Escherichia coli DH5α and E. coli BL21 DE3 were used for cloning and initial protein verification, respectively. Both were cultured in LB media (FisherSci, BP1427-500) overnight in a shaker at 250 rpm and 37°C. The antibiotics used for plasmid selection were 50 μg/mL kanamycin for pET28a and 250 μg/mL erythromycin for pLSVN3 and pLSVN7 (Table 1). L. casei and L. lactis were cultured statically in MRS media (FisherSci, CM0359) at 37°C and 5% CO2 for 2–3 days until reaching log phase. For engineered bacterial culturing, 5 μg/mL erythromycin was added to the MRS media for applying selection pressure. HEK 293 T and TZM-bl mammalian cell lines were grown in Dulbecco Minimum Essential Media (DMEM) (Gibco, 11,965,092) supplemented with 10% FBS (Gibco, 10,082), 1 mM L-glutamine (Gibco, 25,030), and 100 μg/mL penicillin/streptomycin (Gibco, 15,140). Cells were cultured in a humidified cell culture incubator at 37°C with 5% CO2 using T-75 flasks.

The scytovirin (SVN) gene sequence (GenBank: P86041.1) (Bokesch et al., 2003) with an added E-tag sequence (total 348 bp) was synthesized with SacI/XhoI restriction sites by GenScript and inserted into the vector pET28a (Novagen, Inc.) to test protein expression in E. coli BL21 (DE3) bacterial cells. SVN was overexpressed in E. coli previously by Xiong et al. (2006). SVN expression was induced by IPTG (1 mM) to increase protein production which was verified by Western blotting using an anti-E-tag (ab3397, Abcam), anti-His-tag (HRP-66005, Proteintech) or anti-SVN (A64238-050, Epigentek) polyclonal antibodies.

To express the SVN protein on the surface of LAB, constructs were created which included the Lactate dehydrogenase promoter (Pldh), signal peptide (SP), cell membrane anchor protein (ANC), E-tag marker (E), and the protein marker (GFP). The constructs were built up based on our previous plasmid pWZ486 (Wei et al., 2019) derived from pTRKH3-ldhGFP (Addgene). The SVN-E-tag-GFP fusion sequence replaced the CD4 gene sequence in pWZ486 using the SacI/XhoI restriction sites. The constructs pLSVN3 and pLSVN7 were verified by DNA sequencing and PCR using the specific primers (Table 1) marked in the construct maps.

Plasmids were transformed into the Lactic acid bacteria (LAB) by electroporation as previously described (Wei et al., 2019; Welker et al., 2019). Briefly, overnight cultures of LAB cells were diluted (1:50) into fresh MRS media with 1% glycine and incubated at 37°C without shaking for 2 h. Cells were harvested and treated with 50 mM EDTA (pH 8.0) for 15 min, followed by two washes with ice-cold electroporation buffer (0.5 M sucrose) and resuspended in electroporation buffer (1/100 volume of the initial culture). 50 μL of cells were mixed with plasmid DNA and incubated on ice for 15 min. The mixture was added to an ice-cold 0.2 cm GenePulser (Biorad) cuvette and pulse was immediately applied at the conditions of 10 KV/cm, 200 Ω, and 25 μF. Cells were suspended in 1 mL MRS broth with 2 mM CaCl2 and 20 mM MgCl2 and then incubated at 37°C for 4 h. Cells were pooled on MRS plates with 5 μg/mL erythromycin and cultured as described above. To verify transformation, single colonies were picked and added to 25 μL PCR master mix (Promega, M791B) containing 1 μM forward and reverse primers (Table 1). PCR amplification occurred in a SimpliAmp thermocycler (Applied BioSystems, A24811) under the following conditions: Stage 1; 95°C for 5 min, Stage 2 (35 cycles); 95°C for 30 s, 55°C for 30 s, 72°C for 1 min, Stage 3; 72°C for 7 min. PCR products were visualized on 1% agarose gel electrophoresis.

For the pLSVN3 construct transformed into L. casei, the fusion protein was detected based on the GFP fluorescence. Bacteria were washed three times with PBS and analyzed on a BD FACSAria using a 488 nm laser. For the pLSVN7 construct transformed into L. lactis, the bacteria were first stained for 1 h with two primary antibodies: mouse monoclonal anti E-tag (Novus NBP2-67081) and rabbit polyclonal anti SVN (A64238-050, Epigentek). Goat anti mouse conjugated with AlexaFluor488 (A-11011, ThermoFisher) and goat anti rabbit conjugated with AlexaFluor594 (A-11012, ThermoFisher) were used as secondary antibodies, respectively. Bacteria were analyzed on a Beckman Coulter CytoFLEX FX at 488 nm and 561 nm. Unstained bacteria and stained wild-type bacteria were used in all experiments for an appropriate gating strategy.

