Glycoengineered Outer Membrane Vesicles: A Novel Platform for Bacterial Vaccines

The World Health Organization has indicated that we are entering into a post-antibiotic era in which infections that were routinely and successfully treated with antibiotics can now be lethal due to the global dissemination of multidrug resistant strains. Conjugate vaccines are an effective way to create a long-lasting immune response against bacteria. However, these vaccines present many drawbacks such as slow development, high price, and batch-to-batch inconsistencies. Alternate approaches for vaccine development are urgently needed. Here we present a new vaccine consisting of glycoengineered outer membrane vesicles (geOMVs). This platform exploits the fact that the initial steps in the biosynthesis of most bacterial glycans are similar. Therefore, it is possible to easily engineer non-pathogenic Escherichia coli lab strains to produce geOMVs displaying the glycan of the pathogen of interest. In this work we demonstrate the versatility of this platform by showing the efficacy of geOMVs as vaccines against Streptococcus pneumoniae in mice, and against Campylobacter jejuni in chicken. This cost-effective platform could be employed to generate vaccines to prevent infections caused by a wide variety of microbial agents in human and animals.


Results
Generation of geOMV displaying S. pneumoniae serotype 14 capsule (CPS14). The general strategy for the production of geOMV is represented in Fig. 1. In principle, our platform can be applied to any surface glycan, such as O antigens, capsules, exopolysaccharides, or glycans present in N-and O-glycoproteins. The initial steps in the biosynthesis of these glycans are common. These include the assembly of the glycan onto the undecaprenylpyrophosphate carrier and the flipping of the lipid-linked sugars across the inner membrane of Gramnegative or the cell membrane in Gram-positive bacteria. If these pathways are reconstituted in E. coli, the WaaL ligase transfers the glycan from the carrier to the lipid A, which is synthesized independently, and exported to the bacterial surface. OMVs displaying such antigens are then naturally produced. The biosynthetic pathways of type I capsular polysaccharides and O antigens only diverge once the polysaccharide has been translocated across the inner membrane ( Supplementary Data, Fig. S1a). WaaL is not glycan-specific and can transfer a variety of sugars onto the lipid A core 27 . We predicted that expressing the S. pneumoniae capsule in an E. coli strain lacking its own O antigen would result in an LPS consisting of capsular polysaccharide attached to the lipid A core. To obtain such a strain, we constructed an inducible plasmid (pNLP80) which expresses the capsule synthesis cps locus and includes all essential glycosyltransferases, the flippase (Wzx), and the polymerase (Wzy) required for the synthesis of the S. pneumoniae serotype 14 capsule (CPS14). The expression of CPS14 was evaluated in two different E. coli mutant strains, CLM37 28 and CLM24 29 (Fig. 2a). CLM37 harbours a mutation which interrupts the initiator glycosyltransferase (wecA) for O antigen and Enterobacterial common antigen (ECA) synthesis 28 . In Streptococcus capsule synthesis the initiating glycosyltransferase is WchA, which transfers a glucose residue to the Und-PP carrier. WecA and WchA attach different sugars to the lipid and therefore would compete for the carrier. We reasoned that a wecA mutant strain, such as CLM37, in which competition for Und-PP is minimized, should produce more CPS14 than the wild-type strain. To demonstrate the attachment of the CPS14 to lipid A, we employed strain CLM24 (waaL mutant) 29 . In this strain the glycan remains attached to the lipid carrier and CPS14 should not be transferred to the lipid A. LPS extractions and OMV obtained from the two E. coli mutants show the presence of CPS14 in the WecA mutant but not in the WaaL ligase mutant strain, which indicates that CPS14 is attached to the lipid A core and directed to the vesicles in strains such as CLM37 that carry an intact WaaL (Fig. 2a). The signal detected in the WaaL mutant strain corresponds to UndPP-linked CPS14. The geOMVs were visualized by