Design
Overview
As allergies affect one billion people worldwide and can be life threatening, it is important they are well diagnosed before the patient experiences serious allergic reactions. However, the current detection tests are limited in the number of allergens per test and some allergies are hard to diagnose.
This is why we want to create a new method to detect the sensitivity of patients to any allergen. Here is how we have imagined this innovative method.
Antibodies are able to bring cells together
All allergic patients possess specific Immunoglobulins E antibodies (IgE; Figure 1), able to bind to allergens in their blood. When they do so, the allergic reaction is triggered. The detection of IgE is a relevant way to diagnose allergies. This requires knowing their structure and function.
Figure 1: General structure of IgE antibodies. IgE are composed of a heavy 𝛆 chain (in dark blue) and a light chain (in light blue) with a variable fragment (in orange).
There are different classes of antibodies, differentiated by their heavy chain (Stavnezer. 1996). IgE are the class possessing constant 𝛆 heavy chain (in deep blue on Figure 1). Antibodies also have a variable fragment which recognizes a specific antigen (Hiemstra et al. 1998). In the case of IgE, the variable fragment binds to a specific allergen.
Antibodies are selected in the body to have nanomolar to picomolar affinity to molecules perceived by the body as pathogens. For example, IgG, another class of antibodies, can bind to most viruses and pathogen microorganisms (Figure 2). Once their variable fragment is linked to the pathogen, IgG can bind to a phagocyte by their constant fragment and the phagocytosis is triggered to destroy the pathogen through phagocytosis (Hiemstra et al., 1998).
Figure 2: Phagocytosis mechanism. (Pieter S. et al. 1998)
This means one property of IgG is to bring cells close together. A similar mechanism exists with IgE but the antibody binds allergens instead of microorganisms and, once linked to the immune cell, triggers the allergic reaction instead of phagocytosis.
Bacterial aggregation could attest for the presence of IgE
What if the presence of IgE could be assessed by clumping engineered cells together, like IgG could do? This biomimetic idea was the core of our project.
Using bacterial aggregation as a molecular marker has already been performed (Kylilis et al. 2019). For example, synthetic biology was used to display anti-GFP nanobodies at the surface of E. coli (Kylilis et al. 2019). This way, the bacteria aggregated in the presence of extracellular dimeric GFP. This technique can be used with a variety of biomarkers as long as the binding site of the biomarker is displayed at the surface of the bacteria (Figure 3).
Figure 3: low-cost protein biomarker test for point-of-care medical diagnostics. E. coli cells with surface-displayed nanobodies are mixed with a biological sample (e.g., blood) from a patient. If the patient sample includes the protein biomarker indicative of disease state, this will cause cross-linking between bacterial cells and biomarker molecules that leads to cell agglutination and changes the visual appearance of the test sample (Kylilis et al. 2019).
From this, we imagined a system where the receptor of the IgE constant fragment is displayed at the surface of E. coli mixed with E. coli cells displaying allergens. These two strains (respectively named strain D and A), should aggregate in presence of IgE, as one strain binds the constant fragment and the other binds the variable fragment of the same IgE (Figure 4). This aggregation should occur only in presence of IgE with a specificity for the displayed allergen. Knowing which allergen is expressed at the surface of the second strain, the aggregation could be used to characterize the specificity of the IgE. Using a blood sample (or serum sample), this would be an innovative way to demonstrate the presence of any IgE in the blood, their ability to bind to allergens, and hence, a proof of the allergy predisposition of the blood-donor.
Figure 4: theoretical behavior of strains A and D in presence of IgE recognizing the allergens displayed by strain A (A), in absence of IgE (B). The bacteria aggregate only when there is IgE (in blue and orange) to bind to the proteins displayed by strain A (orange) and strain D (blue).
Recombinant allergens offer many possibilities for research and diagnosis
Most current allergenic tests use purified natural extract. In our new idea, the detection of IgE is based on the display of recombinant allergens on the surface of E. coli. It turns out that using recombinant allergens has many advantages.
First of all, natural extracts are mixtures of several allergens. An unique allergenic source indeed contains several different allergens. For example, there are 18 different allergens in peanuts (WHO/IUIS Allergen Nomenclature home page) and some could be sensitive to the purification process and the storage condition (Ansotegui, 2020). In contrast, recombinant allergen should allow a molecular and robust identification of the incriminated allergen.
