Design and Immunogenicity of a Nanoparticle Combination Vaccine for Salmonella in Poultry

ABSTRACT

Salmonella enterica serovars (Salmonella) are common causes of bacterial gastroenteritis (salmonellosis) often associated with the consumption of poultry products. The control of Salmonella in broiler chickens is difficult because colonized birds may carry this pathogen undetected into the slaughterhouse and cause meat contamination. Poultry vaccines can mitigate colonization, and one promising production method is via plants. Iron is an essential nutrient for Salmonella; therefore, our goal was to exploit iron uptake systems as vaccine targets that could starve this pathogen of iron and thus prevent it from colonizing chickens. One challenge with this approach is that Salmonella possesses several iron acquisition proteins that are redundant in function. Therefore, we created nanoparticle‐based plant‐produced vaccine candidates against four different Salmonella iron acquisition proteins to create a multivalent vaccine that may be more effective than targeting a single protein. Extracellular antigenic peptides from each of these proteins were fused in tandem to the N‐terminus of the self‐assembling Brucella spp. lumazine synthase. These fusion constructs were then expressed in Nicotiana benthamiana plants. The resulting recombinant proteins were extracted, purified and characterized. The vaccine candidates administered individually or in combination elicited specific antibodies in mice and bound to the surface of Salmonella, indicating their potential in preventing Salmonella colonization in poultry.

1. Introduction

Contamination of poultry products by Salmonella represents a significant food safety and public health threat and is an important cause of economic losses for the poultry industry. Effective pre‐harvest control of Salmonella in poultry is crucial to reduce its impacts (Mouttotou et al. 2017). Salmonella vaccines for poultry present a cost‐effective and prompt strategy, given the lower costs and regulatory complexities associated with animal vaccines (FAO and WHO 2023).

Subunit Salmonella vaccines, composed of surface‐exposed proteins as antigens crucial for bacterial survival, offer potential advantages in immune recognition compared to internal antigens. Studies utilizing peptides from surface‐associated proteins linked to the virulence of Salmonella such as the FliC protein, type I fimbriae and type III secretion systems have shown some promising results but have limited efficacy (Ijaz et al. 2025; Toyota‐Hanatani et al. 2009; Wisner et al. 2011). Thus, there is still a need for more effective Salmonella vaccines for broiler chickens.

Salmonella displays TonB‐dependent outer membrane transporters (TBDTs) linked to the acquisition of iron, an essential element for bacterial survival and virulence. These TBDTs, crucial for iron uptake, are potential vaccine antigen targets. Pathogens, such as Salmonella, employ sophisticated mechanisms, employing siderophores to acquire iron. These siderophores, characterized by their low molecular weight and strong affinity for ferric iron (Fe³⁺), are categorized as catecholate, hydroxamate, α‐hydroxycarboxylate and mixed types. Iron‐regulated TBDTs play a key role in transporting Fe³⁺‐loaded siderophores across the bacterial outer membrane (Baumler et al. 1998; Wilson et al. 2016).

Enterobacteria primarily produce enterobactin as their main siderophore, which belongs to the catecholate type. In Salmonella, two TBDTs, namely FepA and IroN, facilitate the transport of enterobactin‐Fe³⁺ and its glucosylated version, salmochelin‐Fe³⁺, respectively, across the outer membrane (Baumler et al. 1998; Muller et al. 2009; Rabsch et al. 1999). Another TBDT, CirA, is specialized in transporting monomers, dimers and linear trimers of the siderophore 2,3‐dihydroxybenzoylserine, along with colicins and microcins (Maurakis et al. 2025). FhuA, identified as ferric hydroxamate uptake protein A, is another TBDT capable of mediating the uptake of Fe³⁺ bound to the siderophore ferrichrome, which is of the hydroxamate type (Buchanan et al. 2007; Wang et al. 2018).

Chekabab and colleagues examined the roles of catecholate‐iron and hydroxamate‐iron transporters, including IroN, FepA and FhuA, in the growth and survival of S. enterica serovars Enteritidis and Typhimurium (Chekabab et al. 2019). Deletion mutants of these transporters were generated and compared to wild‐type strains. Results showed that IroN and FepA, both catecholate‐iron transporters, contribute to bacterial growth in iron‐poor conditions, with double deletions of fepA and iroN resulting in reduced siderophore production and impaired growth. Meanwhile, FhuA, a hydroxamate‐iron transporter, plays a crucial role in iron uptake and transport, particularly in ΔiroNΔfepA mutants, where it facilitates the uptake of hydroxamate Fe³⁺.

