Characterization of Cell-Envelope Proteinases from Two Lacticaseibacillus casei Strains Isolated from Parmigiano Reggiano Cheese

Simple Summary Lactic acid bacteria are nutritionally fastidious microorganisms typically used in the production of fermented dairy foods. The lactic acid bacteria’s proteolytic system is crucial for their growth in milk, and plays a paramount role in developing the organoleptic and healthy properties of fermented dairy foods. Cell-envelope proteinases are the first component of this system, responsible for the degradation of caseins into short peptides. In the present work, the cell-envelope proteinases of two highly proteolytic Lacticaseibacillus casei strains, previously isolated from ripened Parmigiano Reggiano cheese, were characterized from a genetic and biochemical point of view. Two different prt genes existed in the genomes of both strains, but only one, named prtR1, was expressed. PrtR1 proteins extracted from both strains displayed the highest activity at 40 °C and pH 7. Interestingly, PrtR1 extracted from Lacticaseibacillus casei PRA205 retained some residual activity at 5 °C and at pH 4. These important biotechnological features can be exploited in the production of fermented dairy products. Peptidomic analysis revealed that both proteinases were able to release β- and αS1-casein-derived bioactive peptides, suggesting that Lacticaseibacillus casei can be a source of new proteinases that can be exploited for the formulation of dairy beverages or hydrolysates enriched in bioactive peptides. Abstract In the present work, two cell-envelope proteinases (CEPs) from Lacticaseibacillus casei strains PRA205 and 2006 were characterized at both the biochemical and genetic levels. The genomes of both L. casei strains included two putative CEPs genes prtP2 and prtR1, but only prtR1 was transcribed. The extracted PrtR1 proteinases were serine proteinases with optimal activity at 40 °C and pH 7.5, and were activated by Ca2+ ions. Interestingly, PrtR1 from L. casei PRA205 exhibited high residual activity at pH 4 and at 5 °C, suggesting its possible exploitation for fermented food production. The caseinolytic activity against αS1- and β-casein indicated that both PrtR1s belonged to the PI/PIII type. These PrtR1s cleaved β-casein peptide bonds preferentially when amino acid M or N was present at the P1 subsite and amino acids A and D were at the P1′ subsite. Several bioactive peptides were found to be released from PrtR1 after αs1- and β-casein hydrolysis.


Introduction
Lactic acid bacteria (LAB) are a group of microorganisms generally found in nutrientrich environments and are commonly used in the manufacturing of fermented dairy foods.
L. casei DSMZ20011 T , L. rhamnosus PRA161, and L. rhamnosus 0503 [2,3]. All of the reference strains were confirmed by mutL multiplex PCR [12]. Details of the methods used in species confirmation are reported in the Supplementary Materials. Stocks of cultures were stored frozen at −80 • C in de Man-Rogosa-Sharpe (MRS) medium (Oxoid, Basingstoke, Hampshire, UK) supplemented with 25% (w/v) glycerol. Before experimental use, all of the strains were twice propagated in MRS broth at 37 • C for 24 h, under anaerobic conditions. For the entire duration of the experiments, the reference strains were maintained in MRS medium supplemented with 7% (w/v) agar at 4 • C. All media and chemicals used in this study were purchased from Sigma-Aldrich (St. Louis, MO, USA), unless otherwise indicated. Primers and sequencing services were provided by Bio-Fab Research (Rome, Italy).

DNA Extraction
Genomic DNA (gDNA) was extracted from late exponential cultures grown in MRS, as described by Gala et al. [13]. Briefly, cells (1.5 mL) were centrifuged at 6000× g for 10 min, washed with 500 µL of TE buffer (10 mmol/L Tris-HCl, 1 mmol/L EDTA, pH 8.0), and re-suspended in 200 µL of TE buffer with glass beads (diameter 0.106 mm). Subsequently, the cell suspension was vortexed with Vortex-Genie 2 (MoBio, Jefferson City, MO, USA) for 4 min (two rounds of 2 min at the maximum speed, with 1 min in ice, and then 2 min at maximum speed). Next, 15 µL of proteinase K (20 mg/mL) was added, and the mixture was incubated at 60 • C for 1 h. After incubation, 40 µL of 20% SDS was added, and the samples were incubated at 65 • C for 15 min. After cooling at room temperature, 90 µL of refrigerated 5 mol/L potassium acetate was added, and the mixture was centrifuged at 10,000× g for 10 min. After performing phenol-chloroform extraction and ethanol precipitation, DNA samples were suspended in 50 µL of TE buffer. Then, the suspension was mixed with 1.5 µL of RNAse (10 mg/mL) and incubated at 37 • C for 2 h. The concentration and purity of gDNA samples were determined using a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA), while gDNA quality was evaluated by electrophoresis on 0.8% (w/v) agarose gel containing ethidium bromide (0.5 µg/mL) in 0.5X TBE buffer (45 mmol/L Tris-HCl, 45 mmol/L boric acid, and 1 mmol/L EDTA, pH 8.0). gDNA samples were diluted to 50 ng/µL in ddH 2 O and stored at −20 • C for subsequent analysis.