Producing recombinant allergens also allows to mutate randomly the gene coding for an allergen and analyze how the mutated allergen influences the allergic reaction. Today, more than 700 genes coding for allergens have been identified (Bannon, 2014). It means that 700 recombinant allergens could theoretically be produced. Banks of allergens exist but using E. coli for our test offers the promising possibility to test an infinity of mutated allergens in order to find the ones with the highest specificity toward the patient IgE. This could be of dramatic importance for the identification of allergens highly specific to each individual patient and a major breakthrough in the development of efficient custom-made desensitization treatment.
To sum up, the objective of the DAISY project is to create two strains of E. coli:
- One strain displaying a receptor to the constant fragment of the IgE on its surface, called strain D.
- One strain displaying a recombinant allergen on its surface, called strain A. Different allergen displaying strains will be created as a proof of concept of an E. coli library of allergens.
Here are explained our technical choices for the creation of these strains.
Strain A, displaying allergens on its surface
Plasmid and system of expression
First, we wanted to control the expression level of the allergen to be displayed in order to maximize the number of allergens on the outer membrane without impacting E. coli metabolism or producing inclusion bodies. We chose to use the T7 promoter which is inducibled with Isopropyl β-D-1-thiogalactopyranoside (IPTG) and widely used in E. coli (Weng et al., 2010). In this scope, the pET21b (+) plasmid (Figure 5) was selected since it contains a T7 promoter and terminator as well as ampicillin resistance for easy selection (gift from Dr. Ambre Jousselin). We chose to work with E. coli Tuner (DE3) strain to control level expression of recombinant proteins by tuning the IPTG concentration. As every (DE3) E. coli strain, Tuner expresses the T7 ARN polymerase required to use the T7 promoter.
Figure 5: Structure of the pET21b(+) plasmid. It allows for ampicillin resistance for positive selection and protein expression regulated by the T7 IPTG inducible promoter.
Allergen display for strain A
To display allergens on the surface of E. coli, it was chosen to fuse them to the transmembrane protein OmpA. OmpA is the most abundant membrane protein in E. coli, with 100,000 copies on average (Ortiz-Suarez et al., 2016) and it possesses convenient loops outside the outer membrane (shown in red on figure 6).
If the allergen is fused to a loop outside the outer membrane, it should be displayed at the external bacterial surface. We chose a chimeric OmpA which is slightly shorter and previously used for a similar application (Yang et al., 2016). A GISS linker has been added between OmpA and the allergen to allow a better accessibility of the allergen to the IgE for aggregation.
Figure 6: 3D structure of OmpA from E. coli(PDB: 1G90 ). The arrows represent 𝛽-sheets which here form a 𝛽-barrel across the outer membrane (Ortiz-Suarez et al., 2016). The loops in red face the extracellular media while the blue-ones face the periplasm.
We then added other elements to the OmpA-allergen fusion in order to perform displaying controls (Figure 7).
Figure 7: final construction to express allergen at E. coli surface. RBS: ribosome binding site present on the plasmid. OmpA to display the allergen at the cell surface. GISS: linker to improve the accessibility of the displayed proteins. TEV: binding site for TEV protease. HIS tag : Histidine tag.
A 6xHis-tag was added at the end of the fusion protein to check that the allergen is successfully displayed at the surface of the bacteria with anti-His-tag fluorescent antibodies and to facilitate further protein purification. A TEV protease site (Parks et al. 1994) was inserted between OmpA and the allergen sequence to give the possibility to remove the allergen sequence for control purposes.
The choice of allergens
In the idea of building a proof-of-concept library of allergens, we found 6 different allergens that have already been expressed in E. coli and demonstrated to be recognized by their corresponding human IgE (Table 1).