Considering the redundancy of TBDTs and their varied functions in iron uptake, it is reasonable to suggest that selectively blocking individual TBDTs may be insufficient to impede the overall iron acquisition crucial for the growth and survival of Salmonella. Thus, this study strategically targeted four TBDTs—CirA, FepA, FhuA and IroN—as antigens for vaccine development, aiming to hinder overall iron acquisition of Salmonella. TBDTs, like other membrane proteins, often possess intricate structures involving multiple transmembrane domains and complex tertiary structures, rendering their extraction and recombinant expression challenging (Grisshammer 2006; Shamriz, Kaldis, et al. 2026). Earlier studies indicated that the extracellular loops of TBDTs contain highly conserved antigenic peptides (T and B cell epitopes) (Zarate‐Bonilla et al. 2014).

We previously produced extracellular loops of FepA fused to Brucella sp. lumazine synthase (BLS) (Shamriz et al., submitted). BLS is an enzyme involved in riboflavin biosynthesis that self assembles into decameric nanoparticles (Bacher et al. 1980; Wei et al. 2018). The potential of BLS as a scaffold for antigen display has been demonstrated (Alfano et al. 2015; Cristofalo et al. 2025; Hiriart et al. 2017). This repetitive presentation of antigens enhances the likelihood of multiple B‐cell receptors binding to the antigens simultaneously, triggering a robust immune response (Ra et al. 2014). We produced the fusion protein in Nicotiana benthamiana plants and showed that BLS assembles as a decamer, and the resulting protein induces the production of antibodies specific to FepA (Shamriz et al., submitted).

In this research, antigenic peptides within the extracellular loops of CirA, FhuA and IroN were identified. To concurrently present these predicted antigenic peptides to the immune system, we fused the extracellular antigenic peptides to BLS to create vaccine candidates targeting these three TBDTs. These three proteins as well as FepA‐BLS were produced in N. benthamiana and assessed for immunogenicity individually as well as in a mix. We found that antibodies produced against all four TBDTs bind to Salmonella enterica Enteritidis.

2. Results

2.1. The Design, Expression and Purification of Vaccine Candidates

The 3D structures of CirA, FhuA and IroN, predicted by Phyre2, are illustrated in Figure 1. Each protein is predicted to fold as a 22‐stranded anti‐parallel β‐barrel, with an N‐terminal plug domain inserted within the barrel itself. While the overall fold of each protein is similar, significant sequence and structural variability exists with the surface‐exposed outer loops of each protein. The outer‐facing loops are numbered according to their position within the structure, with loop 7 of CirA (420‐APSLLQLSPDWATNSCRGGCRIVGSPDLKP—449), for example, connecting β‐strands 13 and 14 within the β‐barrel. Antigenic peptides, identified using DNAStar software's Protean 3D package (Table 1), were identified on the loops of the respective proteins and were designated with ‘L’, representing Loop, along with a numerical identifier corresponding to the specific loops where they are located. This surface localization is essential for accessibility to the immune system, enabling recognition by antibodies. Several criteria, including sequence length, hydrophilicity/hydrophobicity, surface orientation and flexibility, B‐cell epitopes, T‐cell epitopes and MHC II epitopes, were analysed to enhance surface exposure and antigenicity.

FIGURE 1.

Predicted Salmonella CirA, FhuA and IroN homology models generated using the Phyre2 server. The figure depicts the extracellular space at the top and the periplasmic space at the bottom. The 22‐stranded β‐barrel, mainly embedded in the membrane bilayer and the N‐terminal domain within it, is represented in grey. Antigenic peptides, corresponding to specific extracellular loops, are mapped onto the structure and coloured.

TABLE 1.

The predicted antigenic peptides of each TonB‐Dependent Outer Membrane Transporter (TBDT).

TBDT	Loop No.	Predicted antigenic peptide sequence
CirA	L7	420‐APSLLQLSPDWATNSCRGGCRIVGSPDLKP‐449
	L8	488‐ISRTPDVNAAPGYSNFVGFETNSRGQRVPVFRY‐520
FhuA	L3	272‐PETGYYGWLPKEGTVEPLPNGKRLPTDFNEGAKNN‐306
	L4	348‐YGVCSDPANRDSKQCAALAPADKGH‐372
	L5	426‐YDDSVPLLDLYNPVYTDFDFASRDPAT‐452
	L7	535‐PASQTGENGKIFAP‐548
	L8	582‐LMADPAGSFFSVEG‐595
	L9	631‐DTTYKGNTPAQVPK‐644
	L10	673‐GSSYGDPANSFKVGS‐687
	L11	715‐NLFDREYVASCFQTYGCFWGAER‐737
IroN	L3	276‐IYAGDTQNSTSNAVTKSLAQSGRET‐300
	L4	335‐MNEGLSGGGEGRITNDQTF‐353
	L5	392‐DPSSTSLTVKDSNIAGIPGSAANRSS‐417
	L7	482‐APNLYQTSEGYLLYSKGNGCPKDITSGG‐509
	L8	550‐NKIVAGDQIIGRSASGAYVLQ‐570
	L10	650‐PRTHAESRSEETKGLSGK‐667