In Silico Analysis, PCR Screening, and Phylogenetic Tree Construction
The in silico search for putative prt genes in L. casei genomes was performed using the BLASTp algorithm against the NCBI database, using L. casei PrtP (AFJ15093.1) as the query sequence. A curated dataset containing 44 proteins from 23 strains was built, and sequences were aligned with the constraint-based alignment tool (COBALT) method [14], using the default settings (Supplementary Table S2). Conserved protein domains were analyzed via NCBI Batch CD-Search (https://www.ncbi.nlm.nih.gov/Structure/bwrpsb/bwrpsb.cgi; last access on 14 June 2021; E-Value threshold 0.01, max. hits 500) using the CDD database.
Primer pairs were designed on the conserved regions within or surrounding the catalytic domains of prt genes, using Primer3 [15] (Supplementary Table S3). All PCR reactions were carried out in a T100 thermal cycler (Bio-Rad, Hercules, CA, USA) with DreamTaq DNA Polymerase (Thermo Scientific, Waltham, MA, USA), following the manufacturer's instructions. PCR products were resolved by 1.2% (w/v) agarose gel electrophoresis stained with ethidium bromide (5 µg/mL) and, when required, PCR amplicons were digested with the endonuclease EcoRV (Thermo Scientific, Waltham, MA, USA) in a 20 µL final volume, according to the manufacturer's instructions. PCR products were purified with the DNA Clean & Concentrator™-5 Kit (Zymo Research, Irvine, CA, USA), and were sequenced on both strands through a DNA Sanger dideoxy sequencing process using both external and internal primers (Supplementary Table S3). Sequences were assembled in DNASTAR (DNASTAR, Inc. Madison, WI, USA) and trimmed on both ends to remove primer sequences. The nucleic acid sequences of the prtP and prtR partial genes were deposited at GenBank database under the accession numbers MZ606853 to MZ606856. Phylogenetic relationships were inferred using the gamma distribution (shape parameter = 1) model and the neighbor-joining (NJ) method. Bootstrap support values were calculated from 1000 replicates in MEGA6 [16]. All trees were visualized using the interactive tree of life (iTOL) v5.2 [17].

Induction of Cell-Envelope Proteinases (CEPs)
To induce the expression of CEPs, the strains were pre-cultured in 50 mL of MRS broth and incubated for 72 h at 37 • C under anaerobic conditions. After centrifugation at 10,000× g for 20 min at 4 • C, cells were re-suspended in physiological solution (0.9% NaCl) at a final OD600nm value of 10 10 CFU/mL, and then spread on milk-citrate-agar (MCA) plates (4.4% re-suspended skimmed milk, 0.1% Na citrate, 0.1% yeast extract, 0.5% glucose, and 1.5% agar) in triplicate. After incubation for 48 h at 37 • C, cells were recovered from plates and re-suspended in physiological solution at a final concentration of 10 10 CFU/mL. Samples were used for both RNA extraction and biochemical characterization.

RNA Extraction and RT-PCR
For RNA extraction, glassware was sterilized at 180 • C for at least 4 h to degrade RNases, and all the solutions were prepared with diethyl pyrocarbonate (DEPC)-treated water. MCA-induced cells were washed twice with DEPC-treated TE buffer (100 mmol/L Tris-HCl, 50 mmol/L EDTA, pH 8.0), and cell pellets were maintained at −80 • C until thawed with 1 mL of Tri-reagent using the Direct-zol RNA MiniPrep Kit (Zymo Research, Irvine, CA, USA). Mechanical lysis was performed using a Vortex-Genie 2 for two rounds of 20 min at the highest speed, alternated with 3 min on ice. The quantity of total RNA was measured spectrophotometrically using a NanoDrop ND-1000 (NanoDrop Technology, Wilmington, DE, USA), while the integrity was checked via denaturing gel electrophoresis on a 0.9% (w/v) agarose gel with formaldehyde (10 mL of 10× MOPS running buffer) and 18 mL of 37% formaldehyde (12 mol/L) on a pH 7.0 1× MOPS running buffer (0.4 mol/L MOPS, 1 mol/L sodium acetate, and 0.01 mol/L EDTA), after RNA treatment at 65 • C for 10 min.
Contaminating DNA was removed by treating 2 µg of RNA sample with dsDNase (Thermo Scientific, Waltham, MA, USA) at 37 • C for 2 min in a preheated thermocycler with the lid temperature adjusted to 37 • C. After being chilled on ice and briefly centrifuged, 12.5 µL treated RNA samples were used as templates for first-strand cDNA synthesis with random hexamers (Thermo Scientific, Waltham, MA, USA), using RevertAid Reverse Transcriptase (Thermo Scientific, Waltham, MA, USA), according to the manufacturer's instructions. CEP-specific RT-PCR assays were performed with the primers reported above (Supplementary Table S3). RT-PCR of the 16S rRNA gene was carried out as previously reported, and used as the positive control [2].

CEP Extraction and Protein Concentration Determination
Following induction of CEPs, cells were centrifuged at 12,000× g for 10 min and re-suspended in 100 µL of Ca 2+ -free buffer (Tris-Cl 50 mmol/L; pH 7.5) at a final concentration of 10 10 CFU/mL. After 30 min of incubation at 37 • C under agitation, extracts were centrifuged (12,000× g for 10 min), the supernatants collected, and the pellets re-suspended in the same volume of the Ca 2+ -free buffer for three additional cycles of incubation [18]. The collected supernatants containing the crude enzyme extracts were pooled and stored at −20 • C, awaiting biochemical assays. The protein concentration was assessed using the Bradford Assay Kit according to the manufacturer's instructions (Sigma-Aldrich, St. Louis, MO, USA). The release of lactate dehydrogenase was measured in order to monitor the membrane integrity and the cell lysis during the extraction phase, as previously described [19].

CEPs' Enzymatic Activity Assay
The proteinase activity of crude extracts was evaluated via a chromogenic assay. The activity was measured using the specific substrate succinyl-alanyl-alanyl-prolyl-phenylalaninep-nitroanilide. The assay mixture, containing 107 µL of 50 mmol/L sodium phosphate buffer (pH 7), 56 µL of 5 mol/L NaCl, 9.5 µL of 20 mmol/L of substrate and 15 µL of extract, was incubated at 40 • C for 120 min in a covered water bath. The reaction was stopped by adding 94 µL of 80% acetic acid. The released p-nitroanilide was measured at 410 nm using a microplate reader. Control reactions with the substrate but without CEPs were also prepared [20]. One unit of proteinase activity was defined as the amount required to liberate 1 µmol of p-nitroanilide per minute. Specific activity was expressed as units of proteinase activity per mg of protein.

