Bacterial strains and growth conditions

A schematic overview of the bioengineering of probiotic E. coli cells expressing exogenous polysaccharides is shown (Fig. 1). In brief, a flagella-deficient clone of E. coli strain Nissle 1917 (EcNΔflhD) was used as “a mini factory” for MV production16,17. EcNΔflhD was further transformed by a low-copy plasmid pWSK129 (vector control)45 or the derivative plasmid pNLP80 which contains an entire locus responsible for pneumococcal CPS14 biogenesis10. Both pWSK129 and pNLP80 carry a kanamycin-resistance marker and an isopropyl-β-d(-)-thiogalactopyranoside (IPTG)-driven lac promoter. All E. coli strains were grown either in LB broth or on LB agar at 37 °C in aerobic conditions. Kanamycin sulfate (Fujifilm Wako Pure Chemical, Osaka, Japan) and IPTG (Fujifilm Wako Pure Chemical) were also supplemented at 50 µg/mL and 0.01 mM, respectively, when required. Two strains of Streptococcus pneumoniae, ATCC 700676 and KSP1094 were also used as representative serotype-14 strains of laboratory and clinical strains, respectively, to evaluate the reactivity of antibodies produced in immunized mice. Both strains were grown on brain heart infusion-based blood agar at 37 °C in an 5% CO2 incubator or in an anaerobic chamber (miniMACS, Don Whitley Scientific Ltd., Shipley, UK) in 80% N2, 10% H2, and 10% CO2.

Preparation and chemical analysis of LPS

LPS was isolated from freeze-dried E. coli cells by a hot-phenol extraction method46, with some modifications. In brief, E coli cells were collected from 250-mL culture for 16 h at 37 °C in aerobic conditions, and freeze-dried. The freeze-dried cells suspended with 10 mL of distilled water was mixed with 10-mL of 90% phenol, and stirred for 30 min at 68 °C. The water-phase was collected after centrifugation at 12,000 × g, and dialyzed against distilled water for 3 days. The dialyzed LPS solution was freeze-dried (crude LPS). The crude LPS was washed once with distilled water by ultracentrifugation at 100,000 × g to prepare purified LPS for chemical analysis. Analysis of sugar and fatty acid composition was performed by gas-liquid chromatography (GC-1014s, Shimadzu, Kyoto, Japan)47.

MV isolation

Bacterial culture supernatant was collected by centrifugation at 7190 × g for 30 min at 4 °C from 16-h culture of E. coli grown in LB broth supplemented with 1% glycine. The collected supernatant was filtered through a 0.22-µm-pore PVDF membrane to eliminate contaminated cells, and was subsequently ultracentrifuged at 100,000 × g for 2 h to yield MVs as sediments. Thanks to the peptidoglycan-weakened effect of glycine, supplementation with 1% glycine to LB broth resulted in an ~70-fold increase in the yield of MVs as assessed by protein amount when compared to MVs not induced with glycine16,17. On the other hand, the endotoxin activity of glycine-induced MVs was ~8-fold reduced compared to that of non-induced MVs16,17. The detoxified MVs were resuspended in 20 mM Tris-HCl (pH 8.0) and stored at −20 °C. The protein concentration of MVs was measured by Bradford assay48 using bovine serum albumin (BSA) as a standard.

Thermostability of CPS14+MVs

CPS14+MVs standardized at a concentration of 1 mg/mL with PBS (pH 7.4) were treated without or with heating at 100 °C for 30 min. Following ultracentrifugation of both non- and heat-treated CPS14+MVs, the resultant sediment and supernatant after ultracentrifugation were subjected to SDS-PAGE analysis. The morphology of CPS14+MVs without or with heat treatment was examined by FE-SEM and a Zetasizer (Zetasizer Pro, Malvern, UK). MV samples without or with heat treatment were also used for in vivo immunization experiments. The timeline is shown in Fig. S8D.

