(Design → Build → Test → Learn)_n

DESIGN

Our goal is to use synthetic biology to produce an edible vaccine by combining the natural ability of bees to defend against bacteria and fungi with our ability to modify bacteria.

(drawing of bee immunity and bacteria modification)

Combining Bacteria with Deformed Wing Virus Proteins

Vitellogenin is vital for the initiation of TGIP due to its binding specificity to bacterial cell walls . ^{generating...} This in turn helps to activate the honey bee’s basal immune responses.

However, this does not normally work for viruses. As a result, feeding a deactivated version of DWV to bees would not elicit an immune response and injecting them with DWV antibodies is not a feasible option for real-world applications. ^{generating...} , ^{generating...} (refer to description)

To accomplish this, we needed to find a way of attaching a DWV capsid protein (something that the bee can recognize when infected) to the cell surface of a bacterial cell to use it as a shuttle!

The aim is that vitellogenin will bind as usual to the peptidoglycan cell wall on the bacteria and carry with it our engineered viral protein around the bee’s body. ^{generating...}

This attachment aims to facilitate the passage of these proteins from workers, to the queen bee, and to her offspring, allowing the latter to develop an immune response against DWV in the way it would with bacteria and fungi. ^{generating...}

Chassis Selection

In a paper from Salmela, Amdam, and Freitak; Vg binding, essential for TGIP, was seen to have a higher binding affinity to different types of pathogenic organisms. In particular, the peptidoglycan of gram positive bacteria, the lipopolysaccharides of gram negative bacteria, and the zymosan of yeast. In this particular study, they saw a much higher binding affinity of Vg to peptidoglycan cell walls. ^{generating...} From this, we decided that a gram-positive bacteria would be a good choice to increase our chances of transporting our vaccine into the eggs of the queen.

Extensive research and consultation with experts led us to choose Bacillus subtilis WB800N as our gram-positive chassis for extracellular protein expression. ^{generating...}

This strain, with eight protease knockouts, boasts a higher protein production capacity than standard E. coli strains used in protein production, enhancing our vaccine’s potential. This KO mutant will also allow us to be more confident about the final structure of the engineered cell wall attached viral protein as it is known that WT B. subtilis can degrade surface display proteins. ^{generating...} B. subtilis is also the most commonly used gram-positive bacteria in synthetic biology which enabled us to find more standardized parts and protocols that we could use in our project. ^{generating...}

Protein Selection

DWV is an RNA virus from the iflaviridae family. ^{generating...} Its genome is composed of a single polyprotein typical of its family with four capsid proteins (VP1, VP2, VP3, and VP4), a helicase, a genome-linked viral protein (VPG), an RNA-dependant RNA polymerase (RdRp), a leader polypeptide (Lp) and a 3C-protease (3C-pro). ^{generating...}

DWV Genome diagram (5’-3’)

Among the DWV capsid proteins, VP3 is identified as a protrusion domain protein and was found to be the most critical for infection. ^{generating...}

In experiments where bees were injected with antibodies against these capsid proteins, protection against VP3 significantly increased their survival chances. ^{generating...} Therefore, we believe that honey bees will have higher chances of protecting themselves against this virus by detecting this capsid protein during a DWV infection. Thus, VP3 became the protein of interest in our vaccine! We extracted the sequence from the genome referencing protease cutting sites that the virus uses on its own poluprotein as well as comparing them with partial sequences available.

Cell Surface Display

Now that we have a bacterial chassis and a target protein to put in its exterior, we need to figure out a way to display it on the outside.

In our literature search, we found that there are several methods of surface display in Synbio:

Class	Method Name	Description	Aspects
Lipoprotein Attachment	Lipoprotein Attachment	A lipoprotein signal or "lipobox" (L-X-Y-C) sequence gets fused to the heterologous protein which sends it to the lipoprotein pathway. 1. Protein Synthesis 2. Signal Peptide Prossesing: Protein is guided through the Sec pathway and signal peptide is cleaved 3. Lipid modification: Lipd moiety is added to exposed cisteine 4. Membrane Insertion: Lipid moiety anchors the protein	Gram Positive & Gram Negative Keeps protein native conformation for accesible interactions: - Surface display systems - Vaccines - Enzyme engineering
Outer membrane protein insertion	Outer Membrane Vesicles	Fuse protein to "outer membrane domains" (e.g. b-Barrel proteins [BAM Complex])	Coding sequence for hyperfolder yellow fluorescent protein (hfYFP). Engineered to withstand denaturing conditions. Codon optmized for B. Subtilis.
Outer membrane protein insertion	Ice Nucleation Protein Display System	Fuse protein to "auto-transporter" system (e.g. tergetting signals). In this case to the Ice Nucleation Protein (INP) INP Secretes, guides and anchors the heteroloous protein	Gram Positive only. Ideal for large proteins. Stable foreign protein expression (avoids cytosolic protease activity). May affect disulfide bonds, and thus protein structure.
Cell Wall Anchoring	Sortase Mediated Anchoring	Fusing heteologous protein with a cell wall-binding motif/domain displaying it on the cell wall.	Gram Positive. Used for antigen presentation Sortase can be co-expressed for better results.
Cell Wall Anchoring	B. subtilis Spore Display System	Fuses heterologous protein to a B. subtilis spore coat protein via a linker. Many proteins are used with varying effects (e.g. CotB, CotC, CotG, OxdD, etc).	Gram Positive. Used for antigen presentation Sortase can be co-expressed for better results.

