Structural Elucidation of Agrochemicals and Related Derivatives Using Infrared Ion Spectroscopy

A Machine-Learned "Chemical Intuition" to Overcome Spectroscopic Data Scarcity

Machine learning models for predicting IR spectra of molecular ions (infrared ion spectroscopy, IRIS) have yet to be reported owing to the relatively sparse experimental data sets available. To overcome this limitation, we employ the Graphormer-IR model for neutral molecules as a knowledgeable starting point and then employ transfer learning to refine the model to predict the spectra of gaseous ions. A library of 10,336 computed spectra and a small data set of 312 experimental IRIS spectra is used for model fine-tuning. Nonspecific global graph encodings that describe the molecular charge state (i.e., (de)protonation, sodiation), combined with an additional transfer learning step that considers computed spectra for ions, improved model performance. The resulting Graphormer-IRIS model yields spectra that are 21% more accurate than those produced by commonly employed DFT quantum chemical models, while capturing subtle phenomena such as spectral red-shifts due to sodiation. Dimensionality reduction of model embeddings demonstrates derived "chemical intuition" of functional groups, trends in molecular electron density, and the location of charge sites. Our approach will enable fast IRIS predictions for determining the structures of unknown small molecule analytes (e.g., metabolites, lipids) present in biological samples.

Structural Elucidation of Agrochemical Metabolic Transformation Products Based on Infrared Ion Spectroscopy to Improve In Silico Toxicity Assessment

Toxicological assessments of newly developed agrochemical agents consider chemical modifications and their metabolic and biotransformation products. To carry out an in silico hazard assessment, understanding the type of chemical modification and its location on the original compound can greatly enhance the reliability of the evaluation. Here, we present and apply a method based on liquid chromatography-mass spectrometry (LC-MS) enhanced with infrared ion spectroscopy (IRIS) to better delineate the molecular structures of transformation products before in silico toxicology evaluation. IRIS facilitates the recording of IR spectra directly in the mass spectrometer for features selected by retention time and mass-to-charge ratio. By utilizing quantum-chemically predicted IR spectra for candidate molecular structures, one can either derive the actual structure or significantly reduce the number of (isomeric) candidate structures. This approach can assist in making informed decisions. We apply this method to a plant growth stimulant, digeraniol sinapoyl malate (DGSM), that is currently under development. Incubation of the compound in Caco-2 and HepaRG cell lines in multiwell plates and analysis by LC-MS reveals oxidation, glucuronidation, and sulfonation metabolic products, whose structures were elucidated by IRIS and used as input for an in silico toxicology assessment. The toxicity of isomeric metabolites predicted by in silico tools was also assessed, which revealed that assigning the right metabolite structure is an important step in the overall toxicity assessment of the agrochemical. We believe this identification approach can be advantageous when specific isomers are significantly more hazardous than others and can help better understand metabolic pathways.

An expeditive and green chemo-enzymatic route to diester sinapoyl-l-malate analogues: sustainable bioinspired and biosourced UV filters and molecular heaters

Sinapoyl malate, naturally present in plants, has proved to be an exceptional UV filter and molecular heater for plants. Although there are nowadays industrially relevant sustainable synthetic routes to sinapoyl malate, its incorporation into certain cosmetic formulations, as well as its adsorption on plant leaves, is limited by its hydrophilicity. To overcome these obstacles, it is important to find a way to effectively control the hydrophilic-lipophilic balance of sinapoyl malate to make it readily compatible with the cosmetic formulations and stick on the waxy cuticle of leaves. To this end, herein, we describe a highly regioselective chemo-enzymatic synthesis of sinapoyl malate analogues possessing fatty aliphatic chains of variable length, enabling the lipophilicity of the compounds to be modulated. The potential toxicity (i.e., mutagenicity, carcinogenicity, endocrine disruption, acute and repeated-dose toxicity), bioaccumulation, persistence and biodegradability potential of these new analogues were evaluated in silico, along with the study of their transient absorption spectroscopy, their photostability as well as their photodegradation products.

Identification of organic micro-pollutants in surface water using MS-based infrared ion spectroscopy

Comprehensive monitoring of organic micro-pollutants (OMPs) in drinking water sources relies on non-target screening (NTS) using liquid-chromatography and high-resolution mass spectrometry (LC-HRMS). Identification of OMPs is typically based on accurate mass and tandem mass spectrometry (MS/MS) data by matching against entries in compound databases and MS/MS spectral libraries. MS/MS spectra are, however, not always diagnostic for the full molecular structure and, moreover, emerging OMPs or OMP transformation products may not be present in libraries. Here we demonstrate how infrared ion spectroscopy (IRIS), an emerging MS-based method for structural elucidation, can aid in the identification of OMPs. IRIS measures the IR spectrum of an m/z-isolated ion in a mass spectrometer, providing an orthogonal diagnostic for molecular identification. Here, we demonstrate the workflow for identification of OMPs in river water and show how quantum-chemically predicted IR spectra can be used to screen potential candidates and suggest structural assignments. A crucial step herein is to define a set of candidate structures, presumably including the actual OMP, for which we present several strategies based on domain knowledge, the IR spectrum and MS/MS spectrum.