Effects of Temperature, pH, Metal Ions, and Inhibitors on CEPs' Activity
The effect of temperature on the CEPs' activity was measured as described above, modifying the incubation temperature. The assay was carried out at four different temperatures (5 • C, 35 • C, 40 • C, and 45 • C), at a constant optimal pH value of 7. The effect of pH on enzyme activity was evaluated by modifying the pH of the reaction buffer, keeping the temperature constant at 40 • C. Sodium acetate buffer (50 mmol/L) was utilized for reactions carried out at pH 4-6 whereas sodium phosphate buffer (50 mmol/L) was used for the reaction performed at pH 8. To analyze the influence of metal ions on proteinase activity, KCl, CaCl 2 , or MgCl 2 was added to the reaction mixture at a final concentration of 2 mmol/L (pH 7, 40 • C). Similarly, in order to evaluate the effects of inhibitors on proteinase activity, EDTA or PMSF was added to the mix, at a final concentration of 0.5 mmol/L (pH 7, 40 • C).

Casein Hydrolysis
To test the caseinolytic activity of CEPs, 5 mg/mL of αS1-or β-casein solution, dissolved in sodium phosphate buffer (pH 7.0; 100 mmol/L), was mixed at a 1:1 volume ratio with the crude extracts containing 1.5 U/mL of CEP activity. The mixtures were incubated at 40 • C and, after various intervals (0, 8,24,30,48, and 56 h), aliquots of samples were withdrawn and stored at −20 • C for further analysis.

SDS-PAGE
The casein breakdown pattern was assessed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Samples collected during the caseinolytic test described above were diluted in Laemmli buffer in order to load 2.5 µg of total protein per lane. Denaturation was completed by boiling the samples for 3 min. SDS-PAGE was carried out on 12% polyacrylamide gels on vertical electrophoresis cells for 1 h at 200 V. Gels were stained with the Coomassie blue staining method (0.1% Coomassie brilliant blue in 100 mL of 40% methanol, 10% glacial acetic acid, 50% H 2 O) for 1 h under stirring. Subsequently, gels were de-stained with the de-staining solution (40% methanol, 10% glacial acetic acid, 50% H 2 O) for 30 min under stirring. The de-staining step was repeated four times.

Identification of Peptides by Ultrahigh-Performance Liquid Chromatography/High-Resolution Mass Spectrometry (UHPLC/HRMS)
Samples collected during the casein hydrolysis assay were mixed with 1% trifluoroacetic acid at a 1:1 volume ratio and submitted to UHPLC/HRMS analysis for peptide identification. Chromatographic separation was carried out with UHPLC (UHPLC Ultimate 3000 Separation Module, Thermo Fisher Scientific, San Jose, CA, USA) equipped with a C18 column (Acquity UPLC HSS C18 Reversed Phase, 2.1 × 100 mm, 1.8 µm particle size, Waters, Milford, MA, USA). Mass spectrometry (MS) and tandem MS experiments were performed on a Q Exactive Hybrid Quadrupole-Orbitrap Mass Spectrometer (Thermo Fisher Scientific, San Jose, CA, USA). The full description of the gradient, flow rate, and MS and MS/MS parameters was reported by Martini et al. [21].
Peptide sequencing was carried out by using MASCOT (Matrix Science, Boston, MA, USA) protein identification software, with the following search parameters: enzyme, none; peptide mass tolerance, ±5 ppm; fragment mass tolerance, ±0.12 Da; variable modification, oxidation (M) and phosphorylation (ST); maximal number of post-translational modifications permitted in a single peptide, 4. The assignment procedure was confirmed by the manual verification of MS/MS spectra.

Identification of Bioactive Peptides
The peptides identified via mass spectrometry analysis were investigated in relation to previously identified bioactive peptides. The identification of bioactive peptides was carried out by using the Milk Bioactive Peptide Database (MBPDB)-an online database of human milk-and dairy-derived bioactive peptides [22].

Calculation of the Cleavage Specificity
The cleavage probability and the positive or negative influence on the cleavage of a specific amino acid in the P1 and P1 subsites were calculated as described by Solieri et al. [23]. The amino acid residues were designated as P1 in the N-terminal direction (on the left of the sequence) and P1 in the C-terminal direction (on the right of the sequence) from the cleaved bond. The residue P1 interacts with the subsite S1 in the enzyme active site, whereas the residue P1 interacts with the subsite S1 in the enzyme active site. The peptide bond cleaved by the CEPs is defined as the P1-P1 bond.
The amino acid residue in position P1 or P1 influences the CEP cleavage probability. If the amino acid residue A is in the position n (P1 or P1 subsite), the cleavage probability of the P1-P1 bond will be as follows: total amino acid A cleaved in position n total amino acid A in protein x 100 (1) The mean cleavage probability was defined by the following formula: The positive or negative influence of an amino acid residue A in the P1 and P1 subsites was quantified by the coefficient Kn: Kn values > 0 indicate a favorable influence of the amino acid A in the specific subsite on the cleavage of the P1-P1 bond, whereas Kn values < 0 suggest a negative effect on the cleavage.

Statistical Analysis
Data are shown as the mean ± standard deviation (SD) for three replicates. Univariate analysis of variance (ANOVA) with Tukey's post hoc test was applied using GraphPad Prism 6.0 (GraphPad Software, San Diego, CA, USA). The differences were considered significant at p < 0.05.

