HS-SPME/GC–MS and chemometric approach for the study of volatile profile in X-ray irradiated mozzarella cheese

HS-SPME/GC–MS and chemometric approach for the study of volatile profile in X-ray irradiated mozzarella cheese Rosalia Zianni1, Annalisa Mentana2, Michele Tomaiuolo2, Maria Campaniello2, Marco Iammarino2, Diego Centonze3, Carmen Palermo1 1Università di Foggia, Dipartimento di Medicina Clinica e Sperimentale, Via Napoli 25 - 71122 Foggia, Italy 2Istituto Zooprofilattico Sperimentale della Puglia e della Basilicata, Laboratorio Nazionale di Riferimento per il trattamento degli alimenti e dei loro ingredienti con radiazioni ionizzanti, Via Manfredonia, 20 - 71121 Foggia, Italy 3 Università di Foggia, Dipartimento di Scienze Mediche e Chirurgiche, Via Napoli 25 - 71122 Foggia, Italy Summary: The optimization of HS-SPME for the analysis of the X-ray irradiated mozzarella cheese volatile profile using a central composite design and a response surface methodology, was performed. The optimised conditions were applied to irradiated samples at different dose levels and the variations have been evaluated by a chemometric approach. Keywords: HS-SPME/GC-MS, chemometric analysis, Food irradiation 1 Introduction Food irradiation is a process in which food products are exposed to ionising radiation, such as X-rays, to destroy and inactivate pathogenic and spoilage microorganisms [1]. In this study, X-ray irradiation was applied to mozzarella cheese produced with cow milk and the modifications in the composition of volatile organic compounds (VOCs) have been investigated. Headspace solid-phase microextraction (HS-SPME) technique coupled with gas chromatography mass (GC-MS) was used to extract, isolate and enrich the volatile fraction from the sample matrix [2]. Design of experiments (DOEs) and Response Surface Methodology (RSM) were used for the optimization of the HS-SPME process. Five parameters have been chosen: type of fibre, X-ray irradiation dose (kGy), extraction temperature (°C), extraction time (min) and sample amount (g). Type of fibre and X-ray irradiation dose level were evaluated through a first screening step. A central Composite Experimental Design (CCD) was used to optimise the remaining three factors, thus selected as independent variables. The influence of key parameters was evaluated on the total area and total number of VOCs. The optimised HS-SPME conditions were used to analyse a representative number of non-irradiated and X-ray irradiated mozzarella samples at three dose levels, i.e., 1.0, 2.0 and 3.0 kGy. The collected data were elaborated by Principal Component Analysis (PCA), Linear Discriminant Analysis (LDA) and Partial Least Square-Discriminant Analysis (PLS-DA) to discriminate the variation of volatile profiles among non-irradiated and irradiated samples. 2 Experimental The X-ray radiation treatment was performed using a low-energy X-ray irradiator (RS-2400, Radsource Inc., Texas, USA). To maximise chemical information of irradiated mozzarella volatiles, the extraction performances of the four different SPME fibres (PDMS, DVB/CAR/PDMS CAR/PDMS and PDMS/DVB) were evaluated at the same analytical conditions. Then, a CCD of 17 runs was implemented, selecting three factors: sample amount, extraction time and temperature, assessing them at three different experimental levels. VOCs were analysed using a 6890N gas chromatograph (Little Falls, DE, USA) coupled with an Agilent 5975 mass selective detector, equipped with a Gerstel MPS auto-sampler (Gerstel, Baltimore, MD, USA). X-ray irradiated samples at three dose levels of 1.0, 2.0 and 3.0 kGy and the non-irradiated were analysed in the optimised HS-SPME conditions. Chemometric analyses including PCA, PLS-DA and LDA were performed using free software R version 4.1.1. These classifiers were chosen because they have been widely applied in similar contexts and they generally provided accurate results [3]. 3 Results The DVB/CAR/PDMS was the best fibre in terms of sensitivity for the analysis of X-ray irradiated samples and then it was used in the CCD analyses. The optimum HS-SPME conditions were estimated by means of the desirability function method, used for simultaneous optimization of the multiple responses. The surface responses, reported in Figure 1, shows as the optimised combination of extraction temperature, extraction time and sample amount, corresponded to the following values: 75 °C, 75 min and 5 g. A total of 11 classes of compounds, including alcohols, aldehydes, alkanes, alkenes, aromatic compounds, carboxylic acids, esters, ketones, methyl esters and oxygen and sulphur-containing compounds were identified. A PCA preliminary data exploration was performed and showed how irradiated and non-irradiated samples were grouped in the PC subspace, also based on dose levels. The score dispersions at the three doses were influenced by the specific classes, i.e., from alcohols at 1.0 kGy, alkanes and alkenes at 2.0 kGy, ketones and aldehydes at 3.0 kGy. The Volcano plot and Variable Importance in Projection (VIP) were used for identifying and assessing the discriminant volatiles from the volatolomic dataset [4]. The selected VOCs were used for PLS-DA and LDA. Both discriminant approaches were applied in double cross validation scheme [4] and the results highlighted the strong discriminating capacity of the PLS-DA and LDA algorithms in distinguishing irradiated samples from non-irradiated ones. 4 Conclusions The effect of HS-SPME parameters in terms of fibre type, extraction temperature, extraction time and sample amount were optimised to obtain the maximum total area and number of VOCs of X-ray irradiated mozzarella samples. The best parameters of HS-SPME were used to investigate non-irradiated samples and the irradiated ones at three different dose levels. Some classes of compounds as hydrocarbons, oxygen-containing compounds, alcohols, aldehydes and ketones increased in irradiated samples, due to possible oxidation mechanisms induced by the irradiation treatment, as already reported for other matrices [5]. The classification by two discriminant approaches, LDA and PLS-DA, was effective in the indication of irradiation treatment and the dose levels too. Fig. 1. Responses surface plots for significant effect of extraction temperature (°C) and extraction time (min) on a) number of VOCs and of extraction temperature (°C) and sample amount (g) on b) total VOCs area References 1. O. B. Odueke, K.W. Farag R.N. Baines, S.A. Chadd, Food and Bioprocess Technology, 9 (2016), pp 751-767. 2. A.S. Bertuzzi, P.L.H. McSweeney, M.C. Rea, K.N. Kilcawley, Comprehensive Reviews in Food Science and Food Safety 17 (2018), pp 371-390. 3. F. Di Donato, A. Biancolillo, D. Mazzulli, L. Rossi, A. A. D’Archivio, Microchemical Journal 165, (2021), 106133. 4. E. Szymańska, E.Saccenti, A. K. Smilde, J. A. Westerhuis, Metabolomics, 8 (2012), pp 3-16. 5. D.U. Ahn, C. Jo, D.G. Olson, Meat Science, 54 (2000), pp 209-215. This work was supported by the Italian Ministry of Health who funded the Project code GR-2018-12367064.