MALDI-TOF-MS Platform for Integrated Proteomic and Peptidomic

Subscriber access provided by NEW YORK UNIV

Article

A MALDI-TOF-MS platform for integrated proteomic and peptidomic profiling of milk samples allows rapid detection of food adulterations Mauro Sassi, Simona Arena, and Andrea Scaloni J. Agric. Food Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jafc.5b02384 • Publication Date (Web): 22 Jun 2015 Downloaded from http://pubs.acs.org on June 24, 2015

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of Agricultural and Food Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 35

Journal of Agricultural and Food Chemistry

A MALDI-TOF-MS platform for integrated proteomic and peptidomic profiling of milk samples allows rapid detection of food adulterations

Mauro Sassi *, Simona Arena * and Andrea Scaloni

Proteomics & Mass Spectrometry Laboratory, ISPAAM, National Research Council, Naples, Italy

* These authors gave an equal contribution to this work

Keywords: MALDI-TOF-MS; protein profiling, peptide profiling, milk analysis, speciation, thermal treatment.

Corresponding author:

Andrea Scaloni Proteomics & Mass Spectrometry Laboratory ISPAAM, National Research Council via Argine 1085, Naples, I-80147, Italy Email: [email protected] Phone: +39 (081) 5966006 Fax: +39 (081) 5965291

1

ACS Paragon Plus Environment


Page 2 of 35

Abstract Adulteration of ovine, caprine and buffalo milks with more common bovine material occurs for economic reasons and seasonal availability. Frauds are also associated with the use of powdered milk instead of declared, fresh material. In this context, various analytical methods have been adapted to dairy science applications with the aim to evaluate adulteration of milk samples, although time-consuming, suitable only for speciation or thermal treatment analysis, or useful for a specific fraud type. An integrated MALDI-TOF-MS platform for the combined peptidomic and proteomic profiling of milk samples is here presented, which allows rapid detection of illegal adulterations due to the addition of either non-declared bovine material to water buffalo, goat and ovine milks, or of powdered bovine milk to the fresh counterpart. Peptide and protein markers of each animal milk were identified after direct analysis of a large number of diluted skimmed and/or enriched diluted skimmed filtrate samples. In parallel, markers of thermal treatment were characterized in different types of commercial milks. Principal components scores of ad hoc prepared species- or thermal treatment-associated adulterated milk samples were subjected to partial least squares regression, permitting a fast accurate estimate of the fraud extents in test samples either at protein and peptide level. With respect to previous reports on MALDI-TOF-MS protein profiling methodologies for milk speciation, this study extends this approach to the analysis of the thermal treatment and introduces and independent, complementary peptide profiling measurement, which integrates protein data with additional information on peptides, validating final results and ultimately broadening the method applicability.

2


Page 3 of 35


Introduction Raw milk is one of the main constituent of the human diet, being an essential source of nutrients for the newborns and an important food for adult individuals. It is also used to prepare a multitude of dairy products that, depending on the making dairy animal species and transformation procedures, are characterized by different organoleptic/nutritional properties and economic values 1,2. In this context, the exclusive use of a specific, pure raw milk is mandatory in the manufacturing of traditional high-grade European Protected Designation of Origin cheeses, such as the water buffalo Mozzarella, Feta, Roquefort, Idiazabal, Pecorino Romano and Fiore Sardo. However, it has been estimated that buffalo, goat and sheep milk only represent 11, 2 and 1.4% of the raw material produced worldwide and used for dairy productions, compared to the bovine counterpart (about 85%) 3. The adulteration of ovine, caprine and buffalo milk with common bovine material occurs often because of the higher prices of the first ones, and the fluctuation in their availability due to the corresponding seasonal shortages. Fraudulent activities are also associated with the addition of powdered/frozen material to fresh milk in the period of low lactation 4,5. Adulteration practices mentioned above determined an ultimate need of rapid, sensitive, and reliable methods to determine the nature of raw milk (and of the corresponding dairy products). To this purpose, various analytical techniques and deriving methods were used and/or adapted to dairy science applications with the aim to evaluate the nature of milk biomolecules. In this context, near-infrared (NIR), infrared (IR) and nuclear magnetic resonance (NMR) spectroscopic, chromatographic, electrophoretic, immunoenzymatic, sensoristic, LC-electrospray (ESI)-mass spectrometric (MS) and polymerase chain reaction-based assays were used for milk speciation and thermal treatment (TT) determination (see Suppl. Tab. S1 for application details and related literature). Above-mentioned approaches have the disadvantage of being sometimes time-consuming as result of different preparation steps or long instrument acquisition times, suitable only for a specific milk assay (speciation or TT analysis), useful for a specific adulteration type and/or eventually not adaptable for global screening of milk samples. For example, the EU reference methodology to detect bovine proteins in dairy products is based on gel isoelectric focusing of -caseins after plasminolysis (EC Regulation No. EC 273/2008) 6. However, this method suffers of interpretation difficulties due to overlapping of species-specific bands and may request integration with laborious immunoblotting steps 7. Recently, matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS)-based approaches have been developed for biological fluid profiling, showing their potential in the quali-quantitative

3



Page 4 of 35

analysis of milk proteins and peptides 8. MALDI-TOF-MS has the advantage of requiring small amounts of even heterogeneous samples, permitting their analysis in a very short time. In particular, detection of -lactalbumin, lactoglobulin and 2/3-casein isoforms during protein profiling (PR-P) of whey and skimmed milk samples has been used to reveal definite species-associated milk frauds

9-13

. Linear regression analysis of the resulting mass spectra

quantified non-declared milk material even in low amounts 13. Recently, MALDI-TOF-MS PR-P of skimmed samples combined with multivariate calculation techniques allowed an accurate estimate of the levels of adulteration in bovine-goat-sheep milk mixtures14. No similar applications were developed to assess fraudulent addition of powdered material to fresh milk. Direct serum or skimmed milk PR-P applications did not find a counterpart at peptide level. In fact, peptide profiling (PE-P) approaches were developed only when contemplating additional experimental steps consisting in the enzymatic digestion (with trypsin or plasmin) of the whole milk samples or of selected proteins preventively resolved by electrophoresis

4,5,15-20

. This experimental set up severely affected the time necessary for

entire analysis and highly complicated the nature of the resulting MALDI-TOF spectra, generally determining the need of additional chromatographic steps for an accurate quantification of the adulteration percentages. In this study, an integrated platform for the combined peptidomic and proteomic profiling of milk samples has been developed to rapidly reveal illegal adulterations due to the addition of either non-declared bovine material to water buffalo, goat and ovine milks, or of powdered bovine material to the fresh counterpart. To this purpose, specific peptide and protein markers of fresh bovine milk (BM), water buffalo milk (WM), ovine milk (OM) and goat milk (GM) were identified after direct analysis of skimmed samples. Markers of TT were characterized in fresh bovine milk (BM), pasteurized milk (PM), UHT milk (UM) and powdered milk (IM) samples. Principal components (PCs) scores of had hoc prepared species- or TT-associated adulterated milk samples were subjected to partial least squares (PLS) regression, allowing an accurate estimate of the fraud extents in test samples either at protein and peptide level.