Surface roughness

MV surface topography was evaluated by an indirect, SEM micrograph-based profilometry by Gwyddion software ver. 2.58 as reported previously49 with some modifications. In brief, for analysis of MV surface morphology in detail, field emission scanning electron microscopy (FE-SEM) was operated at an accelerating voltage of 5 kV, a short working distance of 2.0- or 2.1-mm working distance, 106-fold magnification. The edge detection method was performed by cropping the center areas (30 × 30 nm2) of MVs with the sizes of 60 ± 10 nm in diameter. Fifteen images were randomly selected and analyzed in both MVs (vector control) and CPS14+MVs, respectively. In the roughness detection, the different values from adjacent pixels were obtained regardless of background and curved surfaces like MVs. Since the zero-crossing often highlights noise, the processing includes smoothing by the Laplacian of Gaussian and noise reduction by setting thresholds as follows: the full width at half maximum of Gaussian is 10 pixels, and the threshold amount of normalized root-mean-square error is 1.0. The surface roughness of MVs was calculated as a percentage of outline/total number of pixels.

Nano-flow cytometry (nFCM)

The size and concentration of MV samples were analyzed using a NanoFCM (NanoFCM Inc., Xiamen, China) according to the manufacturer’s instructions. Briefly, a cocktail of silica nanospheres (SiNPs) with four different diameters (68, 91, 113, and 155 nm) (NanoFCM Inc.) was used for calibration before nFCM analysis. PBS (buffer alone) was also analyzed as a background signal. MV concentration and size distribution were calculated using the nFCM software package, NanoFCM v. 2.0.

Quantitative dot-blot (QDB) analysis of pneumococcal CPS14 and E. coli O6 polysaccharide antigens

The amounts of pneumococcal CPS14 and E. coli O6 polysaccharides in whole cells and MVs were quantified by quantitative dot-blot (QDB) analysis. Bacterial cells were heat inactivated at 100 °C for 15 min, standardized at OD600 = 2.0, and serially diluted 2-fold with PBS (pH 7.4). MVs were standardized at 100 ng/µL (calculated as protein equivalent) were serially diluted 2-fold with 20 mM Tris-HCl (pH 8.0). For QDB of pneumococcal CPS14, purified CPS14 was used as the standard (#76943, Statens Serum Institut [SSI], Hillerød, Denmark). The CPS14 prepared at the concentration of 10 ng/mL were 2-fold serially diluted. For QDB analysis of E. coli O6 polysaccharides, heat-inactivated whole cells of the E. coli O6 reference strain50 were used as the standard. Whole cells standardized at OD600 = 2.0 were serially diluted 2-fold with PBS (pH 7.4). The amount of O6 contained in 1 CFU of the O6 reference strain was defined as one unit of the O6 amount. The signal intensity of each sample was quantified based on standard curves for CPS14 and O6, which were generated based on the average of signal intensities obtained in duplicate assays of purified CPS14 and the E. coli O6 reference strain, respectively, on the same membrane. Ten microliter of each solution was blotted onto PVDF membrane, and evaporated at 50 °C for 5 min. The blotted membrane was blocked with 1% skim milk (SM) in PBS containing 0.05% Tween 20 (PBST) for 16 h at 4 °C or for 2 h at 37 °C, and then was incubated with anti-CPS14 antiserum (#16753, SSI) diluted at 1:10,000 dilution. Horseradish peroxidase (HRP)-labeled anti-rabbit IgG (GE Healthcare, Buckinghamshire, UK) was used as the secondary antibody at 1:200,000 dilution. Chemiluminescence after addition of HRP substrate with high sensitivity (Western BLoT Hyper HRP Substrate, Takara-bio, Shiga, Japan) or with ultra-high sensitivity (Immobilon ECL Ultra Western HRP Substrate, Merck, Darmstadt, Germany) was quantified by a densitometry program of Fusion solo (Vilber Lourmat, Marne-la-Vallée, France). Relative CPS14 amounts in MV samples were calculated based on the standard curve of CPS14 standards. To define the amount of CPS or O6 per cell, the number of EcNΔflhD cells was estimated based on the turbidity of the bacterial culture after determining that the OD600 at 0.1 corresponded to 5 × 107 CFU. To define the amount of CPS or O6 per single MV particle, the number of EcNΔflhD MVs was estimated based on the total protein amounts of MVs, using data obtained from nano-flow cytometry (Table 1), in which one nanogram of total protein of the MVs (vector control) and CPS14+MVs corresponded to 2.11 × 106 and 2.38 × 106 particles (mean ± SD), respectively.