In Silico Survey of Putative prt Genes in the L. casei Genome
Most LAB are thought to possess only one type of CEP [4], but it has been proven that more than one prt gene exists in the genomes of several LAB species, and that the pattern of prt genes is highly variable at the inter-strain level. For instance, L. helveticus possesses four unique types of CEP-namely prtH1, prtH2, prtH3, and prtH4-while L. rhamnosus CGMCC11055 possesses both prtP and prtR genes, although prtP was initially detected only in L. lactis and L. paracasei/L. casei species [4,19,[24][25][26].
Using the L. casei PrtP (AFJ15093.1) amino acid sequence as a query in BLASTp search, we built a dataset consisting of 44 putative Prt-encoding genes in L. casei. Protein alignment with COBALT showed that 22 proteins exhibited the insert domain, and were categorized as PrtP (data not shown). The remaining 22 proteins lacked the insert domain, and were categorized as PrtR, with three different lengths of approximately 1500, 1800, and 2200 amino acids (data not shown). This result suggests that both prtP and prtR paralogs exist in several L. casei genomes, including in the type strain ATCC 393 T . Clustering analysis showed two distinct clades for L. casei PrtP proteins, referred to as PrtP1 and PrtP2, while PrtR proteins were grouped into three different clusters, referred to as PrtR1, PrtR2, and PrtR3 ( Figure 1).
Most LAB are thought to possess only one type of CEP [4], but it has been prove that more than one prt gene exists in the genomes of several LAB species, and that th pattern of prt genes is highly variable at the inter-strain level. For instance, L. helveticu possesses four unique types of CEP-namely prtH1, prtH2, prtH3, and prtH4-while rhamnosus CGMCC11055 possesses both prtP and prtR genes, although prtP was initial detected only in L. lactis and L. paracasei/L. casei species [4,19,[24][25][26].
Using the L. casei PrtP (AFJ15093.1) amino acid sequence as a query in BLAST search, we built a dataset consisting of 44 putative Prt-encoding genes in L. casei. Prote alignment with COBALT showed that 22 proteins exhibited the insert domain, and we categorized as PrtP (data not shown). The remaining 22 proteins lacked the inse domain, and were categorized as PrtR, with three different lengths of approximate 1500, 1800, and 2200 amino acids (data not shown). This result suggests that both prt and prtR paralogs exist in several L. casei genomes, including in the type strain ATC 393 T . Clustering analysis showed two distinct clades for L. casei PrtP proteins, referred as PrtP1 and PrtP2, while PrtR proteins were grouped into three different cluster referred to as PrtR1, PrtR2, and PrtR3 ( Figure 1).  . Sequences were aligned using the COBALT tool with default parameters, while the fast minimum evolution algorithm was used to determine the distance for tree construction. Labels represent protein IDs and strain codes. The image was created using the interactive tree of life (iTOL) v5.2.
"CD-search" in the Conserved Domain Database (CDD) predicted that PrtP1 and PrtP2 would share the same functional domains, except for the FIVAR domain (pfam07554)which is present in PrtP2, but not in PrtP1-and the CHB_HEX_C domain, which is present in PrtP1, but not in PrtP2 (Supplementary Figure S4). PrtR proteins shared the subtilisin S8 family domain annotated in the Pfam database (PF00082), but mainly differed from one another at the C-terminus (Supplementary Figure S4).

PCR Screening and Phylogenetic Analysis
Based on the survey of putative prt genes in 44 L. casei genomes, CEP-specific primer pairs were designed across the putative active domains of five different Prt-encoding genes, and were used to assess the type and distribution of prtP and prtR paralogs in L. casei PRA205 and 2006. Both strains were positive for PrtP2-and PrtR1-specific PCR assays, as they exhibited two bands with the expected lengths of approximately 2630 bp and 1723 bp, respectively, while no PCR products were obtained with the primer pairs targeting the prtP1, prtR2, and prtR3 genes ( Figure 2). This pattern of prt genes resembled the one we found in the genome of L. casei LC5, used as the reference strain in the subsequent in silico analyses. Recently, Asahina et al. [27] also identified both prtR and prtP genes in a highly proteolytic L. paracasei strain used as an adjunct starter in Gouda-type cheese production, whilst Vukotić et al. [26] identified three distinct proteinase genes-two prtP-type genes and one prtR-type-in Lacticaseibacillus zeae LMG17315. Considering the importance of the proteolytic system in assuring survival in highly protein-rich environments, the observed diversity of CEPs in mesophilic lactobacilli could arise from horizontal gene transfer, since the gain and loss of genes is a driving force in the LAB genome's adaptive evolution [28].
"CD-search" in the Conserved Domain Database (CDD) predicted that PrtP1 and PrtP2 would share the same functional domains, except for the FIVAR domain (pfam07554)-which is present in PrtP2, but not in PrtP1-and the CHB_HEX_C domain which is present in PrtP1, but not in PrtP2 (Supplementary Figure S4). PrtR proteins shared the subtilisin S8 family domain annotated in the Pfam database (PF00082), but mainly differed from one another at the C-terminus (Supplementary Figure S4).