transmission electron microscopy (TEM) (Fig. 2b). geOMV displaying S. pneumoniae CPS14 raise specific antibodies and are effective in OPA assays. To determine whether geOMVs from CLM37 strain displaying the CPS14 would generate an immunogenic response in a murine model, we injected 2 μ g of geOMVs per mouse (n = 10). Control groups immunized with either geOMVs lacking the capsule (empty geOMVs) or the commercially available Prevnar 13 ® , were included for comparison. Booster doses of geOMVs and Prevnar ® were administered on Days 14 and 28 and sera was collected weekly over a 42 day period. Collected sera of the immunized mice were assayed via Western blot analysis (Fig. 3a). Initial immunization with geOMVs did not elicit an IgG immunogenic response. However one week after the first booster dose (Day 21), high IgG titers in response to CPS14 were generated (Fig. 3a). The IgG response continued to be observed over the remaining course of the immunization schedule. To determine whether the immunogenic response was specific for serotype 14, we performed ELISAs with whole pneumococcal cells of serotypes 14 and 9 V (Supplementary Data, Fig. S1b,c). Sera from mice injected with geOMVs with CPS14 reacted only with S. pneumoniae serotype 14, thus demonstrating that the immunogenic response is specific for the glycoengineered serotype. Furthermore, at day 42, the levels of the IgG response of mice injected with pneumococcal geOMVs were comparable to the immunogenic response of mice injected with the commercially marketed pneumococcal vaccine Prevnar 13 ® (Fig. 3b) which suggested that the level of the protective effect could also be similar.  Opsonophagocytosis assays (OPA) are accepted as one the most reliable ways to evaluate the efficacy of pneumococcal vaccines 30 . OPA were performed to evaluate the efficacy of geOMVs against an S. pneumoniae serotype 14 infection. Day 42 sera obtained from the three groups of immunized mice (i.e., placebo geOMVs, pneumococcal geOMVs, and Prevnar 13 ® ) were employed. OPA showed a significantly increased killing effect using pneumococcal geOMVs compared to placebo geOMVs (Fig. 4a), which suggests that isolated pneumococcal geOMVs elicit a protective effect in vitro. In addition, varying mouse serum concentrations (5% and 20%) showed a similar killing effect between pneumococcal geOMVs and the Prevnar 13 pneumococcal vaccine (Fig. 4b), thus demonstrating that geOMVs have a similar protective effect when compared to the commercially available Prevnar 13 ® vaccine. geOMV reduce chicken colonization by C. jejuni. We next evaluated the use of geOMVs to vaccinate chickens against C. jejuni. We introduced a plasmid (pACYCpglBmut) that expresses the C. jejuni heptasaccharide N-glycan into E. coli EVV11. This strain carries a mutation in Wzy and therefore does not polymerize the C. jejuni glycan as the wild-type strain does. We isolated geOMVs from this strain and analyzed them by Western-blot employing an N-glycan specific antibody. An immunoreactive signal migrating around 15 kDa that was absent in control OMVs clearly indicates that the C. jejuni N-glycan was displayed in geOMVs ( Fig. 5a). To analyze the efficacy of the geOMVs containing the C. jejuni N-glycan, we tested the protective effect of these geOMVs in a C. jejuni chicken challenge model. Four groups of chickens (n = 8 for the positive control, n = 6 for the negative and experimental groups) were vaccinated orally with PBS, placebo geOMVs or Campylobacter geOMVs, and challenged with oral doses of PBS (negative control) or C. jejuni strain 81-176. Birds were monitored for 7 days before being euthanized and the amount of C. jejuni isolated from the cecum was determined for each bird. Birds that received PBS (Fig. 5b, group 2) and placebo geOMVs (Fig. 5b, group 3) showed similar levels of C. jejuni colonization after challenge. Chickens vaccinated with the N-glycan-containing geOMVs exhibited an almost 10 4 -fold reduction in C. jejuni colonization after challenge when compared to the naïve and placebo groups (Fig. 5b, group 4). In agreement with this data, the IgY levels of chicken that received the geOMV containing the glycan were higher than the ones that received the empty OMV (Fig. S1d).