Antigenic peptides typically fell within the 10–20 residue range, balancing immunogenicity and specificity. In this study, antigens that were shorter than 10 residues were ignored. For CirA, L7 and L8 were genetically fused in tandem to the N‐terminus of BLS. For FhuA, all the antigenic peptides except for L1, L2 and L6 were fused in tandem to the N‐terminus of BLS. For IroN, all the antigenic peptides except for L1, L2, L6 and L9 were fused in tandem to the N‐terminus of BLS (Figure 2, Table 1). These vaccine candidates were called CLB (CirA Loops—BLS), FLB (FhuA Loops—BLS) and ILB (IroN Loops—BLS).

FIGURE 2.

Vaccine candidate design. Predicted antigenic peptides, colour‐coded based on their respective loops as shown in Figure 1, were connected to each other and the N‐terminus of BLS (PDB 1 T13) using flexible and rigid linkers. BLS is depicted in light orange. c‐Myc tags at the C‐terminus and HA and His‐tags at the N‐terminus were included for detection and purification purposes.

All three constructs were codon optimized for transient expression in N. benthamiana. FLB was cloned into plant expression vectors targeting the recombinant protein to five subcellular compartments (cytosol, chloroplast, apoplast, endoplasmic reticulum and vacuole) to determine where it would accumulate to the highest level. Both target antigenic peptides and BLS originate from bacteria and, as a result, do not undergo glycosylation in their host organisms. The in silico analysis indicated the absence of any potential N‐glycosylation sites in all three proteins. Therefore, their expression in plant cells capable of glycosylating proteins should not pose any issues. Western blot analysis confirmed the expression of FLB in all five compartments of N. benthamiana (Figure 3a) with highest levels in the apoplast, ER and vacuole compartments of the secretory pathway by 8 dpi. Nevertheless, to avoid the addition of an N‐terminal signal peptide and C‐terminal peptides for retaining the protein in the ER or targeting it to the vacuole, we decided to target this protein to the cytosol. Quantification against known amounts of a standard protein showed the highest level of accumulation at day 8 post infiltration of up to 0.81 mg/g fresh weight (FW) in the cytosol (Figure 4). CLB was successfully cloned only in the chloroplast‐targeting expression vector, and accumulated well in that compartment up to 0.52 mg/g FW (Figure 3b, Figure 4), while ILB was only successfully cloned in the ER‐targeting vector and also accumulated well in that compartment up to 0.99 mg/g FW (Figure 3c, Figure 4). Subsequent purification of CLB, FLB and ILB was achieved through IMAC.

FIGURE 3.

Recombinant expression of FLB (a), CLB (b) and ILB (c) in N. benthamiana. Leaf tissue was collected at 4, 6 and 8 days post‐infiltration (DPI). Twenty μl crude extract was loaded in each well. Anti c‐myc antibody was used to detect the protein. FLB, CLB and ILB are approximately 47, 28 and 37 kDa, respectively. p19; protein extracted from plants infiltrated with p19 as negative control. APO, apoplast; CHL, chloroplast; CYT, cytosol; ER, endoplasmic reticulum; VAC, vacuole. The green arrow points to the recombinant protein band.

FIGURE 4.

Accumulation of CLB in the chloroplast, FLB in the cytosol and ILB in the endoplasmic reticulum. Three leaf discs from each of five different plants were taken on 4, 6 and 8 days post‐infiltration (DPI) and pooled. Protein amounts were quantified with GelQuant software using a standard of known concentration as a reference. Protein accumulation levels were calculated based on plant tissue mass, amount of soluble plant protein extraction buffer used, and the amount of extract loaded such that accumulation would be represented as mg of recombinant protein per gram of plant tissue fresh weight (FW).

