Overview

Our solution for AHPND (Acute Hepatopancreatic Necrosis Disease) caused by Vibrio parahaemolyticus (Vp for short) can be divided into three sections: detecting the pathogen, treating the disease and preventing the outbreak of infection. We organized our experimental results in the order of Detection, Treatment and Prevention. Besides, we demonstrated the feasibility of the kill switch, which is also an important part for the sake of biosafety and biocontainment of OMEGA. The fundamental aspects of our projects are composed of all the results we obtained.

Detection

Since the AHPND-causing pathogen Vibrio parahaemolyticus can secrete toxins PirA and PirB, we decided to choose the toxin encoding genes, pirA and pirB, as the target for developing our detection system in OMEGA. In order to overcome the limitations of the whole-cell biosensor (1-4) in which the commonly used input, quorum-sensing signals, are so complicated among Vibrio spp. (5, 6) and to bypass the extensive use of technology like lateral flow assay (LFA), the cell-free system combining post-transcriptional regulation was considered as the strategy to detect the pathogen-derived nucleic acids with leveraging the wisdom of synthetic biology. The total workflow of the detection system can be divided into three sections: nucleic acids enrichment, recombinase polymerase amplification (RPA), in vitro transcription, and translation. As described in detail in Design, the ribozyme-enabled detection of RNA (RENDR) (7) plays a significant role in signal output, and therefore we constructed an array of expression circuits to test the performance of different split modes combing different reporters in vivo in particular, before the real implementation on the cell-free system.

Highlight

Confirm the toxin genes can be enriched from the samples and amplified via RPA
Verify the performance of various guides with three reporters in vivo characterization

Introduction

Firstly, to guarantee the specificity and accuracy of the RENDR system, the conserved regions of pirA and pirB were analyzed by using the Nucleotide BLAST of NCBI. And the transcripts of the conserved regions of pirA and pirB were set as input signals. While the RNA guide sequences (guides, for short) linked with split ribozymes are carefully designed to be complementary to the RNA input signals. Secondly, the NUPACK (8) web-server was used to calculate the free energy differences between the seed complex and three single-stranded RNAs, which was compared with the free energy differences of the circuit used in the reference (7). Two input-guided pairs for each toxin gene, whose free energy differences were greater than or equal to that in the reference were selected and named “pirA(B)_g1” and “pirA(B)_g2”, respectively. Additionally, we adopted two ribozymes split modes for constructing the detection circuits (split site at 15 nt (named “α”) and split site at 402 (named “β”)), which are reported to be highly functional. And therefore, the gene circuits expressing guides were all written in the following style, T7-pirA_g1α-T7t, which means that the guides are designed to target pirA with the pairing type g1 and ribozyme-split mode α under the control of T7-transcriptional system.

Another important part that should be mentioned is the reporter, which determines the feature of the output signal (9). We have attempted to apply three different types of reporters for constructing the detection system, including fluorescence (GFP), colorimetry (AmpC), and bioluminescence (NanoLuc luciferase), which were all tested for characterization in vivo. While for the implementation in vitro and related examination in our home-built hardware, only colorimetric and bioluminescent reporters were tested in accordance with our Design. The following figure (Fig. 1) presents the different elements composing of the toxin gene sensing circuit.

Fig. 1 Illustration of the elements involved in the design of the guides circuit of the Detection part.

Results

Toxin genes can be extracted from water, enriched on fiber filter paper, and amplified via RPA

Show Details

Fiber filter paper enriches DNA

Filter paper can enrich DNA from plasmid extract

To test the enrichment effect of fiber filter paper, BBa_K4195179_pSB1C3 (expressing the conserved region of pirA) or BBa_K4195180_pSB1C3 (expressing the conserved region of pirB) were used. 300 μL plasmid solution with different concentrations (1 ng/mL, 5 ng/mL, and 10 ng/mL) were added into the 1.5-mL tube containing a circular filter paper with different sizes (diameter was 3 mm, 4 mm, 5 mm, and 6 mm). After being incubated for 1 min, the paper was washed with wash buffer and directly amplified using PCR. In this case, deionized water and plasmid solution with the same concentration were used as negative and positive controls, respectively. More details of the reagents we used are shown in our Experiments page.

As shown in DNA gel electrophoresis of the PCR products (Fig. 2), target bands (567 bp and 569 bp) can be observed at the position between 750 bp and 500 bp. The result proved that the filter paper is able to enrich the plasmid. Based on the result, we select the 3 mm filter paper, which had the best enrichment effect even at the plasmid concentration of 1 ng/mL, for further experiments.

Fig. 2 DNA gel electrophoresis of the fiber paper enriched PCR products from plasmid extract.

Filter paper can enrich DNA from bacteria lysate

The overnight culture of colonies containing plasmid BBa_K4195179_pSB1C3 or BBa_K4195180_pSB1C3 was diluted in 10-fold and 100-fold. The bacterial culture was lysed by extraction buffer, and the plasmids are enriched by the filter paper placed into the bacteria lysate. After enrichment for a certain time, the paper was washed with wash buffer and directly amplified using PCR. Deionized water and plasmid solution were used as negative and positive controls, respectively.

