Collection of allergens (strains A)

The leading idea of our project is to screen hundreds of allergens. As a proof of concept, we chose to focus on engineering a small number of allergens: Gal d 2, Ara h 2, Der p 1 and Ana o 3 (see Design).

---

Gal d 2 strain

Figure 1: OmpA_Gal d 2 fragment.The expression of OmpA_Gal d 2 construction is controlled by the T7 promoter. OmpA was linked to TEV+Gal d 2 by a GISS linker. Gal d 2 was linked to the HISTag by the same linker.

OmpA_Gal d 2 fragment from IDT gblock (shown above in Figure 1) was amplified by PCR using the high fidelity Phusion polymerase with primers IF3_allergen-F and IF4_Gal D2/DARPin-R (see Primers). Expected size of the amplicon was 1675 bp.

pET-21 b (+) vector was linearized by PCR using the high fidelity Phusion polymerase with primers IF1_GalD2/DARPin-F and IF2_plasmid-R (see Primers). Expected size of the amplicon was 5442 bp.

Amplification product sizes were checked on EtBr stained agarose electrophoresis gel (see Electrophoresis) (Figure 2).

Figure 2: pET-21 b (+) linearized (A) and Gal d 2 amplified fragment (B). Expected sizes of the amplicons were 5442 bp (A) and 1675 bp (B). PCR amplicon sizes of pET-21 b (+) (A) and Gal d 2 (B) were checked with agarose electrophoresis gel and revealed with EtBr. A theoretical gel is presented with each gel and the NEB 1 kb DNA ladder is used for the experimental gels. (note that a different ladder is presented on the theoretical gel).

Amplification products matched the expected size, they were further purified from the gel (see Gel extraction).

The Gal d 2-OmpA construction was then inserted into pET-21 b (+) by In-Fusion®. The resulting products were transformed into Stellar competent cells. Transformants were selected on LB-agar-ampicillin plates (see Cultures). 20 transformants were screened by colony PCR with primer pairs flanking the insertion zone ( screening_inserts-F and screening_inserts-R, expected size of the amplicon : 2092 bp) (see Primers). 2 positive transformants were detected (Figure 3).

Figure 3: identifying fragments that bear pET21 b (+)_Ompa_Gal d 2 by colony PCR. Expected size of the amplicon was 2092 bp. The positive clones were colonies 17 and 24. PCR amplicon sizes of colonies with Gal d 2 plasmid were checked with agarose electrophoresis gel and revealed with EtBr. A theoretical gel is presented with each gel and the NEB 1 kb DNA ladder is used for the experimental gels (note that a different ladder is presented on the theoretical gel).

These transformants had their plasmid extracted by Miniprep and digested by EcoRV to assess the assembly (expected fragments at 4332 bp and 2785 bp, see Figure 4).

Figure 4: restriction profile of pET-21 b (+)_OmpA_Gal d 2 final construction. Enzyme used was EcoRV. Expected sizes of the amplicons were 4332 bp and 2785 bp. Plasmids were checked with agarose electrophoresis gel and revealed with EtBr. A theoretical gel is presented with each gel and the NEB 1 kb DNA ladder is used for the experimental gels (note that a different ladder is presented on the theoretical gel).

The correct restriction maps were obtained and the insert sequence was further validated by sequencing. The plasmid was named pET-21 b (+)_OmpA_Gal d 2. The plasmids were eventually used to transform E. coli Tuner cells in order to express the OmpA_Gal d 2 construction at the cell membrane (see Preparation of chemically competent cells E. coli).

After cloning this first allergen, the plasmid obtained was used as a basis to build the other allergen constructions of our bank: Ara h 2, Der p 1 and Ana o 3.

---

Ara h 2 strain

Figure 5: OmpA_Ara h 2 fragment. The expression of OmpA_Ara h 2 construction is controlled by the T7 promoter. OmpA was linked to TEV+Ara h 2 by a GISS linker. Ara h 2 was linked to the HISTag by the same linker.

Ara h 2 gene ordered on IDT (shown above in Figure 5) was amplified by PCR using the high fidelity Phusion polymerase with primers IF3_allergen F and IF4_Ara h 2 (see Primers). Expected size of the amplicon was 567 bp.

In parallel, pET-21 b (+)_Gal d 2_OmpA vector was linearized, excluding the Gal d 2 fragment by PCR. The primers used were IF1_allergen and IF2_plasmid (see Primers). Expected size of the amplicon was 5924 bp.

Amplification product sizes were checked on EtBr stained agarose gel (Figure 6).

Figure 6: pET-21 b (+)_OmpA linearized with Gal d 2 exclusion (A) and Ara h 2 amplified fragment (B). Expected sizes of the amplicons were 5924 bp (A) and 567 bp (B). PCR amplicon sizes of pET-21 b (+)_OmpA (A) and Ara h 2 (B) were checked with agarose electrophoresis gel and revealed with EtBr. A theoretical gel is presented with each gel and the NEB 1 kb DNA ladder is used for the experimental gels (note that a different ladder is presented on the theoretical gel).

Amplification products matched expected sizes, they were further purified from the gel.

The Ara h 2 fragment was then inserted into pET-21 b (+)_OmpA by In-Fusion. The In-Fusion assemby reaction was transformed into Stellar competent cells. Transformants were selected on LB-ampicillin plates. 15 transformants were then screened by colony PCR with primer pairs flanking the insertion zone (primers used: screening_inserts-F and screening_inserts-R, expected size of the amplicons: 1450 bp) (see Primers). 13 positive transformants were detected (Figure 7).

Figure 7: Identifying strains that bear pET21 b (+)_OmpA_Ara h 2 by colony PCR. The expected size of the amplicons was 1450 bp. 13 positive transformants were detected. PCR amplicon sizes of colonies with Ara h 2 plasmid were checked with agarose electrophoresis gel and revealed with EtBr. A theoretical gel is presented with each gel and the NEB 1 kb DNA ladder is used for the experimental gels (note that a different ladder is presented on the theoretical gel).

Four of these transformants (colonies 21, 22, 23, 24) had their plasmid extracted by Miniprep. Sequences were all validated by Sanger sequencing. The plasmid was named pET-21 b (+)_OmpA_Ara h 2.

The plasmid was finally used to transform E. coli Tuner cells to express OmpA_Ara h 2 at the cell membrane.

---

Der p 1 strain

Figure 8: OmpA_Der p 1 fragment. The expression of OmpA_Der p 1 construction was controlled by the T7 promoter. OmpA was linked to TEV+Der p 1 by a GISS linker. Der p 1 was linked to the HISTag by the same linker.

Der p 1 construction from IDT gblock (shown above in Figure 8) was amplified by PCR using the high fidelity Phusion polymerase with the primers IF3_allergen F and IF4_Der p 1 (see Primers). Expected size of the amplicon was 994 bp.

Amplification product sizes were checked on EtBr stained agarose gel (Figure 9).

Figure 9: Der p 1 amplified fragment. Expected size of the amplicons was 994 bp. PCR amplicon sizes Der p 1 were checked with agarose electrophoresis gel and revealed with EtBr. A theoretical gel is presented with each gel and the NEB 1 kb DNA ladder is used for the experimental gels (note that a different ladder is presented on the theoretical gel).

PCR products matched expected sizes and amplicons were further purified from the gel. The Der p 1 construction was inserted into linearized pET-21 b (+)_OmpA (see Figure 6) by In-Fusion.

The In-Fusion assemby reaction was transformed into Stellar competent cells. Transformants were selected on LB-ampicillin plates. 15 transformants were screened by colony PCR with primer pairs flanking the insertion zone ( screening_inserts-F and screening_inserts-R; expected size of the amplicons: 1894 bp) (see Primers). 12 positive transformants were detected (Figure 10).

Figure 10: identifying strains that bear pET21 b (+)_OmpA_Der p 1 by colony PCR. The expected size of the amplicons was 1894 bp. 12 positive clones were detected. PCR amplicon sizes of colonies with Der p 1 plasmid were checked with agarose electrophoresis gel and revealed with EtBr. A theoretical gel is presented with each gel and the NEB 1 kb DNA ladder is used for the experimental gels (note that a different ladder is presented on the theoretical gel).