Confocal microscopy was performed using the same fluorescent fusion protein or antibody combinations as described for flow cytometry (above). Following the antibody staining, bacteria were pelleted, resuspended in 10 μL PBS, and transferred to a microscope slide with cover slip. Images were captured using a Nikon A1R-Ti2 (Nikon Instruments, NY, USA) inverted confocal system.

The Ebola pseudotyped viruses were made from a HIV-1 backbone plasmid, pSG3^ΔEnv (NIH HIV Reagent Program). The Ebola Envelope gene (GP, Zaire ebolavirus, GenBank: AIO11753.1), was synthesized by GenScript and cloned into pcDNA3.1(+). Both plasmids were co-transfected into 293 T cells in a 10 cm plate using transfection reagent polyethyleneimine (PEI). Three days post transfection, the medium was harvested and centrifuged at 500 g to remove cell debris, and then the supernatants were stored at −80°C (Platt et al., 2009; Wang et al., 2022). The viral titers were determined by reverse transcriptase assay.

The titers of pseudotyped viruses were determined by Reverse transcriptase assay (RTA) (Wei et al., 2019). 500 μL of pseudotyped virus stock was spun at 14,000 g for 2 h at 4°C to precipitate the virus. The viral pellet was resuspended in a Triton X-100-based suspension buffer and vortexed, followed by three rapid freeze–thaw cycles to lyse the viruses. 50 μL of reaction mix [Oligo-dT Poly-A and ³H-dTTP (PerkinElmer)] was added, and the samples were incubated at 37°C for 1 h in a heating block. Then, the samples were pipetted onto DEAE Filter mat circle papers (PerkinElmer), followed by three 10 min washes in 2X SSC buffer, and one 10 s wash in 100% ethanol. The filters were dried at room temperature and analyzed using a scintillation counter to measure the incorporation of ³H-dTTP into cDNA. The average CPM values from duplicates were determined.

Pseudotyped Ebola virus stocks were mixed with wild-type bacteria or engineered bacteria (~5×10⁷/mL) in 1.5 mL micro-centrifuge tubes. The mixtures of bacteria and viruses were incubated for 1 h at room temperature. Then the tubes were spun for 1 min at a 13,000 g to remove the bacteria and bound pseudotyped virus. The supernatants were collected, and the viral titers determined by RTA.

For the neutralization assay, pseudotyped Ebola virus was mixed with wild-type or engineered bacteria in the same manner as the adsorption assay described above. After centrifugation to remove the bacteria and bound pseudovirus, the remaining supernatants were applied to TZM-bl cells which were used as the target cells due to their ability to express luciferase when infected (Platt et al., 2009). The TZM-bl cells were set at a density of 6.0 × 10³ per well in a 96-well plate. Each neutralization assay was performed in triplicate with 5,000 RT units of pseudovirus per well used as the starting titer. Two days post-infection, the supernatants were removed, the cells were washed once with PBS, lysed in 1x Passive Lysis Buffer, and frozen at −80°C. The plates were then thawed, and luciferase activity was measured using beetle luciferin substrate (Promega) in a Veritas Luminometer.

Statistical analyses were conducted for virus adsorption and virus neutralization data using GraphPad Prism software (version 9.0). The significances were determined by using unpaired two-tailed Student’s t-test at p-value ≤0.05.

The scytovirin gene was initially synthesized by adding an E-tag at the N-terminus and cloned into pET28a with SacI/XhoI sites. The SVN plasmid was transformed to E. coli BL21 (DE3) cells for protein expression. The SVN fusion protein was observed under induction of 1 mM IPTG by Coomassie blue staining. The induced band of ~16kD fusion protein was noticed clearly (Figure 1A). To further confirm this protein, Western blotting was carried out by using E-tag, His-tag, and SVN specific antibodies. Positive bands were observed from all three different specific antibodies (Figure 1B), suggesting that the ~16kD SVN-E-tag fusion protein is expressed correctly. Figure 1C shows the construct design including the two His-tags encoded by the pET28a vector, as well the SVN ribbon structure with two highly similar domains (D1 and D2) for carbohydrate binding (Moulaei et al., 2007). Thus, the SVN-E-tag expressing construct can be further utilized for following bacterial engineering.