As shown in DNA gel electrophoresis of the PCR products (Fig. 3), enrichment of DNA was accomplished from bacteria lysate dilutions. The ability of the filter paper to enrich DNA had been verified through PCR and DNA gel electrophoresis.

Fig. 3 DNA gel electrophoresis of the fiber paper enriched PCR products from bacteria lysate.

RPA: Recombinase Polymerase Amplification

To enhance the input signals and reduce background interference, Recombinase Polymerase Amplification (RPA) is implemented to amplify the target toxin genes specifically. After obtaining the enriched nucleic acid sample, the next step is to verify whether this enrichment method can be used for subsequent RPA systems. To obtain the single-stranded RNA (ssRNA) which can be recognized by RENDR, the sequence of the T7 promoter is added to the upstream of the forward primer to finally enable the transcription of amplified genes by T7 RNA polymerase.

RPA can distinguish the level of target gene

The activation of protein expression in cell-free system is quantitative with respect to the abundance of target RNA, and increased levels of target will result in increased activation and accumulation of signal outputs. RPA, as with many isothermal amplification methods, undergoes asynchronous amplification with the potential to saturate, which will prevent accurate quantitation. With the intention of the quantitation of target level, amplification time for target gene was optimized. Plasmid (input) with different concentrations (10 ng/mL, 100 ng/mL, 1000 ng/mL and 10000 ng/μL) were amplified using RPA, which was performed through the TwistDx kit with different amplification times (40 min, 30 min, 20 min and 10 min). As shown in Fig. 4, we found that as the amplification time decreased, the differences of the target bands’ gray scales among different input concentrations can be obviously distinguished, especially the 10 min-amplification (Fig. 4d).

Fig. 4 Amplification time optimization. a RPA reactions with different target levels were run at 37 °C for 40 minutes. b RPA reactions with different target levels were run at 37 °C for 30 minutes. c RPA reactions with different target levels were run at 37 °C for 20 minutes. d RPA reactions with different target levels were run at 37 °C for 10 minutes.

RPA can work in the home-built hardware

We have designed and built a thermostatic box to conduct RPA reactions. After being incubated at 37 °C for 30 min using a functional incubation reaction chamber in our hardware, RPA products were verified through DNA gel electrophoresis. The products of RPA reaction incubated at 37 °C in a constant temperature incubator were set as positive control. As shown in Fig. 5, target bands can be observed in all lanes of the reaction in home-built hardware, which suggests that RPA can work well in the hardware we designed.

Fig. 5 RPA reactions in the home-built hardware or a constant temperature incubator. Lane 1, 2: RPA products of the home-built hardware, lane 3, 4: RPA products of the 37 °C constant temperature incubator.

In vivo Verification

Show Details

We hope to achieve the detection of target toxin genes collected from the water sample using RENDR-based system. The RENDR system designed for pirA/B is firstly verified in vivo, with GFP, β-lactamase (AmpC), and NanoLuc luciferase being chosen as the reporters. RNA input signals were transformed into fluorescent, colorimetric, and bioluminescent signals, to verify the feasibility of the detection system (Fig. 6).

Fig. 6 Schematic illustration of RENDR, in which the light bulb represents different output signals contributed by different reporters.

Ribozymes have the ability to catalyze RNA splicing in gene expression in vivo

Molecular cloning

The circuit used in the reference (7) was set as a control circuit. The sequence of self-splicing ribozyme is inserted within the coding sequence (CDS) of a super folder green fluorescence protein (sfGFP) gene such that the translation of the full-length protein is disrupted in the absence of splicing. Upon transcription and splice, the ribozyme removes itself from the flanking exons and forms a functional sfGFP mRNA, which is then translated to produce a fluorescent protein as output. The circuit was assembled into the vector pSB1C3 by standard BioBrick assembly. The constructed plasmids were transformed into E. coli BL21(DE3), then the positive transformants were selected by chloramphenicol and confirmed by colony PCR (Fig. 7) and sequencing.

Fig. 7 DNA gel electrophoresis of the colony PCR products of BBa_K4195155 in E. coli BL21(DE3). Target bands (1595 bp) can be observed at the position around 1500 bp.

Fluorescence intensity measurement

Colonies harboring the correct plasmid were cultivated and induced. The expression behavior of sfGFP was observed under blue-light gel imager, in which the green fluorescence intensity of the group sfG(R)FP after induction was even stronger than that of the positive control (EYFP) (Fig. 8). Thus, we demonstrated that ribozymes have the ability to catalyze RNA splicing in gene expression in vivo.

Fig. 8 The image of different groups for testing the ribozyme’s activity in vivo. Bacteria harboring BBa_K914003 (pRha) and pT7-B0034-eyfp-B0015_pSB1C3 (pT7-EYFP) was set as negative control and positive control, respectively. sfG(R)FP represents the CDS of sfGFP inserted with ribozyme sequence.