Experimental Section Sample preparation Fresh raw BM, WM, OM and GM samples were collected from local farms after morning milking, stored at 4 °C, and assayed within 5 h. Bovine PM, UM and IM samples from different (20 in number) commercial brands were

4


Page 5 of 35


purchased from supermarkets and used before their expiration date (Suppl. Tab. S2). In all cases, 25 different samples were analyzed for each type of milk. For adulteration studies, binary mixtures of BM in WM, OM and GM counterparts, or IM in BM were prepared at 1, 2, 5, 10, 15, 20, 30, 40 and 50% v/v values. Three independent samples were prepared for each specific binary mixture. All milk samples or their binary mixtures were defatted by centrifugation at 3000 rpm, for 30 min, at 4 °C. A part of the corresponding skimmed material (20 µl) was directly analyzed by MALDI-TOF-MS for the generation of the corresponding protein profiles. The remaining material (200 µl) was diluted 1:1 v/v with water, transferred onto Amicon Ultra devices (3 kDa cut-off) (Millipore) and filtrated by centrifugation at 10,000 rpm, at 4 °C. Two technical replicates were prepared for each filtrate. Peptides from filtrates were directly analyzed by MALDI-TOF-MS or extracted with C18ZipTip devices (Millipore) and then subjected to MS analysis.

MALDI-TOF-MS profiling measurements Before MALDI-TOF-MS PR-P, skimmed milk samples were diluted 1:100 with water. Each sample (0.5 μl) was mixed with 0.5 μl of a solution of sinapinic acid (10 mg/ml) (Bruker Daltonics, Bremen, Germany) in 0.1% v/v trifluoroacetic acid (TFA), 30% v/v acetonitrile (ACN), placed onto a Ground Steel Target (Bruker Daltonics) and dried at room temperature. Before MALDI-TOF-MS PE-P, peptide samples (0.5 μl) were mixed with 0.5 μl of a solution of α-cyano-4-hydroxycinnamic acid (25 mg/ml) (Bruker Daltonics) in 0.1% v/v TFA, 30% v/v ACN, placed onto the instrument target reported above and dried at room temperature. Protein and peptide samples were analyzed by MALDI-TOF-MS using an UltraflexExtreme mass spectrometer (Bruker Daltonics) equipped with the FlexControl software package (version 3.4, Bruker Daltonics)

21

. In all cases,

spectral acquisition methods were developed to maximize the number of signals present in the mass spectra and the corresponding signal to noise ratios. In the case of PR-P measurements, spectra were recorded in the positive linear mode (LM) (laser frequency, 1000 Hz; ion source 1 voltage, 25.19 kV; ion source 2 voltage, 23.59 kV; lens voltage, 7.50 kV; sample rate, 0.31; mass range, 5000-20,000 Da). Five independent spectra (1000 shots at random positions on the same target place, for spectrum) were automatically collected, calibrated externally by using the Protein Calibration Standard 1 (Bruker Daltonics) and subsequently analyzed. In the case of PE-P measurements, spectra were

5



Page 6 of 35

recorded both in the positive LM (laser frequency, 1000 Hz; ion source 1 voltage, 25.19 kV; ion source 2 voltage, 23.94 kV; lens voltage, 6.50 kV; sample rate, 0.16; mass range, 500-5000 Da), or in the positive reflector mode (RM) (laser frequency, 1000 Hz; ion source 1 voltage, 25.19 kV; ion source 2 voltage, 22.50 kV; lens voltage, 8.50 kV; reflector, 26.83 kV; reflector 2, 13.75 kV; sample rate, 0.63; mass range, 500 to 5000 Da). Ten independent spectra (500 shots at random positions on the same target place, for spectrum) for each peptide mixture were automatically collected, externally calibrated by using the Peptide Calibration Standard 2 (Bruker Daltonics) and subsequently analyzed.

Bioinformatics and statistics FlexAnalysis (version 3.4) and ClinProt Tools (version 2.2) software packages (Bruker Daltonics) were used for the analysis of all MALDI-TOF-MS data, which included spectral mass adjustment (compression by a factor of 10 in the total mass range), optional smoothing (using the Savitsky–Golay algorithm with a frame size of 25 Da), spectral baseline subtraction, normalization, internal peak alignment and peak picking. Pretreated data were then subjected to visualization and statistical analysis. Peptides or proteins showing a statistically significant difference in signal intensity or mass value were determined by means of Wilcoxon- (PWKW), Anderson-Darling- (PAD) and t- (PTTA) test. Class prediction model was set up by Genetic Alghorithms (GA). Discriminant peaks were considered those presenting at least a p-value < 0.000001 among the PWKW, PAD and PTTA ones. Finally, it was performed a principal component analysis (PCA) of the spectra, which was carried out by an external MATLAB software tool integrated into ClinPro Tools software. Regarding adulterated milk, corresponding principal components (PCs) scores were subjected to PLS regression by XLSTAT software (Microsoft). Adulteration prediction at two points (5 and 15% v/v) was obtained by the forecast function of Excel software (Microsoft), using PLS results as matrix reference data.

Results and Discussion In order to define an optimal method for the suitable discrimination of milk samples by MALDI-TOF-MS profiling, dedicated procedures were set up for the combined analysis of the corresponding protein and peptide components. Due to the composition complexity of each milk type, various experimental protocols were developed to this purpose,

6


Page 7 of 35


which were compared a posteriori for their discrimination performances. PR-P assays were performed directly on diluted skimmed milk samples

9-14

. On the other hand, PE-P measurements were realized on diluted low-molecular

mass skimmed milk filtrates that were then straightly analyzed by MALDI-TOF-MS or extracted with concentration devices before MS analysis. Attempts to assay peptides directly in diluted skimmed milk samples were unsuccessful in terms of recognition performances (data not shown). Different instrument acquisition ranges were assessed for protein and peptide analysis; in the latter case, the effect of a spectral acquisition in LM and RM was also evaluated (see below). Optimized procedures were used either for milk speciation, for the characterization of commercial bovine milk samples subjected to various TTs and finally for the detection of adulterated test samples as result of the addition of non-declared material. In the whole, a dataset of 3500 MALDI-TOF mass spectra was recorded for the speciation of raw milk from four animal species. In particular, 500 mass spectra were acquired for the corresponding PR-P (5 spectral replicas for 25 milk samples from totally different individuals of each animal species); a representative example is reported in Fig. 1A-D. A thousand mass spectra were acquired for the corresponding PE-P of milk filtrates (10 spectral replicas for 25 milk samples from totally different individuals of each animal species) either in LM and RM; representative examples are reported in Suppl. Fig. S1A-D and Suppl. Fig. S2A-D. Five hundred mass spectra were acquired for the corresponding PE-P of milk filtrates subjected to C18ZipTip enrichment (5 spectral replicas for 25 milk samples from totally different individuals of each animal species) either in LM and RM; representative examples are reported in Suppl. Fig. S3A-D and Fig. 2A-D. A dataset analogous to that reported above was recorded in the case of BM and commercial thermally-treated PM, UM and IM samples. A representative example of the corresponding PR-P spectra is reported in Fig. 3A-D. Similarly, exemplificative PE-P spectra of milk filtrates acquired either in LM and RM are shown in Suppl. Fig. S4A-D and Suppl. Fig. S5A-D, respectively. Peptide MALDI-TOF spectra of milk filtrates subjected to C18ZipTip enrichment recorded either in LM and RM are presented in Suppl. Fig. S6A-D and Fig. 4A-D, respectively. Finally, a dataset of 3780 MALDI-TOF mass spectra was recorded for the detection of adulterated milk samples as result of the addition of non-declared material. Experiments were performed either to identify adulteration from raw BM added to WM, OM, or GM, and from IM added to BM. In particular, 540 mass spectra were acquired for the corresponding PR-P (5 spectral replicas of the 9 percentage values assayed for each kind of adulteration; experiments

