(4S)-4-hydroxy-L-argininium

Introduction

Recently, we discovered the biosynthetic pathway of guanitoxin, a potent freshwater neurotoxin from the cyanobacterium Sphaerospermopsis torques-reginae ITEP-024 which possesses a cyclic guanidine moiety in its highly polar structure (B). Previously named anatoxin-a(s), guanitoxin exerts its neurotoxicity through the irreversible inhibition of acetylcholinesterase within the peripheral nervous system. If consumed via contaminated water, guanitoxin can lead to an overexcitation of cholinergic neurons and downstream muscle spasms, respiratory arrest, and potentially death. Within the nine-enzyme biosynthetic pathway that transforms l-arginine into the mature neurotoxin, the pyridoxal-5′-phosphate (PLP)-dependent enzyme GntC constructs the key cyclic guanidine moiety upon which the anticholinesterase O-methylphosphate pharmacophore is assembled. Specifically, GntC performs a cyclodehydration of substrate (S)-γ-hydroxy-l-arginine (1) to l-enduracididine (2), cleanly inverting the stereochemistry at the γ-position. This reaction is highly diastereoselective under in vitro enzyme assay conditions following 1-fluoro-2,4-dinitrophenyl-5-l-alanine amide (l-FDAA, Marfey’s reagent) derivatization and ultra performance liquid chromatography mass spectrometry (UPLC-MS) analysis. In our initial characterization of GntC, we failed to observe any enzymatic production of allo-l-enduracididine diastereomer (2′) nor did we observe any conversion of substrate epimer (R)-γ-hydroxy-l-arginine (1′) within our limits of detection. While the GntC reaction occurs relatively early in the guanitoxin biogenesis, the biochemical validation was foundational for establishing the biosynthetic pathway. Both 1 and 2 were previously isolated from guanitoxin-producing Anabaena flos-aquae cyanobacteria, and the incorporation of backbone-deuterated 1 into the mature toxin was supported via in vivo feeding experiments., Although these stable isotope experiments directly implicated 1 in guanitoxin biogenesis, an intriguing reintroduction of hydrogen atoms at the former β-position implied that additional deprotonation/reprotonation chemistry was involved in toxin production (). Following full pathway characterization, one of these β-position reprotonations could be justified using the retro-aldol (GntG) and transamination (GntE) reactions identified in this study. However, we propose that a mechanistic interrogation into GntC-mediated cyclization of 1 could help further rationalize this in vivo result.

GntC PLP Dependency

To fuel our investigation, substrate (1, 1′) and product (2, 2′) diastereomers were synthesized following previously established synthetic protocols.,Sphaerospermopsis torques-reginae ITEP-024 GntC was heterologously expressed as an Escherichia coli codon-optimized N-terminal His6-tagged construct and purified to near homogeneity using Ni-NTA affinity chromatography.In vitro GntC assays were setup in potassium phosphate (KPi) buffer, amine-containing components were derivatized with l-FDAA, and reaction mixtures were analyzed by UPLC-MS as previously reported. Because GntC co-purifies with its requisite PLP cofactor, we had yet to firmly confirm its dependency in 2 production. Treatment of purified GntC with PLP inhibitor hydroxylamine followed by buffer exchange abolished any catalytic activity with 1 in the apo-GntC form (). Reintroduction of two molar equivalents of PLP to the apoenzyme restored 2 production, validating the cofactor dependence of the intramolecular cyclase GntC.

GntC In Vitro Assay Optimization

We initially investigated a range of pHs in KPi buffer to identify ideal conditions for 2 production. Under a 20:1 substrate/enzyme ratio, relative product formation was compared to an l-FDAA-derivatized glycine standard to correct for any derivatization discrepancies, and we determined that GntC behaved optimally at pH 8.0 (). Previous reports identified that divalent metal cations can aid PLP-dependent enzyme catalysis through the promotion of electron displacement, and we wanted to determine if a similar observation occurred with GntC. Counterintuitively, 2 production was not enhanced in the presence of any 1 mM divalent metal salts; instead, incubation with chelator EDTA showed comparable activity to wild-type GntC. Most other ionic additives (Mg2+, Ca2+, Mn2+, Fe2+, Zn2+, Co2+) showed similar or reduced in vitro2 formation, with the addition of Cu2+ completely abolishing production (). We finally investigated the impact of the His6 tag on in vitro2 production, given the significance of the N-terminus on homodimerization, shaping the active site pocket and establishing catalysis in other PLP-dependent enzymes., We recloned GntC as a C-terminal His6 construct and found a slight reduction in catalytic activity compared to the original N-terminal His6 construct (). In spite of additional optimization efforts, in vitro GntC assays were done under identical reaction conditions to those previously described. Following the development of an l-FDAA-derivatized 2 standard curve (), we determined the steady-state kinetic parameters of GntC toward substrate 1 (kcat = 0.0061 ± 0.008 min–1, KM = 150 ± 80 μM ()).