2.2. Evaluating the Immunogenicity of IMAC‐Purified Plant‐Produced Vaccine Candidates

To assess the immunogenicity of IMAC‐purified plant‐produced CLB, FLB, ILB (Figure 5) and FR‐BLS, a protein we previously produced that contains FepA antigenic peptides fused to BLS (Shamriz, Mak, et al. 2026), mice were immunized with CLB, FR‐BLS, FLB and ILB either individually or in a combination of all four proteins. Subsequently, the presence of IgG antibodies was quantitatively analysed in serum samples obtained from immunized BALB/c mice at 0, 35 and 49 days of age (Figure 6a). To ensure that the immune response was directed against the Salmonella antigenic peptides rather than against BLS, ELISA plates for CLB, ILB and FR‐BLS were coated with encapsulin‐fusion constructs displaying antigenic peptides of Salmonella CirA, IroN and FepA (Charron, Kaldis, Shamriz, Renaud, et al. 2026). For FLB, ELISA plates were coated with whole Salmonella Enteritidis cells. Animals immunized with recombinant CLB, FR‐BLS, FLB or ILB demonstrated the capacity to induce IgG antibodies specific to the antigenic peptides (Figure 6b, Figure S1). In all cases, the results of both the test bleed and final bleed were significantly higher than those of the respective PBS control (Figure 6b).

FIGURE 5.

Analysis of IMAC‐Purified proteins by SDS‐PAGE. The IMAC‐purified FLB, CLB and ILB were analysed using SDS‐PAGE. FLB, CLB and ILB (green arrows) are approximately 47, 28 and 37 kDa, respectively. P19; protein extracted from plants infiltrated with P19, was used as negative control. In the crude extract (CE), flow through (FT) and wash fractions (Wash), as well as in the negative control P19, bands corresponding to RBC‐L were observed (ca. 53 kDa) (black arrows).

FIGURE 6.

Mouse sera were analysed using indirect ELISA with predicted antigenic peptides as the antigens. (a) Schematic overview of the immunization workflow. The diagram illustrates the timeline and experimental design used for mouse immunizations. The schematic applies to all treatment groups (CLB, FR‐BLS, FLB, ILB, Mix and PBS). (b) Antibody responses were assessed before and after immunizing mouse groups (n = 6 per group) with (1) CLB, (2) FR‐BLS, (3) FLB and (4) ILB. Mice were also immunized with a mix of all four proteins and PBS was used as a control. The antibody response to each recombinant protein's antigenic peptides within the mix and PBS was evaluated separately and shown in the corresponding graph. Mouse sera collected at pre‐immunization (PI; blue), test bleed (TB; orange) and final bleed (FB; grey) were diluted 1:8000 and tested for the presence of antibodies against the predicted antigenic peptides. Statistical significance was defined as p < 0.05, and the p‐value is shown above arrows comparing treatment pairs. Error bars show standard deviation.

To assess the collective immunogenicity of plant‐derived CLB, FR‐BLS, FLB and ILB administered as a mix, the antibodies produced against each protein were assessed separately. It is important to note that the mix contained one quarter of the amount of each protein as the individual immunizations to keep the injection volume consistent. While the test bleed generally revealed a weaker response against each of the proteins in the mix than the individual vaccinations, the final bleed revealed a remarkable increase in the immune response, resulting in no statistically significant differences between the final bleed of the mix group and the final bleed of the corresponding individual proteins, except for FR‐BLS (Figure 6b). As well, there was no significant difference between FLB alone and the mix in the test bleed, which indicates that the reduced amount of FLB in the mix was sufficient for generating a good immune response. On the other hand, the reduced dose of the three other proteins in the mix was not able to produce a good immune response after the first boost. To achieve a good immune response in the mix after one boost, it may be necessary to increase the dose of CLB, FR‐BLS and ILB.

To further assess whether the antibodies produced against the Salmonella TBDT extracellular antigens would bind to the wild‐type TBDTs displayed on the surface of Salmonella, a Salmonella binding assay was performed using final bleed serum samples from mice immunized with vaccine candidates CLB, FLB, FR‐BLS and ILB, as well as the control group injected with PBS. We observed consistent co‐localization of sera from mice immunized with recombinant proteins and Salmonella cells (Figure 7) while no signal was detected from sera of mice injected with PBS as a negative control. These findings emphasize the ability of CLB, FLB, FR‐BLS and ILB to generate targeted immune responses, validating their potential as promising vaccine candidates against Salmonella colonization.

FIGURE 7.

Antibodies produced against antigenic peptides bind S. enteritidis . Mouse sera samples collected during the final bleed were tested for antibodies capable of binding to S. enteritidis . Binding is visualized by Alexa Fluor 594‐conjugated goat anti‐mouse IgG (H + L) (Red). DAPI (green) is used to visualize bacterial cells. A merged image shows an overlay of the green and red channels (shown in yellow) used to visualize DAPI and Alexa Fluor 594 co‐localization. Size bar = 10 μm.