Discussion
Conjugate vaccines are an effective way to create a long-lasting IgG immune response. However, besides being expensive to produce, these vaccines present a series of drawbacks. Large volumes of pathogenic cells need to be cultured to obtain the polysaccharides and the chemical crosslinking of the sugars to the protein is complex, with large batch to batch variability. The development of novel conjugate vaccines is very slow. Diverse serotypes not included in Prevnar 13 ® are prevalently being isolated in different geographical locations, and therefore novel technologies that can accelerate the development of multivalent vaccines are required. The generation of conjugate vaccines through exploitation of bacterial protein glycosylation systems is a very promising alternative, but not all the glycan chains are efficiently attached to proteins by the oligosaccharyltransferase PglB, which requires a HexNAc residue at the reducing end of the sugar chain 31 . PglL, the N. meningitidis O-oligosaccharyltransferase, can recognize a galactose but not glucose residue at the reducing end (Feldman, manuscript in preparation). Interestingly, about 90% of the S. pneumoniae strains contain glucose at the reducing end. Here we demonstrated that OMVs produced by glycoengineered E. coli expressing a glycan from unrelated bacterial pathogens, such as the CPS14 capsule from S. pneumoniae or the heptasaccharide derived from N-liked glycans from C. jejuni, can be effective in producing a significant immune response. In the case of S. pneumoniae, geOMVs induced an immune response, as measured by serum IgG levels and efficacy in OPA tests, which was similar to the one generated by the most widely used commercial conjugate vaccine. In the future, a multivalent vaccine could be generated by mixing geOMVs carrying capsules from different serotypes. Furthermore, geOMVs could complement the current conjugate vaccines, especially for serotypes corresponding to glycan structures for which the conjugation has not been solved. Although the Western blots together with the specific antibody response to CPS14 and the OPA assays suggest that the right structure of the CPS14 has been displayed in the geOMVs, future work will confirm the exact structure attached to lipid A. In the case of the C. jejuni geOMV vaccine candidate, the 4 log reduction in chicken colonization is, to our knowledge, unprecedented for this microorganism. However, the power of the vaccine could be increased if, in addition to engineering E. coli to produce pathogenic bacterial surface glycans, the geOMVs were modified to express antigenic membrane proteins to be directed to the vesicles. Previous studies have already explored incorporating antigenic proteins from pathogenic species (Neisseria, Streptococcus, Leishmania, Vibrio, and Yersinia) into vesicles derived from laboratory E. coli and Salmonella strains 17,32-35 . Furthermore, OMV produced in Salmonella carrying pneumococcal protein antigens showed promise in murine models 36 . The geOMV platform would enable the glycans and proteins to synergistically increase the vaccine immunogenicity capacity. This can also be important for cases like in S. pneumoniae, in which a protein antigen could expand the protection to serotypes not included in the vaccines.
One of the main concerns with vesicle based vaccines is the safety issues as OMVs contain endotoxic LPS 37 . LPS lipid A has been shown to provoke severe/lethal inflammatory responses in the host [38][39][40] . The OMV vaccines employed in humans were derived from Neisseria meningitidis. Several studies have analyzed the effect of modifying the lipid A to abrogate its interactions to obtain OMVs with LPS preparations tailored for human vaccine development [41][42][43] . For example, N. meningitidis strains lacking LpxM or LpxL also render a lipid A with minimal toxicity 41,44 . It has been shown that OMVs produced in E. coli can also be detoxified through modifications of the lipid A 37 . These modifications can be carried out through the action of lipid A deacylases, such as PagL 45 . The overexpression of B. pertussis PagL resulted in OMVs with lower endotoxic activity compared to wild type B. pertussis OMVs. Monophosphorylated lipid A species recently became the first new Food and Drug Administration-approved adjuvant in several decades 46 . Therefore, it might be possible to generate geOMVs with a perfect balance between reduced toxicity and optimal adjuvanticity by generating strains containing modifications in the levels of lipid A acylation, phosphorylation, and/or other modifications. A statistically significant reduction in colonization was observed in group 4 birds that received the Cj-N-glycan presenting OMVs when compared to either group 3 (*p-value = 0.045) or group 2 (**p-value = 0.032). No Campylobacter cells were detected in group 1, the detection limit based on the experimental setup was 200 cfu/gram cecal content. P-values were calculated by the one-way ANOVA test.