Four of these transformants (colonies 2, 6, 11 and 13) had their plasmid extracted by Miniprep and digested by SalI (expected size of the fragments: 6 kb and 1 kb) or double-digested by HindIII and EcoRI-HF (expected size of the fragments: 5.5 kb and 1.5 kb) to assess the assembly (Figure 11).

Figure 11: restriction profile of pET-21 b (+)_OmpA_Der p 1 for one of the clones. Enzymes were SalI for the simple digestion (expected sizes of the fragments were 6.0 kb and 1.0 kb) and EcoRI and HindIII for the double digestion (expected sizes of the fragments were 5.5 kb and 1.5 kb). Plasmids were checked with agarose electrophoresis gel and revealed with EtBr. A theoretical gel is presented with each gel and the NEB 1 kb DNA ladder is used for the experimental gels (note that a different ladder is presented on the theoretical gel).

The correct restriction maps were obtained for the clones which were further validated by sequencing. The plasmid was named pET-21 b (+)_OmpA_Der p 1.

The plasmids were finally used to transform E. coli Tuner cells to express the Ompa_Der p 1 construction at the cell membrane.

---

Ana o 3 strain

Figure 12: OmpA_Ana o 3 fragment. The expression of OmpA_Ana o 3 construction was controlled by the T7 promoter. OmpA was linked to TEV+Ana o 3 by a GISS linker. Ana o 3 was linked to the HISTag by the same linker.

Ana o 3 construction from IDT gblock (shown above in Figure 12) was amplified by PCR using the high fidelity Phusion polymerase with the primers IF3_allergen F and IF4 Ana o 3 (see Primers). Expected size of the amplicon was 479 bp.

Amplification product sizes were checked on EtBr stained agarose gel (Figure 13).

Figure 13: Ana o 3 amplified fragment. Expected size of the amplicons was 479 bp. PCR amplicon sizes Ana o 3 were checked with agarose electrophoresis gel and revealed with EtBr. A theoretical gel is presented with each gel and the NEB 1 kb DNA ladder is used for the experimental gels (note that a different ladder is presented on the theoretical gel).

The products matched expected sizes and amplicons were further purified from the gel. The Ana o 3 construction was inserted into pET-21 b (+)_OmpA linearized (see Figure 6) by In-Fusion.

In-Fusion assemby reaction was transformed into Stellar competent cells. Transformants were selected on LB-ampicillin plates. 15 transformants were screened by colony PCR with primer pairs flanking the insertion zone (primers used: screening_inserts-F and screening_inserts-R) (see Primers). 4 positive transformants were detected (Figure 14).

Figure 14: pET21 b (+)_OmpA_Ana o 3 construction fragments from colony PCR. PCR amplicon sizes of colonies with Ana o 3 plasmid were checked with agarose electrophoresis gel and revealed with EtBr. A theoretical gel is presented with each gel and the NEB 1 kb DNA ladder is used for the experimental gels (note that a different ladder is presented on the theoretical gel).

These transformants (colonies 1, 3 and 9) had their plasmid extracted by Miniprep and digested by Eco-RI and Eco-RV (expected size of the fragments: 5052 bp and 3211 pb) to assess the assembly (Figure 15).

Figure 15: restriction profile of pET-21 b (+)_Ana o 3 final construction. Enzymes were EcoRI and EcoRV. Plasmids were checked with agarose electrophoresis gel and revealed with EtBr. A theoretical gel is presented with each gel and the NEB 1 kb DNA ladder is used for the experimental gels (note that a different ladder is presented on the theoretical gel).

The correct restriction maps were observed and these clones were further validated by sequencing. The plasmid was named pET-21 b (+)_OmpA_Ana o 3.

The plasmid was finally used to transform E. coli Tuner cells to express the OmpA_Ana o 3 construction at the cell membrane.

Achievements so far: we managed to have all our allergen construction correctly cloned.

---

Controls of correct protein expression

We first assessed the expression of the allergens at the outer surface of the bacteria, an important control experiment prior to the aggregation experiments. As only anti-Ara h2 and anti-Der p 1 IgE were available, we prioritized the verification of these allergen expressions first. Two experiments were conducted to do so: one with a fluorescent anti-HIS-tag IgG and one with an IgE coupled to a fluorescent anti-IgE IgG.

Test of Ara h 2 protein expression by HIS-tag detection

As the protein was HIS-tagged, we first tried to detect it by binding a fluorescent anti-HIS-tag IgG to the cell membrane and measuring fluorescence (see HIS-tag detection by immunofluorescence). Cells without IgG were observed as negative control. Microscope observations are shown in Figure 16. The microscope parameters to observe GFP were the following: excitation filter 360/40, dichroic beamsplitter 400 nm, emission filter 470/40.

Figure 16: fluorescence of E. coli Tuner cells containing the pET-21 b (+)_OmpA_Ara h 2 vector mixed with 1:100 anti-HIS-tag IgG (A), or without IgG for negative control (B). Each image represents an observation in phase contrast (on the left) and in fluorescence (on the right). Observations were made at the microscope parameters for GFP.

No difference in fluorescence was visible between the control and the sample. The only fluorescence observed was due to autofluorescence of cells, so we could not conclude if the protein was expressed at the membrane, especially without a positive control (like binding the protein to a HIS-tag resin and observing fluorescence).

Test of the allergens functionality by double antibody recognition (ELISA test)

We assessed the functionality of the allergens by binding specific IgE to them (anti-Ara h 2 and anti-Der p 1), and then binding a fluorescent anti-IgE IgG to the complex (see Allergen detection and recognition on the surface by immunofluorescence). No fluorescence was observed in either case (data not shown). Without a positive control, we could not conclude between an incorrect exposition of the protein (not facing the outer surface), or a mere problem during the experimental setup.

Construction of strain D

To be able to screen all these allergens, we also needed to construct our strain D, displaying DARPin proteins on its surface. The first step was the construction of a DARPin-sfGFP strain to prove the correct expression of our protein and to evaluate its subcellular localization. The sfGFP sequence was later deleted to obtain the final strain D.

---

DARPin-sfGFP strain

A second strain was generated with the OmpA_DARPin-sfGFP construction in order to evaluate the outer exposition of the fusion protein at the cell membrane. The aim was to later delete the sfGFP sequence from the plasmid to obtain the DARPin strain.

Vector construction

Figure 17: DARPin-sfGFP construction. The sfGFP served as a marker to prove the correct expression of the protein.

OmpA_DARPin-sfGFP fragment from IDT (shown above in Figure 17) was amplified by PCR using the high fidelity Phusion DNA polymerase with primers IF3_allergen-F and IF4_Gal d 2/DARPin-R (see Primers). The expected size of the amplicon was 1468 bp.

The pET-21 b (+) vector was linearized by PCR using the high fidelity Phusion DNA polymerase with primers IF1_Gal d 2/DARPin-F and IF2_plasmid-R (see Primers). The expected size of the amplicon was 5442 bp.

Amplification product sizes were checked on EtBr stained agarose gel. Amplicons were further purified from the gel. Figure 18 shows the amplification product of the OmpA_DARPin-sfGFP construct matching the expected size. For the plasmid related amplicon, see Figure 2.

Figure 18: OmpA_DARPin-sfGFP fragment amplified by PCR. The expected size of the amplicon was 1468 bp. The PCR amplicon size was checked with agarose electrophoresis gel and revealed with EtBr. A theoretical gel is presented on the left and the NEB 1 kb DNA ladder was employed for the experimental gel. (note that a different ladder is presented on the theoretical gel).

The DARPin-sfGFP construction was then inserted into our linearized pET-21 b (+) by In-Fusion to assemble the pET-21 b (+)_OmpA_DARPin-sfGFP plasmid.

The In-Fusion mixture was first transformed into chemically competent E. coli Stellar cells. Transformants were selected on LB-ampicillin plates. 42 transformants by colony PCR with a primer pair flanking the insertion zone (primers used: screening_inserts-F and screening_inserts-R, expected size of the amplicon : 1885 bp) (see Primers). 18 positive transformants were detected. Figure 19 shows an example of such positive clones.