RENDR can regulate the translation of bacterial messenger RNA in response to the presence or absence of any RNA input in vivo.

pirA toxin gene detection: molecular cloning

For in vivo verification of toxin pirA, we designed BBa_K4195141, BBa_K4195142, BBa_K4195145, BBa_K4195156, BBa_K4195157, BBa_K4195160, BBa_K4195161, BBa_K4195168, BBa_K4195169, BBa_K4195172, BBa_K4195173 (different guides with different reporter genes). These parts were all separately assembled into the vector pSB3K3 by standard BioBrick assembly. Each constructed plasmid and plasmid BBa_K4195179_pSB1C3 (input of pirA) were co-transformed into E. coli BL21(DE3). Then the positive transformants were selected by kanamycin and chloramphenicol. Other related parts are listed as followings: BBa_K4195151, BBa_K4195183, BBa_K4195184, BBa_K4195185, BBa_K4195186, which were assembled into pSB1C3 plasmid backbone, respectively. The constructed plasmids were transformed into E. coli BL21(DE3), then the positive transformants were selected by chloramphenicol and confirmed by colony PCR and sequencing. The results of colony PCR were recorded on the corresponding parts page.

pirA toxin gene detection: RENDR with a GFP output

Colonies harboring the correct plasmid(s) were cultivated and induced. The expression behavior of GFP is represented by normalized RFU_GFP (RFU_GFP/OD₆₀₀), using a microplate reader. The circuit, ori-g1α with input sequence ( BBa_K4195176) used in the reference (7) was used as the positive control. After evaluating the RFU_GFP/OD₆₀₀ value in each group (Fig. 9), pirA_g2 α and pirA_g2β exhibited better performances in both single plasmid system and dual plasmid system, especially the group of pirA_g2α in dual plasmid system (Fig. 9b).

Fig. 9 The results of reporter GFP for pirA toxin gene detection in RENDR-based system. a Single plasmid-based characterization. b Dual plasmid-based characterization. RFU: relative fluorescence unit.

pirA toxin gene detection: RENDR with a β-lactamase (AmpC) output

Colonies harboring the correct plasmid(s) were cultivated and induced. The expression behavior of β-lactamase (AmpC) is represented by normalized RAU_AmpC (RAU_AmpC/OD₆₀₀), after measuring the absorbance in 490 nm as time progressed using a microplate reader. After evaluating the RAU_AmpC/OD₆₀₀ value in each group (Fig. 10), pirA_g2α (Fig. 10b) and pirA_g2β (Fig. 10d) exhibited better performances in dual plasmid system, which could be selected as the candidates for following experiments if necessary.

Fig. 10 The results of reporter AmpC for pirA toxin gene detection in RENDR-based system as time progressed of (a) pirA_g1α, (b) pirA_g1β, (c) pirA_g2α and (d) pirA_g2β. RAU: relative absorbance unit.

pirA toxin gene detection: RENDR with a NanoLuc output

Colonies harboring the correct plasmid(s) were cultivated and induced. The expression behavior of NanoLuc is represented by normalized RLU_NanoLuc (RLU_NanoLuc/OD₆₀₀), after measuring the bioluminescence as time progressed using a microplate reader. After evaluating the RLU_NanoLuc/OD₆₀₀ value in each group (Fig. 11), pirA_g1α (Fig. 11a) and pirA_g1β (Fig. 11c) in dual plasmid system exhibited better performances, which could be selected as the candidates for following experiments if necessary.

Fig. 11 The results of reporter NanoLuc luciferase for pirA toxin gene detection in RENDR-based system as time progressed of (a) pirA_g1α, (b) pirA_g1β, (c) pirA_g2α and (d) pirA_g2β. RLU: relative luminescence unit.

Thus, based on the results from Fig. 9, Fig. 10, and Fig. 11, we concluded that pirA_g1α and pirA_g2β showed better performances in vivo, which may be further used in the cell-free system in vitro.

pirB toxin gene detection: Molecular cloning

For toxin pirB, we designed BBa_K4195143, BBa_K4195144, BBa_K4195158, BBa_K4195170, BBa_K4195171, BBa_K4195174, BBa_K4195175. These parts were respectively assembled into the vector pSB3K3 by standard BioBrick assembly. The result of colony PCR was recorded on the corresponding parts page. Each constructed plasmid and plasmid BBa_K4195180_pSB1C3 (input of pirB) were transformed into E. coli BL21(DE3). The positive transformants were selected by kanamycin and chloramphenicol. Other related parts are listed as following: BBa_K4195149, BBa_K4195150, BBa_K4195164, BBa_K4195165, BBa_K4195166, BBa_K4195187, BBa_K4195188, BBa_K4195189, BBa_K4195190.These parts were respectively assembled into the pSB1C3 plasmid backbone. The constructed plasmids were transformed into E. coli BL21(DE3), then the positive transformants were selected by chloramphenicol and confirmed by colony PCR and sequencing. The result of colony PCR was recorded on the corresponding parts page.

pirB toxin gene detection: RENDR with a GFP output

Colonies harboring the correct plasmid(s) were cultivated and induced. The expression behavior of GFP is represented by normalized RFU_NanoLuc (RFU_NanoLuc/OD₆₀₀), after measuring the bioluminescence as time progressed using a microplate reader. The circuit, ori-g1α with input sequence (BBa_K4195176) used in the reference (7) was used as the positive control. After evaluating the RFU_GFP/OD₆₀₀ value in each group (Fig. 12), pirB_g2α in both system and pirB_g1α/β in single plasmid system exhibited better performances, which could be selected as the candidates for following experiments if necessary.