Table 1: List of the ordered allergen sequences
| Chosen allergen | Accession number on NCBI database | Allergenic source Number |
References about the successful expression of the allergen in E. coli |
|---|---|---|---|
| Gal d 2 | V00383.1 | Hen’s egg | Dhanapala et al. 2014 |
| Ana o 3 | AAL91665.1 | Cashew nut | Robotham et al. 2005 |
| Der p 1 | U11695.1 | Acarid | Bussière et al. 2010; Pulsawat et al. 2010 |
| Der p 2 | AF276239.1 | Acarid | Bussière et al. 2010 |
| Ara h 2 | AY158467.1 | Peanut | Lehmann et al. 2003 |
| Fel d 4 | AAS77253.1 | Cat | Smith et al. 2004 |
These allergens are issued from different sources such as egg, nuts, acarid or cat. We only bought the purified IgE targeting Der p 1 and Ara h 2 since they are as expensive as a participation to iGEM. The other allergens are thus intended to be used as negative control for the aggregation tests. The DNA construction to express and display these allergen was built to be generic for the whole library, requiring only to replace the allergen by In-Fusion to construct a new plasmid for each allergen (Figure 8).
Figure 8: Experimental design to replace the allergen in the plasmid of strain A. Gal d 2 replacement by Ana o 3 was used to illustrate the strategy.
Strain D displaying DARPin on its surface
To obtain IgE linked by their constant fragment to E. coli cells, we first thought about expressing Fc𝜀RI, the human receptor for IgE, on the surface of E. coli (Blank et al. 2003; Gasser et al. 2018; Figure 9). However, agencing the 4 chains of the human protein in E. coli appeared as a too difficult challenge during a short iGEM project.
Figure 9: Two Fc𝜀RI at the surface of a mast cell. Fc𝜀RI is usually displayed as a dimer. When two IgE bind the receptors, the allergic reaction is triggered.
Looking for an alternative, we found DARPin synthetic proteins (acronym for Designed Ankyrin Repeat Proteins, in blue in figure 10). These proteins have been designed to have a high affinity for a variety of ligands (Fornasiero et al., 2014). Among them, DARPin e2_79 has the better affinity and specificity for IgE (Baumann et al., 2010). It is much shorter than the complete Fc𝜀RI IgE receptor (0.375 kb versus 1.7 kb; Uniprot: P12319, P30273, Q01362), has been successfully expressed in E. coli, and has a better affinity for IgE as, in equimolar proportion of DARPin e2_79 and Fc𝜀RI, 80 % of IgE binds to the DARPin (Baumann et al., 2010).
Figure 10: 3D structure of DARPin e2_79 and IgE.The DARPin (in blue) is composed of a repetition of 𝞪-helix. The IgE constant fragment (in orange) has high affinity for the DARPin. (PDB: 4GRG).
The same system of expression as for the allergens of strain A was used with an OmpA fusion to display the DARPin on the cell surface. Likewise, we used the same GISS linker for accessibility, TEV site and His-tag for control purposes (Figure 11). For additional control, a gene coding for sfGFP was added at the N-terminus of the DARPin to report for both its expression and display. sfGFP is a mutant of the GFP, the Green Fluorescent Protein, displaying a higher folding ability in external medium, so well suited for our purpose.
Figure 11: Construction added to E. coli to display DARPin_sfGFP fusion protein at the membrane surface.RBS: ribosome binding site. OmpA: display of the allergen at the cell surface. GISS: a linker to improve accessibility. TEV: a binding site for TEV protease. HIS tag : Histidine tag. sfGFP: reporter of the expression and display.
Our two strains A and D are designed to favor agglutination in presence of specific IgE. This method was conceived not only to test single allergen but also to be compatible with high-throughput screening approaches. Both strategies are described below.
Targeted detection of allergen predisposition through microscopy approach
Thanks to strains A and D, aggregates should form only in presence of IgE with a variable fragment specific to the allergen displayed by strain A. This point needs to be first validated using commercially available IgE. We supposed the aggregation will be affected by a variety of parameters (cell concentrations, IgE concentration, temperature, stirring, etc.). This prompted us to develop an innovative model to pre-determine the most promising conditions while in the same time assessing many of them at the bench. To detect agglutination, we first opted for a simple approach by confocal microscopy. However, because microscopy may present limitations in the number of allergens to be tested, high-throughput screening technologies would be required.
High-throughput approaches
Our design allows us to create a library of E. coli strain A displaying each of the 700 known allergens or even more with degenerated sequences. Thus, a screening procedure would result in an automatic mixing of cells from strain A with a sample of the patient’s blood in presence of bacteria from strain D. Aggregation would result from the presence of an IgE in the considered blood sample (Figure 12). With this method, the sensitivity of the patient for any possible allergen could be detected in the same test and with only one sample of the patient’s blood.