Figure 19: identifying strains that bear pET-21 b (+)_Ompa_DARPin-sfGFP by colony PCR. The expected size of the amplicon was 1885 bp. The positive clones were colonies 13, 14, 15 and 16. The PCR amplicon size was checked with agarose electrophoresis gel and revealed with EtBr. A theoretical gel is presented on the left and the NEB 1 kb DNA ladder was employed for the experimental gels (note that a different ladder is presented on the theoretical gel).

Five of those transformants had their plasmid extracted by Miniprep and digested with the restriction enzyme SalI to assess the assembly. The expected fragments were expected at 5574 bp and 1336 bp. Figure 20 shows correct restriction maps for all clones. The assembled plasmids were further validated by Sanger sequencing.

Figure 20: restriction profile of the pET-21 b (+)_OmpA_DARPin-sfGFP final construction. The enzyme used was SalI. The expected sizes of the fragments were 5574 bp and 1336 bp. All clones had correct restriction profiles. Plasmids were checked with agarose electrophoresis gel and revealed with EtBr. A theoretical gel is presented on the left and the NEB 1 kb DNA ladder is used for the experimental gel (note that a different ladder is presented on the theoretical gel).

The plasmids were finally transformed into chemically competent E. coli Tuner cells that were used to verify the OmpA_DARPin-sfGFP construction. It was then necessary to identify the best induction conditions and to assess whether our fusion protein was well presented onto the outer surface of the cell membrane.

Controls to verify the correct protein expression

As the presence of the OmpA_DARPin(-sfGFP) fusion protein at the outer surface was key for our aggregation experiments, we tried to assess its cellular localization through several experiments. First of all, we needed to verify whether the protein was well expressed by the bacteria. To do so, we performed observations under an epifluorescent microscope. The goal was then to verify precisely the OmpA_DARPin-sfGFP localization , which we did by fractionating the bacterial cells and monitoring fluorescence emission with a microplate reader. These experiments are detailed hereafter.

• Validation of the protein expression by microscopy

The expression of the OmpA_DARPin-sfGFP encoding gene in E. coli Tuner cells was induced with a concentration of 50 µM of IPTG. The empty pET-21 b (+) plasmid was used as a negative control here. After 4 hours of incubation at 37°C, fluorescence was observed on an epifluorescence microscope as shown on Fig. 21.

Figure 21: fluorescence of E. coli Tuner cells containing the pET-21 b (+)_OmpA_DARPin-sfGFP vector (A) or the empty pET-21 b (+) plasmid (B) as a negative control. Each image represents an observation in phase contrast (on the left) and in the GFP fluorescence channel (microscope parameters for GFP).

Some fluorescence emission was clearly observed with pET-21 b (+)_OmpA_DARPin-sfGFP whereas none was seen with the empty plasmid. This suggests that the construction indeed allowed expressing the OmpA_DARPin-sfGFP fusion. The fluorescence seemed to be localized in the cytoplasm as it colored the entire cell. However, the resolution of the microscope could not allow us to determine if the membrane was fluorescent as well.

To determine the optimal concentration of IPTG to express our construction, we next assessed the expression after induction with different concentrations: 10 µM, 25 µM and 100 µM (see Induction with IPTG on liquid medium). After 4 hours of incubation at 37°C, fluorescence was observed on an epifluorescence microscope. We also added an assay with 10 µM of IPTG at 22°C to test the influence of temperature. All microscope observations are shown in Figure 22.

Figure 22: fluorescence of E. coli Tuner cells containing the pET-21 b (+)_OmpA_DARPin-sfGFP vector, with induction at different concentrations of IPTG: 10 µM (A), 25 µM (B), and 100 µM (C). Another assay was induced at 10 µM of IPTG and incubated 4 hours at 22°C (D). Each image represents an observation in bright field (on the left) and in the GFP fluorescence channel (microscope parameters for GFP). All observations were compared to an empty plasmid which showed no fluorescence (data not shown).

With 10 µM of IPTG at 37°C, fluorescence was low and a phenomenon of bleaching quickly occured, so it was difficult to take pictures. With 25 µM of IPTG at 37°C, fluorescence intensity was higher and it was easier to take pictures. With 100 µM of IPTG at 37°C, lots of inclusion bodies were visible, indicating that the expression levels were too high for proper protein folding.

With 10 µM of IPTG at 22°C, some bacteria had a filamentous shape, maybe due to the low temperature that caused some forms of abiotic stress.

Based on our results, we concluded that the best condition was an induction with 25 µM of IPTG at 37°C as the corresponding colonies displayed very few inclusion bodies but high levels of fluorescence. These conditions were chosen for the other experiments.

• Validation of the expression at the membrane by cell fractionation

To determine where the OmpA_DARPin-sfGFP fusion proteins were situated in the E. coli Tuner cells, we designed a fractionation protocol that allowed the separation of the different parts of the cells (cytoplasm and periplasm versus membrane). Briefly, after induction with 25 µM of IPTG at 37°C, we used sonication to break up the cells, resuspension in separating buffers and differential centrifugation steps (see Cell fractionation and surface control). We then measured the fluorescence emission from each fraction on a microplate reader. A strain with an empty plasmid was induced at 50 µM of IPTG as a negative control. Curves of fluorescence depending on the dilution factor were established as shown on Figure 23.

Figure 23: fluorescence of the cytoplasm/periplasm and membrane fractions. The negative control (empty plasmid induced at 50 µM of IPTG) is represented in blue and the actual assay (construction induced at 25 µM of IPTG) in orange. The excitation wavelength of the microplate reader was 485 nm. Emission was observed at a wavelength of 528 nm.

As expected, the control fractions were considerably less fluorescent than the samples (approximately a hundred times). This indicates that the OmpA_DARPin-sfGFP fusion protein was present in both the cytoplasm and surrounding membranes. The fluorescence in the membranes was twice lower than in the cytoplasm.

We concluded that the protein was mainly present in the cytoplasm, but also in the membranes as we wanted. We hypothesized that our pET-21 b (+)_OmpA_DARPin-sfGFP plasmid led to too high expression levels, leading to saturation of the membrane and expression in the cytoplasm as well.

Even if we confirmed the presence of at least some OmpA_DARPin-sfGFP protein in the membrane fraction, we could not be sure if the protein was exposed at the very surface of the bacteria or inside the membrane (facing inwards). We conducted another experiment to investigate this last point.

• Validation of the outer expression at the membrane by cell fractionation and TEV treatment

To make sure that the OmpA_DARPin-sfGFP was exposed at the surface of the cells and to quantify more precisely the fluorescence emission from each fraction, the experiment was later repeated with the addition of a TEV treatment after breaking the cells. The TEV protease should allow releasing the DARPin-sfGFP fusion from OmpA (Fig.17), meaning that after the TEV treatment, the fluorescence emission from the membrane fraction should decrease. Results are shown in Figure 24. This time again we included an empty vector control.

Figure 24: fluorescence of the cytoplasm/periplasm and membrane fractions. The negative control (empty plasmid induced at 50 µM of IPTG) is represented in blue and the assay (construction induced at 25 µM of IPTG) in orange. Each fraction was treated with TEV protease and compared with a non-treated sample. The excitation wavelength of the microplate reader was 485 nm. Emission was observed at a wavelength of 528 nm.

Similarly to the previous experiment, the control fractions were considerably less fluorescent than the samples (approximately a hundred times), indicating that GFP was well present in both the cytoplasm and membrane fractions of the induced sample. The fluorescence in the membrane appeared to be 2 to 3 times lower than in the cytoplasm, corroborating results from the previous experiment.

There was no difference between samples whether any TEV treatment was included or not. This could have been expected since we treated here the proteins alone and not bound to the membrane, meaning that we compared DARPin-sfGFP alone isolated after TEV treatment versus OmpA_DARPin-sfGFP, i.e, the same quantity of sfGFP in both. A more meaningful experiment would have been to purify the membrane before or after TEV treatment and to compare sfGFP fluorescence.