Fig. 12 The results of reporter GFP for pirB toxin gene detection in RENDR-based system. a Single plasmid-based characterization. b Dual plasmid-based characterization. RFU: relative fluorescence unit.

pirB toxin gene detection: RENDR with a β-lactamase (AmpC) output

Colonies harboring the correct plasmid(s) were cultivated and induced. The expression behavior of β-lactamase (AmpC) is represented by normalized RAU_AmpC (RAU_AmpC/OD₆₀₀), after measuring the absorbance in 490 nm as time progressed using a microplate reader. After evaluating the RAU_AmpC/OD₆₀₀ value in each group (Fig. 13), piB_g1α in dual plasmid system(Fig. 13a), pirB_g1β (Fig. 13c) and pirB_g2β (Fig. 13d) exhibited better performances, which could be selected as the candidates for following experiments if necessary.

Fig. 13 The results of reporter AmpC for pirB toxin gene detection in RENDR-based system as time progressed of (a) pirB_g1α, (b) pirB_g1β, (c) pirB_g2α and (d) pirB_g2β. RAU: relative absorbance unit.

pirB toxin gene detection: RENDR with a NanoLuc output

Colonies harboring the correct plasmid(s) were cultivated and induced. The expression behavior of NanoLuc is represented by normalized RLU_NanoLuc (RLU_NanoLuc/OD₆₀₀), after measuring the bioluminescence as time progressed using a microplate reader. After evaluating the RLU_NanoLuc/OD₆₀₀ value in each group (Fig. 14), pirB_g2α in dual plasmid system (Fig. 14b) and pirB_g1β in single plasmid system (Fig. 14c) exhibited better performances, which could be selected as the candidates for following experiments if necessary.

Fig. 14 The results of reporter NanoLuc luciferase for pirB toxin gene detection in RENDR-based system as time progressed of (a) pirB_g1α, (b) pirB_g1β, (c) pirB_g2α and (d) pirB_g2β. RLU: relative luminescence unit.

Thus, based on the results from Fig. 12, Fig. 13, and Fig. 14, we concluded that pirB_g1β, pirB_g1α, and pirB_g2α showed better performances in vivo, which could be further used in the cell-free system.

Conclusion

Fig. 15 The heat map of the fold change (with input/without input) of output signals of different guides with diverse reporters.

Through experiments, we have verified the capability of our hardware to operate the RPA amplification, and the feasibility of the RENDR-based detection system to sense the toxin gene input in vivo (Fig. 15). Actually, some guides circuits showed a certain level of leakage in absence of input and the output signals were even lower upon adding the input. This might be attributed to the concentration differences of transcripts of guides and input, which is related to the direction of reaction equilibrium of self-splicing. This was the reason why the tests of some circuits were implemented in dual plasmid system, since the copy number of pSB3K3 that harbors the guides is smaller than that of pSB1C3 for harboring input sequence. As for some guides circuits that showed awful behaviors all the time, we speculated that it may be caused by the unfavorable structure of the transcripts for ribozyme self-splicing. Based on the in vivo characterization, we continued to perform the tests in cell-free expression system, which is closer to the Implementation form of our project.

Future prospects

Many experiments were carried out to achieve the detection targeted to the toxic gene of pirA and pirB within a limited time. However, the cause of some inverse results upon adding the input sequences should be further investigated. And the combinations that the constitutively expressed guides sequences with the input under the control of an orthogonal promoter may be tested for decreasing the scramble of RNA polymerases.

Treatment

Our strategy for treating AHPND is to kill the pathogenic Vibrio parahaemolyticus by a specific plasmid-encoded endolysin (10), which can be delivered via the engineered outer membrane vesicles (OMVs). For selectively targeting, we determined to decorate the surface of OMVs with two ligands, TTPA and TTPB, which were identified from the phage OWB (11) of V. parahaemolyticus that could bind to Vp0980 receptor on V. parahaemolyticus. In this case, two surface display systems, ClyA (12) and INPNC (13) were chosen to anchor the ligands on the surface membrane of engineered bacteria, thus resulting in the surface-modified OMVs. In addition, we also implemented RNA interference (RNAi) technology to silence ompA and lpp and overexpression of mepS for hypervesiculation of the engineered bacteria.

Highlight

Confirm the targeted binding of TTPA to Vp0980
Verify the function of RNAi and MepS for hypervesiculation of the engineered bacteria
Achieve the killing of endolysin to Vibrio alginolyticus

Introduction

Successful implementation of the surface display system in bacteria should be carefully verified through two steps, the first is to ensure that the cargo protein can be anchored on the bacterial surface, and the second is the verification of cargo protein’s normal function as its free state. Many methods have been developed to characterize the successful location of cargo protein when fused with the anchor, including flow cytometry (11, 14-17), western blot (14, 15, 18-20), enzyme linked immunosorbent assay (ELISA) (18), fluorescence microscope imaging (11, 14, 21) and so on. Among these characterization means, immuno-based technology was the most implemented due to the specificity and accessibility of antibodies. Therefore, given the laboratory resources and the complexity of operations, we finally decided to use the immunofluorescence (IF) technology for testing whether the cargo proteins are displayed or not and further probing the binding event of the protein-protein interaction (PPI) pairs on the surface of bacteria (Fig. 16). This approach was implemented both in the experiments of the Treatment part and Prevention part.