Subcellular fractionation

Subcellular fractionation of E. coli cells was performed by a standard protocol of a different solubilization technique by using 2% N-lauryl sarcosyl51.

Limulus assay

To quantify LPS, a limulus assay was performed using an Endospecy ES-50M kit (Seikagaku Co., Tokyo, Japan), according to the manufacturer’s instructions, with LPS from Escherichia coli O111:B4 (Sigma) as the standard.

Cryo-X-ray photoelectron spectroscopy (Cryo-XPS)

The chemical surface composition of isolated MVs was analyzed by using a multivariate curve resolution analysis of carbon 1 s spectra obtained from Cryo-XPS analyses52,53. Briefly, cryo-XPS spectra were recorded from frozen suspensions of OMVs in PBS, at liquid nitrogen temperature (-160 ̊C). Thereafter, C 1 s spectra for each sample were fitted using Matlab (Mathworks Inc) with spectral envelopes previously determined and published representing the profiles for three major substance groups, i.e., peptides (proteins or peptidoglycan), lipids or polysaccharides. Using this methodology the content of C atoms from these three building blocks could be predicted at the outermost portion (less than ~10 nm depth) of the MV surface.

SDS-PAGE and western blot

MV and whole-cell samples were separated by Tris-glycine SDS-PAGE using 12.5% polyacrylamide gels and stained with Coomassie brilliant blue (CBB). For Western blotting, gels were electroblotted onto PVDF membranes. Rabbit polyclonal antibodies against FliC flagellin54 were used at 1:2000 dilution. Rabbit polyclonal antibodies against CPS14 (#16753, SSI) were used at 1:10,000 dilution. Mouse monoclonal antibodies against MBP (E8032, New England Biolabs, Ipswich, UK) were used at a dilution of 1:10,000. Mouse monoclonal antibodies against lipid A-core oligosaccharide (Clone WN1 222-5, Hycult Biotech, Uden, Nederlands) was used at a dilution of 1:1000. Rabbit polyclonal antibodies against OmpA, DsbA, RodZ, and Crp were used as previously described55. HRP-labeled anti-rabbit IgG (GE Healthcare) was used as the secondary antibody at 1:200,000 dilution. Following addition of Western BLoT Hyper HRP Substrate (Takara-bio) or Immobilon ECL Ultra Western HRP Substrate (Merck, Darmstadt, Germany), chemiluminescence was visualized with Fusion solo (Vilber Lourmat). For detection of both lipid A-core oligosaccharide and CPS14 on the same membrane, fluorescence western blotting was performed using StarBright Blue (SBB) 520-labeled goat anti-rabbit IgG (Bio-Rad, Hercules, CA, USA) and SBB700-labeled goat anti-mouse IgG (Bio-Rad) as the secondary antibodies, in place of HRP-labeled antibodies. Fluorescence was visualized with Amersham ImageQuant 800 Fluor (Cytiva, Marlborough, MA, USA). Excitation was performed by use of a 460-nm blue light laser for both SBB520 and SBB700, and fluorescence emission was collected in the green range of the spectrum using a band pass filter of 525 ± 10 nm (emission maximum: 520 nm) for the SBB520 dye and in the infrared range of the spectrum using a band pass filter of 715 ± 15 nm (emission maximum: 700 nm) for the SBB700 dye.

