An MPL ingredient originally isolated from MFGM beta-serum powder was generously donated from Fonterra Co-operative Group Ltd. (PL700 Concentrate, Auckland, NZ). The major phospholipid classes and fatty acid composition of this ingredient are described in Supplementary Table S1. For bacterial growth, de Man Rogosa Sharpe (MRS) broth (Sigma Aldrich, St. Louis, MO) supplemented with 0.05% L-cysteine (Thermo Fisher, Waltham, MA; referred to as MRSC) was prepared with or without 0.5% MPL added prior to autoclaving based on preliminary experiments, which determined 0.5% was suitable for bacterial growth. MRSC agar plates were also prepared using 1.5% agar.

Bifidobacterium longum subsp. infantis ATCC 15697 (B. infantis) was purchased from the American Type Culture Collection (ATCC, Manassas, VA). The strain was inoculated from glycerol stocks stored at −80°C to MRSC broth and covered with sterile mineral oil (Bio-Rad, Hercules, CA) for anaerobiosis at 37°C. Following two subcultures, the bacteria was adjusted to an optical density (OD₆₀₀) of 0.1 ± 0.01 corresponding to ~10⁸–10⁹ CFU/mL on a microplate reader (Multiskan GO, Thermo Fisher) and added at 10% of the final volume to MRSC with or without MPL (0.5% w/v). Bifidobacterium infantis was incubated with or without MPL in anaerobic conditions at 37°C for approximately 15 h to reach the late exponential growth phase. After treatment, each strain was washed 3× using sterile saline solution (0.85% NaCl (w/v), pH 7.0) and the suspension was again adjusted to OD₆₀₀ = 0.1 ± 0.01.

The effect of MPL on the survival of B. infantis during simulated digestion was tested following the static INFOGEST 2.0 digestion model and guidelines consisting of oral, gastric, and intestinal phases (23). For the oral phase, 5 mL of bacterial suspension (OD₆₀₀ = 0.1) was mixed with 4 mL of simulated salivary fluid, 1.5 mM CaCl₂, 75 U/mL salivary α-amylase, and distilled water up to a final volume of 10 mL. The oral phase was incubated at 37°C for 2 min with mixing on a rocking platform (VWR, Radnor, PA) at 95 rpm. For the gastric phase, 8 mL of simulated gastric fluid was added to the same tube and pH was adjusted to 3.0 using 5 M HCl. Then, 2,000 U/mL pepsin, 0.15 mM CaCl₂, and distilled water up to final volume of 20 mL were added and tubes were incubated at 37°C for 2 h with mixing. For the intestinal phase, 8 mL of simulated intestinal fluid was added to each tube after the oral and gastric phases and pH was adjusted to 7.0 using 5 M NaOH. Tubes were incubated with porcine bile salts (10 mM) for 30 min with rocking. Following bile salt incubation, pancreatin (trypsin activity 100 U/mL), 0.6 mM CaCl₂, and water up to a final volume of 40 mL were added and incubated for 2 h at 37°C. Preliminary experiments were conducted to determine the appropriate volumes of 5 M HCl and 5 M NaOH for pH adjustments. Throughout the simulated digestion, partially anaerobic conditions were maintained using a flush of N₂ gas prior to incubation. Individual tubes were prepared for each timepoint to reduce the effects of repeated sampling. At the end of incubation, the digesta was immediately centrifuged (3,500 × g, 4°C, 10 min), followed by removal of the supernatant, resuspension of the pellet in 5 mL of saline solution, and plating on MRSC agar plates using serial dilutions. All digestion reagents were purchased from Sigma.

The ability of MPL-treated B. infantis to adhere to human goblet-like cells was investigated using the HT29-MTX-E12 goblet-like cell line (European Collection of Authenticated Cell Cultures, Salisbury, United Kingdom). HT29-MTX cells were cultured in 25 mM high glucose Dulbecco’s Modified Eagle Medium (DMEM) containing 10% heat-inactivated fetal bovine serum, 1% penicillin–streptomycin (100 units/mL penicillin and 100 units/mL streptomycin), 1% non-essential amino acids (100X), and 1% 200 mM L-glutamine with media replacement every 48 h. Intestinal cells were plated in 12-well plates using a seeding density of 5 × 10⁵ cells/well and incubated at 37°C in a 5% CO₂ humidified atmosphere for 21 days post-confluency. Media was replaced with serum- and antibiotic-free media 1 day prior to all experiments. All cell culture reagents were purchased from Thermo Fisher (Gibco).