3. Discussion

In recent years, there has been a growing trend in utilizing antigen‐display technology to produce particulate antigens as vaccine candidates. Self‐assembled nanoparticles have emerged as a remarkably versatile and efficient platform in the development of these candidates (Lopez‐Sagaseta et al. 2016; Shi et al. 2023). The benefits of employing self‐assembled protein nanoparticles as vaccines, rather than soluble subunit antigens, stem largely from their size, ranging from 20 to 200 nm. This particulate presentation facilitates efficient uptake by antigen‐presenting cells. Additionally, self‐assembled nanoparticles feature densely repetitive surface structures that enhance the activation of immunological responses (Morales‐Hernandez et al. 2022).

We have previously demonstrated that BLS displaying antigenic loops from FepA assembles into a decamer when expressed in plants (Shamriz, Mak, et al. 2026). In this study, additional iron regulated proteins, namely CirA, FhuA and IroN, were selected, and vaccine candidates based on the BLS nanoparticle were designed and produced in plants. The yield of recombinant protein expression in N. benthamiana can vary widely based on several factors including the specific protein being expressed, the promoter and expression system used, as well as the plant growth conditions. The economically viable level of protein accumulation for vaccine development in N. benthamiana plants is proposed to be 0.1 mg/g (Rybicki 2010). In this study, all vaccine candidates achieved protein accumulation levels surpassing 0.1 mg/g in N. benthamiana.

Zarate‐Bonilla et al. (2014) utilized classical methods and physicochemical properties to predict antigenic regions in Salmonella spp. in chickens. Their analysis involved assigning values to amino acids and determining flexibility profiles using the Bhaskaran scale (Zarate‐Bonilla et al. 2014). By predicting the 3D structures of IroN, FepA and CirA proteins, they identified significant antigenic peptides on the proteins' external surfaces. These peptides were earmarked as potential candidates for future recombinant or synthetic peptide vaccines. The authors suggested the possibility of cross‐immunity against various Enterobacteria, drawing parallels with iron regulated outer membrane proteins (IROMPs) from Salmonella spp. and E. coli (Zarate‐Bonilla et al. 2014).

In our study, a different approach led to the prediction of additional antigenic peptides, distinct from those reported previously. These predictions were based on specific factors considered in our analysis. The accuracy of these predictions and their viability as vaccine candidates can only be confirmed through production and subsequent immunogenicity studies in poultry. While the study by Zarate‐Bonilla et al. (2014) provided valuable insights into the role of TBDTs and surface‐bound antigenic peptides in vaccine development, it is crucial to highlight that the predicted peptides were not experimentally produced or tested (Zarate‐Bonilla et al. 2014). Considering several factors such as sequence length, hydrophilicity/hydrophobicity, surface orientation, flexibility, B‐cell epitopes, T‐cell epitopes and MHC II epitopes (DNAStar software, Protean 3D package), we predicted antigenic peptides and assessed their antigenicity.

Subunit vaccines often suffer from low immunogenicity, susceptibility to degradation and limited ability to activate the innate immune system. Consequently, higher doses, multiple injections and strong adjuvants are typically necessary to ensure a potent immune response. To mitigate these disadvantages, the antigenic peptides derived from CirA, FhuA and IroN were presented on the BLS self‐assembling nanoparticle in a comparable manner as previously done with FepA antigenic peptides (Shamriz, Mak, et al. 2026). This allowed us to overcome challenges associated with membrane protein expression, as these three full‐length proteins are embedded in the bacterial outer membrane. The feasibility of expressing BLS‐based antigens in plants has been demonstrated by Alfano et al. (2015), who produced a BLS‐VP8d fusion in tobacco plastids at remarkably high levels, reaching ~40% of total soluble protein. Their study also showed that the fusion protein remained soluble and stable throughout plant development and even in lyophilized leaves, enabling low‐cost storage and downstream processing (Alfano et al. 2015). The suitability of plant expression systems for nanoparticle‐based vaccine production is further supported by several studies demonstrating robust accumulation and proper assembly of recombinant proteins in Nicotiana species. Charron and co‐workers expressed multiple FepA‐encapsulin fusion proteins in N. benthamiana and observed substantially higher accumulation when targeted to the chloroplast, with levels exceeding 0.7 mg/g fresh weight, and up to 2.4 mg/g in transplastomic N. tabacum . These studies confirmed correct nanoparticle assembly in planta, including mosaic particle formation and multiple antigens displayed on a single particle (Charron, Kaldis, Shamriz, Diarra, et al. 2026; Charron, Kaldis, Shamriz, Renaud, et al. 2026). We also demonstrated that chloroplasts can successfully express bacterial membrane proteins such as TBDTs, highlighting the potential of plastids for producing structurally complex antigens relevant to Salmonella vaccines (Shamriz, Kaldis, et al. 2026).