7



Page 8 of 35

were performed on triplicate samples) (data not shown); 1080 mass spectra were acquired for the corresponding PE-P of milk filtrates (10 spectral replicas of the 9 percentage values assayed for each kind of adulteration; experiments were performed on triplicate samples) either in LM and RM (data not shown); 540 mass spectra were acquired for the corresponding PE-P of milk filtrates subjected to C18ZipTip enrichment (5 spectral replicas of the 9 percentage values assayed for each kind of adulteration; experiments were performed on triplicate samples) either in LM and RM (data not shown). In each case, statistical analysis of mass spectra of pure milk samples allowed to identify discriminant signals based on either intensity or mass values. Class prediction model was set up by GA; marker proteins or peptides were identified by means of their PWKW, PAD and and PTTA p-values. Corresponding data for milk speciation are shown in Suppl. Tab. S3 (PR-P - 23 markers), Suppl. Tab. S4 (PE-P in LM - 65 markers), Suppl. Tab. S5 (PE-P in RM - 106 markers), Suppl. Tab. S6 (PE-P in LM after C18ZipTip enrichment - 63 markers) and Suppl. Tab. S7 (PE-P in RM after C18ZipTip enrichment - 95 markers), together with additional statistical parameters. Corresponding data for the characterization of thermally-treated commercial bovine milks are shown in Suppl. Tab. S8 (PR-P - 10 markers), Suppl. Tab. S9 (PE-P in LM - 27 markers), Suppl. Tab. S10 (PE-P in RM - 47 markers), Suppl. Tab. S11 (PE-P in LM after C18ZipTip enrichment - 46 markers) and Suppl. Tab. S12 (PE-P in RM after C18ZipTip enrichment - 53 markers), together with additional statistical parameters. As expected, a number of common peptide markers occurred between analyses performed in LM and RM. In particular, 43, 39, 18 and 23 common peptide markers were observed during PE-P on milk filtrates for speciation, on C18ZipTip-enriched milk samples for speciation, on milk filtrates for TT characterization, on C18ZipTip-enriched milk samples for TT characterization, respectively. Importantly, data from PWKW, PAD and PTTA analysis defined the recognition capability of each specific assay; corresponding results are reported in Tab. 1 and 2. They demonstrate that milk speciation (100% recognition capability) was guaranteed in the case of MALDI-TOF-MS profiling measurements performed either on proteins from milk filtrates, on peptides from milk filtrates analyzed in RM, and on peptides from C18ZipTip-enriched milk filtrates analyzed both in LM and RM, notwithstanding the heterogeneous polypeptide composition of this biological fluid

16,18,22

(Tab. 1). Conversely, the

characterization of the TT of commercial bovine milks was ensured only in the case of peptides from C18ZipTipenriched milk filtrates analyzed in RM (Tab. 2). The latter result is probably dependent on the higher complexity of the protein and peptide mixtures present in commercial milk samples 5,15,17,19,20,23-26; it also underlines that the method

8


Page 9 of 35


used for spectral acquisition can highly affect the recognition capability of the whole platform. On this basis, particular care is necessary for the initial set up of the PR-P and PE-P experimental conditions to guarantee the maximum number of signals present in the mass spectra and of the corresponding signal to noise ratios. These instrumental setting have to be maintained during the whole analysis of the different milk samples with the aim of not compromising the recognition capability of the whole platform. As expected, PCA of the data reported in Suppl. Tab. S3-12 were in line with the recognition capability values reported in Tab. 1 and 2. For example, Fig. 5A and 5B show the PCA 3D scatter plots of results from BM, WM, GM and OM proteins analyzed in LM and from BM, WM, GM and OM C18ZipTip-enriched peptides in RM, respectively, which were used for milk speciation. A good separation of the colored spots was evident in both cases. Similarly, Fig. 6A and 6B show the PCA 3D scatter plots of results from BM, PM, UM and IM proteins analyzed in LM, and from BM, PM, UM and IM C18ZipTip-enriched peptides in RM, respectively, which were used for the recognition of the material thermal treatment. Although more appreciable in a three-dimensional representation, well-resolved colored spots were evident also in these cases. Best dot separation occurred in the case of peptides. Less resolved PCA 3D scatter plots were obtained in the case of the remaining methods used for milk speciation (Suppl. Fig. S7A-C) and for the recognition of milk TT (Suppl. Fig. S8A-C). On the basis of the data reported in Tab. 1-2, Fig. 5-6, Suppl. Fig. S7-S8 A-C and the capability of this MS platform of performing independent analyses on different molecules (proteins and peptides), we conclude that complementary information associated with the highest recognition capability values for milk speciation and for thermal treatment characterization are obtained when diluted skimmed milk samples and C18ZipTip-enriched diluted skimmed milk filtrates are analyzed for the respective protein and peptide content in LM and RM, respectively.

Marker assignment Marker protein signals recognized for milk speciation (Suppl. Tab. S3 and Fig. 1) and for TT recognition (Suppl. Tab. S8 and Fig. 3) were associated with specific components based on literature data

9-14,17

and mass calculations. As

expected, most of the signals were assigned to -Lactalbumin, -Lactoglobulin, as well as to γ2/3-caseins and proteoso peptones resulting from the activity of milk plasmin on β-Casein

17,27-29

. In particular, γ2-caseins and 3-

caseins were identified as marker signals in BM, WM, GM and OM; -Lactalbumins were assigned markers in BM,

9



Page 10 of 35

WM and GM; β-CN fragments (98-207) were discriminant signals in GM and OM; -Lactoglobulin, proteoso peptones p.p.8.I., and β-Casein fragments (1-68) and (69-209) were discriminant molecules in GM, BM and WM, respectively (Tab. 3). The latter complementary peptides originated after specific proteolysis at K68; this lysine is exclusively present within the water buffalo β-Casein sequence 4,17. When compared to data reported above, PR-P of thermally-treated milk samples identified a reduced number of protein markers, i.e. proteoso peptones p.p.8.I. (BM, PM and UM), -Lactoglobulin (BM and PM), -Lactalbumin adduct with lactose (m = +324) (IF), γ2-casein (UM) and its adduct with lactose (IM) (Tab. 4). Various non-assigned marker species were specific for UM and IM; the corresponding hypothetical lactose adducts were present only in IM. Finally, marker peptide signals from C18ZipTip-enriched diluted skimmed milk filtrates recognized for speciation (Suppl. Tab. S7 and Fig. 2) and for TT recognition (Suppl. Tab. S12 and Fig. 4) were associated with specific components based on MALDI-TOF/TOF and nanoLC-ESI-LIT-MS/MS analyses (see Supporting Information for experimental details). Fragmentation data were obtained from peptide mixtures either loaded on the MALDI plates also used for PE-P experiments and directly analyzed by nanoLC-ESI-LIT-MS/MS analysis; in the latter case, a specific data searching for peptides having a mass value coincident with that of the marker species was realized. This combined approach identified 38 marker peptides for milk speciation as deriving from -Casein (BM, WM, GM and OM), -S1-Casein (BM and OM), -S2-Casein (GM), Glycosylation-dependent cell adhesion molecule 1 (BM and WM), -Casein (WM, GM and OM), Cytoglobin isoform X1 (WM), Serum amyloid A protein (GM and OM), Butyrophilin subfamily 1 member A1 (GM), Lactoperoxidase (GM) and Cohesin subunit SA-3 (OM), among which some bearing specific modifications (Tab. 5). Similarly, 28 marker peptides for TT recognition were identified as deriving from -Casein (BM, PM, UM and IM), -S1-Casein (BM, PM, UM and IM), -S2-Casein (PM, UM and IM), Glycosylation-dependent cell adhesion molecule 1 (BM) and Polymeric immunoglobulin receptor (PM) (Tab. 6). Either in the case of milk speciation and thermal treatment recognition, the slight difference in the nature of the marker molecules identified with PR-P and PE-P experiments should be associated with the molecular mass value of intact proteins, which may exceed the instrument mass range we used, and the resistance of specific proteins to proteases that may hamper the production of the corresponding peptides.

