**1. Introduction**

Fishery products are defined by Regulation (EC) N 853/2004 of the European Parliament and the Council as "all seawater or freshwater animals (except for live bivalve mollusks, live echinoderms, live tunicates and live marine gastropods, and all mammals, reptiles and frogs) whether wild or farmed and including all edible forms, parts and products of such animals" [1]. The description includes all fishes (Osteichthyes, bony fishes, and Chondrichthyes, cartilaginous fishes), shellfish, and clams.

The fishery products are classified in four commercial categories: Fresh (no manipulation), prepared (any operations that affect the anatomical wholeness of the animal), frozen, and processed (any operation that transform the product, such as smoking, marinating, salting). The freshness of

a fish product is the most important commercial quality factor for the consumer, because the safety is an essential prerequisite without which food cannot be placed on the market or further transformed. The Regulation (EC) N 2406/96 of the European Parliament and the Council defines four categories for fresh fish products (extra, a, b, and c); fishery products grouped in the last class must be judged as not suitable for human consumption and must be removed from the market [2].

Fishery products are perishable due to the action of many enzymes, both endogenous, located in the fish muscles, and exogenous, produced by intrinsic bacteria that are present and can contaminate in the products [3]. Consequently, the exponential growth of bacteria that triggers oxidative chemical reactions causes fading and opacification of the product, with appearance of *<sup>o</sup>*ff*-flavor*, that is an unpleased flavor caused by chemical lipid oxidation or non-protein nitrogen (NPN) degradation [3,4]. NPN in fishery products are distinguished in two structural categories: Volatile and nonvolatile compounds. Nonvolatile NPN compounds are mostly represented by heterocyclic metabolites while volatile ones own low molecular weight and are represented in fishery products mainly by ammonia, trimethylamine (TMA, C3H9N), and dimethylamine (DMA, C2H7N). TMA is an endogenous compound abundant in fishery products: It is a post-mortem product, deriving from trimethylamine-N-oxide (TMAO, C3H9NO) by an enzymatic activity [5–8].

TMAO is an amine oxide, less volatile, and less basic with respect to TMA, due to its oxidation. Under enzymatic activity, TMAO could generate several chemical compounds: DMA and formaldehyde (FA) from endogenous (aquatic environment) muscle bacteria (*Pseudomonas* and *Alteromonas*) activity, and TMA from exogenous (which accumulates in fish products after capture and is typical of the terrestrial environment) bacteria (*Salmonella*, *Vibrio*) activity [9]. Only a limited population of bacteria can cause deterioration of fishery products: They are called specific spoiling micro-organism (SSO) and are mostly Gram-negative micro-organisms. After fishing, SSO can contaminate fishery products on the surface and grow even at low temperatures [10–13]. DMA, FA, and TMA are products of enzymatic degradation by intrinsic bacteria together with biogenic amines (histamine, tyramine, phenylethylamine, putrescine, and cadaverine), thiols and hydrogen sulfide (H2S). Moreover, in mollusks such as squid, the SSO enzymes can produce hypoxanthine and acetic acids salts, that contribute to the appearance of *o*ff*-flavor* [14,15].

The organo-nitrogen compounds derived from degradation of NPN and protein are quantifiable as a total volatile basic nitrogen TVB-N; for fresh fish products the TVB-N amount should be minor of 10 mg/kg, for aged ones the quantity ranges normally between 300–350 mg/kg (Regulation (EC) N 853/2004) [16–18]. Due to the unhealthiness, the *o*ff*-flavor* appearance, and color changes, aged fishery products must be retired from the market. However, some illicit treatments on these products might simulate a "false" freshness and one of these treatments is the use of hydrogen peroxide (H2O2). The illicit treatment with 0.5%–0.8% hydrogen peroxide water solutions is known and was reported in the literature [19]. Residues of hydrogen peroxide in food may be of toxicology concern.

Hydrogen peroxide is both an oxidant in aqueous solution with acidic pH and a reductant in alkaline water solution. When it is used as an illicit treatment for fishery products, the oxidizing properties are exploited. H2O2 can indeed convert TMA, a degradation product, to TMAO, the amine oxide naturally present in living fishes. TMAO is odorless and it has oxidizing properties giving to the fishery substrate high redox potential. This grea<sup>t</sup> redox potential is typical of muscle tissue of fresh fishery products, and decreases rapidly when TMAO is reduced to TMA by enzymatic activity [20–22].

If the TMAO amount is increased by H2O2 treatment, the proteins are stabilized [23]. For example, the amount of mucins, the main glycoproteins of mucus, is reduced on the fish skin after H2O2 treatment because of chemical degradation. The decrease in mucins concentration reduces viscosity and slows down the appearance of *o*ff*-flavor* [24].

Finally, the illicit treatment with H2O2 can cause a whitening and "re-freshing" e ffect on fishery products due to its oxidative properties (peroxidation of double bonds present in chromophores) [25,26].

The aim of the present study was the development of mass spectrometry (MS) based methods to evaluate the concentration of di fferent compounds related to the illicit treatment of fish food based on the use of hydrogen peroxide (whitening and "re-freshing"). For this purpose, both liquid chromatography and gas chromatography hyphenated mass spectrometers were used. Mass spectrometry is recognized as one of the powerful and sensitive analytical techniques to identify, characterize, and quantify small molecules, such as amines and ethers. The use of this kind of MS based techniques is worthwhile in food analysis requiring complementary approaches to the detection of chemicals with di fferent physical–chemical properties such as in the present work [27–29].