---

DARPin strain

After successful assembly of the OmpA_DARPin-sfGFP construct and assessment of the fusion protein localization, the next step was to delete the sfGFP encoding part. The final OmpA_DARPin construction (shown below in Figure 25) would later be used for the captation of IgE and aggregation experiments.

Figure 25: OmpA_DARPin construction. This construction was used to transform the final strain D, used for aggregation experiments.

Vector construction

To delete the sfGFP fragment, pET-21 b (+)_OmpA_DARPin-sfGFP was linearized by PCR using the high fidelity Phusion DNA polymerase with primers “exclusion sfGFP by In-Fusion (F)” and “exclusion sfGFP by In-Fusion (R)”, which excluded the sfGFP sequence (see Primers). The expected size of the amplicon was 6202 bp.

Amplification product sizes were checked on EtBr stained agarose gel. Amplicons were further purified from the gel. Figure 26 shows an amplification product matching the expected size.

Figure 26: pET-21 b (+)_OmpA_DARPin linearized. The expected size of the amplicons was 6202 bp. PCR amplicon sizes were checked with agarose electrophoresis gel and revealed with EtBr. A theoretical gel is presented on the left and the NEB 1 kb DNA ladder is used for the experimental gels. (note that a different ladder is presented on the theoretical gel).

The linearized plasmid was then recircularized by In-Fusion to form the pET-21 b (+)_OmpA_DARPin vector.

pET-21 b (+)_OmpA_DARPin was transformed into E. coli Stellar competent cells. Transformants were selected on LB-ampicillin plates. We screened 12 transformants by colony PCR with a primer pair flanking the zone where the sfGFP encoding fragment was deleted (primers used: screening_inserts-F and screening_inserts-R, expected size of the amplicon: 1162 bp) (see Primers). We detected 10 positive transformants. Figure 27 shows an example of such positive clones.

Figure 27: identifying the strains that bear pET21 b (+)_Ompa_DARPin by colony PCR. The expected size of the amplicon was 1162 bp. Only two clones were not positive (D.3 and D.6). All clones were compared to a control with a linearized plasmid still containing the sfGFP fragment. PCR amplicon sizes were checked with agarose electrophoresis gel and revealed with EtBr. A theoretical gel is presented on the left and the NEB 1 kb DNA ladder is used for the experimental gels (note that a different ladder is presented on the theoretical gel).

Four of those transformants had their plasmid extracted by Miniprep and digested by the restriction enzymes EcoRV and SalI to verify the assembly. The expected sizes of the fragments were 4853 bp and 1334 bp. Figure 28 shows correct restriction profiles for all the clones. The plasmids from these clones were further confirmed by Sanger Sequencing.

Figure 28: restriction profile of the pET-21 b (+)_OmpA_DARPin final construction. The enzymes used were EcoRV and SalI. The expected sizes of the fragments were 4853 bp and 1334 bp. All clones had correct restriction maps. Plasmids were checked with agarose electrophoresis gel and revealed with EtBr. A theoretical gel is presented on the left and the NEB 1 kb DNA ladder is used for the experimental gels (note that a different ladder is presented on the theoretical gel).

Controls of correct protein expression

We then tried to assess the expression and functionality of the DARPin, i.e. its capacity to bind IgE. Two experiments were conducted for this purpose. As the DARPin protein was HIS-tagged, we first tried to detect it by binding a fluorescent anti-HIS-tag IgG to the cell membrane and measuring fluorescence. No fluorescence was observed (data not shown), so we could not conclude if the protein was expressed at the membrane, especially without a positive control.

We then tried to assess the functionality of the DARPin by binding an IgE to it (the DARPin being able to link to the constant part of IgE), and then binding a fluorescent anti-IgE IgG to the complex. No fluorescence was observed (data not shown). Without a positive control, we cannot conclude between an incorrect exposition of the protein (not facing the outer surface), or a mere problem during the setup of the experiment.

First aggregation experiments

The required E. coli A and D strains having been constructed, a first set of aggregation experiments was conducted. We aimed at capturing specific IgE in order to detect them via the formation of bacterial aggregates (see Aggregation).

Here, our A strain produced the Ara h 2 allergen. It was mixed with the D strain designed to display DARPin proteins on its surface. Finally, we added an anti-Ara h 2 IgE to the mixture. We conducted several experiments to determine the optimal parameters (cellular concentration of each strain, strain ratio, IgE concentration, influence of agitation, incubation temperature…).

• Influence of IgE concentration

The first parameter tested was the concentration of IgE in our reaction mix. Three different IgE dilutions were chosen: 1:100 (final IgE concentration: 10 µg/mL), 1:250 (4 µg/mL) and 1:1000 (1 µg/mL). Equal quantities of strains A and D were mixed, reaching an OD600 of 3.2 after induction. Incubation was made overnight at room temperature. Three replicates were included to evaluate repeatability, as shown in Figure 29.

Figure 29: test of aggregation with strains A and D mixed with different concentrations of anti-Ara h 2 IgE: 1:100 (A), 1:250 (B), and 1:1000 (C). A mixture comprising strains A and D without any added IgE was used as negative control (D). Three replicates were made. Observations were made in phase contrast for replicates 1 and 2, and in bright field for replicate 3.

The first replicate presented bacterial aggregates for the three tested IgE dilutions, while the negative control showed no aggregates. There was no visible difference in the proportion or size of the aggregates between the three dilutions, so the IgE concentration did not have a strong impact. While these first observations looked promising, two other replicates performed later did not show any aggregate neither for the samples nor for the negative controls. As the experiment did not seem reproducible, we wanted to determine if the observed aggregates in replicate 1 could have stemmed from non-specific interactions between the bacterial cells, without any contribution from the IgE.

• Influence of cell concentration

We hypothesized that a high cell density may lead to non-specific aggregates. We therefore repeated the experiment but with a tenth of the previous cellular concentration. After adding the anti Ara h 2 IgE, incubation was made overnight at room temperature. Results for the 1:100 IgE dilution are shown in Figure 30.

Figure 30: test of aggregation with strains A and D mixed with 1:100 anti-Ara h 2 IgE, with cells at OD600 = 3.2 (A) or ten times more diluted (B). A mix with strains A and D without IgE was used as negative control (C). Observations were made in bright field.

As expected, cell density was considerably lower. We did not observe any aggregate formation, neither in the samples nor in the negative control. There was thus no obvious link between cell density and aggregation state.

• Influence of the reaction mix dilutions

For the next test, we decided to keep the same proportion of cells and IgE as in the first experiment but with a greater dilution of the total reaction mix to avoid non-specific aggregation. We compared 1:1, 1:2 and 1:10 mix dilutions. Incubation was made overnight at room temperature. Results for the 1:100 IgE dilution are shown in Figure 31.

Figure 31: test of aggregation with strains A and D mixed with 1:100 anti-Ara h 2 IgE, at different dilutions of the reaction mix: 1:1 (A), 1:2 (B), and 1:10 (C). A mix with strains A and D without IgE was used as negative control (D). Observations were made in phase contrast.

This time again we did not observe any aggregate, neither in the samples nor in the negative control. Our results overall indicate that no aggregate formation occured when mixing strain A, strain D and the corresponding IgE in these conditions.

• Influence of agitation

We hypothesized that adding some agitation during the incubation may increase interactions between the strains and IgE. The experiment was repeated but this time the mixture of strain A, strain D and the IgE was shaken with a benchwaver type agitator at 80 rpm. Incubation was made overnight at room temperature. Results for the 1:100 IgE dilution are shown in Figure 32.

Figure 32: test of aggregation with strains A and D mixed with 1:100 anti-Ara h 2 IgE, without agitation (A) or with agitation with a benchwaver type agitator at 80 rpm (B). Mixtures with strains A and D without IgE were used as negative controls, non-agitated (C) and agitated (D). Observations were made in phase contrast.

There were aggregates in the non-agitated sample as well as in its negative control but none in both agitated sample and negative control, so it seemed that the agitation prevented aggregation. These aggregates were certainly nonspecific as they appeared in the non-agitated negative control as well.