One of the most common methods in B. subtilis is fusing a protein of interest to a spore coat protein via a peptide linker. This is a relatively common technique for a new line of oral vaccines using B. subtilis spores thanks to their resistance to the GI tract environment. ^{generating...} The problem is that spores are surrounded by protein and lack the peptidoglycan needed for vitellogenin to attach. ^{generating...}

In 2010, Dr. Meimei Han et al., ^{generating...} found that these surface display proteins can be used for vegetative B. subtilis cells (non-spore) when expressed with the Middle Wall Protein promoter (MWPp) that usually controls the expression of the middle wall-associated proteins in vegetative B. brevis. ^{generating...} In his studies he used CotB, a common B. subtilis’ spore coat protein, to display a cholera viral particle on the exterior of the cells which were also confirmed with correct folding by immunofluorescence. ^{generating...}

Therefore, when expressed in vegetative cells, CotB is exported to the cell surface, allowing it to transport attached proteins with it!

Assembly By MoClo

We designed our DNA constructs using the MoClo standard (using the Type IIS Restriction enzymes BsaI and BbsI). ^{generating...} This was chosen as we are familiar with the method and it allows us to implement a high throughput iteration workflow. Once confirmed that the idea works, this will allow for seamless replacement of linkers and proteins exported to the exterior!

We needed to create a system that allows for the expression of CotB attached via a linker to our protein of interest (POI) controlled by the MWPp. Therefore we designed the following general construct:

To make it highly modular, each part was made a “level -1” part extracted from papers, NCBI, and the genome of DWV digitally. Illegal sites (BsaI, BbsI, BamHI, XhoI, BglII, and EcorI) were deleted by synonym mutations to allow for MoClo assembly and PCR modification into BioBrick standard if needed. Coding sequences were also codon optimized for B. subtilis using IDT’s web service for Bacilus subtilis.

Parts selection

We created various constructs centred around CotB as our surface display protein, VP3 as our virus protein, and MWPp as our promoter.

The MWPp is attached to an RBS “AAAGGAGG” determined to be effective in Bacillus. spp and we also selected a B. subtilis compatible double terminator from the iGEM registry. ^{generating...}

Linkers

We developed a rigid 20 amino acid long linker (EAAAKEAAAKEAAAK) and two cleavable linkers to experimentally confirm the successful localization of our proteins on the cell surface. This was done because we lacked the time and resources to get a VP3-specific antibody for immunofluorescence studies.

Being able to detach the protein from the bacteria through methods that affect only the bacteria’s exterior (by cleaving the linker) and to analyze supernatant versus cell debrie is our best second option.

We designed a photo-cleavable and a thrombin-cleavable linker. The photo-cleavable linker is able to cleave at a 400 nm wavelength. ^{generating...}

^{generating...} The thrombin-cleavable linker is able to cleave in the presence of the protease thrombin. ^{generating...} Thereby, if CotB is successfully on the outside of our B. subtilis cells and we treated the cells with a method to cleave the linker then VP3 would end up in the lysate of our cells. Then we could concentrate the VP3 proteins in our lysate for verification purposes.

Detection

To be able to detect if our protein is in fact sent to the outside we needed an economical and easy way of visualizing it. The first thing that came to our mind was an SDS-PAGE. The problem is that the protein is not very well characterized and we run the risk of having a false negative if the protein concentration cleaved is not enough for detection under the control of MWPp.

His Tag

For this, we created a His(8) tag Lv0 part flanked at the N-terminal with glycines (3) that attaches at the C-terminal of VP3 to decrease chances of interaction with the viral protein in case we want to use this tag in our final design. These histidines would be exposed once the linkers get cleaved and it allows us to purify the displayed protein and increase our detection capabilities. ^{generating...} We did not want to display the histidines on the N-terminal portion of our protein as we wanted to minimize any modifications to our “immune-elicitor”.

Florescent tag

After consulting with our advisors, considering we still lacked experience carrying out an SDS-PAGE protocol, and as a mode of easier confirmation, we decided later in the season to add a fluorescent tag.