PCR Screening and Phylogenetic Analysis
Based on the survey of putative prt genes in 44 L. casei genomes, CEP-specific primer pairs were designed across the putative active domains of five different Prt-encoding genes, and were used to assess the type and distribution of prtP and prtR paralogs in L casei PRA205 and 2006. Both strains were positive for PrtP2-and PrtR1-specific PCR assays, as they exhibited two bands with the expected lengths of approximately 2630 bp and 1723 bp, respectively, while no PCR products were obtained with the primer pairs targeting the prtP1, prtR2, and prtR3 genes ( Figure 2). This pattern of prt genes resembled the one we found in the genome of L. casei LC5, used as the reference strain in the subsequent in silico analyses. Recently, Asahina et al. [27] also identified both prtR and prtP genes in a highly proteolytic L. paracasei strain used as an adjunct starter in Gouda-type cheese production, whilst Vukotić et al. [26] identified three distinct proteinase genes-two prtP-type genes and one prtR-type-in Lacticaseibacillus zeae LMG17315. Considering the importance of the proteolytic system in assuring survival in highly protein-rich environments, the observed diversity of CEPs in mesophilic lactobacilli could arise from horizontal gene transfer, since the gain and loss of genes is a driving force in the LAB genome's adaptive evolution [28]. To exclude the possibility that putative SNPs prevented the annealing of prtP1-specific primers, an additional primer set was designed on the conserved regions of the prtP1 and prtP2 genes. Restriction analysis of the resulting PCR amplicons with the diagnostic endonuclease EcoRV confirmed the presence of the prtP2 gene (Supplementary Figure S5).
We sequenced the prtP2 and prtR1 PCR amplicons from L. casei PRA205 and 2006 for comparative analysis. A BLASTn search showed that the PRA205 and 2006 prtP2 nucleotide sequences were 93.75 and 93.71% identical to the L. casei ATCC 393 T prtP2 To exclude the possibility that putative SNPs prevented the annealing of prtP1-specific primers, an additional primer set was designed on the conserved regions of the prtP1 and prtP2 genes. Restriction analysis of the resulting PCR amplicons with the diagnostic endonuclease EcoRV confirmed the presence of the prtP2 gene (Supplementary Figure S5).
Proteinase genes (prtP and prtM) were also proven to be plasmid-encoded in Lactococcus lactis [31]. In S. thermophilus, the prtS gene is part of a genomic island flanked by conserved insertion sequence (IS) elements [32,33]. This mobile island has been shown to trigger gene gain and loss recombination events, which could be responsible for the . Phylogenetic trees of PrtP2 (A) and PrtR1 (B) amino acid partial sequences. The evolutionary histories were inferred using the neighbor-joining method. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates) are shown next to the branches (values > 0.5). The evolutionary distances were computed using the Poisson correction method, and are given as the number of amino acid substitutions per site. The rate variation between sites was modelled with a gamma distribution (shape parameter = 1). Evolutionary analyses were conducted in MEGA6.
Proteinase genes (prtP and prtM) were also proven to be plasmid-encoded in Lactococcus lactis [31]. In S. thermophilus, the prtS gene is part of a genomic island flanked by conserved insertion sequence (IS) elements [32,33]. This mobile island has been shown to trigger gene gain and loss recombination events, which could be responsible for the huge interspecies and intraspecies variability in proteolytic activity and prt gene patterns observed in lactobacilli. Remarkably, PRA205 and 2006 PrtR1 partial proteins diverged from those of the L. casei group, and clustered separately ( Figure 3B). Further experiments are required in order to establish whether the prtR1 gene is plasmid-encoded in L. casei PRA205 and 2006, or whether it belongs to a genomic island flanked by IS regions.

CEP Extraction and prt Gene Expression Profile
Peptide-rich media generally repress proteinase activities in several lactobacilli, including L. casei [34,35]. Therefore, we decided to assess the CEP activities of the strains PRA205 and 2006, grown until the stationary phase on the MCA medium. Extraction of CEPs was performed using Ca 2+ -free buffer, and the specific enzyme activity was determined with the chromogenic substrate succinyl-alanyl-alanyl-prolyl-phenylalanine-p-nitroanilide. Strain 2006 showed a specific CEP activity of 11.34 ± 0.45 U/mg, whereas for PRA205 it was 3.73 ± 0.03 U/mg. A lactate dehydrogenase activity assay was performed in order to check the membrane integrity during the extraction process. In the crude extracts, the LDH activity was undetectable (data not shown), proving that the proteolytic activity reported was due to the action of CEPs rather than intracellular peptidases. However, we cannot exclude the possible occurrence of cell-wall-associated peptidases. The presence of cell-wall aminopeptidases has been reported for L. lactis, L. helveticus, and Streptococcus cremoris [36][37][38]. However, El Soda et al. [39] did not detect any aminopeptidase, dipeptidase, or carboxypeptidase activity in the cell-wall-associated proteinase fraction obtained from L. casei.
To qualitative establish which prt gene is actively transcribed in MCA-grown cells, we carried out an RT-PCR assay. Figure 4A,B show that both strains (PRA205 and 2006) transcribed prtR1, but not prtP2, suggesting that the CEP activity detected above mainly came from PrtR1 proteinase when the strains were grown on MCA medium. This result differed from that found in L. rhamnosus strain CGMCC11055, which has both prtP and prtR genes, but only prtP gene transcription was detected when grown in a synthetic medium [19]. Moreover, the results are in disagreement with those reported by Pastar et al. [40], which showed that PrtR from L. rhamnosus EN1 was not released following incubation with Ca 2+ -free buffer.
huge interspecies and intraspecies variability in proteolytic activity and prt gene patterns observed in lactobacilli. Remarkably, PRA205 and 2006 PrtR1 partial proteins diverged from those of the L. casei group, and clustered separately ( Figure 3B). Further experiments are required in order to establish whether the prtR1 gene is plasmid-encoded in L. casei PRA205 and 2006, or whether it belongs to a genomic island flanked by IS regions.