To conclude, the aggregation experiment may have succeeded once (first replicate of the test of IgE concentration) but could not be repeated. All the other experiments showed either nonspecific aggregates or no aggregates at all. Some troubleshooting had to be done to analyze the interaction between proteins and IgE on a molecular scale and draw hypotheses as to why the aggregation failed. The lack of positive controls was also a problem to identify what went wrong between the design and the experimental conditions. This also highlighted the plethora of parameters that could impact the aggregation phenomenon. This prompted us to build a model to optimize in silico these parameters and help in rationalizing the conditions to be assessed (see Modeling).

Correction of strain D

Vector construction

A PCR Assembly was conducted to assemble the missing fragment. Primers 1_PCA60, 2_PCA60, and 3_PCA60 (see Primers) and the high fidelity Phusion polymerase were used, with two elongation cycles to enable complete synthesis of the fragment (Figure 35).

Figure 35: Principle of PCR Assembly to build the missing part of the DARPin. Three primers (1_PCA60, 2_PCA60 and 3_PCA60) were used to build the complete missing sequence. These three primers were mixed in the same PCR mix. After a first elongation with 1_PCA60 and 2_PCA60, 3_PCA60 binds to the generated amplicon to end up the elongation.

The fragment was then amplified by PCR using the high fidelity Phusion polymerase with the primers IF3_ampli_DARPin (F) and IF4_ampli_DARPin (R) (see Primers). The expected size of the amplicon was 135 bp. Amplification product size was checked on EtBr stained agarose gel. Figure 36 shows the amplification product matching the expected size.

Figure 36: DARPin* missing fragment amplified. Expected size of the amplicon was 135 bp. PCR amplicon size was checked with agarose electrophoresis gel and revealed with EtBr. A theoretical gel is presented on the left and the NEB 1 kb DNA ladder is used for the experimental gel (note that a different ladder is presented on the theoretical gel).

Next, the pET-21 b (+)_OmpA_DARPin-sfGFP and pET-21 b (+)_OmpA_DARPin plasmids were linearized by PCR using the high fidelity Phusion polymerase with primers IF1_plasmid_DARPin (F) and IF2_plasmid_DARPin (R) (see Primers). The expected sizes of the amplicons were 6910 bp and 6187 bp, respectively. Amplification product sizes were checked on EtBr stained agarose gel. Figure 37 shows amplification products matching expected sizes.

Figure 37: pET-21 b (+)_OmpA_DARPin and pET-21 b (+)_OmpA_DARPin-sfGFP linearized fragments. Expected sizes of the amplicons were respectively 6187 bp and 6910 bp. PCR amplicon sizes were checked with agarose electrophoresis gel and revealed with EtBr. A theoretical gel is presented with each gel and the NEB 1 kb DNA ladder is used for the experimental gels (note that a different ladder is presented on the theoretical gel).

The missing DARPin fragment was then inserted into each of these plasmids by In-Fusion. The constructions obtained were first transformed into Stellar competent cells. Transformants were selected on LB-ampicillin plates, and were further screened by colony PCR with primer pairs flanking the insertion zone (primers used: screening_inserts-F and screening_inserts-R) (see Primers).

Plasmid from positive transformants were extracted by Miniprep and digested to assess the assembly (data not shown). Sequences of all the clones were validated by sequencing, completing the pET-21 b (+)_OmpA_DARPin* and pET-21 b (+)_OmpA_DARPin*-sfGFP constructions.

The plasmids were finally used to transform E. coli Tuner cells to hopefully express the DARPin* and DARPin*-sfGFP constructions at the cell membrane.

Controls to verify correct protein expression

As the insertion of the missing fragment could have changed the expression of the protein, it was necessary to verify it again. To assess the subcellular localization of the new fusion protein in vivo, we repeated the fractionation of the cells and fluorescence intensity measurement with a microplate reader, coupled to TEV treatment. This experiment, done with the DARPin*-sfGFP strain, is detailed hereafter.

To determine where the DARPin*-sfGFP fusion proteins were expressed in the E. coli Tuner cells, we conducted the same fractionation protocol as earlier. We measured the fluorescence of each fraction on a microplate reader (Figure 38). The DARPin* strain without sfGFP was induced with 25 µM of IPTG as a negative control.

Figure 38: fluorescence of cytoplasm/periplasm and membrane fractions. The negative control (DARPin* strain inducted at 25 µM of IPTG) is represented in blue and the assay (construction inducted at 25 µM of IPTG) in orange. Each fraction was treated with TEV protease and compared with a non-treated sample. The excitation wavelength of the microplate reader was 485 nm. Emission was observed at a wavelength of 528 nm.

The results were the same as before the insertion of the missing fragment: the control fractions were considerably less fluorescent than the samples, indicating the accumulation of the fusion protein in both the cytoplasm and membrane. The fluorescence in the membrane still appeared to be 2 to 3 times lower than in the cytoplasm. There still was no difference in samples whether the TEV treatment was applied or not, so we could not conclude whether the protein was exposed at the outer surface of the cell.

In conclusion, the integration of the missing DARPin fragment did not alter the expression of the protein, which was still expressed in the membrane as wanted. We still could not determine if it was exposed correctly at the surface or inside the membrane.

Now that the corrected strain D (expressing the DARPin* protein) was constructed, we conducted another set of aggregation experiments with that strain.

Detecting aggregates using microscopy

As the OmpA_DARPin* construction was corrected, new aggregation assays were conducted to determine if the new fusion protein was able to form aggregates to capture IgE.
The process was similar as in “First tries of aggregation”, with strain A displaying the Ara h 2 allergen for aggregation. The reaction mix included this strain A, along with strain D with the corrected DARPin gene and anti-Ara h 2 IgE. Three parameters were tested: the influence of IgE concentration, agitation and temperature of incubation, over aggregation.

• Influence of IgE concentration

Likewise to the first aggregation experiments, we tested three different IgE dilutions ( overnight incubation at room temperature with a benchwaver type agitator at 80 rpm; Figure 39).

Figure 39: test of aggregation with strains A and D mixed with different concentrations of anti-Ara h 2 IgE: 1:100 (A), 1:250 (B), and 1:1000 (C). A mix with strains A and D without IgE was used as negative control (D). The correct DARPin* on the strain D was used.

This aggregation assay did not show any difference between each sample and the negative control.

• Influence of incubation temperature

To compare the influence of temperature for two overnight incubation conditions tested (4°C and room temperature), for reminder, the incubation is the time spent between adding IgE and the observation on a microscope. The assay with 1:250 dilution of IgE anti Ara h 2 was shown in Figure 40.

Figure 40: test of aggregation with strains A and D incubated overnight with a dilution of IgE anti Ara h 2 of 1: 250 at different temperatures: 4°C (A) and Room temperature (B), negative control without anti Ara h 2 IgE incubated at 4°C (C). The correct DARPin* on the strain D was used.

The comparison does not show any incidence of the tested temperatures on the behavior of bacteria for an IgE dilution of 1:250. Therefore, the temperature of incubation did not trigger unpecific aggregates.

• Influence of stirring

Then the agitation condition was tested with a benchwaver type agitator at 80 rpm at room temperature for overnight incubation. Results are shown for 1:250 IgE anti-Ara h 2 dilution in Figure 41.

Figure 41: test of aggregation with strains A and D incubated overnight with a dilution of IgE anti Ara h 2 of 1: 250 at room temperature: With stirring (A) Without stirring (B) Negative control without IgE Anti-Ara h 2 with stirring (C). The correct DARPin* on the strain D was used. The agitation was made at room temperature with a benchwaver type agitator.

Agitation did not show any incidence on the aggregation capacity.

Despite the troubleshooting and the reconstruction of the DARPin, we could not detect any phenomenon of aggregation. As it happens, a new loop of the DBTL cycle (see Engineering success) has to be implemented to test a larger range of conditions (temperature, stirring, IgE concentrations, medium, other types of IgE and corresponding strain A, etc). For that, we can prospect two outlooks of the DBTL cycle. First we can deepen our efforts on surface control experiments to try to understand how our complexes are positioned on the membrane of strains A and D, and modify it if necessary. Second, our modeling effort (see Modeling) has been achieved at this step of the project and could eventually be used to optimize our set-up of aggregation assays, with a better control of parameters. The model will allow us to save a lot of time in the wet lab and resources for experiments.