Material and Methods
Animal ethics statement. All experiments involving mice were conducted in accordance with the guidelines and policies of the Canadian Council on Animal Care and the principles set forth in the Guide for the Care and Use of Laboratory Animals. The animal protocol was approved by the Animal Welfare Committee of the University of Montreal (protocol #RECH-1523). Studies involving chickens were carried out in accordance with the protocol approved by the Animal Care and Use Committee at the University of Alberta using a 35 day challenge protocol.

Construction of pneumococcal glycan expression plasmid.
The cps gene cluster responsible for the synthesis of S. pneumoniae capsule serotype 14, CPS14, (excluding the regulatory genes) was cloned via a 3-way ligation method. Two separate fragments of the CPS14 locus were amplified using a high fidelity polymerase (iProof ™ , BioRad). Fragment 1 amplified the wchA-wchM region using the forward primer (5′-ATAGAGCTCATGGATAAAAAAGGATTGGAAAT-3′) and reverse primer (5′-TTAAGAAATTCATCCT CATACAA-3′). Fragment 2 amplified the wchM-wciY region using the forward primer (5′-CTGGTCAAC AAATATTAGAAAAA-3′) and reverse primer (5′ATACTCGAGATTCTTTCTGTAAACTCCAAAAA-3′). The underlined sequences denotes SacI and XhoI restriction enzyme sites, respectively. The PCR fragments 1 and 2 were both designed to include the native HindIII site within the wchM gene, thus, after a HindIII RE digest, the two fragments were compatible and reconstituted the wchM gene after ligation. Both fragments were successfully amplified and digested with SacI/HindIII (fragment 1) or HindIII/XhoI (fragment 2) and cloned into the plasmid vector pBBR1MCS-2, which was double digested with SacI and XhoI, generating pBBR1MCS-CPS14. The cloned CPS14 locus was confirmed by restriction enzyme digests and sequencing. For proper expression of the CPS14 locus in E. coli, the CPS14 locus was subcloned into pWSK129. The wchA-wciY CPS14 fragment was cleaved from pBBR1MCS-CPS14 using SacI and XhoI restriction enzymes. pWSK129 was digested with SacI and SalI restriction enzymes. As XhoI and SalI generate compatible sticky ends, the wchA-wciY fragment was ligated into pWSK129. The resulting plasmid, named pNLP80 was introduced in E. coli strains.
Isolation of geOMVs from E. coli expressing bacterial glycans. geOMVs were isolated following the protocol described by Haurat et al. 47 . Briefly, E. coli strains harbouring plasmids expressing genes involved in glycan synthesis (i.e., pNLP80, pEQ3, pACYCpglBmut) and the respective vector control plasmids were grown at 37 °C overnight with shaking in 50 mL of Luria-Bertani (LB) broth plus appropriate antibiotics. The following day, each culture was subcultured (1:100 dilution) into 1 L LB or Terrific Broth (TB) plus antibiotics and grown at 37 °C with shaking for 2 h (OD 600 ~ 0.2). Inductions of glycans were performed as follows: IPTG was added to a final concentration of 0.1 mM to induce expression of Streptococcus CPS14 (pNLP80) while the Campylobacter heptasaccharide biosynthetic gene cluster on pACYCpglBmut is constitutively expressed. The cultures were grown for an additional 26 h at 37 °C with shaking. After incubation, cells were harvested by centrifugation at 23,000 × g for 15 min and the supernatant was collected. The supernantants were filtered twice through 0.44 μ m and 0.2 μ m filters to remove any residual intact cells. The filtered supernatants were ultracentrifuged at 210,000 × g for 3 h at 4 °C. The geOMV pellets were collected and resuspended in buffered saline (PBS).