10


Page 11 of 35


Adulteration Platform capability to recognize milk adulteration due to the addition of either non-declared BM to WM, GM, and OM, or of IM to fresh BM was verified with PR-P and PE-P by analyzing nine impure milk samples, which were prepared in triplicate following the addition of different BM volumes into WM, GM and OM to generate 1, 2, 5, 10, 15, 20, 30, 40, and 50% v/v binary mixtures. Similarly, adulterated samples having identical mixed concentrations were prepared when IM was added to BM. After spectral processing according to the procedures already set up for pure materials, MALDI-TOF-MS data of these binary mixtures were subjected to PCA analysis. Resulting PCA data also included variance, PC score and coefficient values. These information were then used for PLS regression calculations. Fig. 7 shows the PLS regression plot of each binary mixture, as deriving from PR-P and PE-P analysis. An optimal linearity of the data was observed in each case. Normalized values for two adulterated test samples analyzed in triplicate were then used to get a prediction of the corresponding percentages of adulteration. Plots of the predicted adulteration levels, as estimated from the above-mentioned PLS model based on the corresponding PR-P and PE-P data, are reported in Fig. 8. Data from PR-P and PE-P analyses matched each other and were in good agreement with real adulteration percentages (5 and 15%) within the experimental error. These results demonstrated the good capability of this platform to predict illegal adulteration levels of unknown milk samples, when they are processed together with proper adulteration standards. They also demonstrated the importance of having analytical system that provides simultaneously two independent adulteration measurements realized on diluted skimmed milk samples and C18ZipTip-enriched diluted skimmed milk filtrates, which are analyzed for the respective protein and peptide content, respectively. As example, Fig. 9 shows the simultaneous detection of bovine protein and peptide markers in the PR-P and PE-P spectra of an adulterated GM sample containing 5% v/v BM.

Conclusions The development of a versatile and fast method for the detection of fraudulent adulterations in milk is a very important issue, due to the economic impact that these frauds have in dairy productions 6. As mentioned before, a number of analytical methods have been developed to this purpose, which have the disadvantage of being sometimes time-consuming, suitable only for a specific milk assay (speciation or TT analysis), useful for a unique animal species to animal species milk discrimination and/or eventually not adaptable for global screening of milk samples (Suppl.

11



Page 12 of 35

Tab. S1). To overcome these limits and taking advantage of previous applications of the ClinProt technology for biomarker recognition and disease prediction in humans

30-31

, a MALDI-TOF-MS platform for combined PR-P and

PE-P of milk samples has been here developed, which recognizes markers either for speciation in BM, WM, GM and OM, as well as for thermal treatment in BM, PM, UM and IM, quickly assigning the nature of a pure unknown material. This platform was developed taking advantage of other applicative studies in the biomedical research field 32-34

. Platform capability to recognize milk adulteration due to the addition of small volumes of either non-declared

BM to WM, GM, and OM, or of IM to fresh BM was verified by analyzing different ad hoc prepared binary mixtures and specific test samples. In the latter case, concentration values were predicted according to independent PR-P and PE-P analyses and verified a posteriori to be coincident with real data. With respect to previous reports describing the use of PR-P analysis on diluted skimmed samples for milk speciation 9-14

, this study extends this method to the analysis of TT and introduces and independent, complementary

measurement, based on an ad hoc developed PE-P approach that has been set up to integrate protein data with additional information on other molecules, i.e. peptides, validating final results. When coupled to a robust statistical evaluation of the resulting data, this platform ultimately permits the typing of milk and the correlation of MALDI spectra with quantitative adulteration values. By introducing PE-P analysis in the sample processing pipeline, this integrated profiling approach also permits the detection of adulterations due to the addition of powdered material into fresh milk, broadening the applicative scenario of the MALDI profiling techniques in dairy science. The highthroughput character of the MALDI ionization technique and its versatility in answering to a number of issues related to the certification of raw materials used in dairy industry suggest the massive use of this platform in milk analysis with the aim to preserve product tradition and value.

Acknowledgments This work was supported by funds from the Italian Ministry of Economy and Finance to CNR for the project "Innovazione e sviluppo del Mezzogiorno - Conoscenze Integrate per Sostenibilità ed Innovazione del Made in Italy Agroalimentare" - Legge n.191/2009, and from Regione Campania for the Project “QUARC - Qualità delle produzioni tipiche campane ed il suo territorio: approcci innovativi ed integrati per rafforzare la competitività del sistema Agroalimentare” (A.G.C.6 - settore 1), bando CAMPUS.