Fig. 16 Illustration of the immunofluorescence (IF) technology applied in the experiments of the Treatment part and Prevention part.

As for OMVs, common extraction and characterization methods were used, including ultracentrifugation, transmission electron microscope (TEM) imaging, and bicinchoninic acid (BCA) assay. Through those experiments, we could characterize the effect of RNAi and overexpression of the mepS gene.

While for the test of endolysin (Lysqdvp001), the killing efficiency was explored by performing the spot assay after incubating the purified endolysin with the V. parahaemolyticus-alternative strain V. alginolyticus.

Results

Promoters induction performance verification

Show Details

L-rhamnose-inducible promoter induction performance verification

We improved the original promoter (BBa_K914003) and compared the expression level between two genetic circuits (Fig. 17). Then, the effects of different inducers on the improved L-rhamnose-inducible promoter were also investigated. As shown in Fig. 18, the improved L-rhamnose-inducible promoter under the induction of L-rhamnose possessed the advantage of a stronger induction effect and lower leakage, which we will choose finally.

Fig. 17 The comparison of normalized fluorescence intensity between BBa_K4195109 and BBa_K4195191.

Fig. 18 Characterization of the eyfp expression and bacterial growth in different inducer concentrations (a: L-rhamnose, b: L-mannose).

Ligands can be anchored onto the bacterial surface

Show Details

ClyA/INPNC-TTPB-his and ClyA/INPNC-TTPA-his

Molecular cloning

Fig. 19 Graphic description of the expression gene circuits for display cassette designed in the Treatment part.

In order to verify whether ClyA or INPNC can display heterologous proteins on the surface of the engineered bacteria or not, a His-tag (6×His) was fused to the C-terminal of the two ligands (TTPA and TTPB). We used both BBa_I0500 (araC/pBAD) and BBa_B0034 to construct the expression system and obtained the four composite parts BBa_K4195100 (for INPNC-TTPB-his), BBa_K4195101 (for ClyA-TTPB-his), BBa_K4195120 (for ClyA-TTPA-his) and BBa_K4195122 (for INPNC-TTPA-his) (Fig. 19), which are respectively assembled into the vector pSB1C3 by standard BioBrick assembly. The constructed plasmid was transformed into E. coli BL21(DE3), then the positive transformants were selected by chloramphenicol and confirmed by colony PCR (Fig. 20) and sequencing.

Fig. 20 DNA gel electrophoresis of the colony PCR products. a BBa_K4195100_pSB1C3 in E. coli BL21(DE3). Target band (5020 bp) can be observed at the position around 5000 bp. b BBa_K4195101_pSB1C3 in E. coli BL21(DE3). Target band (5002 bp) can be observed at the position around 5000 bp. c BBa_K4195120_pSB1C3 in E. coli BL21(DE3). Target band (3220 bp) can be observed at the position around 3000 bp. d BBa_K4195122_pSB1C3 in E. coli BL21(DE3). Target bands (3238 bp) can be observed at the position around 3000 bp.

Immunofluorescence (IF)

The bacterial culture was cultivated overnight after L-arabinose induction. Then the FITC-labeled anti-His-Tag antibody was used to target the fused His-tag (6×His) displayed via ClyA and INPNC, followed by measuring the fluorescence intensity and OD₆₀₀ of the culture.

Fig. 21 The results of immunofluorescence to characterize the function of the two surface display systems. a Fluorescence intensity/OD₆₀₀ of E. coli whether TTPB fused to anchors or not. The left is for BBa_K4195100 (p = 0.00277) and the right is for BBa_K4195101 (p = 0.0019). b Fluorescence intensity/OD₆₀₀ of E. coli whether TTPA fused to anchors or not. The left is for BBa_K4195120 (p = 0.0126) and the right is for BBa_K4195122 (p = 0.0024).

The results showed that the ratio of fluorescence intensity to OD₆₀₀ of positive control (bacteria harboring surface display system) is higher than that of negative control (bacteria without surface display system) (Fig. 21), which indicates that both of the two ligands can be successfully located on the surface of engineered bacteria.

The function of ligands can be retained after displayed on the bacterial surface

Show Details

Vp0980 is a membrane protein

Molecular cloning and purification of Vp0980

In order to purify Vp0980, a His-tag (6×His) was fused to the N-terminal of it. L-arabinose-inducible expression system was also constructed and we obtained the composite part BBa_K4195110 at pSB1C3. The plasmid was transformed into E. coli Shuffle T7, then the positive transformants were selected by chloramphenicol and confirmed by colony PCR (Fig. 22a) and sequencing.

Fig. 22 DNA gel electrophoresis of the colony PCR products of BBa_K4195110_pSB1C3 (a) and BBa_K4195006_pET-28a(+) (b). a Target bands (2218 bp) can be observed at the position between 3000 bp and 2000 bp. b Target band (716 bp) can be observed at the position around 750 bp.