Two analytical methods were developed: The first one was a direct LC-MS/MS method for the determination of various amines and trimethylamine-oxide (TMAO); the second one was an indirect SPME-GC-MS method for the determination of residues of H2O2 on di fferent fishery products matrices, by the hydroxylation reaction of anisole to 2-hydroxyanisole (guaiacol).

Another aim of this work was the application of the developed methods to investigate about H2O2 fish products treatment. The consequent alteration of the concentration ratio of TMAO and trimethylamine (TMA) which is a known fish product freshness parameter was evaluated.

#### **2. Materials and Methods**

All solvents and analytical standards of dimethylamine (DMA), trimethylamine (TMA), trimethylamine-N-oxide (TMAO), cadaverine (CAD), hydrogen peroxide solution, anisole, guaiacol, and p-xylene-d10 were purchased from Sigma Merck (Merck, Milan, MI, Italy). High-performance liquid chromatography (HPLC)-grade water was obtained from a MilliQ Academic water purification system (Millipore, Milan, Italy). Before use, solvents were filtered through a 0.45 μm filter and degassed for 10 min in an ultrasonic bath.

Fishery products samples were: Atlantic bonito (*Sarda sarda*) mackerel-like fish of the family Scombridae; European squid (*Loligo vulgaris*) belonging to the family Loliginidae. Atlantic bonito and European squid samples were purchased in a local market or caught fresh in Tyrrhenian Sea and brought in the laboratory no later than 5 h after fishing, in ice.

#### *2.1. UHPLC-Tandem Mass Analysis of NPN*

A Nexera (Shimadzu, Milan, MI, Italy) UHPLC (Ultra-high-performance liquid chromatography) coupled through an ESI source with a QTRAP-5500 tandem mass analyzer (Sciex, Milan, MI, Italy) was used to quantify the NPN in fishery samples.

The chromatographic separation was achieved using a RP C18 column (Kinetex EVO, 5 μm, 150 × 2.1 mm, Phenomenex, Castel Maggiore, BO, Italy) and heptafluoro-butanoic acid 10 mM in water (eluent A) and in methanol (eluent B). The gradient run started with 1% B, increased to 35% B in 8 min, up to 100% in 3 min, followed by reconditioning time. Flow rate and injection volume were 200 μL min−<sup>1</sup> and 5 μL, respectively.

The triple quadrupole was used in a MRM positive ion mode with the following source parameters: Curtain gas (arbitrary unit, arb), 25; spray voltage (V), 5500; gas1 (arb), 35; gas2 (arb), 40; capillary temperature (◦C), 400. Nitrogen was used as curtain gas, gas 1, and gas 2. The MRM transitions, potentials, and collision energies were listed in Table 1.

**Table 1.** MRM (multiple reaction monitoring) parameters and voltages for tandem mass analysis in a positive ion mode of dimethylamine (DMA), trimethylamine (TMA), trimethylamine-N-oxide (TMAO), and cadaverine (CAD). DP: Declustering potential; EP: Entrance potential; CE: Collision energy; CXP: Collision exit potential.


The developed analytical method was validated evaluating selectivity, linearity, accuracy and precision, and lower limit of quantitation following the FDA guidelines [30].

#### *2.2. Gas Chromatography–Mass Spectrometry Analysis of H2O2 Residue*

For the GC-MS analysis we used a Varian Saturn 3900 (Agilent, Milan, MI, Italy) system, equipped with a 1177 injector. The separation column was a Zebron ZB-624 30 m, i.d. 0.25 mm (Phenomenex, Castel Maggiore, BO, Italy) applying a temperature gradient from 40 to 240 ◦C in 16 min. Injector temperature was 240 ◦C; split mode was employed and helium gas flow was 1.2 mL min−1.

The mass spectrometry was a Varian Saturn 2100 T ion trap analyzer (Agilent, Milan, MI, Italy) equipped with an EI (electron ionization) source. The full mass acquisition range was from 40 to 500 *<sup>m</sup>*/*<sup>z</sup>*.

The developed methodology was based on the paper of Tanaka et al. [31] calibrating the modified procedure in order to measure H2O2 concentration in the range 0.05 to 1.00 μg/mL.

#### *2.3. Sample Preparation for UHPLC*/*GC-MS Analysis of NPN*

Samples of fish products muscles were weighted, minced, and extracted with pH 2.5, 0.1 M phosphate buffer: 40 mL of buffer were used for 8 g of fish sample; the suspensions were then centrifuged (2300 *g* for 10 min) and filtered (0.45 μm).

Some samples were treated with hydrogen peroxide to simulate the illicit treatment: After a total immersion of 8 g of fish products in H2O2 solution (0.8%) for 2 min, the liquid was removed and samples were rinsed with fresh water. Then, treated fishery products samples were extracted as just described.

For UHPLC-MS, 1 mL of extracted solution was diluted using the starting eluent mixture, placed in a vial, and analyzed in a MRM mode using the triple quadrupole in a positive ion mode.

For GC-MS, 2 mL of extracted solution were placed in a vial for headspace solid-phase microextraction (HS SPME) and added with 100 μL of potassium ferricyanide (K8Fe(CN)6) as catalyzer and 2 μL of anisole. The solution was heated at 60 ◦C in an oil bath for 1 h; the fiber for head space analysis was exposed for 30 min. A Supelco 75 μm Carboxen™-PDMS (polydimethylsiloxane) (Merck, Milan, MI, Italy) fiber was used. The extraction recovery of anisole was checked to be 80% by the use of p-xylene-d10 as an internal standard for 30 min fiber exposition.