Ten mL of diluted bacterial suspension (OD₆₀₀ = 0.1) was transferred to a sterile tube, centrifuged (3,500 × g, 4°C, 10 min), and resuspended in an equivalent volume of serum- and antibiotic-free cell culture medium. HT29-MTX cells were incubated with 1.5 mL of this suspension for 3 h to mimic gastrointestinal transit time. Bacteria control samples without HT29-MTX cells were simultaneously incubated to account for changes in bacterial growth resulting from incubation in high glucose media. Intestinal cells were washed 4X with sterile PBS to remove non-adherent bacteria followed by addition of 1% Triton X-100 (v/v) at 4°C for 30 min to lyse intestinal cells and release adherent bacteria. Samples from each well were centrifuged (8,000 × g, 4°C, 10 min) and washed twice before resuspending in 1.5 mL of saline solution. Control bacteria and adhered bacteria were plated on MRSC agar plates to determine the proportion of adhered bacteria using the following equation: % adhesion = (CFU/mL of adhered bacteria/CFU of control bacteria) × 100%. Adhesion assays were carried out in 3 independent experiments, each containing 4 replicates per sample.

The ζ-potential, or surface charge, of the bacteria was measured as previously described by Ortega-Anaya et al. (24). Briefly, measurements of the bacterial cells (OD₆₀₀ = 0.1) in sterile PBS were recorded using a NanoBrook 90 Plus instrument (Brookhaven Instruments, Holtsville, NY). The refractive indices (RI) for gram-positive Bacillus thuringiensis bacterial cells (1.528) and RI for water (1.33) were used (25). Ten measurements per sample from 3 independent experiments were collected at 25°C using the Smoluchowski approximation.

The bacterial cells with or without MPL were prepared and washed as previously described. A concentrated suspension was prepared in PBS that was further mixed with sterile melted low-gelling temperature agarose solution (4% at 60°C). After solidification, small squares (1–2 mm²) were cut using a scalpel and were further fixed with 3% formaldehyde and incubated for 24 h. After carefully removing the supernatant, the samples were washed 5 times with PBS under gentle agitation (5 min each). Transmission electron microscopy (TEM) analysis was performed at the Campus Microscopy and Imaging Facility (CMIF) of The Ohio State University (OSU). Each sample was fixed with 2.5% glutaraldehyde for 2 h, post-fixed with 1% osmium tetroxide in 0.1 M phosphate buffer for 2 h, and then stained with 1% uranyl acetate for 1 h at room temperature. After dehydration in a graded ethanol series (50, 70, 80, 90, 100), they were embedded in Eponate 12 resin. The samples were sectioned at 70 nm using an ultramicrotome (EM UC7; Leica Microsystems, Vienna, Austria) and stained with 2% uranyl acetate and Reynold’s lead citrate. Then, they were visualized using a FEI Tecnai G2 Spirit microscope (FEI Company, United States) operated at 80 kV.

To determine the effect of MPL on bound polysaccharide production during bacterial growth, 20-μL of bacterial suspension in saline solution (OD₆₀₀ = 0.1) was added to 180 μL MRSC with or without MPL in triplicate in a sterile 96-well plate. Each well was covered with 100 μL of sterile mineral oil for anaerobiosis. The OD₆₀₀ was recorded every 2 h over 24 h in a 37°C microplate reader with 5 s of low-speed shaking prior to each read. Concurrently, 1 mL of the OD-adjusted bacterial suspensions were inoculated into 9 mL of MRSC with or without MPL under anaerobic conditions at 37°C for quantification of bound polysaccharides at 0, 4, 8, 12, 20, and 24 h. Separate tubes were prepared for each time point. At the end of incubation, the tubes were centrifuged (3,500 × g, 4°C, 10 min) and the pellets were washed with saline. Each pellet was subject to alkaline treatment using 2 mL of 2 M NaOH and agitated overnight at 25°C on a rocking platform to remove the bound polysaccharides while rendering the bacterial cells intact (26). The samples were centrifuged (7,000 × g, 4°C, 30 min) and the polysaccharides (supernatant) were precipitated with 2 volumes of EtOH with overnight incubation at −20°C. After centrifugation (9,000 × g, 4°C, 30 min), the crude polysaccharide pellets were resuspended in 0.5 mL MilliQ ultrapure water (18.2 MΩ • cm) and subjected to total sugar and protein quantification using the phenol sulfuric acid assay (27) and a Pierce micro-BCA kit (Thermo Fisher), respectively.