In this study, we used BLS to display extracellular antigenic peptides from S. enteritidis . BLS has been shown to assemble into a decameric nanoparticle, to display antigens on its surface, and to induce a strong protective immune response toward the target pathogens (Craig et al. 2005; Rosas et al. 2006). Furthermore, BLS was shown to function as an adjuvant when administered orally either in combination or in fusion with an antigenic peptide (Fragoso et al. 2011; Verma et al. 2023). Here we showed that the plant‐produced CLB, FR‐BLS, FLB and ILB vaccine candidates elicit a potent immune response against target Salmonella TBDTs in BALB/c mice. This strong immune response is likely due to the multivalent presentation of the antigens to the immune system.

TBDTs exhibit redundancy and perform various functions in iron uptake. Consequently, blocking specific TBDTs may not sufficiently impede overall iron acquisition by Salmonella. To address this challenge, a combined vaccine targeting four TBDTs, CirA, FepA, FhuA and IroN, was evaluated. The rationale behind this approach is to hinder overall iron acquisition, thus limiting Salmonella's growth and survival and its ability to colonize the GI tract of poultry. Initial results showed a lower immune response to each antigen in the combined vaccine compared to individual formulations. However, after a second boost, a significant recovery was observed in immune response in the final bleed samples. These findings underscore the efficacy of the combined vaccine approach, suggesting its potential superiority over separate vaccines in combating Salmonella infection.

Our approach to producing nanoparticle vaccines in plants leverages the scalability, cost‐effectiveness and safety advantages of plant‐based protein production, and adds to our toolbox for animal vaccine production. Additionally, as plants are part of animals' diets, they offer the prospect of producing oral therapeutic products and vaccines for animals without the need for expensive purification processes (Topp et al. 2016). Successful development of these plant‐produced vaccine candidates against Salmonella infection not only addresses a global food safety issue, but also sets a precedent for the future of vaccine design and production. Our study points to innovative solutions in the fight against infectious diseases, underscoring the importance of continued exploration and advancement in the field of vaccine technology.

4. Materials and Methods

4.1. Vaccine Construct Design

The gene sequences for cirA, fhuA and iroN were extracted from the complete genome of S. enterica subsp. enterica serovar Enteritidis str. P125109, hereafter referred to as S. enteritidis , with the GenBank Accession Number AM933172.1. Homology models for these proteins were generated using the Phyre2 server (Figure 1). Antigenic peptides were predicted using DNAStar software's Protean 3D package, considering factors such as sequence length, hydrophilicity/hydrophobicity, surface orientation and flexibility. Additionally, B‐cell epitope, T‐cell epitope and MHC II epitope predictions were considered. The BLASTp program was utilized to confirm that the predicted antigenic peptides could not be found in domain regions of other proteins and bacterial strains. To create a multimeric virus‐like presentation of the predicted antigenic peptides to the immune system, they were genetically fused in tandem to the N‐terminus of Brucella sp. lumazine synthase (PDB 1 T13) using both rigid (EAAAK) and flexible (GGGGS) linkers to balance structural stability with sufficient epitope mobility and spacing for optimal antigen presentation, and predicted their tertiary structures using Phyre2 (Figure 2). The constructs were modified multiple times to reduce repetitive motifs, adjust linker composition and improve overall sequence manufacturability. The final designs reflect both structural considerations and the practical constraints required to obtain a sequence that could be successfully synthesized. Glycosylation sites on the designed constructs were predicted through in silico analysis utilizing the NetNGlyc 1.0 Server (http://www.cbs.dtu.dk/services/NetNGlyc/).

4.2. Cloning and Transient Expression in N. benthamiana

The designed constructs, codon optimized for nuclear expression in Nicotiana benthamiana, and featuring flanking BsaI recombination sites, were synthesized by Bio Basic Inc. (Markham, ON, Canada). These constructs were then cloned into plant expression vectors suitable for targeting the recombinant proteins to five distinct subcellular compartments. These vectors, derived from pCaMGate (Pereira et al. 2014) and referred to as pCLGG‐X (VanderBurgt et al. 2023), employ the double‐enhanced constitutive Cauliflower Mosaic Virus 35S promoter and Nopaline synthase terminator and are compatible with GoldenGate cloning (Marillonnet and Werner 2015). Following cloning of each construct into the expression vectors, Agrobacterium tumefaciens EHA105 was transformed through electroporation. N. benthamiana plants, grown for 7 weeks in a growth chamber at 22°C with 65% relative humidity and a 16‐h photoperiod at a light density of ~100 μmol m⁻² s⁻¹, were utilized as the expression platform. A. tumefaciens cultures carrying either the fusion or the suppressor of posttranscriptional gene silencing p19 from the cymbidium ringspot tombusvirus (Silhavy et al. 2002) were cultured separately to an OD₆₀₀ of 0.5–1 and then combined. The combined cultures were infiltrated into leaves either by injection or vacuum. Sampling was conducted at 4‐, 6‐ and 8‐days post‐infiltration, and consisted in collecting and pooling three leaf discs from each of five infiltrated plants for a total of fifteen leaf discs/tube. Negative control experiments involved infiltrating plants solely with the p19 expression vector.

