Cheese production and sterilization process
Pressed, cooked cheese was produced by using Streptococcus thermophilus FTS2, Lactococcus subsp. lactis SRTA 2006 and Lacticaseibacillus paracasei ATCC 334 as a starter, and they were partially ionized as previously described. Non-ionized and ionized cheese bars (125 g, vacuum packed) were stored at −20 °C and thawed at room temperature before administration to animals. Cheese bacteria viability was quantified by bacterial culture as described in Ibarlucea-Jerez et al.10The concentration of culturable LAB after frozen storage was 1.0 × 109 CFU/g for Lactococcus1.7 × 108 CFU/g for Streptococciand 1.7×108 CFU/g for LactobacillusThe bacterial counts in the ionized cheese reached less than 100 CFU/g for each bacterial strain.
Animal testing design
The experiments were conducted on ten-week-old male Wistar rats (N = 81) (Janvier, Le Genest Saint Isle, France). Animals were housed individually in polycarbonate cages and acclimated for two weeks to the following housing conditions: 22 ± 1 °C, 55 – 60% humidity, and reversed light cycle (50 lux, 12 h: 12-hour dark/light cycle from 3 a.m. to 3 p.m.). All animals received water ad libitum and autoclaved regular chow (Ssniff Spezialdiäten GmbH, Soest, Germany). The experimental design complied with relevant guidelines and regulations and received approval from the local ethics committee of the University of Burgundy (Dijon, France) and the French Ministry of Higher Education and Research (DAP 27646-2020101214594619).
Rats were randomly assigned to one of three diet groups (n = 27/group): standard diet group (SD), ionized cheese diet group (ICD), and cheese diet group (CD). Each group underwent the following three feeding periods: a pre-cheese diet period (2 weeks), a cheese diet period (3 weeks), and a post-cheese diet period (2 weeks). During these periods, the three groups were given food ad libitum from 3:00 p.m. to 9:30 a.m. (the next day) and were not accessible for the rest of the day. During the cheese diet period, the ICD and CD groups each received 4 g of ionized and non-ionized cheese (Fig. 5). Food was not given until both groups had finished eating the piece of cheese. Each period was analyzed independently of the others.
The standard diet group (SD), the ionized cheese diet group (ICD) and the cheese diet group (CD) received food during the three periods (n = 27/group). The ICD and CD groups received one serving of 4 g each of ionized and non-ionized cheese during the “cheese” diet period only. Saliva samples were collected from all animals at the end of each period.
Body weight and food intake were recorded every 2 days during the three periods (Supplementary Figure S1). At the end of each period, salivary pH was measured using pH strips (Merck, Darmstadt, Germany) (Supplementary Table S1), and saliva samples were collected at 1 p.m. from conscious animals using sterile Hydra-Flock ultrafine swabs (Puritan, Guilford, Maine). Saliva samples40 were collected at least 23 hours after the last cheese consumption at the end of the cheese diet period. Saliva-containing swabs were frozen at −80 °C until analysis.
Bacterial profiles in saliva
The total volume of saliva collected by swab generally did not exceed 20 µl. Saliva DNA extraction was performed using the ZymoBIOMICS™ DNA Microprep Kit (Zymo Research, Irvine, California).
qPCR amplifications were performed in duplicate using 15 μl reaction mixtures. Each reaction mixture consisted of 3 μl DNA extract, 800 nM of each primer, and 7.5 μl SsoAdvanced Universal SYBR Green Supermix (Bio-Rad, Marnes-la-Coquette, France) in a CFX96 Real-Time PCR Detection System (Bio-Rad, Marnes-la Coquette, France). Thermal cycling conditions included an initial denaturation at 98 °C for 3 minutes, followed by 40 cycles of 95 °C for 15 seconds and 60 °C for 30 seconds. Primers used for relative gene quantification of the three bacterial species were: L. lactis tuff Gen (LLTUF forward: TGAAGAATTGATGGAACTCG / LLTUF reverse CATTGTGGTTCACCGTTC)41, L. paracasei leuS Gene (LCLEUS forward: GCTGGATGCTGGTATTGCTT LCLEUS reverse ATCACGCAGTTTGCCTTCAT)40, S. thermophilus panE Gen (STPANE forward: CCTCGGTGCTCAAGTGGATT STPANE reverse TGGGTCTGAATGCTAATTGGA)42. The quantifications obtained with the cheese LAB-specific primers were compared with the quantifications obtained with all bacterial primers targeting the 16S gene (SPU926 forward: AAACTCAAAKGAATTGACGG, SPU1062 reverse: CTCACRRCACGAGCTGAC).43The use of these primers as well as the following equations to calculate the relative percentage of each LAB bacterial species were described in Ibarlucea-Jerez et al.10.
The hypervariable region V3-V4 of the 16 S rRNA gene was amplified using Phanta® Max Super-Fidelity DNA polymerase (Vazyme) and primers V3PCR1F_460 bp: 5′-ACGGGAGGCAGCAG-3′ and V4PCR1R_460 bp: 5′-TACCAGGGTATCTAATCC-3′. The PCR mix contained 50 ng DNA, 1 µl dNTPs (10 mM), 1 µl of each primer (20 µM), and 1 µl Taq polymerase in a total volume of 50 µl. The program used used the following parameters: 95 °C for 3 min, followed by 35 cycles of 95 °C for 15 s, 65 °C for 15 s, 72 °C for 60 s, and a final extension at 72 °C for 5 min. Since DNA concentrations were low, a second PCR was required. For this PCR procedure, 5 µl of the previously purified DNA was taken and multiplexed using the same primers and under the same conditions as the PCR mentioned above. The purified amplicons were sequenced using MiSeq sequencing technology (Illumina, San Diego, CA, USA) on the @BRIDGe platform (INRAE, Jouy-en-Josas, France). The paired-end reads obtained from MiSeq sequencing were treated as described in Lecomte et al.44.