We initially wanted to use fuGFP as it was provided in the distribution kit (as a biobrick that we ended up PCRing to make MoClo standard):

Nevertheless, after consulting bee experts and our PI (expert in microscopy), we decided to use a different fluorescent protein. This is due to bee tissue’s natural autofluorescence! ^{generating...} Using a green fluorescent protein would made it extremely difficult to discern signals. Therefore we decided on a yellow fluorescent protein (YFP) as it is distinguishable enough and would allow for future immunofluorescent tests of the displayed protein using a blue fluorescence (even more discernible). ^{generating...}

Searching through the literature we found a hyperfolder yellow fluorescent protein (hfYFP) which seemed suitable for our external display fusion use! ^{generating...}

An added benefit is that the yellow fluorescence from the outside of the B. sutbilis cells can be detected by confocal microscopy. ^{generating...} Thereby, if our vaccine is fed to honey bees we can track the transport of our B. subtilis cells containing our VP3-fluorescent proteins across the honey bee body. Suppose we are able to show the localization of our B. sutbilis cells with VP3 in the hypopharyngeal glands or ovaries of our bees would suggest that our cells have been successfully translocated by Vg and would have the potential to be incorporated into the queen ovaries if fed to a worker or queen bee respectively.

The only catch is that a fluorescent protein may have higher chances of altering the folding of the expressed protein (extremely important in vaccine production) and the surface display efficiency due to its size.

That is why both His tag and hfYFP are interchangeable parts allowing for users to choose which detection method is optimal for their protein of interest!

In summary, we have the following parts with their respective flanks for our constructs:

Summary

Our honey bee vaccine comprises of deactivated B. subtilis WB800N cells with VP3 expressed on the cell surface (which initially may be tested alive). These modified cells can be used to vaccinate honey bees by trans-generational immune priming. By expressing the key protein necessary for DWV infection on the surface of B. subtilis and feeding it to honey bees, we aim to enhance the honey bee’s natural defences against DWV, ultimately contributing to the protection and survival of honey bee populations.

Build

We used Golden Gate assembly to put all of our parts together. They were designed as mentioned above and ordered via IDT and Twist as Lv -1 parts. Our backbones are from the MoClo kit. ^{generating...}

Lv-1 were assembled into pAGM9121 (Universal acceptor vector) using BbsI to make Lv0 parts.

Then the parts were assembled into a Lv1 part using pICH47732 (Acceptor Vector A) and BsaI. ^{generating...}

• Middle Wall Promoter Protein (MWPp): Our promoter was used in a paper that used CotB for vegetative protein display. ^{generating...}

• VP3: Our virus protein is a capsid protein responsible for host infection.

^{generating...}

• Linkers: Allow us for the attachment of our virus protein to our surface display protein.

  • Rigid Linker: Short rigid linker. Used for final vaccine product. ^{generating...}

  • Thrombin Cleavable Linker: Rigid linker with a site recognized and cleaved by thrombin. Used for surface display verification. ^{generating...}

  • Photo Cleavable Linker: rigid linker with a site cleavable by ---nm wavelength. Used for surface display verification. ^{generating...}

• CotB: Our cell surface display is a spore coat protein, CotB, which can bring proteins to the cell surface in vegetative B. subtilis cells. ^{generating...}

• Double Terminator: Our terminator is one of the only known double terminators found in B. subtilis constructs (BBa_B0015). ^{generating...}

After learning(make “learning” hyperlinked to the learn section) that our constructs needed to be in a suitable expression vector (pSEVA3b67Rb) different from the Lv2 backbones we had, we then needed to add our Lv1 transcriptional unit (TU) to this new vector.

For that we made 2 primers to clone and modify the ends of the TU of our Lv1s to be cloneable into the new backbone:

We had to find restriction enzymes that were not located in either of our parts (while maintaining the ability to be excised and be used in a Lv2 backbone). We chose XbaI and SacI. ^{generating...} Using classical cloning techniques we inserted our TUs in pSEVA3b67Rb.

—

Blue-white screening and colony PCR were used to select and confirm our constructs. Our constructs with pSEVA3b67Rb were randomly picked from the colonies showing red colour (pseudo-red-white screening).

Test

[We designed a series of tests for verification, including colony PCR, and transformation into B. subtilis WB800N cells, and surface display confirmation through cleavable linkers and localization with fluorescence in bee tissues.]