CEP Extraction and prt Gene Expression Profile
Peptide-rich media generally repress proteinase activities in several lactobacilli, including L. casei [34,35]. Therefore, we decided to assess the CEP activities of the strains PRA205 and 2006, grown until the stationary phase on the MCA medium. Extraction of CEPs was performed using Ca 2+ -free buffer, and the specific enzyme activity was determined with the chromogenic substrate succinyl-alanyl-alanyl-prolyl-phenylalanine-p-nitroanilide. Strain 2006 showed a specific CEP activity of 11.34 ± 0.45 U/mg, whereas for PRA205 it was 3.73 ± 0.03 U/mg. A lactate dehydrogenase activity assay was performed in order to check the membrane integrity during the extraction process. In the crude extracts, the LDH activity was undetectable (data not shown), proving that the proteolytic activity reported was due to the action of CEPs rather than intracellular peptidases. However, we cannot exclude the possible occurrence of cell-wall-associated peptidases. The presence of cell-wall aminopeptidases has been reported for L. lactis, L. helveticus, and Streptococcus cremoris [36][37][38]. However, El Soda et al. [39] did not detect any aminopeptidase, dipeptidase, or carboxypeptidase activity in the cell-wall-associated proteinase fraction obtained from L. casei.
To qualitative establish which prt gene is actively transcribed in MCA-grown cells, we carried out an RT-PCR assay. Figure 4A,B show that both strains (PRA205 and 2006) transcribed prtR1, but not prtP2, suggesting that the CEP activity detected above mainly came from PrtR1 proteinase when the strains were grown on MCA medium. This result differed from that found in L. rhamnosus strain CGMCC11055, which has both prtP and prtR genes, but only prtP gene transcription was detected when grown in a synthetic medium [19]. Moreover, the results are in disagreement with those reported by Pastar et al. [40], which showed that PrtR from L. rhamnosus EN1 was not released following incubation with Ca 2+ -free buffer.

Biochemical Characterization of L. casei PRA205 and 2006 Cell-Envelope Proteinases' PrtR1
The proteolytic activity of PrtR1 was analyzed by modifying the incubation temperature and the pH of the reaction buffer in order to determine the best conditions of substrate hydrolysis. The PrtR1 extracted from each strain displayed the highest activity at 40 • C, and the enzymatic activity at 40 • C was taken as the reference (100%) for the calculation of residual (%) activity ( Figure 5A). A slight decrease in the PrtR1 catalytic activity was observed at 35 • C, whereas at 5 • C the residual activity rapidly decreased for PrtR1 extracted from L. casei 2006, whereas the proteinase extracted from L. casei PRA205 retained 48% of its activity. The high proteolytic activity at 5 • C seems to be a peculiar feature of L. casei PRA205 PrtR1, since the majority of previously characterized CEPs almost completely lost their activity at temperatures around 20 • C [7,8,19,41]. The effect of pH on the enzyme activity was analyzed by varying the pH of the reaction buffer from 4 to 8, at a constant temperature of 40 • C ( Figure 5B). All of the crude extracts achieved the highest activity at an optimal pH value of 7, which was used as the reference (100%) for the calculation of residual (%) activity. In particular, the residual activity gradually decreased at pH below and above 7 for both strains. However, when the PrtR1 activity was tested at pH 4, proteinase extracted from L. casei PRA205 showed a significantly higher residual proteolytic activity compared to that extracted from L. casei 2006.

Biochemical Characterization of L. casei PRA205 and 2006 Cell-Envelope Proteinases' P
The proteolytic activity of PrtR1 was analyzed by modifying the incuba temperature and the pH of the reaction buffer in order to determine the best conditio substrate hydrolysis. The PrtR1 extracted from each strain displayed the highest act at 40 °C , and the enzymatic activity at 40 °C was taken as the reference (100%) fo calculation of residual (%) activity ( Figure 5A). A slight decrease in the PrtR1 cata activity was observed at 35 °C, whereas at 5 °C the residual activity rapidly decrease PrtR1 extracted from L. casei 2006, whereas the proteinase extracted from L. casei PRA retained 48% of its activity. The high proteolytic activity at 5 °C seems to be a pec feature of L. casei PRA205 PrtR1, since the majority of previously characterized C almost completely lost their activity at temperatures around 20 °C [7,8,19,41]. The e of pH on the enzyme activity was analyzed by varying the pH of the reaction buffer 4 to 8, at a constant temperature of 40 °C ( Figure 5B). All of the crude extracts achi the highest activity at an optimal pH value of 7, which was used as the reference (1 for the calculation of residual (%) activity. In particular, the residual activity gradu decreased at pH below and above 7 for both strains. However, when the PrtR1 act was tested at pH 4, proteinase extracted from L. casei PRA205 showed a significa higher residual proteolytic activity compared to that extracted from L. casei 2006. This property could be of paramount importance because it suggests a pos exploitation of L. casei PRA205 in the production of bioactive peptides in fermented d foods such as cheese, yogurt, and fermented milk, which are characterized by low values. This observation could also explain the fact that L. casei PRA205 whole presented higher proteolytic activity in fermented milk than L. casei 2006 whole despite the latter strain showing the highest CEP specific activity [2,42]. Moreover fact that L. casei PRA205 PrtR1 retained almost 50% of its activity at low tempera suggests that the enzyme might also be active during the cold storage of fermented d foods, promoting the continuous release of bioactive peptides from caseins during storage.
The proteolytic activity of both PrtR1s was inhibited by K + (71.2 ± 1.5% and 6 6.1% of residual activity for L. casei PRA205 and L. casei 2006 PrtR1, respectively) Mg 2+ (80.9 ± 2.4% and 70.5 ± 6.4% of residual activity for L. casei PRA205 and L. casei PrtR1, respectively) ions when added at 2 mmol/L, and vice versa-PrtR1 activity enhanced by 2 mmol/L Ca 2+ ions (143.5 ± 5.3% and 183.0 ± 2% of residual activity f casei PRA205 and L. casei 2006 PrtR1, respectively), as reported for numerous o This property could be of paramount importance because it suggests a possible exploitation of L. casei PRA205 in the production of bioactive peptides in fermented dairy foods such as cheese, yogurt, and fermented milk, which are characterized by low pH values. This observation could also explain the fact that L. casei PRA205 whole cells presented higher proteolytic activity in fermented milk than L. casei 2006 whole cells, despite the latter strain showing the highest CEP specific activity [2,42]. Moreover, the fact that L. casei PRA205 PrtR1 retained almost 50% of its activity at low temperatures suggests that the enzyme might also be active during the cold storage of fermented dairy foods, promoting the continuous release of bioactive peptides from caseins during cold storage.

