd.) was used for site‐directed mutagenesis. The primers used for each mutant are listed in Table S1. The reaction mixture for activity measurements (200 μL) contained 2.5 μg of the wild‐type or mutant GGP1, 0.5–12.5 mm GSH, 200 mm HEPES‐NaOH (pH 7.5), and 200 mm NaCl. To measure the activity of the mutants, 5 mm of GSH was used. The reaction mixtures were incubated at 25 °C and the reaction was terminated by heating at 95 °C. The solutions were centrifuged at 15 000 ** g ** for 15 min at 4 °C. The released Cys‐Gly in the supernatants was quantified using HPLC after derivatization with monobromobimane, as previously described [12].
Measurement of CD spectra and melting curve
CD spectra were measured using a J‐820 spectropolarimeter (JASCO Corp., Hachioji, Tokyo, Japan) equipped with a temperature controller in a 1‐cm path length cuvette. Protein solutions (0.04 mg·mL^−1^) were replaced with 10 mm NaHPO_4_ (pH 7.5) using a PD‐10 column (Sigma‐Aldrich, St. Louis, MI, USA). The averages of the four scans are presented as the molar ellipticity per residue. The secondary structure was estimated using the BeStSel server [52]. For melting curve measurements, CD at 222 nm was recorded in a 5‐mm path length cuvette. Protein samples (0.04 mg·mL^−1^ in 10 mm HEPES‐NaOH (pH 7.5) and 200 mm NaCl) were heated from 25 °C to 70 °C at a rate of 1 °C/min. Melting temperature (T m) was calculated from the thermal denaturation curves using the R software.
Conflict of interest
The authors declare no conflict of interest.
Author contributions
NOO and SF conceived and supervised the study. CY, TK, AM, and SF planned the experiments. KS, TI, and HS performed the experiments. KS performed the protein crystallography, biochemical experiments, and CD measurements. TI and HS performed the biochemical experiments. KS and SF wrote the manuscript. All authors reviewed the final version of the manuscript.
Supporting information
Fig. S1. Electron density map of C100, C154, and flanking residues in the S–S inactive structure. Fig. S2. pH‐activity profile of GGP1. Fig. S3. Melting curves of wild‐type GGP1 and mutants measured by CD at 222 nm. Table S1. Primers used in this study.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Foyer CH & Noctor G (2011) Ascorbate and glutathione: the heart of the redox hub. Plant Physiol 155, 2–18.21205630 10.1104/pp.110.167569 PMC 3075780 · doi ↗ · pubmed ↗
- 2Noctor G , Queval G , Mhamdi A , Chaouch S & Foyer CH (2011) Glutathione. Arabidopsis Book 2011, 1–32.10.1199/tab.0142 PMC 326723922303267 · doi ↗ · pubmed ↗
- 3Noctor G , Mhamdi A , Chaouch S , Han Y , Neukermans J , Marquez‐Garcia B , Queval G & Foyer CH (2012) Glutathione in plants: an integrated overview. Plant Cell Environ 35, 454–484.21777251 10.1111/j.1365-3040.2011.02400.x · doi ↗ · pubmed ↗
- 4Foyer CH & Noctor G (2005) Redox homeostasis and antioxidant signaling: A metabolic interface between stress perception and physiological responses. Plant Cell 17, 1866–1875.15987996 10.1105/tpc.105.033589 PMC 1167537 · doi ↗ · pubmed ↗
- 5Smith IK , Vierheller TL & Thorne CA (1989) Properties and functions of glutathione reductase in plants. Physiol Plant 77, 449–456.
- 6Cassier‐Chauvat C , Marceau F , Farci S , Ouchane S & Chauvat F (2023) The glutathione system: a journey from cyanobacteria to higher eukaryotes. Antioxidants 12, 1199.37371929 10.3390/antiox 12061199 PMC 10295655 · doi ↗ · pubmed ↗
- 7Deshpande AA , Bhatia M , Laxman S & Bachhawat AK (2017) Thiol trapping and metabolic redistribution of sulfur metabolites enable cells to overcome cysteine overload. Microbial Cell 4, 112–126.28435838 10.15698/mic 2017.04.567PMC 5376351 · doi ↗ · pubmed ↗
- 8Ohkama‐Ohtsu N & Wasaki J (2010) Recent progress in plant nutrition research: cross‐talk between nutrients, plant physiology and soil microorganisms. Plant Cell Physiol 51, 1255–1264.20624893 10.1093/pcp/pcq 095 · doi ↗ · pubmed ↗