Colonies with correct sequence were cultivated and induced by L-arabinose to express the protein. Protein purification was carried out through GE AKTA Prime Plus FPLC System, which was verified by sodium dodecyl sulfate (SDS)-12% (wt/vol) polyacrylamide gel electrophoresis. We found that Vp0980 was expressed so poorly (data not shown) that we did not obtain much target protein. Therefore, we turned to reconstruct this part on the common expression vector pET-28a(+) through Gibson assembly and transformed the plasmid into E. coli DH5α and E. coli Origami 2(DE3). The positive transformants were selected by kanamycin and confirmed by colony PCR (Fig. 22b) and sequencing.

Because the Vp0980 is a receptor which has many transmembrane domains, we tried to purify the protein with 1% Triton X-100 and 1 mmol/L PMSF (both final concentrations) added to the buffer before ultrasonication, which is a necessary and crucial step for protein purification. As shown in the gel image of his-Vp0980 (Fig. 23), the target protein (20.0 kDa) can be observed at the position around 20 kDa on the purified protein lanes (FR), although shown with many other protein bands together.

Fig. 23 SDS-PAGE analysis of his-Vp0980 protein. Target bands (20 kDa) can be observed at the position 20 kDa.

We attributed that no obvious protein target band in LS because to that the content of proteins located on the bacterial membrane is minimal.

Immunofluorescence (IF) performed for confirming localization

For further verifying that Vp0980 was localized in membrane, after being induced by 2% arabinose, the E. coli was incubated with FITC-labeled anti-His-tag Antibody at 37 °C for 1 h. The microplate reader was used to measure the value of fluorescence intensity and OD₆₀₀. The fluorescence intensity/OD₆₀₀ value in the group of his-Vp0980 (Fig. 24) was stronger than that of I0500_pSB1C3 (negative control), which demonstrated that his-Vp0980 is located on the outer membrane of E. coli.

Fig. 24 Immunofluorescence (IF) technology was used to probe the localization of Vp0980 (p = 0.0232).

Displayed ligands can bind to Vp0980

IF experiments

Fig. 25 IF experiments were performed to verify whether the ligands displayed can bind to Vp0980 or not. a BBa_K4195121. b BBa_K4195123. c BBa_K4195103. d BBa_K4195102.

Immunofluorescence was used as well to demonstrate the interaction of his-Vp0980 with ClyA/INPNC-TTPA or ClyA/INPNC-TTPB, in which the group without adding his-Vp0980 was set as negative group. After being induced by 2% arabinose, the E. coli (ClyA/INPNC -TTPA or ClyA/INPNC-TTPB) was incubated with his-Vp0980 at 37 °C for 1 h, then incubated with FITC-labeled anti-His-Tag Antibody at 37 °C for 1 h. The microplate reader was used to measure the value of fluorescence intensity and OD₆₀₀. We found that the TTPA, no matter displayed by ClyA or INPNC, could bind to the purified Vp0980 (Fig. 25a & 25b). However, TTPB showed no ligand activity when displayed on the surface of engineered bacteria (Fig. 25c & 25d). We may attributed it to the potential cross-reactivity of our anti-His-Tag antibody, since the signals of both groups with or without purified Vp0980 added were much higher than that of other groups we have tested. Therefore, we could confirm that the TTPA might be further implemented for real applications.

Evaluate the binding affinity of ligands and Vp0980

Molecular cloning and purification of ligands

In order to further quantify the binding affinity of ligands and Vp0980, a myc-tag was fused to the C-terminal of the ligands. We constructed these parts on the common expression vector pET-28a(+) through Gibson assembly and transformed the plasmids into E. coli BL21(DE3). The positive transformants were selected by kanamycin and confirmed by colony PCR (Fig. 26) and sequencing.Colonies with correct sequence were cultivated and induced by IPTG to express the protein. Protein purification was carried out through GE AKTA Prime Plus FPLC System, which was verified by sodium dodecyl sulfate (SDS)-12% (wt/vol) polyacrylamide gel electrophoresis (Fig. 26).

Fig. 26 DNA gel electrophoresis of the colony PCR products and SDS-PAGE analysis of the target proteins of BBa_K4195089_pET-28a(+) (a,b) and BBa_K4195090_pET-28a(+) (c,d). a Target bands (764 bp) can be observed at the position between 750 bp and 1000 bp. b Target band (23.6 kDa) can be observed at the position between 20 kDa and 30 kDa. c Target bands (2546 bp) can be observed at the position between 2000 bp and 3000 bp. d Target band (87.6 kDa) can be observed at the position around 80 kDa.

Dot blot analysis

Dot blot analysis was used to determined the interaction between Vp0980 and its ligands. Firstly, the protein Vp0980 with his-tag was directly spotted onto the nitrocellulose (NC) membrane. Then the NC membrane was incubated with purified ligands with myc-tag and anti-myc-tag antibody in turn and finally probed by the HRP-conjugated secondary antibody. BSA was also directly spotted onto the NC membrane as the negative control. By reading the gray scale value of experimental groups from chemiluminescence imaging results, we quantitatively calculated the binding affinity (K_d) of ligands and Vp0980 (Fig. 27).

Fig. 27 The binding affinity (K_d) of Vp0980 and its ligands (a TTPA. b TTPB.) obtained from dot blot analysis.