Bioinformatics analysis of the 16 S rRNA amplicon was performed using R software v4.2.1 (R Core Team, 2017); more specifically, we used Rstudio and the following R packages: ggplot2 v3.3.545 and phyloseq v1.40.046. Differential frequency tests were performed using DESeq2 v1.32.047.
All samples were rarefied to the same sampling depth before calculating within-sample compositional diversity (observed richness) and between-sample compositional diversity (Bray-Curtis dissimilarities). A principal coordinate analysis (PCoA) was performed for the Bray-Curtis dissimilarities. Raw, undiluted OTU counts were used to construct relative abundance plots. The observed richness data were analyzed using a one-way ANOVA. A permutational multivariate analysis of variance (PCoa) test (PERMANOVA) was performed on the Bray-Curtis matrices using 9999 random permutations. The relative abundances of phylum, family and genus were compared using a Kruskal-Wallis test followed by Dunn’s test. OTUs that had an adjusted p-number were compared with an adjusted p-value.
Saliva proteome analysis
Eighteen saliva samples collected in the cheese diet group during the pre-diet phase were prepared for LC‒MS/MS analyses. The saliva samples were from the saliva of rats with cheese bacterial persistence (101, 107, 115, 116, 120, 121, 123, 124, 126) and from the saliva of rats without bacterial persistence (103, 109, 110, 111, 114, 118, 119, 125, 127). The saliva swabs were centrifuged at 14,000 ×. Gfor 20 min with the swab head facing outward from the tube. Protein concentration was measured using a Nanodrop spectrometer (Thermo Scientific, USA). Additionally, samples were diluted in Milli-Q water (Millipore, Burlington, MA, USA) to adjust all samples to the same protein concentration, after which they were mixed with 1 volume of Laemmli denaturation buffer and heated at 90 °C for 5 min. Samples were loaded onto SDS-PAGE gels containing 12% and 5% acrylamide in separating and stacking gels, respectively. Electrophoresis was performed using a Mini Protean II unit (Bio-Rad, Marnes-La-Coquette, France) at 70 V until the dye front penetrated the separating gel. Gels were fixed in a solution of 65% Milli-Q water, 30% ethanol, and 5% acetic acid for 15 minutes and stained in R-250 Coomassie for 1 hour. After destaining overnight in Milli-Q water, bands were manually excised. Bands were then reduced in 10 mM dithiothreitol in 50 mM ammonium bicarbonate and alkylated in 55 mM iodoacetamide in 50 mM ammonium bicarbonate. Destaining of excised bands was achieved by sequential rinses in 25 mM ammonium bicarbonate/acetonitrile (1:1 v/v). Gel pieces were then dried by incubation in 100% acetonitrile for 10 minutes followed by vacuum drying in a SpeedVac. Finally, the gel pieces were incubated overnight at 37 °C with 30 µL of a trypsin solution (V5111, Promega) at 10 ng/µL in 25 mM ammonium bicarbonate. Peptide extraction was performed by adding 40 µL of 100% acetonitrile and 0.5% formic acid, followed by sonication for 15 minutes. The trypsin digests were vacuum dried in a SpeedVac.
Each sample was solubilized with 20 µl of a 0.1% TFA solution containing 1.2 pmol of Promega isotopologue peptides (Peptide Reference Mix, V7491) and transferred to HPLC vials prior to LC‒MS/MS analysis. Five microliters of each hydrolysate was injected into the nano-HPLC; after a wash and concentration step (Pepmap, 300 µm, 0.5 mm) in 0.05% TFA water, the peptides were separated on an analytical C18 column (Acclaim PepMap, 75 µm, 25 cm, ThermoScientific) by a gradient of acetonitrile (ACN/FA, 99.9/0.1) from 2-25% in 60 minutes. The eluted peptides were nanoelectrosprayed into an Orbitrap Q Exactive HFX mass spectrometer (ThermoScientific) operating in data-dependent mode. The parent ion was selected in the Orbitrap at 60,000 resolution and MS analysis was followed by 18 MS/MS with MS/MS fragment analysis at 15,000 resolution.
The raw data of each analysis were imported into the label-free quantification software LCProQI (Progenesis, Waters). After alignment and peak picking steps that allowed the comparison of identical ion signals in each run, the ion maps were normalized based on the signals of the isotopologue peptides present in each sample at the same amount (300 fmol) (housekeeping method using LCProQI). All detected ions were subsequently quantified and the chosen comparison designs allowed a statistical examination of the entire detected ion map. Identification was performed with MASCOT (2.7, www.matrixscience.com) and PEAKS (XPro, www.bioinfor.com) by comparing the MS/MS (516892 spectra, rank
The identification results were re-imported into LCProQI, allowing the identification of quantified ions. Only proteins with at least two identified peptides were validated as present in the samples (Supplementary Table S2). The relative quantification of each identified protein was calculated by summing the abundance of the most intense unique peptides (Hi-N4 (max), −N3 or −N2 (min) peptides, depending on the protein) that identified a single protein. For the study of oxidized proteins, only modified peptides (+15.99 or 31.9898 (dihydroxylation)) were used to calculate the abundance of the oxidized form of the protein.
The comparison of protein abundance between animals with and without LAB persistence was investigated using ANOVA tests (P-value and Qvalue FDR adjusted) in the WF LCProQI. Proteins with Q-Values p www.pantherdb.org/) and String (https://string-db.org).