Figure 12: Formation of aggregates in presence of IgE, strain D and a mix of strains A displaying a variety of allergens.
As the different strains A are mixed together, a further step is needed to find out which allergen binds the patient’s IgE. The plasmids of the bacteria A which formed aggregates have to be sequenced to find out the allergen they displayed. But first, the aggregates must be sorted and this is why high-throughput technologies are required.
Flow cytometry is a performant technology to screen for aggregates
Flow cytometry is an emerging technology allowing analyzing individual cells from a suspension with a throughput of 10,000 cells per second (Brehm-Stecher, 2014). A flow cytometer aspirates the cells in a passage narrow enough to have the cells placed one behind the other (Figure 13).
Cells of interest express a fluorescent protein, enabling them to be discriminated against by those not expressing it using a laser beam. Some flow cytometers, called cell sorters, are able to separate the cells with a defined fluorescent wavelength and collect them for further experiments (Nexmann et al. 2014).
Figure 13: Cell sorting of a suspension of strains A, strain D and IgE.
Hence, if strain A and D expressed different wavelength fluorescent proteins, then the cell sorter is theoretically able to collect aggregates presenting both fluorescences. Afterward, presence of plasmid could be detected by PCR using dedicated primers and sequenced. In the Daisy project, strain A will express the red fluorescent mRFP1 protein while strain D expresses the blue fluorescent mTagBFP protein (Figure 14). By doing so, the free cells A will be perceived as a red signal, free cells D as a blue signal and the specific aggregates containing cells A and D linked by the patient’s IgE as a blue-red double signal. The cell sorter can then be instructed to sort only the double signals out to isolate the aggregates we are looking for.
Figure 14: A. Construction added to E. coli to display an allergen and constitutively express a red fluorescent protein.The sequence is the one presented in figure 7 with addition of the ihfb800 promoter and mRFP1, a gene coding for a red fluorescent protein.B.Construction added to E. coli to display a DARPin and constitutively express a red fluorescent protein.The sequence is the one presented in figure 11 with a deletion of sfGFP and addition of the ihfb800 promoter and mTagBFP, a gene coding for a blue fluorescent protein
These aggregates will then be collected in separate wells on a microtitration plate and grown on rich media prior to sequencing to find out which allergens are displayed in the cells A.
FACS is effective and has a particularly high-throughput. However, it has limitations. As the aggregates appear in a mix of thousands of different strains A, if the patient has several allergies, the aggregates can contain bacteria displaying different allergens. This may cause difficulties for the sequencing. Furthermore, If the aggregates are too big, they may not be able to fit in the passage of the cell sorter and lead to false negative results.
Microdroplet allows the separation of each allergen
With the willingness to explore the versatility of the DAISY detection method, we contacted Dr. Gabrielle Potocki-Veronesse from TBI to discuss microdroplets technology. Microdroplets are made by encapsulation of an of culture medium droplets of 13 µm diameter by oil. The choice of the input concentration allows us to encapsulate a single cell A into a droplet and some cells D (as consequence, most droplets will be deprived of strain A). Thus, each droplet containing a strain A cell tests a different allergen.
The detection of aggregate will be analyzed by an Artificial Intelligence that we will train to recognize aggregates in the droplets (Figure 15). The droplets containing aggregates will be labeled with fluorescence. This technology is patented by INRAE.
Droplets can then be sorted by FACS to collect only the labeled droplets and to sequence the plasmid of strain A to identify the allergen responsible for the aggregation.
Figure 15: General scheme of the microdroplets strategy used for aggregation detection.IgE (black points) are mixed with strain A (orange with colorful membrane according to the displayed allergen) and strain D (in blue) and encapsulated in oil (beige) as microdroplets. There is one strain A cell maximum per droplet but several strain D cells. An artificial intelligence will be trained to recognize aggregates and, if so, label the droplet. The droplets are then sorted by FACS.
Using droplets presents one disadvantage in comparison to the FACS: the analysis rate of droplets is only 91,000 droplets per hour (Potocki, personal communication, April 14, 2022), and the majority of droplets does not contain strain A due to the low concentration of strain A in the input flow so the actual screening flow rate is 9,000 droplets per hour.
Despite the lower analysis rate of the droplets, the ability to isolate the allergens of the bank make this technology very relevant for our purpose.