OMVs massive release verification

Show Detail

OMVs can be purified via ultracentrifugation

Molecular cloning

The basic part BBa_I0500 was assembled into pSB1C3 plasmid backbone for the purification of OMVs. The plasmid constructed was transformed into E. coli DH5α and E. coli BL21(DE3), and the positive colonies were confirmed by chloramphenicol, colony PCR (Fig. 28), and sequencing. The target band (1524 bp) can be observed at the position around 1500 bp.

Fig. 28 The result of colony PCR. Plasmid pSB1C3.

OMVs purification and transmission electron microscope observation

TThe colony with the corrected sequence was cultivated for OMVs purification. After being collected from the supernatant of culture via ultracentrifugation, the OMVs samples were negatively stained for TEM observation (Fig. 29) to confirm that we purified the OMVs successfully.

Fig. 29 Transmission electron microscopy observation of OMVs.

Construction of hypervesiculation E. coli strain

siRNA design

siRNA sequences were designed through siRCon which is developed by Team Bielefeld-CeBiTec. OmpA 5’UTR and Hfq binding sequence are added to the flanks of siRNA sequences to enhance the binding ability to target sequence as well as increase the half-life of siRNA. Promoter (BBa_I0500) and terminator (BBa_B0015) were used to construct composite parts (Fig. 30).

Fig. 30 The composite parts about siRNA.

Molecular cloning

All related siRNA basic parts and mepS (BBa_K4195004) were assembled into the pSB1C3 plasmid backbone, which was transformed into E. coli DH5α and E. coli BL21(DE3). Positive colonies were confirmed by chloramphenicol, colony PCR (Fig. 31), and sequencing.

Fig. 31 The result of colony PCR. Plasmid pSB1C3. a the target band (1890 bp) of ompA 0.83. b the target band (1890 bp) of ompA 0.88. c the target band (1890 bp) of lpp 0.88. d the target band (1890 bp) of lpp 1.0. e the target band (2254 bp) of mepS.

OMVs purification and BCA assay

OMVs expressed from the gene circuits were purified via ultracentrifugation to verify the successful construction of hypervesiculation E. coli, in which BBa_I0500_pSB1C3 was used as control. Based on the standard curve developed from BCA Kit, the concentration of OMVs can be quantized to exhibit the effect of RNA interference (RNAi) in different gene circuits. As shown in Fig. 32, compared with that to lpp, the interference to ompA did not promote the generation of hypervesiculation. The lower OD₆₀₀ also revealed that RNA interference to lpp can promote the production of OMVs, leading to the destruction of cytomembrane stability.

Fig. 32 Comparison of normalized protein concentration and OD₆₀₀ between the RNAi circuits and negative control. a Verification of BBa_K4195104. b Verification of BBa_K4195105. c Verification of BBa_K4195106. d Verification of BBa_K4195107.

As shown in Fig. 33, Compared with the negative control (BBa_I0500_pSB1C3), MepS also resulted in a higher concentration of OMVs and lower OD₆₀₀ values, which verified its effective function of hypervesiculation.

Fig. 33 Comparison of normalized protein concentration and OD₆₀₀ between BBa_K4195108 and negative control.

Furthermore, about MepS (BBa_K4195108), we also measured the dynamic tendency of normalized protein concentration to time after induction, which indicated the massive secretion of OMVs over time (Fig. 34).

Fig. 34 Normalized protein concentration-time curves of BBa_K4195108 under the induction of L-arabinose.

Lysqdvp001 can kill Vibrio alginolyticus specially

Show Detail

The killing effect of Lysqdvp001

Molecular cloning and protein purification

In order to purify Lysqdvp001, a His-tag (6×His) was fused to the N-terminal of Lysqdvp001. BBa_I0500 and BBa_B0015 were used to construct the expression system, which was assembled on the expression vector pUC57-Simple by the standard assembly. The constructed plasmids were transformed into E. coli DH5α and BL21(DE3), then the positive transformants were selected by ampicillin and confirmed by colony PCR (Fig. 35) and sequencing.

Fig. 35 The result of colony PCR. Plasmid pUC57-Simple.

Colonies with correct sequence were cultivated and induced by L-arabinose to express the protein. Protein purification was carried out through AKTA Prime Plus FPLC System (GE, America), which was verified by sodium dodecyl sulfate (SDS)-12% (wt/vol) polyacrylamide gel electrophoresis (PAGE). As shown in the gel image of the his-Edl060 (Fig. 36), the target protein (26.0 kDa) could be observed at positions between 25 kDa and 30 kDa on the purified protein lanes (FR).

Fig. 36 SDS-PAGE analysis of protein in the lysate of E. coli BL21(DE3) and the elution samples. Target bands (26.0 kDa) can be observed at positions between 25 kDa and 30 kDa.

Spot assay

Three kinds of bacteria (E. coli, Vibrio alginolyticus, and Vibrio natriegens) were diluted from 1/10^-3 to 1/10^-5 and treated with EDTA and sterilized water (negative control, Fig. 37a), alkaline (positive control, Fig. 37b), respectively. As shown in Fig. 37c, three kinds of the bacterium were diluted to 1/10^-5 and treated with EDTA and purified endolysin Lysqdvp001 for 5 min, 10 min, 15 min, and 20 min. After being treated, these bacteria samples were used for spot assay respectively. The result demonstrated that Lysqdvp001 showed not any killing effect on other bacteria, but was highly specific to the Vibrio alginolyticus (Vibrio parahaemolyticus-alternative strain), which was consistent with the killing results that were treated by alkaline lysis.