The lack of time did not allow us to put into practice this new loop of the DBTL cycle in the wet lab, but this new approach of aggregation assay with the new DARPin and the mastering of the aggregation protocol should lead to a better control of assays and more accurate results by limiting non-specific aggregation.

Microdroplets

Microdroplets is a state-of-the-art microfluidics approach to parallelize a multiplicity of tests. For DAISY, it offers the possibility to test a multitude of strain A/strain D/IgE combinations without having any aggregates containing a mix of only strains A. It is then possible to select microdroplets presenting aggregation and hence, highlight the presence of allergenic IgE which could be further identified by sequencing the allergen sequence of the strain A. Even if the aggregation in microtubes did not work as expected, we decided to continue so that we could discover this microfluidics method.
The strategy was to encapsulate strains A, strain D and IgE in 18 µm diameter microdroplets. This way, each droplet acts like an individual reactor in which cells can aggregate independently. To do so, we force two orthogonal flows to cross each other. One of aqueous medium containing the cells and antibody, the other of oil. When they meet, the oil “squeezes” the aqueous medium and so, generates tiny droplets (see Design).

As we aim later to display a bank of allergen on the surface of strain A and isolate each cell in a different droplet, we chose to use a very diluted suspension of strain A (OD600= 0.01). On the contrary, with the aim of maximizing the size of the aggregates, we worked with a higher density of strain D (OD600= 4) to have 10 cells per droplet on average.

Strains A (displaying Ara h 2 allergen) and D were grown, induced for the expression and display of the allergen and DARPin, resuspended in PBS and encapsulated in microdroplets. After incubation overnight at room temperature, the droplets were observed with a microscope, as shown on Figure 42.

Figure 42: first aggregation assay in microdroplets. Observation in bright field, after overnight incubation in PBS at room temperature of droplets containing either (A) strain D with 10 cells on average, (B) strain D with 10 cells on average and maximum one cell of strain A ara h 2, (C) strain D with 10 cells on average, maximum one cell of strain A ara h 2 and IgE anti Ara h 2 at 1 µg/mL, (D) strain D with 10 cells on average, maximum one cell of strain A ara h 2 and IgE anti Ara h 2 at 10 µg/mL. The bacteria were mixed before encapsulation and the IgE were added during the droplet generation.

Aggregates were visible in most droplets, no matter the conditions considered. It means the aggregates were not due to the presence of IgE. The cells were not very mobiles. As the bacteria were resuspended in PBS 48h before the observation, we assumed some bacteria lysed because of the lack of substrate and released their DNA which glued the other bacteria together.

To avoid the lysis of the bacteria during incubation in PBS, we performed another experiment using LB medium instead of PBS. Plus, we calculated that a concentration of 1µg/mL of IgE corresponded to 43 000 IgE per droplet. As there are 10-11 bacteria per droplet on average, this number of IgE may saturate all the binding sites and inhibit the aggregation so we chose a concentration of 10 ng/mL. Finally, as we were running low on anti-Ara h 2 IgE, we used strain A Der p 1 and anti-Der p 1 IgE. Microscope observations of these microdroplets are shown in Figure 43.

Figure 43: second aggregation assay in microdroplets. Observation in phase contrast, after overnight incubation in LB at room temperature of droplets containing (A) strain A Der p 1 with maximum 1 cell, (B) strain D with 5 cells on average and maximum one cell of strain A Der p 1, (C) strain D with 5 cells on average, maximum one cell of strain A Der p 1 and anti-Der p 1 IgE at 10 ng/mL, (D) strain D with 5 cells on average, (E) strain D with 5 cells on average and IgE anti Der p 1 at 10 ng/mL. Strain D was mixed IgE before encapsulation and strain A Der p 1 was added during the droplet generation.

There were a lot more than 5 bacteria per droplet, as there should have been just after the encapsulation. This is likely due to the chosen rich medium, allowing the growth of the bacteria overnight. Aggregates were again visible in most droplets, no matter the conditions considered. As the bacteria were moving fastly, it is unlikely that some bacteria lyzed and glued the other ones. The nonspecific aggregation may be due to the high cell density of the physical medium properties.

With more time, it would have been valuable to optimize the cell concentration or physical parameter of the medium. Then the next step would have been to train the IA to identify aggregation in the droplets. The final step would have been running the experiment with labeling of the IA chosen droplets, sorting of the labeled droplet by FACS, and subsequent detection of the allergen expressed by strain A.

FACS

The second high-throughput approach we wanted to try was the FACS (Fluorescence-Activated Cell Sorting). The FACS platform was chosen for its capacity to sort cells by discriminating them according to their fluorescence.

---

Creating fluorescent strains for FACS application

To identify cellular aggregates including both strain D and strain A with FACS, we needed to discriminate between both strains. In this scope, we chose to render cells fluorescent with the constitutive expression of blue mTagBFP for strain D and red mRFP1 for strains A (expressing either allergens Ana o 3, Der p 1, Gal d 2 and Ara h 2). The fluorescent reporters were obtained from Manon Barthe (Toulouse White Biotechnology and former iGEM Toulouse 2016 member) who used the followings Parts:

• - BBa_J06504 for mRFP1
• - BBa_K592100 for mTagBFP

From her experience, both fluorescent reporters can be used without signal overlapping in cytometric approaches. She also advised us to use the E. coli ihfB promoter with 800 bp before the translation start site for optimal expression (Zhou et al., 2011 ; Barthe et al., 2020).

Construction of strain D expressing blue fluorescence

The ihfb800-mTagBFP construction provided by Manon Barthe was amplified by PCR with a high-fidelity Phusion polymerase using BFP/RFP IF+BglII R site R and BFP/RFP IF+BglII R site F primers (see Primers). The expected size of the amplicon was 1595 bp which matches with the obtained fragment (Figure 44).

Figure 44: PCR amplification of ihfb800-BFP from the psb1c3 plasmid. ihfb800-BFP PCR amplicon size of 1595 bp was checked with 0.6% agarose electrophoresis gel and revealed with EtBr. The elongation time used is 2’ and the hybridization temperature was tested from 55°C to 62°C. A theoretical gel is presented with each gel and the NEB 1 kb DNA ladder is used for the experimental gels. (note that a different ladder is presented on the theoretical gel).

To insert the fragment obtained in the DARPin-expressing plasmid, a linearization has been conducted with primers: planB_FPIF_pET21b(+)_R, planB_FPIF_pET21b(+)_F (see Primers). Expected size was 6317 bp which matches with the obtained result (Figure 45).

Figure 45: PCR linearization of pET-21 b (+)_OmpA_DARPin. The PCR product of linearization was checked with 0.8% agarose electrophoresis gel and revealed with EtBr. The hybridization temperature was 58°C and the elongation time was 7 min. A theoretical gel is presented with each gel and the NEB 1 kb DNA ladder is used for the experimental gels (note that a different ladder is presented on the theoretical gel).

The linearized pET-21 b (+)_OmpA_DARPin plasmid and the ihfb800-BFP fragment were extracted from gel and purified with the PCR clean-up protocol, then assembled by In-Fusion. The product was transformed in Stellar cells and transformants were selected on Ampicillin. Plasmids from the resulting colonies were extracted by Miniprep. The presence of BFP was checked by PCR screening with screening_insert-R and screening_insert-F primers (see Primers). The PCR product containing both BFP and DARPin displayed the correct size at 2892 bp only for colony 2 (Figure 46).

Figure 46: PCR screening of pET-21 b (+)_OmpA_DARPin. The electrophoresis gel was revealed with EtBr. A theoretical gel is presented with each gel and the NEB 1 kb DNA ladder is used for the experimental gels (note that a different ladder is presented on the theoretical gel). Colony 2 presented the correct profile.

Colony 2 was selected and an overnight preculture was performed to control the blue fluorescence on a microscope (Figure 47). A colony containing the pET-21 b (+)_OmpA_DARPin without ihfb800-BFP was used as a negative control. The microscope parameters to observe BFP were the following: excitation filter 360/40, dichroic beamsplitter 400 nm, emission filter 470/40.