Analysis and quantification of geOMV. Presence of geOMVs containing Streptococcus CPS14 or
Campylobacter heptasaccharide was confirmed by Western blotting. Quantification of geOMVs was performed by measurement of KDO using a modified protocol described by Lee and Tasi 48 . Commercially available anti-rabbit CPS14 antibody (Statens Serum Institut, Denmark) was used as positive control (1:1000 dilution) to visualize the presence of the pneumococcal capsule.
Transmission electron microscopy. Isolated geOMVs were absorbed onto carbon-coated copper membrane grids for 3 min. Excess geOMVs were blotted away from membrane grid and samples were negatively stained with 2% (w/v) uranyl acetate (3 min). The grids were analyzed for presence of geOMVs using a Morgagni (FEI) transmission electron microscope (Biological Sciences Microscopy Facility, University of Alberta).
Murine studies with Pneumococcal geOMVs. Three groups of BALB/c mice (n = 10 per group, female, 4-6 weeks old, obtained from Charles River, WA) were immunized intraperitoneally. The test group was injected with 2 μ g of geOMVs isolated from E. coli CLM37 strain expressing CPS14 (pNLP80). A placebo control group received geOMVs isolated from E. coli CLM37 strain not expressing CPS14. The third group was injected with commercially available dose of Prevnar 13 (500 μ L as supplied). Sera were collected weekly via tail bleeds over a 42 day period and booster doses were administered on Days 14 and 28. Final bleed (Day 42) was via cardiac puncture. The presence of antibodies against CPS14 was analyzed by Western-blot using Odyssey imaging systems (LI-COR Biosciences, USA). geOMVs preparations digested with proteinase K were analyzed using 1:500 dilutions of the mouse sera.
Whole cell ELISAs. S. pneumoniae (serotype 9 V and 14, Statens Serum Institut, Denmark) was grown overnight on blood agar plates at 37 °C with 5% CO 2 aerobic conditions. The next day the cells were scraped from the agar plate, resuspended in 1× PBS, and heat inactivated by incubation at 60 °C for 2 h. The cells were diluted to OD 600 ~ 0.6/mL in PBS with protease inhibitor and 100 μ L were seeded into ELISA 96 well plates. The plates were incubated overnight at 4 °C. The following day, the wells were washed three times with 1× PBS before blocking with 2.5% skim milk for 2 h. The wells were washed three times with PBS and incubated with 100 μ L of mouse sera (1:500 dilution) was added to each well and the plates were incubated at room temperature for 1 h. After incubation, the wells were washed again three times with PBS and 100 μ L of IgG anti-mouse-alkaline phosphatase antibody was added to each well and incubated at room temperature for 1 h. After the secondary antibody incubation, Scientific RepoRts | 6:24931 | DOI: 10.1038/srep24931 the wells were washed again three times with PBS and 100 μ L of p-nitrophenyl phosphate was added to each well and the plates were incubated at 37 °C for 1 h followed by reading the absorbance at 405 nm on a BioTek ™ plate reader.
Bacterial suspension preparation. Isolated colonies on sheep blood agar plates of S. pneumoniae serotype 14 (Statens Serum Institut, Denmark ) were inoculated in 5 ml of Todd-Hewitt Broth (THB) (Becton Dickinson, Mississauga, ON, Canada) and incubated for 16 h at 37 °C with 5% CO 2 . Working cultures were prepared by transferring 0.1 mL of 16 h-cultures into 10 mL of THB which was incubated for 5 h. Bacteria were washed 3 times and resuspended in PBS to obtain an OD 600 value of 0.6, which corresponds to 2 × 10 8 colony forming units (CFU)/mL. Final bacterial suspension was prepared in complete cell culture medium to obtain a concentration of 6.25 × 10 4 CFU/mL. The number of CFU/mL in the final suspension was determined by plating samples onto Todd-Hewitt Agar (THA) using an Autoplate 4000 Automated Spiral Plater (Spiral Biotech, Norwood, MA).