Using MRSC agar plates with and without MPL, B. infantis was grown under anaerobic conditions for 48 h at 37°C using anaerobic gas packs. A loopful of bacteria was diluted and resuspended in 20 μL of EtOH. Then, 1 μL of this suspension was added to the diamond-ATR crystal of a 4500a Fourier Transform Infrared (FTIR) Spectrometer Portable Unit (Agilent Technologies, Inc., Santa Clara, CA) and dried using vacuum. This process was repeated several times until sufficient sample was applied to the instrument. The spectra of samples between 4,000 and 700 cm⁻¹ were recorded using the MicroLab software (Agilent) using a resolution of 4 cm⁻¹ and 128 scans. The spectra obtained were classified and analyzed using a SIMCA model with 2^nd derivative data deconvolution on Pirouette software V4.5 (Infometrix Inc., Woodville, WA).

Bound polysaccharides were extracted from bacterial pellets of a 200 mL culture in the presence or absence of 0.5% MPL following the procedure of Ferrari et al. (26). Hydrolysis of extracted sugars was executed by resuspending the freeze-dried powder to a concentration of 12.5 mg/mL using sonication (10 min, 37°C). Resuspended sample was mixed with a half volume of 0.5 M H₂SO₄ and heated at 100°C for 40 min. The sample was then adjusted to pH 7 using 5 M NaOH and diluted to a final concentration of 2 mg/mL in reverse osmosis (RO) water. For quantification of monosaccharides by HPLC, 400 μL of sample was further diluted in 600 μL ACN. Separation of neutral monosaccharides was achieved using an XBridge Amide 3.5 μm (4.6 × 250 mm) column (Waters) equipped with a guard column of the same filling on a Dionex Ultimate 3000 HPLC system with a charged aerosol detector (CAD; Thermo Fisher) (28). Solvent A consisted of 50% IPA, 0.2% TEA, and 25 mM ammonium acetate in water and solvent B consisted of 90% ACN, 0.2% TEA, and 25 mM ammonium acetate. A flow rate of 0.5 mL/min, column temperature of 80°C, and isocratic elution of 90% B for 30 min were followed by column washing and re-equilibration. The CAD evaporator temperature was set to 35°C with a data collection rate of 10 Hz and filter constant of 1.0. All analytical standards and HPLC-grade solvents were purchased from Sigma. The final sugar quantification was normalized by CFU/mL.

Surface proteins were extracted from a 30 mL culture of B. infantis in MRSC with or without MPL after washing in saline solution. Six mL of 5 M LiCl at 37°C for 60 min with agitation was used to remove non-covalently bound proteins (29, 30). Samples were centrifuged at 10,000 × g for 20 min at 10°C and the supernatant was subsequently dialyzed in Milli-Q water at 4°C to remove LiCl using cellulose dialysis tubing (MWCO 12 kDa, Sigma). Surface protein extracts were then concentrated using a Vacufuge Plus concentrator (Eppendorf, Enfield, CT) at 45°C and total protein was determined using a micro-BCA kit.