12


Page 13 of 35


References 1. Bertoni, G.; Calamari, L.; Maianti, M.G. Producing specific milks for speciality cheeses. Proc. Nutr. Soc. 2001, 60, 231-246 2. Johnson, M.E.; Lucey, J.A. Major technological advances and trends in cheese. J. Dairy Sci. 2006, 89, 11741178 3. Gerosa, S.; Skoet, J. Milk a vailability: trends in production and demand and medium-term outlook. FAO of the United Nations ESA Working paper 2012, 12-01 4. Di Luccia, A.; Picariello, G.; Trani, A.; Alviti, G.; Loizzo, P.; Faccia, M.; Addeo, F. Occurrence of betacasein fragments in cold-stored and curdled river buffalo (Bubalus bubalis L.) milk. J. Dairy Sci. 2009, 92, 1319-1329 5. Calvano, C.D.; Monopoli, A.; Loizzo, P, Faccia, M., Zambonin, C. Proteomic approach based on MALDITOF MS to detect powdered milk in fresh cow's milk. J. Agric. Food Chem. 2013, 61, 1609-1617 6. Commission Regulation (EU) No. 273/2008, Methods for the analysis and the evaluation of the quality milk and dairy products. Off. J. Eur. Commun. 2008, L88, 1-115 7. Addeo, F.; Garro, G.; Intorcia, N.; Pellegrino, L., Resmini, P., Chianese, L. Gel electrophoresis and immunoblotting for the detection of casein proteolysis in cheese. J. Dairy Res. 1995, 62, 297-309 8. Villar-Garea, A.; Griese, M.; Imhof, A. Biomarker discovery from body fluids using mass spectrometry. J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. 2007, 849, 105-114 9. Angeletti, R.; Gioacchini, A.M.; Seraglia, R.; Piro, R.; Traldi, P. The potential of matrix-assisted laser desorption/ionization mass spectrometry in the quality control of water buffalo mozzarella cheese. J. Mass Spectrom. 1998, 33, 525-553 10. Fanton, C.; Delogu, G.; Maccioni, E.; Podda, G.; Seraglia, R.; Traldi, P. Matrix-assisted laser desorption/ionization mass spectrometry in the dairy industry 2. The protein fingerprint of ewe cheese and its application to detection of adulteration by bovine milk. Rapid Commun. Mass Spectrom. 1998, 12, 1569– 1573 11. Cozzolino, R.; Passalacqua, S.; Salemi, S.; Malvagna, P.; Spina, E.; Garozzo, D. Identification of adulteration in milk by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. J. Mass Spectrom. 2001, 36, 1031–1037 12. Cozzolino, R.; Passalacqua, S.; Salemi, S.; Garozzo, D. Identification of adulteration in water buffalo mozzarella and in ewe cheese by using whey proteins as biomarkers and matrix-assisted laser desorption/ionization mass spectrometry. J. Mass Spectrom. 2002, 37, 985–991 13. Cunsolo, V.; Muccilli, V.; Saletti, R.; Foti, S. MALDI-TOF mass spectrometry for the monitoring of shedonkey’s milk contamination or adulteration. J. Mass Spectrom. 2013, 48, 148–153 14. Nicolaou, N.; Xu, Y.; Goodacre, R. MALDI-MS and multivariate analysis for the detection and quantification of different milk species. Anal. Bioanal. Chem. 2011, 399, 3491-3502 15. Calvano, C.D.; de Ceglie, C.; Monopoli, A.; Zambonin, C.G. Detection of sheep and goat milk adulterations by direct MALDI-TOF MS analysis of milk tryptic digests. J. Mass. Spectrom. 2012, 47, 1141–1149 16. Baum, F.; Fedorova, M.; Ebner, J.; Hoffmann, R.; Pischetsrieder, M. A sensitive and effective proteomic approach to identify she-donkey's and goat's milk adulterations by MALDI-TOF MS fingerprinting. J. Proteome Res. 2013, 12, 5447-5462 17. Somma, A.; Ferranti, P.; Addeo, F.; Mauriello, R.; Chianese, L. Peptidomic approach based on combined capillary isoelectric focusing and mass spectrometry for the characterization of the plasmin primary products from bovine and water buffalo beta-casein. J. Chromatogr. A. 2008, 1192, 294-300 18. Cuollo, M.; Caira, S.; Fierro, O.; Pinto, G.; Picariello, G.; Addeo, F. Toward milk speciation through the monitoring of casein proteotypic peptides. Rapid Commun. Mass Spectrom. 2010, 24, 1687-1696 19. De Simone, C.; Picariello, G.; Mamone, G.; Stiuso, P.; Dicitore, A.; Vanacore, D.; Chianese, L.; Addeo, F.; Ferranti, P. Characterisation and cytomodulatory properties of peptides from Mozzarella di Bufala Campana cheese whey. J. Pept. Sci. 2009, 15, 251-258 20. Holland, J.W.; Gupta, R., Deeth, H.C.; Alewood, P.F. Proteomic analysis of temperature-dependent changes in stored UHT milk. J. Agric. Food Chem. 2011, 59, 1837-1846 21. Mastrogiacomo, R.; D’Ambrosio, C.; Niccolini, A.; Serra, A.; Gazzano, A.; Scaloni, A.; Pelosi, P. An odorant-binding protein is abundantly expressed in the nose and in the seminal fluid of the rabbit. PLoS One 2014, 9, e111932

13



Page 14 of 35

22. Dallas, D.C., Guerrero, A., Parker, E.A., Garay, L.A., Bhandari, A., Lebrilla, C.B., Barile, D., German, J.B. Peptidomic Profile of Milk of Holstein Cows at Peak Lactation. J. Agric. Food Chem. 2014, 62, 58-65 23. Arena, S.; Renzone, G.; Novi, G.; Paffetti, A.; Bernardini, G.; Santucci, A.; Scaloni, A. Modern proteomic methodologies for the characterization of lactosylation protein targets in milk. Proteomics 2010, 10, 34143434 24. Arena, S.; Renzone, G.; Novi, G.; Scaloni, A. Redox proteomics of fat globules unveils broad protein lactosylation and compositional changes in milk samples subjected to various technological procedures. J. Proteomics 2011, 74, 2453-2475 25. Renzone, G.; Arena, S.; Scaloni, A. Proteomic characterization of intermediate and advanced glycation endproducts in commercial milk samples. J. Proteomics 2015, 117, 12-23 26. Meltretter, J.; Wust, J.; Pischetsrieder, M. Modified peptides as indicators for thermal and nonthermal reactions in processed milk. J. Agric. Food Chem. 2014, 62, 10903-10915 27. Eigel, W.N. Identification of proteose-peptone component 5 as a plasmin-derived fragment of bovine betacasein. Int. J. Biochem. 1981, 13, 1081-1086 28. Andrews, A.T. Proteinases in normal bovine milk and their action on caseins. J. Dairy Res. 1983, 50, 45-55 29. Eigel, W.N.; Hofmann, C.J.; Chibber, B.A.K.; Tomik, J.M.; Keenen, T.W.; Mertz, E.T. Plasmin-mediated proteolysis of casein in bovine milk. Proc. Natl. Acad. Sci. USA 1979, 76, 2244-2248 30. Solassol, J.; Jacot, W.; Lhermitte, L.; Boulle, N.; Maudelonde, T.; Mangé, A. Clinical proteomics and mass spectrometry profiling for cancer detection. Expert Rev. Proteomics 2006, 3, 311-320 31. Magni, F.; Van Der Burgt, Y.E.; Chinello, C.; Mainini, V.; Gianazza, E.; Squeo, V.; Deelder, A.M.; Kienle, M.G. Biomarkers discovery by peptide and protein profiling in biological fluids based on functionalized magnetic beads purification and mass spectrometry. Blood Transfus. 2010, 8, Suppl 3:s92-97 32. Terracciano, R.; Preianò, M.; Palladino, G.P.; Carpagnano, G.E.; Barbaro, M.P.; Pelaia, G.; Savino, R.; Maselli, R. Peptidome profiling of induced sputum by mesoporous silica beads and MALDI-TOF MS for non-invasive biomarker discovery of chronic inflammatory lung diseases. Proteomics 2011, 11, 3402-3414 33. Preianò, M.; Pasqua, L.; Gallelli, L.; Galasso, O.; Gasparini, G., Savino, R.; Terracciano, R. Simultaneous extraction and rapid visualization of peptidomic and lipidomic body fluids fingerprints using mesoporous aluminosilicate and MALDI-TOF MS. Proteomics 2012, 12, 3286-3294 34. Preianò, M.; Falcone, D.; Maggisano, G.; Montalcini, T.; Navarra, M.; Paduano, S.; Savino, R.; Terracciano, R. Assessment of pre-analytical and analytical variables affecting peptidome profiling of gingival crevicular fluid by MALDI-TOF mass spectrometry. Clin. Chim. Acta 2014, 437, 120-128 35. Chianese, L.; Caira, S.; Lilla, S.; Pizzolongo, F.; Ferranti, P.; Pugliano, G.; Addeo, F. Primary structure of water buffalo alpha-lactalbumin variants A and B. J. Dairy Res. 2004, 71, 14-19

14


Page 15 of 35


Table 1. ClinProt model generation report for the speciation of milk samples from different animal species (bovine, buffalo, goat and ovine).