Assay. Diluted whole blood (5 × 10 5 total leukocytes) was mixed with 5 × 10 3 CFU of S. pneumoniae (MOI of 0.01) and either 5% or 20% (v/v) of serum from control (placebo) or vaccinated mice in a microtube to a final volume of 0.2 mL. Microtubes were incubated for 2 h or 4 h at 37 °C with 5% CO 2 , with shaking. After incubation, viable bacterial counts were performed on THA using an Autoplate 4000 Automated Spiral Plater. Tubes with addition of naive mouse sera (5% or 20% v/v) or of commercial rabbit anti-S. pneumoniae type 14 serum (20% v/v) (Statens Serum Institut, Denmark), were used as negative and positive controls, respectively. The % of bacteria killed was determined using the following formula: % Bacteria killed = [1 − (bacteria recovered from sample tubes/ bacteria recovered from negative control tubes with naïve sera)] × 100. Final OPA conditions were selected based in several pre-trials using different incubation times and MOIs (data not shown). A more detailed procedure can be found in 49 .
Chicken vaccination and challenge. In general each group contained up to 8 leghorn birds (Poultry Research Facility, University of Alberta) that were randomly tested for the presence of Campylobacter on the day of hatch (Day 1) by plating cloacal swabs onto selective Karmali agar. In all cases no Campylobacter colonies were observed after 48 h of incubation under microaerobic conditions at 37 °C. Vaccination was performed by orally gavaging with 300 μ L of PBS containing 500 ng of geOMVs on Days 7 and 21. Control groups were gavaged with 300 μ L of PBS only.
Birds were challenged on day 28 by oral gavaging with either PBS (negative control) or with 300 μ L PBS containing 10 2 C. jejuni 81-176 cells. The challenge strain was prepared as follows: C. jejuni 81-176 was grown for 18 h on Mueller-Hinton (MH) agar and harvested with cold MH broth. Cells were washed twice with cold PBS and adjusted to an OD 600 of 1.2 (OD 600 of 1.2 equals 3 × 10 8 cell/mL). Serial dilutions in PBS were performed dependent on the final amount of cells that were administered. For example: 3 × 10 2 cells/mL [ = 1 × 10 2 cells per 300 μ L (= 1 dose)]. Cells were maintained on ice until used. Birds were maintained for an additional 7 days after challenge and then euthanized. Ceca were removed and the contents were adjusted to 1 mg cecal content per 1 mL with sterile PBS. Aliquots of 10-fold serial dilutions (in PBS) of the cecal contents were plated onto selective Karmali agar. CFU were determined after incubation of the plates for 48 h under microaerobic conditions. ELISA for C. jejuni N-glycan-specific antibodies. Blood samples were collected on Day 28 (vaccine response prior to challenge) and kept at 37 °C until a firm blood clot was formed. Samples were centrifuged (5 min, 18.000 × g, 4 °C) and the supernatants (sera) were transferred to fresh tubes. After addition of glycerol to a final concentration of 10%, sera were stored at − 20 °C until further use. For ELISA coating, Campylobacter N-glycan compounds (Cj-N-glycan) and their chemical conjugation to the protein carrier bovine serum albumin (BSA) was performed as described 50,51 . Maxisorb plates (Thermo Fisher) were coated with 500 ng of BSA-Cj-N-glycan conjugate per well for 18 h at 4 °C. After removal of unbound antigen the plate was blocked for 1 h at room temperature with PBS-Tween, 5% skim milk. After discarding the blocking solution 100 μ L of chicken sera diluted 1:10 in PBS-Tween, 1% skim milk was added to each well. Plates were incubated for 1 h at room temperature and washed 3 times for 5 min with PBS-Tween. After addition of 100 μ L of 2 nd antibody solution (anti chicken IgY, diluted 1:500 in PBS-Tween, 1% skim milk) and incubation for 1 h at room temperature wells were washed 4-times for 5 min with 100 μ L of PBS-Tween and developed using the 1-Step p-Nitrophenyl Phosphate (PNPP) assay following the instructions of the manufacturer (Thermo Fisher). Immuno-reactivity in each serum was determined after reading the plate at OD 405 in a plate reader.