Figure 47: Observation on microscope of pET-21 b (+)_OmpA_DARPin strain with ihfb800-BFP (A) pET-21 b (+)_OmpA_DARPin strain with ihfb800-BFP were used a negative control (B). Colonies showed significant blue fluorescence. Observations were made in phase contrast (left of each panel) and at the microscope parameters for BFP (right of each panel).

Sequencing revealed a mutation in the promoter but it did not seem to impact mTagBFP expression. This validated the blue-fluorescent strain D.

The construction, however, was obtained with the first DARPin sequence. We therefore had to insert the missing fragment of the protein. The method was the same as followed for the non-fluorescent protein. A PCR Assembly was conducted to assemble the missing fragment. The fragment obtained for the non-fluorescent DARPin construction was used again for this construction (see figure 36). Briefly, pET-21 b (+)_OmpA_DARPin_BFP plasmid was linearized by PCR using the high fidelity Phusion polymerase with the primers IF1_plasmid_DARPin (F) and IF2_plasmid_DARPin (R) (see Primers). Expected size of the amplicons was 7812 bp. Amplification product sizes were checked on EtBr stained agarose gel (figure 48).

Figure 48: pET-21 b (+)_OmpA_DARPin_BFP linearized fragments. Expected size of the amplicon was 7812 bp. PCR amplicon size was checked with agarose electrophoresis gel and revealed with EtBr. A theoretical gel is presented with each gel and the NEB 1 kb DNA ladder is used for the experimental gels. (note that a different ladder is presented on the theoretical gel).

Since it was validated, the missing DARPin fragment was then inserted by In-Fusion,The resulting products were first transformed into Stellar competent cells. Transformants were selected on LB-ampicillin plates. 11 colonies were further screened by colony PCR with primer pairs flanking the insertion zone (primers used: exclusion sfGFP (F) and IF4 ampli DARPin, expected size : 384 pb) (see Primers). 8 PCR products were matching the expected sizes (Figure 49).

Figure 49: PCR screening of pET-21 b (+)_OmpA_DARPin_BFP. Expected size was 384 bp. The electrophoresis gel was revealed with EtBr. A theoretical gel is presented with each gel and the NEB 1 kb DNA ladder is used for the experimental gels (note that a different ladder is presented on the theoretical gel).

Plasmid from positive transformants were extracted by Miniprep and digested to assess the assembly (data not shown). Sequences were validated by sequencing, thus completing the pET-21 b (+)_OmpA_DARPin*_BFP construction.

The plasmids were finally used to transform E. coli Tuner cells to express the pET-21 b (+)_OmpA_DARPin_BFP correct construction at the cell membrane. The construction was finally ready to be used for the FACS.

Construction of strain A expressing red fluorescence with mRFP

The ihfb800-mRFP1 construction from Manon Barthe was amplified by PCR with the high-fidelity Phusion polymerase using BFP/RFP IF+BglII R site R and BFP/RFP IF+BglII R site F primers (see Primers). The expected size of the amplicon is 1622 bp and this was watching the results on the gel (Figure 50).

Figure 50: PCR amplification of ihfb800-RFP from the psb1c3 plasmid. ihfb800-RFP PCR amplicon size of 1622 bp was checked with 0.6% agarose electrophoresis gel and revealed with EtBr. The elongation time used is 2’ and the hybridization temperature was 57°C. A theoretical gel is presented with each gel and the NEB 1 kb DNA ladder is used for the experimental gels. (note that a different ladder is presented on the theoretical gel).

After DpnI digestion, the fragment was recovered by PCR clean-up.
To insert the fragment on allergen-expressing plasmids, linearization of these plasmids has been conducted in parallel (Gal d 2, Ana o 3, Der p 1 and Ara h 2 expressing plasmids) with primers: planB_FPIF_pET21b(+)_R and planB_FPIF_pET21b(+)_F (see Primers), as shown in Figure 51. For each linearization, the expected sizes were:

• - pET-21 b (+)_OmpA_Gal d 2: 7117 bp
• - pET-21 b (+)_OmpA_Ana o 3: 6475 bp
• - pET-21 b (+)_OmpA_Der p 1: 6919 bp
• - pET-21 b (+)_OmpA_Ara h 2: 6475 bp

Figure 51: PCR linearization of Gal d 2, Ana o 3, Der p 1 and Ara h 2. The PCR product of linearization was checked with 0.6% agarose electrophoresis gel and revealed with EtBr. The hybridization temperature was 58°C and the elongation time was 7 min. A theoretical gel is presented with each gel and the NEB 1 kb DNA ladder is used for the experimental gels (note that a different ladder is presented on the theoretical gel).

Except for Der p 1, the analytical gel showed two bands for each PCR product with one for each sample corresponding to the expected size. Since we aimed to obtain at least two constructions with at least one compatible with commercial antibodies we owned, we carried on without Der p 1.

In-Fusion was performed to insert the ihfB800-mRFP1 fragment in the linearized Gal d 2, Ana o 3 and Ara h 2 expressing plasmids. The resulting products were transformed in Stellar cells and transformants were selected on ampicillin. Colonies were screened by PCR using screening_insert_R and screening_insert_F primers (see Primers). The expected sizes of amplicons were: for Ara h 2, 3046 bp with RFP and 1454 bp without; for Ana o 3, 2944 bp with RFP and 1453 bp without; for Gal d 2, 3688 bp with RFP and 2096 bpwithout. Figure 52 shows the corresponding gels. We obtained correct sizes for the three allergen constructions.

Figure 52: verification of the insertion of RFP fragment: Ara h 2 gel (A), Ana o 3 gel (B), Gal d 2 (C). The PCR screening was checked with 0.6% agarose electrophoresis gel and revealed with EtBr. A theoretical gel is presented with each gel and the NEB 1 kb DNA ladder is used for the experimental gels (note that a different ladder is presented on the theoretical gel). Colonies 1, 2 and 3 present the correct size for Ara h 2 ; colony 8 for Ana o 3; colonies 9, 10, 11 for Gal d 2.

Successful colonies were further tested by digestion with NotI which cut both in the plasmid and in the mRFP1 gene (Figure 53). Correct sizes were expected to be 1280 and 6791 bp for Ara h 2, 1178 and 6791 bp for Ana o 3 and 1922 and 6791 bp for Gal d 2.

Figure 53: digestion by NotI of Ara h 2 with mRFP1 insertion (A), of Ana o 3 with mRFP1 insertion(B) and of Gal d 2 with mRFP1 insertion (C). The digestion product was checked with 0.6% agarose electrophoresis gel and revealed with EtBr. A theoretical gel is presented with each gel and the NEB 1 kb DNA ladder is used for the experimental gels (note that a different ladder is presented on the theoretical gel). Colonies 1, 2 and 3 present the correct size for Ara h 2; colonies 21 and 23 present the correct size for Ana o 3; colonies 7 and 13 present the correct size for Gal d 2.

Sequencing of each sample showed the same mutation in the ihfb800 promoter than for the DARPin-mTagBFP construction. This mutation seemed generic and was likely present in the initial sequence. As it did not appear to impact expression, we validated the sequences.

We then assessed the mRFP1 fluorescence on a fluorescent microscope. Very unfortunately, we did not succeed to obtain more fluorescence than with a negative control without the ihfB800-mRFP1 construction (figure not shown). Consequently, we decided to replace the mRFP1 sequence by those of the mScarlet-I fluorescent reporter, whose excitation and emission characteristics are similar.

Construction of strain A expressing red fluorescence with mScarlet-I

As we could not determine if the mRFP1 was indeed expressed in our strain A, we replaced its gene by that of the mScarlet-I (see Figure 54). The mScarlet-I gene was supplied by Maxence Holtz (TWB and former iGEM Toulouse 2021).

Figure 54: allergen construction with mScarlet-I, example for Ara h 2 allergen. Promoter used for the fluorescent protein is the constitutive ihfB800.

We designed primers to replace the mRFP1 coding sequence with those of the mScarlet-I gene. PCR amplification was done using Phusion polymerase and IF3_mSCARLET-I (F) and IF4_mSCARLET-I (R) primers (see Primers). Expected size of the amplicon was 699 bp. The correct size was obtained (Figure 55A).