Caseinolytic Specificity of Extracted PrtR1, and PeptidomicsAnalysis of Casein Hydrolysates
The proteolytic activity of PrtR1 was assessed against αS1and β-caseins as substrates. Aliquots of samples were taken after 0, 8,24,30,48, and 56 h, and analyzed via SDS-PAGE and mass spectrometry experiments. As shown in Figure 6, both of the extracted PrtR1s completely degraded β-casein after 48 h of incubation under the optimal conditions, while αS1-casein was hydrolyzed at a lower rate, with the protein bands still present after 56 h ( Figure 6). 2022, 10, x FOR PEER REVIEW 1 by 1 mmol/L PMSF (86.5% and 99.9% of inhibition for L. casei PRA205 and L. casei PrtR1, respectively), suggesting that both PrtR1s were members of the serine prote family. The chelator EDTA at 1 mmol/L concentration also inhibited the PrtR1 ac (31.5% and 39.2% of inhibition for L. casei PRA205 and L. casei 2006 PrtR1, respectiv indicating that cations (i.e., Ca 2+ ions) are required for their activity.

Caseinolytic Specificity of Extracted PrtR1, and PeptidomicsAnalysis of Casein Hydroly
The proteolytic activity of PrtR1 was assessed against αS1-and β-casein substrates. Aliquots of samples were taken after 0, 8,24,30,48, and 56 h, and anal via SDS-PAGE and mass spectrometry experiments. As shown in Figure 6, both o extracted PrtR1s completely degraded β-casein after 48 h of incubation under the op conditions, while αS1-casein was hydrolyzed at a lower rate, with the protein band present after 56 h ( Figure 6). Based on the specific hydrolysis patterns of the αS1-, β-, and κ-caseins, CEPs classified in three different types: CEPs belonging to the PI type primarily hydr β-casein, which is cleaved into more than 100 different oligopeptides, ranging from 30 amino acid residues [6]. PIII-type CEPs can hydrolyze αS1-, β-, and κ-caseins eq well [6]. There is also a mixed CEP type named PI/PIII that cleaves β-casein in a si Based on the specific hydrolysis patterns of the αS1-, β-, and κ-caseins, CEPs were classified in three different types: CEPs belonging to the PI type primarily hydrolyze β-casein, which is cleaved into more than 100 different oligopeptides, ranging from 4 to 30 amino acid residues [6]. PIII-type CEPs can hydrolyze αS1-, β-, and κ-caseins equally well [6]. There is also a mixed CEP type named PI/PIII that cleaves β-casein in a similar manner to PI-type CEPs, and is also able to cleave αS1-casein, to a lesser extent, [10]. The above reported results indicated that β-casein was the preferential substrate over αS1casein, suggesting that PrtR1 extracted from L. casei PRA205 and L. casei 2006 belonged to the PI/PIII type. These results are inconsistent with the findings of Kojic et al. [18], who characterized a PI-type CEP in L. casei HN14, while they are consistent with the mixed PI/PIII-type CEP isolated from L. casei IFLP 731 [41].
Samples collected during the casein hydrolysis assay were submitted to UHPLC/HRMS analysis for peptide identification in order to study the casein breakdown patterns produced by the PrtR1 extracted from the selected strains. The full list of identified peptides at the different time points for the both βand αS1-caseins, along with the mass spectrometry data, are reported in Supplementary Tables S4-S7.
Crude extracts were incubated with β-casein for up to 48 h, and the peptidomic analysis revealed that a total of 116 and 119 peptides were released by PrtR1 extracted from L. casei PRA205 and L.  Figure S9), 57.7% of peptides released from β-casein and 61.4% of peptides released from αS1-casein were commonly found in the hydrolysates obtained from the two PrtR1s, suggesting a similar cleavage specificity.