We designed our constructs to perform the following tests in our iGEM season:

Test 1 CotB Surface Display By Cleavable Linkers

• cells containing constructs with cleavable linkers will be subjected to either 400 nm wavelengths or thrombin.
• If VP3 is isolated from the lysate then surface display can be confirmed. We designed our VP3 proteins with our cleavable linkers to contain a small His tag section on the end of the linker that is still attached to the VP3 protein after cleaving. Using magnetic his tag protein purification we can concentrate the VP3 proteins in the lysate of our cells and run them in an SDS-PAGE. ^{generating...} If the size of our VP3 bands compares to what is expected in literature and in our Benchling simulations (~28kDa), then we can successfully show that while our cells are intact, CotB connected to VP3, ends up on the cell surface of our B. subtilis. ^{generating...}

We also designed linkers that attach to both a hyper folder yellow fluorescent protein (hfYFP) and the VP3 protein. Thereby, CotB could transport our VP3 protein with a fluorescent signal. If the fluorescence successfully presented itself on the outside of our cells as another confirmation of the localization of our proteins to the cell surface of B. sutbilis.

Test 2 Honey bee feeding and translocation of our vaccine

• modified B. subtilis cells will be killed by sonication.
• dead cells will be mixed with sugar water and fed to honey bees.
• after incubating for a few hours, honey bees will be dissected and tissues will be inspected with fluorescence microscopy.

---

Learn

[Throughout the process, we encountered and addressed various challenges, including issues with level 1 backbones, primer design, and illegal BsaI sites in CotB. We optimized our approach based on lessons learned from each stage of development.]

Our first problem was our trials transforming B. subtilis. None of our attempts worked. We dove into literature and found that our vectors would not work for expression in B. subtilis as they lack the proper origin of replication. We consulted with the author of a paper that caught our attention which also contained a kit of highly biosafe vector plasmids very suitable for live use of modified bacteria in the field. ^{generating...} Dr. Tom Ellis gave us support on the use of this plasmid and confirmed its potential use for our project.

These vectors need specific modified strains, so any implementation using these vectors would have to be done once the proof of concept is done. Nevertheless, they have a shuttle vector called pSEVA3b67Rb which allows for replication in E.coli and B. subtilis while also having a constitutive production of cytosolic RFP. ^{generating...}

This was perfect for us as it would help us solve one of the unknown questions in the bee TGIP mechanism. Is the entire bacteria translocated across the bee’s body or only its membrane? By comparing signals of RFP vs hfYFP (or VP3 immunofluorescence) we could answer this and learn more about the bee immune system.

But for this we needed to further engineer our constructs as mentioned in design.

A second problem we faced was continuous difficulties getting transformants in our first Lv1 assemblies. We found later checking our sequences with Snapgene that CotB had an illegal BsaI site in the middle. As CotB was in most of our constructs, this explained our low yields as it increased the number of inserts, affecting our mole ratio calculations. A new version was ordered from IDT as directed mutagenesis was not in our expertise. This, nevertheless, caused a delay in our constructs as we received it later in the season.

A third problem was that some of our Lv0 parts were smaller than 100bp. Our primers for our MoClo parts are standardized to use a single primer for all Lv0s but this produced small amplified fragments (<150bp). This is quite difficult to discern with our gel composition and DNA ladder (1Kb Plus from NEB). For this, we designed a new primer for smaller fragments producing 1.5Kb amplification bands. As our second primer is sequence-specific, we only needed to see a band of the correct size to have a good confirmation.

Test 1 Confirmation of level 0 parts

• forward and reverse primers were created for colony PCR verification of our constructs in E. coli DH5a

Test 2 Level 1 assembly and confirmation

• forward and reverse primers were created for colony PCR verification of our constructs in E. coli DH5a

Test 3 Digestion and ligation of our level 1’s into pSEVA3b67Rb backbone

• forward and reverse primers were created for verification of our level 1 assemblies in the B. subtilis compatible backbone, pSEVA3b67Rb

Test 4 Transformation into B. subtilis WB800N cells

• pSEVA3b67Rb contains a cytosolic red fluorescent protein so we can verify that cells were successfully transformed by the colour of the colony. Colonies with RFP should appear pink.
• Colony PCR verifies the correct assembly of the insert into the backbone

Test 2 and 3 have been partially completed and verified. Repeated errors in our assemblies and lack of colonies in our transformations led us to tackle our assemblies with another approach as well. We created assemblies that allowed us to build our level 1 transcriptional units with 1 golden gate reaction. These transcriptional units are assembled as linear pieces of DNA that can then be digested to create SacI and XbaI sticky ends that are compatible with our B. subtilis backbone, pSEVA3b67Rb. A simple ligation of our transcriptional units into the backbone can be performed with one simple reaction. We have verified 2 of our 4 ordered constructs. We are now able to attempt these transformations in the B. subtilis cells.

Test 4 was originally attempted with an unknown B. subtilis strain supplied by our university science department. Several different protocols were attempted for transformation by electroporation, however, no transformations were successful. We even attempted an obscure chocolate milk-mediated transformation protocol from 1998. ^{generating...} Towards the end of the iGEM season, we were forced to purchase the expensive but ideal B. subtilis strain WB800N to attempt the transformation of our constructs. Transformations are still underway.