Milk speciation Molecular targets/Acquisition method

Cross validation (%)

Recognition capability (%)

Proteins from diluted samples/LM

95.11

100.00

Peptides from C18ZipTip-enriched diluted milk filtrates/RM

89.43

100.00

Peptides from diluted milk filtrates/RM

90.10

100.00

Peptides from C18ZipTip-enriched diluted milk filtrates/LM

89.11

100.00

Peptides from diluted milk filtrates/LM

77.47

97.39

15



Page 16 of 35

Table 2. ClinProt model generation report for the recognition of thermally-treated commercial milk samples (raw, pasteurized, UHT and powdered). Recognition of milk thermal treatment Molecular targets/Acquisition method

Cross validation (%)

Recognition capability (%)

Proteins from diluted samples/LM

87.82

97.19

Peptides from C18ZipTip-enriched diluted milk filtrates/RM

89.69

100.00

Peptides from diluted milk filtrates/RM

90.07

96.46

Peptides from C18ZipTip-enriched diluted milk filtrates/LM

86.94

98.04

Peptides from diluted milk filtrates/LM

82.44

98.00

16


Page 17 of 35


Table 3. Marker species recognized during MALDI-TOF-MS protein profiling for milk speciation. MS experiments were performed in linear mode as reported in the experimental section. Difference value between the maximal and the minimal average peak/area intensity of all classes (DAve) (see ClinProt Tools software for details), protein/polypeptide name, and experimental/theoretical MH+ (average) value are reported. Theoretical MH+ values were calculated according to the amino acid sequence entries present in the UniProt Knowledge Database and the corresponding post-translational modifications. P, phosphate; f, fragment. Lactalbumin variant B mass value was calculated according to dedicated studies 35.

Bovine milk DAve

Description

Exp. MH+

Theor. MH+ (UniProt Code)

36.26

-Lactalbumin

14181

14179 (P00711)

29.46

2-Casein/-Casein f(106–209)

11827

11825 (P02666)

17.55


11560

11560 (P02666)

15.97

Unknown

12006

10.29

Proteoso peptone p.p.8.I/-Casein f(29–105)

8640

8638 (P02666)

33.66

-Casein f(69–209)

15749

15750 (Q9TSI0)

30.90

Unknown

11991

30.39


11521

11520 (Q9TSI0)

27.62

-Casein f(1–68)P5

8251

8252 (Q9TSI0)

20.56


11786

11785 (Q9TSI0)

16.63

Lactalbumin variant B

14236

14236

36.26

-Lactalbumin variant A

14186

14187 (P00712)

29.02

Unknown

11882

21.33


11288

16.48

Unknown

11922

15.17

Unknown

5891

14.06


11554

11553 (Q95L76)

9.48

-Casein f(98–207)

12466

12466 (Q95L76)

8.66

-Lactoglobulin

18185

18187 (P02756)

11287 (Q95L76)

Ovine milk 36.25


11536

11535 (P11839)

31.26

3-casein/-Casein f(108–207)

11271

11270 (P11839)

22.59

Unknown

8573

14.40

-Casein f(98–207)

12449

9.48

Unknown

9712

12448 (P11839)

17



Page 18 of 35

Table 4. Marker species recognized during MALDI-TOF-MS protein profiling of thermally-treated commercial milk samples. MS experiments were performed in linear mode as reported in the experimental section. Difference value between the maximal and the minimal average peak/area intensity of all classes (DAve) (see ClinProt Tools software for details), protein/polypeptide name, and experimental/theoretical MH+ (average) value are reported. Theoretical MH+ values were calculated according to the bovine amino acid sequence entries present in the UniProt Knowledge Database and the corresponding post-translational modifications. The Amadori adduct is indicated with + Lactose; f, fragment.

Raw milk DAve

Description

Exp. MH+

Theor. MH+ (UniProt Code)

10.88


8639

8638 (P02666)

4.46

-Lactoglobulin variant A

18362

18364 (P02754)

Pasteurized milk 10.88


8639

8638 (P02666)

4.46

-Lactoglobulin variant A

18362

18364 (P02754)

UHT milk 38.27

Unknown

8007

24.64

Unknown

5326

10.88

Proteoso peptone p.p.8.I/-Casein f(29–105)

8639

8638 (P02666)

6.19

2-Casein + Lactose/-Casein f(106–209) + Lactose

12150

12149 (P02666)

Powdered milk 40.19

Unknown

6787

24.82

Unknown

6754

22.77

Unknown at m/z 6787 + Lactose

7111

16.40

Unknown at m/z 6754 + Lactose

7079

7.51

-Lactalbumin + Lactose

14502

14503 (P00711)

6.19

2-Casein/-Casein f(106–209) + Lactose

12150

12149 (P02666)

18


Page 19 of 35


Table 5. Marker species recognized during MALDI-TOF-MS peptide profiling for milk speciation. MS experiments were performed in reflectron mode as reported in the experimental section. Most intense MH+ signals reported in Supplementary Tables S7 were subjected to fragmentation analysis for structural assignment (see Supporting Information for experimental details). Difference value between the maximal and the minimal average peak/area intensity of all classes (DAve) (see ClinProt Tools software for details), UniProt/NCBI sequence code (the latter highlighted in italics), protein name, experimental/theoretical MH+ (monoisotopic) value, modifications(s), mass error (in Da), peptide sequence, MASCOT/SEQUEST score (the latter highlighted in italics), and the mass spectrometry method used for peptide assignment (LIFT or CID) are reported. Theoretical MH+ values were calculated according to the indicated UniProt/NCBI sequence entries. N.d., not determined.

Bovine milk DAve

Accession code

Description

Exp. MH+

Theor. MH+

709.31

P02666

-Casein

1589.97

476.00

P02666

-Casein

1264.74

Unknown

2132.00

-S1-Casein

1250.67

Unknown

1347.78

57.52 56.66

P02662

56.16

Modification(s)

Error (Da)

Sequence

Score

LIFT

1589.93

-0.04

EPVLGPVRGPFPIIV

54

✓

1264.70

-0.04

EPVLGPVRGPFP

55

N.d. 1250.60

-0.07

QKEDVPSERY

✓ ✓

37

N.d.

✓ ✓

44.51

P80195

Glycosylation-dependent cell adhesion molecule 1

2723.33

2723.36

0.03

ILNKPEDETHLEAQPTDASAQFIR

126

30.65

P02666

-Casein

1414.76

1414.71

-0.05

VPYPQRDMPIQA

47

Unknown

2528.43

Description

Exp. MH+

827.39

Unknown

2541.46

N.d.

✓

564.57

Unknown

2145.91

N.d.

✓


1693.92

Unknown

2555.35

-Casein

2026.95

Unknown

2379.29


2724.35

Unknown

1731.78

11.41

CID

N.d.

✓ ✓ ✓

Buffalo milk DAve

313.56

Accession code

594080212

168.81 119.65

P11840

97.33 95.92 56.19

594080212

Theor. MH+

Modification(s)

1693.89

Error (Da)

-0.03

SSRQPQNQNPKLPLS

Score

75

N.d. 2026.89

Q->pyroE (Nterm); Q->E (at Q3, Q5 or Q7)

-0.06

QEQNQEQPIRCEKEER

2724.30

-0.05

ILNEPEDETHLEAQPTDASAQFIR N.d.