Analysis of the β-Casein Cleavage Site Specificity
The analysis of the β-casein cleavage site specificity revealed the presence of 63 and 66 different cleavage sites in samples hydrolyzed by PrtR1 extracted from L. casei PRA205 and L. casei 2006, respectively ( Figure 7A,B); these represented 30.3% and 31.7% of all peptide bonds present in the proteins, respectively. A total of 54 cleavage sites were in common between the two PrtR1s, representing 85.7 and 81.8% of the total sites found in L. casei PRA205 and L. casei 2006 CEPs, respectively, suggesting that these CEPs had almost the same cleavage specificity.
The time-course analysis indicated that the hydrolysis of β-casein began at the hydrophobic C-terminal region. After 8 h of reaction, 80% and 76% of the cleavage sites produced by L. casei PRA205 and L. casei 2006 PrtR1, respectively, were located between residue 161 and the C-terminal amino acid V at position 209. At this point in time, no cleavage sites were found in the N-terminal region between residues 1 and 72 or in the central region situated between the amino acids at positions 106 and 134. Furthermore, considering the long region between amino acids 1 and 134, only two and four cleavage sites were detected in the L. casei PRA205 and L. casei 2006 PrtR1 hydrolysates, respectively. After 24 h of hydrolysis, the majority (59% for L. casei PRA205 PrtR1 and 62% for L. casei 2006 PrtR1) of the cleavage sites were still in the C-terminal region. At the end of the hydrolysis, the cleavage sites were distributed throughout the entire β-casein sequence for both of the PrtR1s; however, no cleavage sites were identified in the polyphosphorylated region between residues 8 and 28. It is worth noting that the majority of the previously characterized lactobacilli proteinases have a proven preference for hydrolyzing the C-terminal region of β-casein [6,43].  Furthermore, the cleavage probability at subsites P1 and P1 , along with the Kn coefficients, which measure the positive or negative effects of amino acids on the P1-P1 cleavage probability, were calculated (Table 1).  1 One code letter was used for amino acid nomenclature. The amino acid C is not present in the sequence of β-casein, while the amino acid W was omitted from the analysis because it occurs only once in the β-casein sequence. Numbers in brackets are the number of specific residues in the protein sequence. 2 The cleaved bonds are reported in Figure 7. 3 See the Materials and Methods section for the calculation of the %P1 and %P1 cleavage probability and the Kn coefficient. Amino acids with a positive impact on the cleavage probability are shown in bold.
PrtR1 extracted from both L. casei strains showed a marked preference for the amino acid M, for polar uncharged amino acids (S, Q, and N), and for positively charged amino acids (R and K) in the P1 position. In particular, the amino acids M and N showed the highest Kn values for both PrtR1s, therefore exerting the strongest positive effect on cleavage probability at the P1 subsite. The only difference between the two PrtR1s was the strong positive effect of amino acid A on PrtR1 extracted from L. casei 2006. In position P1, the amino acids I, T, P, V, and H, as well as the negatively charged (D and E) and aromatic (F and Y) amino acids, exhibited a strong negative effect on cleavage probability for both of the extracted PrtR1s.
The strongest positive effect on cleavage probability at subsite P1 was found for the small aliphatic amino acids A and G, as well as the amino acids H and D, for both of the extracted PrtR1s. In both cases, the amino acids D and A showed the highest Kn values. The strongest negative effect at position P1 was exhibited by the polar uncharged amino acids (especially P, Q, and T) and by the amino acid E. Some of the identified preferred amino acids have already been reported in previous studies; for example, the positive effect of amino acid Q in position P1 has been already described for CEPs characterized from L. delbrueckii CRL581, L. helveticus BGRA43, L. paracasei BGHN14, L. rhamnosus BGT10 and PRA331, and Lc. lactis NCDO763 [20,23,43,44]. The preference for this amino acid at the P1 subsite seems to be a common feature among the CEPs extracted from lactobacilli belonging to different species. In addition, the preference for N and M at position P1 has already been described for L. lactis NCDO763 and L. rhamnosus PRA331, respectively [23,44]. To the best of our knowledge, the strong positive effects exerted by the positively charged amino acids (R and K) in the P1 position and by the residue D in the P1 position have never been reported in the CEPs of LAB. The profile of cleavage-site specificity was slightly distinct from that previously reported for L. casei PRA205 [23]. These differences may be related to different experimental conditions, e.g., hydrolysis carried out with PRA205 whole cells and milk.

Analysis of the αS1-Casein Cleavage Site Specificity
Unlike β-casein, the hydrolysis of αS1-casein started at the N-terminal part of the protein. After 24 h of incubation, 50% and 45% of the cleavage sites were located in the sequence between residues 1 and 36 for PrtR1 extracted from L. casei PRA205 and L. casei 2006, respectively ( Figure 8A,B). Most of these cleavage sites were positioned in the fragment 1-23. Additional preferred cleavage sites at the beginning of the hydrolysis occurred in the sequence 90-110 for L. casei 2006 PrtR1, and in sequence 140-160 for both extracted CEPs. At the end of hydrolysis, the cleavage sites were mainly concentrated in the fragments 1-40 and 90-160, as already reported for the majority of characterized CEPs [6].

Identification of Bioactive Peptides Using the Milk Bioactive Peptide Database
The identification of bioactive peptides released by PrtR1 was carried out by searching all of the peptides found in the mass spectrometry analysis experiments against the Milk Bioactive Peptide Database (MBPDB). The hydrolysis of β-casein by PrtR1 of L. casei PRA205 and L. casei 2006 produced 14 and 17 functional peptides previously demonstrated to have several bioactivities, respectively ( Table 2).
The PrtR1 of both L. casei strains commonly released a total of 13 bioactive peptides, whereas 4 were uniquely produced by L. casei 2006 PrtR1 and 1 by L. casei PRA205 PrtR1. In contrast, the hydrolysis of αS1-casein by PrtR1 of L. casei PRA205 and L. casei 2006 generated 9 and 11 bioactive peptides, respectively, among which 8 were commonly released by the PrtR1 of both strains.
The identified bioactive peptides showed several functional properties, as 13 peptides were ACE inhibitors, 13 were antimicrobial, 6 were antioxidants, 5 had immunomodulatory activity, 1 showed dipeptidyl peptidase IV (DPPIV)-inhibitory activity, and 1 showed anxiolytic activity. Most of the identified bioactive peptides were tested in vitro for their bioactivity, whereas some of them also exhibited their activities in vivo. For example, the tripeptide LLY demonstrated a positive influence in vivo in Swiss albino mice against ethanol-induced oxidative stress [45]. The β-casein fragment KVLPVPQ showed antihypertensive activity in vivo in spontaneously hypertensive rats (SHRs) [10]. In addition, the bioactive peptide LYQEPVLGPVRGPFPIIV exerted its immunomodulatory activities by stimulating lymphocyte proliferation in rats [46].

Identification of Bioactive Peptides Using the Milk Bioactive Peptide Database
The identification of bioactive peptides released by PrtR1 was carried o searching all of the peptides found in the mass spectrometry analysis experim against the Milk Bioactive Peptide Database (MBPDB). The hydrolysis of β-case PrtR1 of L. casei PRA205 and L. casei 2006 produced 14 and 17 functional pep previously demonstrated to have several bioactivities, respectively (Table 2).  With the sole exception of the αS1-casein-derived peptides LGY and RPKHPIKHQGL PQEVLNENLLRFFVAPFPEVFGKEK, all the other identified bioactive peptides were found at the end of the hydrolysis, suggesting their resistance to further hydrolysis by PrtR1 (Table S3). In addition, 16 bioactive peptides were released only in the final periods of hydrolysis.
To exert their physiological effect, bioactive peptides should be stable under gastrointestinal conditions and resistant to the hydrolysis by gastrointestinal proteases. In a recent study, it was demonstrated that most of the peptides identified after in vitro gastro-pancreatic digestion contained 1-4 proline residues near the carboxylic ends of their