The surface protein profile of B. infantis in response to MPL was visualized using SDS-PAGE by loading 30 μg protein per sample on a Mini-Protean TGX Stain-free gel (4–20%) and applying 90 V for 15 min followed by 180 V for 30 min. Five μL of Precision Plus Unstained Protein Standard was used as a reference. Gels were imaged using a ChemiDoc MP Imaging System and densitometric analyses of the gels were performed using Image Lab Version 6.1 to semi-quantitatively determine differences in bacterial surface proteins. All reagents and equipment used for gel electrophoresis were purchased from Bio-Rad. Bands with the greatest increase or decrease in expression (at least 2-fold change) were excised, digested in gel using trypsin, and sequenced using Capillary-LC/MS/MS at the Campus Chemical Instrument Center (CCIC) Mass Spectrometry and Proteomics Facility at OSU. Data were searched using Mascot Daemon by Matrix Science V2.7.0 (Boston, MA) via Proteome Discoverer V2.4 (Thermo Scientific) and the database searched against the most recent UniProt databases. A decoy database was also searched to determine the false discovery rate (FDR) and peptides were filtered accordingly at 1% FDR. Proteins identified with at least two unique peptides were considered as a reliable identification. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE (31) partner repository with the dataset identifier PXD042186. Although the proteomics was not performed in replicates, we did run 3 independent replicates of the SDS-PAGE gels and the band patterns exhibited were compared using the BioRad ImageLab software. The two bands of interest to us were present in the 3 independent replicates with similar experimental molecular weight.

All experiments were conducted at least in triplicate. Data presented are represented as the mean ± SEM where appropriate. Statistical analysis for ζ-potential, digestion survival, and monosaccharide composition were conducted using a Student’s t-test. Statistical analysis for adhesion was conducted using Mann Whitney U-tests. Differences of p < 0.05 were considered statistically significant. All statistical analysis was performed in GraphPad Prism V9.4 unless otherwise noted.

Under simulated digestion conditions, MPL treatment significantly improved the survival of B. infantis at the end of the intestinal phase (p < 0.05) from 12% to 19% (Figure 1A). However, no changes in survival in the earlier digestive phases (oral and gastric) were observed in the presence of MPL. These findings suggest the role of MPL in mitigating bile-induced stress exhibited in the intestinal phase, while having no effect on low pH tolerance from gastric conditions.

As a follow up to these changes in survival, the influence of MPL bifidobacterial adhesion was tested using a cell culture model, HT29-MTX, which is a mucin-producing clone of HT29 cells and models goblet-like cells of the intestinal epithelium. This cell line permits the potential study of interactions between bacteria and mucus that are resemblant of the intestinal barrier. After MPL treatment, the bacteria were applied to the goblet-like cells for 3 h under aerobic conditions. The 3-h exposure resulted in a significant decrease in the adherence of B. infantis from 42.4 to 6.77%, when previously incubated with MPL (Figure 1B). This decrease corresponds to an approximately 0.2-log reduction. Using the same cell line, four inflammatory cytokines (IL-6, MCP-1, TNFα, and IL-4) were also measured in response to control or MPL-treated B. infantis and decreasing trends of expression were observed for pro-inflammatory mediators; however, no significant differences were detected (data not shown).

The surface charge of gram-positive bacteria is generally attributed to the presence of cell surface proteins and carbohydrates, such as wall teichoic acid (WTA), lipoteichoic acid (LTA), and capsular polysaccharides, among others, that can be induced by changes in the environment and by nutritional stimuli (32). To estimate the surface charge of B. infantis in the presence and absence of MPL, ζ-potential was measured. The ζ-potential of B. infantis was found to be negative and decreased significantly upon exposure to MPL from −17.3 to −24.3 mV (Figure 2A). The decreased, or more negative, surface charge of B. infantis in response to MPL suggests attachment of MPL at the surface or an alteration in the composition of the cell surface, potentially through changes in surface carbohydrates and/or proteins.

TEM was used to investigate whether the observed changes in surface charge impacted the surface morphology of these bacteria. Figure 2 displays the TEM micrographs of B. infantis in the presence and absence of MPL. As indicated by the blue arrow in Figure 2B, the outer surface of B. infantis appears thicker upon MPL treatment, which resembles an increase in bacterial surface polysaccharides.