LIFT

✓ ✓

49

N.d.

19

Sequence

✓ ✓

135

✓ ✓

CID


Page 20 of 35

55.09

Unknown

2237.23

N.d.

✓

52.43

Unknown

2291.17

N.d.

✓

50.79

Q3HW31

-Casein

887.49

887.46

50.35

E1BMU2

Cytoglobin isoform X1

1989.83

1989.96

43.21

594080212


1606.89

25.18

594080212


1519.85

24.69

Unknown

2131.43

N.d.

✓

23.15

Unknown

1603.67

N.d.

✓

-0.03

YPFPGPIP

21

✓

0.13

KVPGEMEIERRERSEE

21

✓

1606.86

-0.03

SRQPQNQNPKLPLS

23

✓

1519.83

-0.02

RQPQNQNPKLPLS

33

✓

M6Ox

19.93

Q3HW31

-Casein

1314.77

1314.74

-0.03

SLVYPFPGPIPK

67

18.00

594080212


1319.73

1319.71

-0.02

RQPQNQNPKLP

24

17.13

Q9TSI0

-Casein

1227.74

1227.71

-0.03

LVYPFPGPIPK

41

10.62

Unknown

2907.30

N.d.

✓

8.79

Unknown

2305.35

N.d.

✓

✓

✓ ✓ ✓

Goat milk DAve

Accession code

Description

Exp. MH+

Theor. MH+

Modification(s)

Error (Da)

Sequence

Score

993.43

P02670

-casein

1957.82

1957.79

Q1->pyroE

-0.03

QEQNQEQPICCEKDER

3.75

858.47

W5PJE0

Serum amyloid A protein

1899.05

1898.94

-0.11

SGKDPNHFRPAGLPDKY

39

✓

576.54

A5JST2


1834.90

1834.85

-0.05

QGWGTFLREAGQGAKDM

110

✓

336.00

P33048

-Casein

1115.61

1115.56

-0.05

REQEELNVV

32

✓

221.76

A5JST2


1703.84

1703.81

Q1->pyroE

-0.03

QGWGTFLREAGQGAKD

92

✓

171.67

P02670

-casein

2922.27

2922.30

Q1->pyroE

0.03

QEQNQEQPICCEKDERFFDDKIAK

4.15

✓

131.71

P02670

-casein

2794.11

2794.20

Q1->pyroE

0.09

QEQNQEQPICCEKDERFFDDKIA

4.33

✓

126.32

P02670

-casein

2251.83

2251.93

Q1->pyroE

0.10

QEQNQEQPICCEKDERFF

3.64

✓

23.52

P02670

-casein

2610.15

22.48

A5JST2


1388.69

17.91

Unknown

947.48

N.d.

✓

16.79

Unknown

1048.54

N.d.

✓

Q1->pyroE

N.d. 1388.66

-0.03

20


AYRDMKEANYK

LIFT

CID ✓

✓ 48

✓

Page 21 of 35


15.96

A3EY52

Butyrophilin member A1

15.35

A3F9D6

15.04 11.37

subfamily

1

1460.72

1461.62

0.90

MGEDSASGDIETLH

77

✓

Lactoperoxidase

1344.76

1344.73

-0.03

GQVWEESLKRL

60

✓

P33049

-S2-casein

1209.68

1209.66

-0.02

TNAIPYVRYL

A5JST2


1850.82

1850.84

0.02

QGWGTFLREAGQGAKDM

Unknown

1189.62

11.10

Q1->pyroE; M17Ox

✓ 102

N.d.

✓ ✓

Ovine milk DAve

Accession code

Description

Exp. MH+

Theor. MH+

Modification(s)

Error (Da)

Sequence

810.55

P02669

-casein

2016.85

2016.84

Q1->pyroE

-0.01

QEQNQEQRICCEKDER

563.21

W5PJR5


1761.85

1761.85

Q1->pyroE

0.00

QGLGTFLREAGQGAKDM

102

✓

455.43

B0LRN2


1625.88

1625.81

D->N D12)

-0.07

DPNHFRPAGLPDKY

60

✓

302.85

P42819

Serum amyloid A

1652.80

1652.81

0.01

DPNHFRPPGLPDKY

57

✓

174.48

P11839

-casein

1630.83

1630.79

-0.04

REQEELNVVGETVE

39

118.53

Unknown

2311.86

N.d.

✓

98.44

Unknown

2853.27

N.d.

✓

(D1

LIFT

✓

W5Q6Z7

Cohesin subunit SA-3

2981.27

2981.24

71.12

B0LRN2


953.49

46.23

P04653

-S1-casein

1529.90

35.47

Unknown

2104.08

N.d.

✓

35.33

Unknown

2002.30

N.d.

✓

25.50

Unknown

1616.68

N.d.

✓

-S1-casein

1430.79

P04653

-0.03

QQELEELLQSSFLDEDEVYSLAATLK

27

✓

953.45

-0.04

DPNHFRPA

40

✓

1529.82

-0.08

SPEVLNENLLRFV

65

✓

1430.76

-0.03

21


SPEVLNENLLRF

CID ✓

89.56

23.68

Q1->pyroE

or

Score

46

✓


Page 22 of 35

Table 6. Marker species recognized during MALDI-TOF-MS peptide profiling of thermally-treated commercial milk samples. MS experiments were performed in reflectron mode as reported in the experimental section. Most intense MH+ signals reported in Supplementary Tables S12 were subjected to fragmentation analysis for structural assignment (see Supporting Information for experimental details). Difference value between the maximal and the minimal average peak/area intensity of all classes (DAve) (see ClinProt Tools software for details), UniProt/NCBI sequence code (the latter highlighted in italics), protein name, experimental/theoretical MH+ (monoisotopic) value, modifications(s), mass error (in Da), peptide sequence, MASCOT/SEQUEST score (the latter highlighted in italics), and the mass spectrometry method used for peptide assignment (LIFT or CID) are reported. Theoretical MH+ values were calculated according to the indicated UniProt/NCBI sequence entries. N.d., not determined.

Raw milk DAve

Accession code

Description

Exp. MH+

Theor. MH+

144.16

P02666

-Casein

1589.97

1589.94

32.55

Unknown

1490.91

N.d.

✓

18.21

Unknown

1979.13

N.d.

✓

-Casein

1151.73

Unknown

2131.84

15.94

P02666

14.19

Modification(s)

1151.69

Error (Da)

Sequence

Score

LIFT

-0.03

EPVLGPVRGPFPIIV

54

✓

-0.04

GPVRGPFPIIV

26

N.d.