We had antibodies against Ara h 2 so we decided to clone mScarlet-I in the pET-21 b (+)_OmpA_Ara h 2 strain for the experiment. We did not have antibodies against Ana o 3 so we decided to use the pET-21 b (+)_OmpA_Ana o 3 strain for the negative control. To amplify these vectors excluding mRFP1 sequence, we used primers IF1_exclusion_RFP (R) and IF2_exclusion_RFP (F) (see Primers). Expected size of the amplicon were 7365 bp for Ara h 2 and 7263 bp for Ana o 3. The correct size was obtained (Figure 55B). Amplicons were further purified from the gel.

Figure 55: mScarlet-I amplified fragment (A), pET-21 b (+)_Ara h 2_OmpA and pET-21 b (+)_Ana o 3_OmpA linearized (B). Expected sizes of the amplicons were 699 bp (A) and 7365 bp and 7263 bp (B). PCR amplicon sizes were checked with agarose electrophoresis gel and revealed with EtBr. A theoretical gel is presented with each gel and the NEB 1 kb DNA ladder is used for the experimental gels. (note that a different ladder is presented on the theoretical gel).

The mScarlet-I gene was inserted into pET-21 b (+)_OmpA_Ara h 2 and pET-21 b (+)_OmpA_Ana o 3 linearized by In-Fusion. In-Fusion assemby reaction was transformed into Stellar competent cells and transformants were selected on LB-ampicillin plates. Two transformants clones were red directly on the plate, straightly indicating the expression of mScarlet-I (Figure 56).

Figure 56: Second plating of pET-21 b (+)_Ara h 2_OmpA_mScarlet-I transformed cells and pET-21 b (+)_Ana o 3_OmpA_mScarlet-I transformed Stellar cells. The colonies were selected on LB-ampicillin plates. Pink coloring indicates the expression of the mScarlet-I.

The plasmids were extracted and used to transform E. coli Tuner cells. They also turned red (data not shown) indicating that the ihfB800 promoter is well functional as well in Stellar as in Tuner E. coli backgrounds.

The colonies containing the pET-21 b (+)_OmpA_Ara h 2_mScarlet-I and pET-21 b (+)_OmpA_Ana o 3_mScarlet-I plasmids were observed under the fluorescence microscope to detect the mScarlet-I fluorescence (Figure 57). The microscope parameters to observe mScarlet-I fluorescence were the following: excitation filter 545/30x, dichroic beamsplitter 570 nm LP, emission filter 610/75.

Figure 57: fluorescence of pET-21 b (+)_Ana o 3_OmpA_mScarlet-I transformed Tuner cells. Cells transformed with plasmids not containing mScarlet-I were used as negative controls. Observations were made at the microscope parameters for mScarlet-I.

---

FACS experiments

As for the microfluidic experiment, we wanted to go through the whole cytometric approach since we had achieved to build the strains and had the material to do so, and even if we did not have enough time to perform all the needed controls to assess allergen and DARPin functions.

Cytometric validation of the fluorescent strains

The first step was to check if the blue and red signals were detectable using a Spectral FACS from Cytek Biosciences. Liquid cultures of DARPin with and without BFP or of both allergens with and without mScarlet-I strains were performed and used to calibrate the FACS machine. (Figure 58).

Figure 58: mTagBFP and mScarlet-I detection with the Spectral FACS. PBS medium was tested on (A), strain A Ana o 3 with constitutive expression of mScarlet-I on (B), Strain A Ara h 2 expressing Ara h 2 with constitutive expression of mScarlet-I on (C), strain A Ara h 2 without expression of mScarlet-I on (D), strain D with constitutive expression of mTagBFP1 on (E) and strain D without expression of mTagBFP1 on (F). 10,000 events were recorded. The population was selected according to the area of their shadow under the laser and their granulometry before selectioning particles corresponding to cell and cellular aggregates. Laser channel V3-A was used for blue signals and YG3-A for red signals.

Nonfluorescent signals from the PBS could be seen on Figure 58A while the specificity of the DARPin/mTagBFP and Allergen/mScarlet-I fluorescent cells was validated on respectively figure 58E, and 58B and 58C, with lost of the signals without the fluorescent reporters (58D and 58F).

Observing aggregation through FACS

We then performed a first preliminary aggregation assay by mixing fluorescent strain A and strain D with or without the IgE corresponding to strain A (Figure 59).

Figure 59: aggregation assay and cytometric analysis. Strain A Ara h 2 with constitutive expression of mScarlet-I was mixed with strain D with constitutive expression of mTagBFP1 in presence of either no IgE (A), non-specific anti-Der p 1 IgE (B) or specific anti-Ara h 2 IgE (C). 100,000 events were recorded for (A) and (C) and 10,000 events were recorded for (B). The population was selected according to the area of their shadow under the laser and their granulometry before selectioning particles corresponding to cell and cellular aggregates. Laser channel V3-A was used for blue signals and YG3-A for red signals.

While it was possible to observe aggregation (double blue-red signals in the upper right sections of the figures 59ABC), these aggregates were not specific to the presence of the corresponding IgE since the same occurrences were obtained without IgE, with a non-specific IgE or with a specific IgE (respectively figures 59A, 59B and 59C). We could only conclude this assay remains to be optimized.

Detecting allergen enrichment

Even if the previous assay pointed out a lack of specificity of the aggregates for the IgE, we chose to go for the final experiment of the project by assessing a putative enrichment in strain A Ara h 2 in the aggregates when in presence of Ara h 2 IgE. To do so, we isolated 96 aggregates on a LB-ampicillin filled microtitration plate using the sorting function of the Spectral FACS. As control, 96 aggregates from the control without IgE were isolated too. The aggregates were sorted in LB medium with ampicillin and left for growth during 72 hours.
Analytical PCR were then performed on 7 aggregates to assess our capacity to identify the allergen carried by strain A for each aggregate (Figure 60). Primers used were screening_insert F/ IF4_Ana o 3 (expected size: 584 bp) and screening_insert F/IF4_Ara h 2 (expected size: 748 bp) (see Primers).

Figure 60: analytical PCR of the aggregates. Strains A with Ara h 2 and Ana o 3 with constitutive expression of mScarlet-I was mixed with strain D with constitutive expression of mTagBFP1 in presence of either specific Ara h 2 IgE (A) or no IgE (B), and sorted by FACS on LB-ampicillin filled microtitration plates. 7 colony PCR were performed using primers screening_insert F and IF4_Ara h 2 to investigate Ara h 2 presence (expected size 584 bp) and primers screening_insert F and IF4_Ana o 3 to investigate Ana o 3 (expected size 748 bp) presence in the aggregate. (A) gel of the PCR product from the assay. (B) gel of the PCR product from the negative control without IgE.

As could have been expected from the cytometric profiles of Figure 59, no specific enrichment was obtained for Ara h 2. Ara h 2 strains were present in 5 over 7 aggregates of the assay (Figure 60A). It would have been of interest to assess if Ana o 3 was present in the 2 other aggregates but the PCR targeting Ana o 3 did not work and there was unfortunately no time left to try it again. Besides, Ara h 2 strains were present in all the negative control aggregates (without the corresponding IgE). We concluded that our sorted aggregates were not specific, and further statistical analysis would be pointless. This is not surprising since the sorting step previously showed no increase in the aggregate number when adding the antibody. However, we were able to prove that our idea was feasible as it is possible to sort aggregates out to identify which strain composed them.

Conclusion

We concluded that the cytometric approach is technically feasible since we were able to create the strains, to detect aggregates with a cytometric approach and to assess which allergen is present in these aggregates. We did not have enough time in the course of the iGEM competition to perform troubleshooting and reproduce the experiment with a variety of conditions. Many controls would have been necessary here, starting with the validation of DARPin and allergen expression at the cellular membrane before the FACS, which could not be achieved precedently (see “Controls to verify correct protein expression” in “Correction of strain D”). Optimization of the clumping conditions according to our modeling results will also be mandatory to improve the assay.