Electron and immuno-electron microscopy

Morphological analysis of MVs was performed by FE-SEM with sub-nanometer resolution (Regulus8220, Hitachi High-Tech, Tokyo, Japan)16, with modified protocol of sample preparation to be optimized for MVs. For FE-SEM analysis, MVs standardized at 100 ng/µL with 20 mM Tris-HCl (pH 8.0) were placed on poly-l-lysine-coated coverslips immobilized in 4-well multi-dishes (#176740, Thermo Fischer Scientific, San Jose, CA, USA) and incubated for 150 min at 15–25 °C to allow the MVs to attach on the coverslips. Subsequently, MVs were fixed with admixture of 2% paraformaldehyde and 2.5% glutaraldehyde for 16 h. The fixed samples were dehydrated in graded acetone solutions, and dried in a critical point dryer using CO2 (CPD 300, Leica Microsystems, Wetzlar, Germany). The samples were then coated with osmium vapor using an osmium plasma coater and finally visualized with FE-SEM (Regulus8220, Hitachi High-Tech, Tokyo, Japan). The 3D structure of MVs was constructed from FE-SEM images by ImageJ software (version 1.44, National Institutes of Health) equipped with interactive 3D Surface plot analysis at the following setting parameters: smoothing 2.0; perspective 0.0; lightning 0.2.

For immuno-FE-SEM, bacterial cells or MVs immobilized on the coverslips were incubated with 200 µL of 1% BSA in PBS for 30 min at 15–25 °C. After washing once with 0.1% BSA in PBS, samples were incubated with CPS14 antibody (SSI) or normal rabbit antibody in 0.1% BSA in PBS for 1 h at 15–25 °C. After washing three times with 0.1% BSA in PBS, samples were incubated with 200 µL of colloidal gold-labeled secondary antibody prepared with 0.1% BSA in PBS for 1 h at 15–25 °C. In the present study, goat anti-rabbit IgG (H + L) secondary antibody conjugated with 12-nm colloidal gold (#111-205-144, Jackson ImmunoResearch, West Grove, PA, USA) and goat anti-mouse IgG (H + L) secondary antibody conjugated with 10-nm colloidal gold (EMGMHL10, BBI solutions, UK) were used for probing primary rabbit and mouse antibodies, respectively. After immunoreaction procedures, fixation, dehydration, drying and osmium plasma coating were sequentially performed in the manner described in the previous paragraph.

For immuno-TEM of E. coli and S. pneumoniae, the cells were fixed with 4% paraformaldehyde for 2 h. After washing three times with 0.1 M phosphate buffer, the pellet of cells was embedded in agar, dehydrated with in a series of ethanol concentrations (50, 70, 80, 90, 95, and 100%), and embedded in LR white resin (Nisshin EM Co. Ltd., Tokyo, Japan). Ultrathin-sections on nickel grid were incubated with 1% BSA in PBS for 30 min at 15–25 °C. After washing once with 0.1% BSA in PBS, the sections were incubated with CPS14 antibody (SSI) or normal rabbit antibody in 0.1% BSA in PBS for 1 h at 37 °C. After washing three times with 0.1% BSA in PBS, the sections were incubated with 12 nm colloidal gold-labeled anti-rabbit IgG antibody (BBI solutions) 0.1% BSA in PBS for 1 h at 37 °C. After immunoreaction, the sections were stained with 4% uranyl acetate and lead citrate, and then analyzed with a transmission electron microscope (H-7700, Hitach High-Tech). For assessment of the immunoreactivity of the serum and BALF samples from mice, FE-SEM images of 100 cells were randomly captured for each group at 5 × 105-fold magnification. A rectangular area defined as 180 × 250 nm2 was cropped from the center of a pneumococcal cell. The number of immuno-gold particles on the surface per cell was counted.