The SVN-E-tag construct was cloned into our previous plasmid (pWZ486) through SacI/XhoI sites to produce a new construct designated as pLSVN3 (Figure 2A). This GFP-E-tag-SVN fusion protein with the anchor has been modeled and shown in Figure 2B, which should be flexible for capturing viral particles. The construct was created first in E. coli DH5α cells and demonstrated the PCR fragment of ~1800 bp of GFP-E-tag-SVN sequence was correctly amplified using primers 154 and 037 (Figure 2C). To engineer the L. casei strain, the SVN construct pLSVN3 was transformed by electroporation into the L. casei for protein expression. The transformants were verified by PCR. A ~ 1800 bp PCR band from the amplification with primers 154 and 037 revealed that the pLSVN3 plasmid was transformed into the L. casei cells (Figure 2D). A ~ 75kD band corresponding to the fusion protein size (GFP-E-SVN-ANC) was detected by Western blotting using specific GFP and SVN antibodies (Figure 2E), suggesting that the SVN-GFP based fusion protein was produced from the L. casei cells. The smaller size band (~43kD) that appeared is the partial fragment of the fusion protein complex without the anchor domain (ANC). Flow cytometry analysis revealed the rate of SVN-GFP positive cells was 33.5% (Figure 2F a and b). To determine whether these protein inhibitors are displayed on the bacterial surface, confocal microscopy was used to visualize these SVN-GFP fused proteins. The pictures from confocal microscopy exhibited the fusion protein in green (GFP) displayed on the surface of bacteria (Figure 2F b). The biological functional studies were conducted for virus binding and neutralization. The engineered bacteria showed a moderate binding activity to the pseudotyped Ebola particles, reduced 37.2% of virus load, but the wild-typed (WT) bacteria also showed weak binding to the pseudotyped viruses because of the unspecific binding, reduced 21.8% (Figure 3A). The virus inhibition assay indicated that SVN-expressing bacteria had a moderate inhibition activity against pseudotyped Ebola virus infection which was 39.1%, but the WT-bacteria also had 23% inhibition (Figure 3B). These moderate functions may be due to the lower positive rate of SVN-expressing cells and the GFP interference of SVN binding in the SVN-GFP-ANC fusion protein complex.

Lactococcus lactis is another type of LAB which has a round shape and is widely used in the production of buttermilk and cheese. Thus, it is a very safe host strain candidate for anti-Ebola infection. We created the construct to remove the protein marker GFP but keep the E-tag marker from the construct pLSVN3, and this new construct was designated as pLSVN7 (Figure 2A). The construct pLSVN7 was transformed into the L. lactis using electroporation method. Positive bacterial colonies in the erythromycin selection plates were picked for further evaluation. The PCR method was first used to confirm the pLSVN7 plasmid was in the bacterium of L. lactis using primers 048 and 199. The expected ~700 bp band was identified (Figure 4A). Western blotting further demonstrated that SVN fusion protein (E-tag-SVN-ANC) was shown and in the correct molecular weight of 55kD both in anti-E-tag and anti-SVN specific antibodies. The 16kD SVN-E-tag fusion protein from the E. coli BL-21 (DE3) lysate as the positive control was also shown in the Western blots (Figures 4B,C). These Western blotting data demonstrated that the E-tag-SVN-ANC fusion protein was expressed in L. lactis. Next, the Confocal microscopy method was used to verify the surface display. Two colors of fluorescence were used for labeling E-tag (green) and SVN (red), respectively. The results demonstrated that the SVN fusion protein is clearly expressed on the surface of the bacterium. The single color, green or red and the merged color of green and red indicated the overlapping presence of fusion protein surface expression (Figure 5). Furthermore, Flow cytometry analysis was also conducted, and the data presented in Figure 6, shows that the positive rate of bacteria has reached 92.4%. The results demonstrated that the SVN-fusion protein is unambiguously expressed on the surface of these bacteria.

Functional studies were performed against the pseudotyped Ebola viruses. First, we used the viral particles adsorption method to evaluate the bacterial ability for capturing the viruses. We then tested the engineered strain activity to inhibit Ebola virus infection using TZM-bl cells. The virus adsorption and neutralization data are shown in Figure 7. Compared to the wild-type control bacteria, L. lactis, SVN engineered bacteria demonstrated improved functional abilities, absorbing 22.1% more pseudoviral particles and reducing infection by 28.5%. Overall, the SVN expressing L. lactis was able to adsorb 37.6% and neutralize 53.5% of the total pseudotyped Ebola virus, reducing infection in half compared to the virus only positive control (Figures 7A,B). In comparison, the wild-type L. lactis also demonstrated some ability to bind pseudotyped Ebola virus and reduce infection (15.5% adsorption and 24.7% neutralization) through non-specific binding, but the specific binding of the engineered bacteria to the viruses and are more than twice as effective, suggesting the SVN-expressing L. lactis bacteria can offer better protection against Ebola virus infection.