An In Silico Infrared Spectral Library of Molecular Ions for Metabolite Identification

Infrared ion spectroscopy (IRIS) continues to see increasing use as an analytical tool for small-molecule identification in conjunction with mass spectrometry (MS). The IR spectrum of an m/z selected population of ions constitutes a unique fingerprint that is specific to the molecular structure. However, direct translation of an IR spectrum to a molecular structure remains challenging, as reference libraries of IR spectra of molecular ions largely do not exist. Quantum-chemically computed spectra can reliably be used as reference, but the challenge of selecting the candidate structures remains. Here, we introduce an in silico library of vibrational spectra of common MS adducts of over 4500 compounds found in the human metabolome database. In total, the library currently contains more than 75,000 spectra computed at the DFT level that can be queried with an experimental IR spectrum. Moreover, we introduce a database of 189 experimental IRIS spectra, which is employed to validate the automated spectral matching routines. This demonstrates that 75% of the metabolites in the experimental data set are correctly identified, based solely on their exact m/z and IRIS spectrum. Additionally, we demonstrate an approach for specifically identifying substructures by performing a search without m/z constraints to find structural analogues. Such an unsupervised search paves the way toward the de novo identification of unknowns that are absent in spectral libraries. We apply the in silico spectral library to identify an unknown in a plasma sample as 3-hydroxyhexanoic acid, highlighting the potential of the method.

Targeted Small-Molecule Identification Using Heartcutting Liquid Chromatography-Infrared Ion Spectroscopy

Infrared ion spectroscopy (IRIS) can be used to identify molecular structures detected in mass spectrometry (MS) experiments and has potential applications in a wide range of analytical fields. However, MS-based approaches are often combined with orthogonal separation techniques, in many cases liquid chromatography (LC). The direct coupling of LC and IRIS is challenging due to the mismatching timescales of the two technologies: an IRIS experiment typically takes several minutes, whereas an LC fraction typically elutes in several seconds. To resolve this discrepancy, we present a heartcutting LC-IRIS approach using a setup consisting of two switching valves and two sample loops as an alternative to direct online LC-IRIS coupling. We show that this automated setup enables us to record multiple IR spectra for two LC-features from a single injection without degrading the LC-separation performance. We demonstrate the setup for application in drug metabolism research by recording six m/z-selective IR spectra for two drug metabolites from a single 2 μL sample of cell incubation extract. Additionally, we measure the IR spectra of two closely eluting diastereomeric biomarkers for the inborn error of metabolism pyridoxine-dependent epilepsy (PDE-ALDH7A1), which shows that the heartcutting LC-IRIS setup has good sensitivity (requiring ∼μL injections of ∼μM samples) and that the separation between closely eluting isomers is maintained. We envision applications in a range of research fields, where the identification of molecular structures detected by LC-MS is required.

Characterization of Solar Radiation-Induced Degradation Products of the Plant Sunscreen Sinapoyl Malate

Agricultural activities at lower temperatures lead to lower yields due to reduced plant growth. Applying photomolecular heater agrochemicals could boost yields under these conditions, but UV-induced degradation of these compounds needs to be assessed. In this study, we employ liquid chromatography-mass spectrometry (LC-MS) coupled with infrared ion spectroscopy (IRIS) to detect and identify the degradation products generated upon simulated solar irradiation of sinapoyl malate, a proposed photomolecular heater/UV filter compound. All major irradiation-induced degradation products are identified in terms of their full molecular structure by comparing the IRIS spectra obtained after LC fractionation and mass isolation with reference IR spectra obtained from quantum-chemical calculations. In cases where physical standards are available, a direct experimental-to-experimental comparison is possible for definitive structure identification. We find that the major degradation products originate from trans-to-cis isomerization, ester cleavage, and esterification reactions of sinapoyl malate. Preliminary in silico toxicity investigations using the VEGAHUB platform suggest no significant concerns for these degradation products' human and environmental safety. The identification workflow presented here can analogously be applied to break down products from other agrochemical compounds. As the method records IR spectra with the sensitivity of LC-MS, application to agricultural samples, e.g., from field trials, is foreseen.