CID

✓ ✓

11.05

P80195


2723.31

2723.36

0.05

ILNKPEDETHLEAQPTDASAQFIR

126

10.03

P02662

-S1-Casein

1250.65

1250.60

-0.05

QKEDVPSERY

37

Error (Da)

Sequence

Score

LIFT

CID

✓ ✓

Pasteurized milk DAve

Accession code

Description

Exp. MH+

Theor. MH+

129.77

P02662

-S1-Casein

1337.71

1337.68

-0.03

HIQKEDVPSER

23

✓

✓

26.06

P02663

-S2-Casein

923.57

923.53

-0.04

IPYVRYL

46

✓

✓

18.96

P81265

Polymeric receptor

768.52

769.42

0.90

AAPAGAAIQ

29

✓

15.94

P02666

-Casein

1151.73

1151.69

-0.04

GPVRGPFPIIV

28

✓

Error (Da)

Sequence

Score

immunoglobulin

Modification(s)

UHT milk DAve

Accession code

Description

Exp. MH+

Theor. MH+

262.34

P02666

-Casein

1284.68

1284.64

-0.04

RELEELNVPGE

59

120.64

P02666

-Casein

1994.16

1994.15

-0.01

LYQEPVLGPVRGPFPIIV

45

76.91

P02663

-S2-Casein

1251.66

1251.75

0.09

TKVIPYVRYL

29

Modification(s)

22


LIFT

CID ✓

✓

✓ ✓

Page 23 of 35

41.20


P02662

-S1-Casein

1689.83

1689.82

-0.01

PIGSENSEKTTMPLW

112

✓

Error (Da)

Sequence

Score

LIFT

-0.03

HIQKEDVPSER

53

✓

-0.03

NENLLRF

39

✓

-0.05

RELEELNVPGE

58

Powdered milk DAve

Accession code

Description

Exp. MH+

Theor. MH+

129.77

P02662

-S1-Casein

1337.71

1337.68

97.04

P02662

-S1-Casein

905.51

905.48

73.78

P02666

72.10

-Casein

1283.70

Unknown

791.45

Modification(s)

1283.65

E->Q (E2, E4, E5 or E11) E->Q (E2, E4, E5, E11 or E14)

N.d.

63.23

P02666

-Casein

1624.87

1624.85

54.19

P02754

-Lactoglobulin

1246.70

1245.58

53.94

P02666

-Casein

1396.77

1396.74

33.77

P02666

-Casein

1068.61

28.44

P02666

-Casein

25.01

P02663

-S2-Casein

✓

CID

✓

✓

-0.02

RELEELNVPGEIVE

51

✓

-1.12

TPEVDDEALEK

77

✓

-0.03

RELEELNVPGEI

30

✓

1068.56

-0.05

YPVEPFTES

26

✓

1311.70

1311.71

0.01

YQEPVLGPVRGP

58

✓

1595.85

1594.71

S7Phospho

-1.14

KTVDMESTEVFTK

98

✓

E->Q (E2, E4, E5 or E11)

-0.04

RELEELNVPGEIV

34

✓

E->Q (E2, E4, E5 or E11)

21.07

P02666

-Casein

1495.84

1495.80

20.13

P02662

-S1-Casein

1099.64

1099.61

-0.03

EVLNENLLR

46

✓

17.41

P02666

-Casein

788.43

788.41

-0.02

RELEEL

27

✓

16.39

P02666

-Casein

1097.63

1097.59

-0.04

RELEELNVP

43

✓

12.48

P02666

-Casein

1154.66

1154.61

-0.05

LTLTDVENLH

67

✓

8.86

P02666

-Casein

1125.63

1125.61

-0.02

VVPPFLQPEV

43

✓

5.69

P02662

-S1-Casein

2012.42

2014.00

1.57

IPNPIGSENSEKTTMPLW

46

✓

E->Q (E2, E4 or E5) Q7->E

23



Page 24 of 35

Legend to Figures

Fig. 1 MALDI-TOF-MS PR-P in LM of diluted milk samples. Spectra refer to BM (A), WM (B), GM (C) and OM (D). MH+ signals corresponding to identified marker molecules are highlighted by reporting the associated mass values.

Fig. 2 MALDI-TOF-MS PE-P in RM of diluted milk filtrates subjected to C18ZipTip enrichment as resulting from BM (A), WM (B), GM (C) and OM (D). MH+ signals corresponding to identified marker molecules are highlighted by reporting the associated mass values.

Fig. 3 MALDI-TOF-MS PR-P in LM of diluted thermally-treated commercial milk samples. Spectra refer to fresh BM (A), PM (B), UM (C) and IM (D). MH+ signals corresponding to identified marker molecules are highlighted by reporting the associated mass values.

Fig. 4 MALDI-TOF-MS PE-P in RM of milk filtrates subjected to C18ZipTip enrichment from thermally-treated commercial milk samples. Spectra refer to fresh BM (A), PM (B), UM (C) and IM (D). MH+ signals corresponding to identified marker molecules are highlighted by reporting the associated mass values.

Fig. 5 PCA scatter plot of protein (A) and peptide (B) MALDI-TOF-MS data in milk samples having different animal origin. Data referring to BM, WM, GM and OM are highlighted in red, green, blue and light blue, respectively. MALDI-TOF-MS PR-P experiments were performed in LM on diluted milk samples. MALDI-TOF-MS PE-P experiments were performed in RM on diluted milk filtrates subjected to C18ZipTip enrichment.

Fig. 6 PCA scatter plot of protein (A) and peptide (B) MALDI-TOF-MS data in milk samples subjected to different thermal treatment. Data referring to fresh BM, PM, UM and IM are highlighted in red, green, blue and light blue, respectively. MALDI-TOF-MS PR-P experiments were performed in LM on diluted milk samples.

24


Page 25 of 35


MALDI-TOF-MS PE-P experiments were performed in RM on diluted milk filtrates subjected to C18ZipTip enrichment.

Fig. 7 PLS Plot of MALDI-TOF-MS data deriving from protein (A, C, E and G) and peptide (B, D, F and H) profiling experiments on adulterated milk samples resulting from the addition of BM to WM (A, B), GM (C, D), and OM (E, F), or of fresh BM to IM (G, H). MALDI-TOF-MS PR-P experiments were performed in LM on diluted milk samples. MALDI-TOF-MS PE-P experiments were performed in RM on diluted milk filtrates subjected to C18ZipTip enrichment. The different dots show the concentration of BM in the mixture in relation to the normalized PC scores.

Fig. 8 Plots showing the predicted levels of milk adulteration after MALDI-TOF-MS protein (A, C, E and G) and peptide (B, D, F and H) profiling experiments. The different panels represent data for adulterated samples resulting from the addition of BM to WM (A, B), GM (C, D), and OM (E, F). Data for adulterated samples resulting from the addition of fresh BM to IM are also shown (G, H). MALDI-TOF-MS PR-P experiments were performed in LM on diluted milk samples. MALDI-TOF-MS PE-P experiments were performed in RM on diluted milk filtrates subjected to C18ZipTip enrichment. Blue squares represent the training data set, while the red rhombus indicate the test set. The predicted levels of milk adulteration are reported ± S.D.

Fig. 9 Cropped regions of the MALDI-TOF-MS PR-P and PE-P spectra of a GM sample containing 5% v/v BM. MALDI-TOF-MS PR-P was performed in LM on diluted milk samples. MALDI-TOF-MS PE-P was performed in RM on diluted milk filtrates subjected to C18ZipTip enrichment. MH+ signals corresponding to GM and BM marker molecules are highlighted by reporting the associated mass values; those associated with BM adulteration are marked with an asterisk.

25



Figure 1 26


Page 26 of 35

Page 27 of 35


Figure 2 27



Page 28 of 35

Figure 3 28


Page 29 of 35


Figure 4

29



Page 30 of 35

Figure 5

30


Page 31 of 35


Figure 6

31



Figure 7

32


Page 32 of 35

Page 33 of 35


Figure 8 33



Page 34 of 35

Figure 9 34


Page 35 of 35


TOC Graphics 266x151mm (96 x 96 DPI)


MALDI-TOF-MS Platform for Integrated Proteomic and Peptidomic

Recommend Documents