Considering the effect of MPL on the surface charge and morphology of B. infantis, an initial screening for features affected by the MPL treatment was conducted using FTIR. Bacterial colonies were resuspended in ethanol to reduce the interference of matrix components on the samples. The second derivative of the raw spectra was obtained to deconvolute overlapping bands while conserving the meaning and interpretation of specific wavenumbers. An assessment by soft independent modeling by class analogy (SIMCA) was used to identify the interclass distance and important bands with high discriminating power between the class “B. infantis” and the class “B. infantis + MPL.” Using three principal components, a Cooman’s plot was constructed to depict the orthogonal differences of each sample to the classification models (Figure 3A). The vertical red line indicates the cutoff point with samples being classified as “B. infantis” if to the left of the line. The horizontal red line indicates the cutoff point with samples being classified as “B. infantis + MPL” if below the line. Samples generally fit well within their respected classes, with a few exceptions belonging to either class (bottom, left quadrant; Figure 3A). Regardless, an interclass distance of 5.8841 was obtained, indicating the FTIR spectra of the classes were significantly different as values above 3 indicate statistical significance. To determine the most interesting bands generating these differences, the magnitude of the discriminating power for each band is represented in Figure 3B. The bands with the greatest discriminating power include 1,730, 1,663, 1,201, and 1,048 cm⁻¹. The assignments and interpretations of these bands are listed in Table 1. The data collectively shows that MPL treatment is associated with changes in the lipid profile, suggesting the accumulation of MPL and changes in the bacterial surface through carbohydrate polymers.

We further explored the production of bound polysaccharides in the presence and absence of MPL by carbohydrate quantification across the growth of B. infantis. Figure 4A represents the quantification of bound polysaccharides over time through monitoring the B. infantis growth curve using OD₆₀₀. Both control and MPL-treated bacteria exhibited similar exponential phase lengths (0 to 8 h); however, the carrying capacity was greater for control bacteria compared to the MPL-treated bacteria represented by the differences in OD₆₀₀ in the stationary phase. During the exponential phase, no differences between control and MPL-bound polysaccharide production were observed. At the start of the early stationary phase (12 h), bound polysaccharide production peaked in both treatments but was higher in MPL-treated bacteria despite the overall lower cell count based on OD₆₀₀ readings. This trend continues until 20 h, when production begins to decline likely due to the lack of remaining available nutrients, and eventually returns to similar levels as the control samples at 24 h.

Since carbohydrates from peptidoglycan and glycoproteins could possibly contribute to the observed increase in bound polysaccharides, the protein content was also monitored and compared to trends of bound polysaccharide production (Figure 4B). The data suggest that the observed increase in bound polysaccharides at 12 h is likely not related to the peptidoglycans and glycoproteins as the increase in protein did not mirror the increase in bound polysaccharides production. These findings were further corroborated by quantification of the neutral sugars galactose (D-Gal) and glucose (D-Glu), which are the principal sugars comprising B. infantis ATCC 15697 polysaccharides, using HPLC-CAD. Total Gal and Glu increased from 63.7 g/100 g powder to 110 g/100 g powder and retained similar D-Gal to D-Glu ratios near 1.6:1 with MPL treatment (Table 2). These data support the role of MPL in enhancing bound polysaccharides production in B. infantis.

For an additional level of understanding how MPL influences bifidobacteria surface properties, the effect of MPL on the surface protein profile was assessed as proteins also modulate bifidobacteria-host interactions within the GIT, and influence stress tolerance and GIT survival (37). Using SDS-PAGE separation (Figure 5), extracted surface proteins of control and MPL samples were compared using a densitometric assessment. Two bands with the greatest relative changes were excised and sequenced, as identified in Figure 5. These proteins were identified as a peptide ABC transporter substrate-binding protein with an experimental molecular weight (MW) of 57 kDa that decreased 2.02-fold in the presence of MPL and a trypsin-like peptidase domain-containing protein with an experimental molecular weight (MW) of 50 kDa that increased 3.36-fold (Table 3). Due to the nature of the SDS-PAGE separation, the discrepancy in experimental and theoretical MW of the trypsin-like peptidase domain-containing protein could be due reduction of disulfide bonds between subunits leading to a 50 kDa experimental MW compared to the theoretical 60 kDa MW. Furthermore, we noticed a tendency in the overlap between the sequenced peptide matches from this band and the accession protein toward the C-terminus.