**1. Introduction**

Clear evidence of animal and human illness and death caused by fungal metabolites have been reported worldwide since the 1970s. The secondary metabolites of fungi growing on agricultural commodities, called mycotoxins, are still considered as a major health concern [1]. Analytical methods, therefore, have been developed and subsequently validated to determine mycotoxins in different food and feed samples [2]. These methods are used in the monitoring laboratories to screen and confirm the samples that are contaminated with mycotoxins. In the European Union (EU), the maximum levels (ML) of regulated mycotoxins in food are in force [3]. For some mycotoxins without any ML, the European Food Safety Authority (EFSA) has published opinions, often indicating the need of harmonized and suitable methods to make sound exposure assessments. According to the EFSA, cereals, vegetables, and oilseeds frequently contain *Alternaria* mycotoxins, which could cause mutagenic, genotoxic, fetotoxic, and teratogenic effects [4]. These agricultural commodities are mostly infected by *Alternaria* species (e.g., *Alternaria alternata*) that produce more than 70 secondary metabolites from which the five most important ones are tenuazonic acid (TEA), altenuene (ALT), alternariol (AOH), tentoxin (TEN), and alternariol monomethyl ether (AME) (Figure 1) [4]. Currently, these *Alternaria* toxins are not regulated because there has not been enough information available to establish ML (i.e., risk assessment). Additionally, the development of a standard method is imperative, which is not available yet [4]. Hence, a few control laboratories have analyzed them on a regular basis using LC-MS/MS technique, which is the most suitable technique according to the EFSA [4]. However, interlaboratory comparisons (ILC) have already been organized for these toxins in order to support the legislation [5,6].

**Figure 1.** Structure and physical-chemical properties of five toxins analyzed in this study.

The first ILC was a proficiency test (PT) for *Alternaria* toxins in tomato juice, organized by the Federal Institute for Risk Assessment (BfR, Berlin, Germany) in 2014 [5]. Afterwards, the European Commission Joint Research Centre, EU Reference Laboratory for Mycotoxins (JRC, Geel, Belgium), performed the method validation study (MVS) to find a candidate LC-MS/MS method as a possible basis for drafting a standard method for *Alternaria* toxins in 2015 [6]. The MVS was started with a pre-trial and included tomato juice as test samples, followed by the final trial with tomato juice, cereals and sunflower seed samples. The reproducibility of the candidate method did not fulfill the requirements of the European Committee for Standardization (CEN). Due to lack of isotopically labeled internal standards (ISTDs), the candidate method showed low interlaboratory precision of some compounds (i.e., TEA and AME) [6]. In 2018, the ISTDs for five *Alternarias* mentioned above became commercially available. Therefore, the MVS was repeated by utilizing the LC-MS/MS determination with isotope dilution (LC-ID-MS/MS) and the obtained results met the requirements [7]. This MVS showed that the application of isotope dilution is critically important for analyzing *Alternaria* toxins in food samples using LC-MS/MS method.

Furthermore, Liu and Rychlik [8] published the advantage of using isotopically labeled TEN and its derivates for quantification of toxins in various food samples. They reported the synthesis and application of TEN-d3, DH-TEN-d3 (dihydrotentoxin-d3), and isoTEN-d3 (isotentoxin-d3) for quantifying native toxins in cereal-, vegetable-, and fruit-based samples and included different types of oils as well [8]. In another research performed by Liu and Rychlik, the biosynthesis of 13C-labeled ISTDs for seven *Alternaria* toxins was described [9]. The application of AOH-13C14, ALT-13C15, and AME-13C15 in the future can further enhance the quantification of toxins in food since the 13C-labeled ISTDs have advantages over the deuterated ISTDs. Namely, there is no retention time difference between the native target compound and its corresponding 13C-labeled analogue. This enables the total compensation of matrix effect (ME) during LC-MS/MS analysis. Furthermore, there will not be substantial overlap between ISTD signals and the isotopic signals of analyte if the molecular mass of the isotopologue is more than 5 mass units. However, the authors reported the isotope effect between AME and AME-13C15 when acetonitrile/2-propanol mixture was used as the organic modifier in the eluent [9]. Therefore, only the deuterated AOH and AME are commercially available so far.

This paper describes the use of a LC-ID-MS/MS method for analyzing five toxins mentioned above in sunflower oil samples for the first time. Even though the high contamination of sunflower seeds with *Alternaria* toxins (TEA: LOQ—5400 μg/kg; AOH: LOQ—1200 μg/kg; TEN: LOQ - 880 μg/kg; AME: LOQ—440 μg/kg) have been recently reported worldwide [4,6,10–12], the existing methods (Table S1) involved mainly the vegetable-, cereal-, fruit-based, and oilseed samples [8–36]. The oil samples got less attention, so far only three studies have included the analysis of this matrix [8,11,12]. The reason for excluding this sample matrix could be the different sample manipulations needed for this lipophilic sample. Thus, this matrix is now in the focus of the current study. The aims of the work presented here were to: (i) set up LC-ID-MS/MS separation process for five *Alternaria* toxins without chemical derivatization; (ii) develop a sample preparation approach, which is suitable for sunflower oil; (iii) fine-tune the LC-ID-MS/MS method to achieve the quantification limit as low as possible; (iv) perform inhouse validation of the method to meet the requirements set by EU; and (v) apply the method for real samples and also naturally contaminated and spiked sunflower seed QC samples.