4.3. Protein Expression Analysis

Total soluble proteins were extracted from the collected leaf discs and analysed as described by Shamriz, Mak, et al. (2026). Briefly, pre‐weighed leaf discs were flash frozen and pulverized. The resulting powder was mixed with protein extraction buffer (consisting of 1× PBS, pH 7.5, 0.1% Tween‐20, 2% polyvinylpolypyrrolidone (PVPP), 1 mM EDTA pH 8.0, 1 mM PMSF, 1 μg/mL Leupeptin and 100 mM Sodium L‐ascorbate) in a 1:3 weight/volume ratio. All samples were vortexed and sonicated, then centrifuged at 20,000 × g for 5 min at 4°C, and the resulting supernatants containing total soluble proteins (TSP) were collected.

TSP were denatured, separated by Sodium Dodecyl Sulphate–Polyacrylamide Gel Electrophoresis (SDS‐PAGE) and transferred onto polyvinylidene difluoride (PVDF) membranes. The membranes were blocked and immunodetected with anti‐c‐myc (Genscript, Piscataway, USA, Cat. No. A00864) or anti‐His (Takara Bio USA Inc. Cat. No. 631212) primary antibodies diluted 1:5000 in 0.5% blocking solution, followed by goat anti‐mouse horseradish peroxidase (HRP)‐conjugated secondary antibody (Bio‐Rad, Hercules, USA, Cat. No. 1706516) diluted at 1:5000 in 0.5% blocking solution. Chemiluminescence was detected using Enhanced the Chemiluminescent detection solution (Bio‐Rad, Hercules, USA, Cat. No. 1705061) and captured using a MicroChemi 4.2 imaging system (DNR Bio‐Imaging Systems Ltd. Aachen, DEU). The accumulation of recombinant proteins was quantified by comparing them to a standard curve generated from known concentrations of an in‐house synthetic protein standard designed with multiple detection tags, using GelQuant software (DNR Bio‐Imaging Systems Ltd. Aachen, DEU). The in‐house synthetic protein used as a standard for quantification on western blots is a fusion of enhanced green fluorescence protein (eGFP) to an elastin‐like polypeptide (ELP—28 repeats of VPGVP) (Conley et al. 2009) and hydrophobin I (HFBI) (Joensuu et al. 2010). N‐terminal tags for c‐Myc, Xpress, FLAG and 6xHis were incorporated upstream of eGFP, while StrepII and hemagglutinin (HA) tags were added to the C‐terminus of HFBI. This DNA construct was cloned into our ER‐targeting vector and transiently expressed in N. benthamiana leaves. The recombinant protein was purified by IMAC then dialyzed into PBS. The synthetic protein was run on SDS‐PAGE and subsequently stained with Gel Code Blue. Its concentration was determined by densitometric comparison to known amounts of a BSA standard curve. Proteins separated by SDS‐PAGE were stained with GelCode Blue Stain Reagent solution (Thermo Scientific, Waltham, USA, Cat. No. 24590) and destained with ultrapure water.

4.4. Protein Purification

Protein purification was performed by the batch IMAC method as described by Shamriz et al. (submitted). Briefly TSP extract was mixed with Ni²⁺ resin (Cytiva, Wilmington, USA, Cat. No. 17531801) and incubated at 4°C on a shaker for 1 h, centrifuged and washed three times with five resin volumes of washing buffer (PBS containing 20 mM imidazole). The Ni²⁺ resin was then packed in a column, and the recombinant protein was eluted in five fractions of one volume elution buffer each (PBS containing 500 mM imidazole). Subsequently, the IMAC‐purified proteins were dialyzed against PBS before animal immunization.

4.5. Animal Trial

Mouse immunization was done by CEDARLANE (Burlington, Ontario, Canada) under the approved Animal Use Protocol AUP117AB. This protocol was reviewed and approved by Cedarlane's institutional Animal Care Committee and was developed in accordance with the Canadian Council on Animal Care (CCAC) guidelines for the ethical use of animals in research.

