2.6.2. LC-ESI-MS/MS Analysis

Two different procedures were followed, route 2.A or 2.B (Figure 1), in order to apply different workflows and instrumental platforms with the aim to validate the final results.

Method 2.A: One g of biscuits' powder was combined with 5 mL of Hexane, left shaking at high speed for 15 min at r.t., centrifuged at 4000 g for 20 min, and the supernatant discarded. This lipid removal procedure was repeated once again and samples were dried by evaporation using a flow of nitrogen. Four mL of extraction buffer [50 mM Trizma base, 2 M Urea, 1% (*w/v*) Octyl ß-D-glucopyranoside] were added to each defatted sample, which was shaken at high speed for 1 h at r.t. After centrifugation, 500 µL of supernatant (protein extracts) were reduced for 60 min, shaking at 1000 rpm at 60 ◦C with 5 mM (final concentration) Tris-(2-carboxyethyl)-phospine, incubated with 25 µL of 100 mM cysteine blocking reagent (Methyl methane-thiosulfonate) for 15 min at r.t., diluted with 425 µL of digestion buffer (100 mM Ammonium bicarbonate, 5 mM Calcium chloride), and digested with 20 µg Trypsin TPCK treated (SCIEX, Redwood City, CA, USA) at 37 ◦C for 16 h. The reaction was stopped with 30 µL of Formic acid (FA) and the digested samples were filtrated using a centrifugal filter unit with 10 kDa MWCO (Merck Millipore). The filtrates were further analysed by mass spectrometry. In order to obtain a calibration curve spanning the concentration range 5–50 part per million (ppm, defined as µg of allergenic ingredient per g of matrix) for quantitative analysis, biscuits fortified with allergen commodity (egg powder and skim milk powder, certified reference material BCR-685, Sigma-Aldrich, Milan, Italy) were prepared and processed as samples.

Digested samples were analyzed using a micro-LC-ESI-Triple Quadrupole (TQ) platform: a M3 MicroLC-TE System interfaced with a QTrap6500<sup>+</sup> mass spectrometer equipped with an IonDrive Turbo V Ion Source (TurboIonSpray probe with 50 µm i.d. electrode; Sciex). In this case, 10 µL of tryptic peptides (~40 µg) were injected onto a ChromXP C18 trap column cartridge (5 µm × 10 mm, 120 Å, 300 µm i.d.; Sciex) for pre-concentration and desalting at a flow rate of 50 µL/min, and subsequently separated using a HALO peptide-ES C18 column (300 µm × 150 mm, 160 Å, 2.7 µm; Advanced Materials Technology, Wilmington, DE, USA) maintained at 40 ◦C. Mobile phase A was H2O + 0.1% FA, and mobile phase B was 0.1% FA in acetonitrile (ACN). Peptides were separated by linear gradient of 2–40% mobile phase B over 11 min at a flow rate of 10 µL/min, followed by a 2 min rinse with 98% mobile phase B. The column was re-equilibrated at the initial conditions for 5.4 min. The QTrap mass spectrometer was operating in positive ESI High Masses Multiple Reaction Monitoring (MRM) mode (Unit Resolution on Q1 and Q2); the data were acquired using Analyst (version 1.6.3, Sciex). Source/Gas parameters were: 20 psi Curtain Gas; Medium Collision Gas; 5500 V IonSpray Voltage; 150 ◦C Temperature; 35 psi Ion Source Gas 1; 20 psi Ion Source Gas 2, and instrumental settings optimized for each individual milk and egg peptide marker are reported in Table S3. Analyses were performed on two different portions of Magretti biscuits used for the OFC and SPT (Table S2).

Method 2.B: A subset of samples (from cookies administered to child patients namely n. 8, 9, 11, 13–14, 16–21, 23, 25–28) underwent a different sample preparation workflow and a MRM method was built up on a different platform: an LX50 UHPLC pump provided with an autosampler and an ESI interface connected to a QSight 220 TQ mass spectrometer (PerkinElmer Inc., Waltham, MA, USA) as already published [20].

Instrumental settings optimized for each individual milk and egg peptide marker are reported in Table S4. All analyses were performed in triplicates.

Allergen-free and incurred biscuits at the highest level of 300 µg allergenic ingredient/g matrix were produced at laboratory scale according to the procedure described elsewhere [21]. Lower concentration levels covering the calibration range as previously indicated in route 2.A were produced from these two stock samples by appropriate dilution with blank matrix powder. The allergen-free biscuit digest was fortified with increasing amount of synthetic standard peptides (GenScript, Piscataway, NJ, USA) specific for milk and egg allergens (in the range 0.0125–0.2500 µg/mL) and calibration curves were prepared by plotting the signal of each candidate peptide against the inclusion level in the biscuit sample. All extracts were submitted to the workflow previously described before its injection (10 µL) in duplicate in the QSight equipment.

For quantification purposes, each synthetic peptide with peptide concentration (expressed as µg/mL) was first converted in protein molarity, assuming that full digestion of the protein took place and then a proper conversion factor was applied for the calculation taking into consideration the mass/volume ratio used for protein extraction.

Quantitative analysis on data obtained by QTrap mass spectrometer was performed by MultiQuant software (version 3.0.2, Sciex) applying MQ4 algorithm for peak integration (minimum Gaussian smooth width of 1 point) and data processing. Calibration curves were generated by plotting peak areas against allergen commodity concentrations, with 1/x fitting. In particular, calibration points were produced spanning one order of magnitude concentration range expressed as µg allergenic ingredient/g matrix.

Peak integration and data processing on QSight 220 spectrometer MS data was performed by using 3Q Simplicity software (version 1.4, Perkin Elmer) applying Moving Average algorithm for peak integration (minimum Gaussian smoothing of 5 point).

The reporting units were converted into total proteins of allergenic ingredient (µg/g) assuming 35.39% and 48.00% of total protein content for milk and egg, respectively, in accordance with what reported by USDA.