Animal experiments

All animal experiments were approved by National Institute of Infectious Diseases (NIID) Institutional Animal Care and Use Committee (Protocol nos. 118149 and 121046) and performed in compliance with the commitee guidelines. In Figs. 4A, 5A, E, 6A, S7A, S8B, and S8D, the experimental overviews are shown with the timelines of immunization and euthanasia. Six-week-old female BALB/c mice (Japan SLC, Inc, Hamamatsu, Japan) were subcutaneously immunized in the back twice with a 3-week interval with isolated MVs at a dose equivalent of 1 µg of protein containing 0.013 µg of CPS14, CPS14 (SSI) at different doses [0.01, 0.1, 1, and 10 µg, CPS14 equivalent] or the combination of MVs and CPS14, for which the injection volume was 0.09 mL. In some experiments, 23-valent pneumococcal polysaccharide vaccines (PPSV23, pneumovax23, Merck) and 13-valent pneumococcal conjugate vaccines (PCV13, prevenar13, Pfizer, Rockville, MD, USA) were subcutaneously administered in the back of female BALB/c mice in the same manner. Different doses of PPSV23 (0.5, 1.5, 4.5 µg, CPS equivalent) and PCV13 (0.043, 0.13, 0.39 µg, CPS equivalent) were administered to the mice. In an experiment with a 1-year follow-up for persistence of antibody production, immunization with PBS, PPSV23 (4.5 µg, CPS equivalent), PCV13 (0.039 µg, CPS equivalent), and CPS14+MVs (0.013 µg, CPS equivalent) was performed for 4 times at the following ages of mice: 6, 9, 12, and 36 weeks. In another experiment, 7-, 19, and 54-week-old female BALB/c mice were subcutaneously immunized 3 times at 3-week intervals in the back with PPSV23 (4.5 µg, CPS equivalent), PCV13 (0.039 µg, CPS equivalent), and CPS14+MVs (0.013 µg, CPS equivalent). Serum and BALF samples were collected 2 weeks after the final immunization. Serum, nasal wash and bronchoalveolar lavage fluid (BALF) samples were collected from mice and used for detection of CPS14-specific antibodies by ELISA. In the procedure of antigen coating, CPS14 antigen was coated at 125 ng per well onto ELISA plates. Alkaline phosphatase (AP)-labeled anti-mouse IgG (H + L) was purchased from Thermo Fischer Scientific, and used at 1:1000 dilution. AP-labeled anti-mouse IgM, IgE, IgA, IgG1, IgG2a, IgG2b, IgG3 were purchased from Southern Biotech (Birmingham, AL, USA), and used at 1:1,000 dilutions. Chromogenic development using para-nitrophenyl phosphate were recorded at absorbance at 405 nm at 15, 30, 60, 120 min with the plate reader Cytation5 (Biotek, Winooski, VT, USA). For T-cell response study, BALB/c mice were subcutaneously immunized in the back twice with a 3-week interval with PBS, PCV13 (0.039 µg, CPS equivalent), and CPS14+MVs (0.013 µg, CPS equivalent). At 3 weeks after the final immunization, the mouse spleens were collected and analyzed.

T-cell assessment using flow cytometry

Splenocyte suspensions were prepared by physically disrupting the spleen capsule and then passing the suspension through nylon mesh filter with a pore size of 70 µm. After removal of red blood cells by treatment with ACK lysis buffer, freshly isolated splenocytes standardized at 2 × 106 cells in RPMI1640 with 10% FBS were stimulated with 5 µL of a cell activation cocktail with brefeldin A (#423303, Biolegend, San Diego, CA, USA) in a CO2 incubator for 6 h. After harvesting stimulated splenocytes, surface staining was performed with the following panel of antibodies and reagents: PerCP anti-mouse CD3ε (clone 145-2C11, #100325, Biolegend), and FITC anti-mouse CD4 (clone GK1.5, #100405, Biolegend). After fixation and permeabilization with a Cyto-Fast Fix/Perm Buffer Set (#426803, Biolegend), intracellular cytokine staining was performed with PE anti-mouse IFN (clone XMG1.2, #505807, Biolegend) and APC anti-mouse IL-4 (clone 11B11, #504105, Biolegend). Stained splenocytes were subjected to flow cytometry analysis using FACS Canto II with FACS Diva software (BD Biosciences, Inc., Franklin Lakes, NJ, USA).

Statistical analysis

Statistical analysis was performed with a Mann–Whitney U-test or one-way analysis of variance (ANOVA), followed by Tukey’s multiple comparison test. p-values ≤ 0.05 were considered to indicate statistical significance.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.