Thirty‐six female BALB/c mice aged 6 to 8 weeks were divided into six distinct groups. Pre‐immune blood samples of 50 μL were collected from each mouse. Four groups were subcutaneously immunized with 50 μg/100 μL/mouse of individual vaccine candidates: CLB, FR‐BLS, FLB and ILB. A fifth group was subcutaneously immunized with a combination of all four vaccine candidates (12.5 μg of each protein/100 μL). The final group, serving as the negative control, received PBS alone. The primary immunization was emulsified with 100 μL of Complete Freund's adjuvant, followed by two boosts on days 21 and 42, mixed with 100 μL of incomplete Freund's adjuvant. The mixture (200 μL) was emulsified, and 50 μL was injected subcutaneously at four different sites in each mouse. Test bleeds (50 μL) were collected at day 35 for titre analysis, and the mice were euthanized at day 49 for final bleeds.

4.6. Enzyme‐Linked Immunosorbent Assay (ELISA)

ELISA MaxiSorp1 plates (Thermo Fisher Scientific Inc., Waltham, MA, USA) were coated with 100 μL of 0.5 mg/μl of encapsulin nanoparticle‐based fusion proteins containing the target loops (Charron, Kaldis, Shamriz, Diarra, et al. 2026; Charron, Kaldis, Shamriz, Renaud, et al. 2026) at 4°C overnight, followed by blocking with 2% BSA‐PBS for 1 h at room temperature. After washing, plates were incubated with serially diluted test bleed and final bleed (1:500, 1:2000, 1:8000, 1:32000, 1:128000 and 1:512000) in 1% BSA‐PBS dilution buffer and run against the pre‐immune bleed for an hour at 37°C. After washing, goat anti‐mouse IgG‐HRP‐conjugated secondary antibody (1:5000; Bio‐Rad, Hercules, CA, USA) was applied for 1 h at room temperature. Following additional washing steps, antigen‐bound antibodies were detected using 2,2′‐azino‐bis (3‐ethylbenzothiazoline‐6‐sulfonic acid) (ABTS, Sigma‐Aldrich, Oakville, ON, Canada) and the colorimetric reaction was measured at 405 nm using an ELISA plate reader (iMark Microplate Absorbance Reader, Bio‐Rad, Feldkirchen, Germany).

The whole‐cell Salmonella ELISA was performed as described by Shamriz, Kaldis, et al. (2026). Briefly, ELISA MaxiSorp1 plates (Thermo Fisher Scientific Inc., Waltham, MA, USA) were treated with Poly‐L‐lysine (Sigma‐Aldrich, Oakville, ON, Canada) to allow binding of S. enteritidis to the plates. After S. enteritidis was bound to the plates, serial dilutions of mouse sera from the different treatments were added, followed by goat anti‐mouse IgG‐HRP‐conjugated secondary antibody (1:5000). Bound antibodies were detected by a colorimetric reaction with ABTS (Sigma‐Aldrich, Oakville, ON, Canada).

4.7. Salmonella Binding Assay

The Salmonella binding assay was performed as described by Shamriz, Mak, et al. (2026); Shamriz, Kaldis, et al. (2026). Briefly, S. enteritidis was mixed with mouse sera from the various treatments (serum from the final bleed was used at a 1:100 dilution in 1% BSA‐PBS) and after incubation, the cells were resuspended in Alexa Fluor 594‐conjugated goat anti‐mouse IgG (H + L) (Abcam, Cambridge, UK, Cat. No. ab150116), then in DAPI (Thermo Fisher Scientific, Cat. No. D1306). The prepared cells were then transferred onto poly‐L‐lysine coated coverslips (Millipore Sigma, St. Louis, USA, Cat. No. S1815), dried and mounted onto glass slides with Aqua‐Poly/Mount (Polyscience Inc., Warrington, PA, United States, Cat. No. 18606). Binding of mouse antibodies to S. enteritidis cells was visualized using an Olympus LSM FV 1200 microscope. Images were captured with a 60× water objective lens. DAPI was imaged with excitation at 350 nm and emission at 455–465 nm. Alexa Fluor 594 was visualized with excitation at 590 nm and emission at 617 nm.

4.8. Statistical Analysis

Statistical analysis of ELISA results was conducted using two‐sample t‐tests for unequal variance (heteroscedastic) and two‐tailed distribution. The resulting p‐values were then adjusted using the Benjamini–Hochberg method to account for the 10 pairwise comparisons examined for each vaccine antigen, separately. Comparisons were made between the pre‐immune bleed of the control to both the individual antigen and mix groups, between each test bleed, and each final bleed. The test and final bleeds were also compared separately within the individual antigen and mix groups.

Funding

This study was supported by Agriculture and Agri‐Food Canada through the Federal Genomics Research and Development Initiative to mitigate antimicrobial resistance (GRDI‐AMR and GRDI‐AMR2) programme.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: ELISA titration curves for CLB, FR‐BLS, and ILB proteins.

Data Availability Statement

The data that supports the findings of this study are available in this article.