Fig. 37 The result of killing effect verification. Grouped by different treatment: a EDTA and sterilized water. b Alkaline Lysis. c EDTA and Lysqdvp001.

Conclusion

The efficiency of the inducible promoters BBa_I0500 and BBa_K914003 was verified, which resulted in high expression with induction and low leakage without induction. We successfully proved that the two ligands (TTPA and TTPB) can be displayed by two surface display systems (ClyA and INPNC) on the surface of engineered bacteria. Besides, the function of TTPA to bind Vp0980 was verified after the ligand was displayed.The practicability of the plasmid transfer killing method was verified as effective. The RNAi and MepS methods were verified effective to promote OMVs secretion.

The experiments on the L-rhamnose-induced promoter can be divided into mainly two parts. Adding a regulatory element in front of the L-rhamnose-induced promoter made its inducing efficiency stronger. Then, a comparative experiment was conducted on the improved promoter with different inducers (L-rhamnose and L-mannose), which demonstrated that L-rhamnose is a better inducer.

In the surface-display system, we display the tail tubular protein TTPA/TTPB of Vibrio parahaemolyticus bacteriophage on the surface of OMVs by which to target Vibrio parahaemolyticus more efficiently. ClyA and INPNC are employed to build a surface display system, through which TTPA/TTPB can be displayed on the surface of OMVs.

For the plasmid transfer killing system, experiments designed for it were partially finished. We verified the killing effect of Lysqdvp001 expressed by edl060. Results showed that the endolysin has a high killing effect on Vibrio alginolyticus (similar to Vibrio parahaemolyticus), which created the precondition for the Proof of Concept verification of Lysqdvp001 for Vibrio alginolyticus killing.

As for the RNAi method, it is proved that the interference to ompA has no obvious effect while the interference to lpp triggered the massive secretion of OMVs. At the same time, MepS method also triggered massive secretion of OMVs. With these results, a kinetic model to characterize an enzymology dynamic curve was developed successfully.

Future prospects

The experiments have been done, and the design has been proven. Given the fact that Vibrio alginolyticus is similar to Vibrio parahaemolyticus, we expect to express the vp0980 gene in Vibrio alginolyticus so that we can simulate the natural environment more realistically. The TTPA or TTPB displayed OMVs containing plasmid (loading the gene edl060 encoding endolysin Lysqdvp001) are wished to incubate with Vibrio alginolyticus expressing Vp0980 to verify the feasibility and effectiveness of our complete design.

Prevention

Our strategy for preventing AHPND mainly caused by the VP_AHPND toxins PirA and PirB is to neutralize the toxins via the receptors that are surface-displayed on OMVs released from the engineered bacteria. To build the fortifications against toxins PirA and PirB, two surface display systems (ClyA and INPNC) were chosen to anchor two receptors (rFET and rLvAPN1) in our Design. At first, we have constructed four expression circuits for the anchor-receptor fusion proteins and verified that the function of receptors is retained on the surface of engineered bacteria. Besides, the interaction between toxins and receptors has been detected in vitro.

Highlight

Confirm the display of the receptor rFET and rLvAPN1
Verify the retained function of displayed receptors
Detect the interaction between rLvAPN1 and PirA in vitro

Introduction

In order to verify whether two receptors can be displayed by the two surface display systems or not and whether the receptors after displayed can still bind the toxins or not, we chose immunofluorescence (IF) technology as our characterization method, which has been mentioned above. Besides, isothermal titration calorimetry (ITC) (1, 2) was used to detect the interaction between receptors and toxins in vitro, which gives out the binding affinities of those protein-protein interaction (PPI) pairs.

Results

Toxin receptors can be anchored onto the bacterial surface

Show Details

ClyA/INPNC-rFET-his and ClyA/INPNC-rLvAPN1-his

Molecular cloning

Fig. 38 Graphic description of the expression gene circuits for display cassette designed in the Prevention part.

In order to verify whether ClyA or INPNC can display heterologous proteins on the surface of the engineered bacteria or not, a His-tag (6×His) was fused to the C-terminal of the two receptors. We used both BBa_I0500 (araC/pBAD) and BBa_B0034 to construct the expression system and obtained the four composite parts BBa_K4195126 (for INPNC-rFET-his), BBa_K4195127 (for ClyA-rFET-his), BBa_K4195131 (for ClyA-rLvAPN1-his) and BBa_K4195133 (for INPNC-rLvAPN1-his) (Fig. 38), which are respectively assembled into the vector pSB1C3 by standard BioBrick assembly. The constructed plasmid was transformed into E. coli BL21(DE3), then the positive transformants were selected by chloramphenicol and confirmed by colony PCR (Fig. 39) and sequencing.

Fig. 39 DNA gel electrophoresis of the colony PCR products. a BBa_K4195127 in E. coli BL21(DE3). Target bands (2484 bp) can be observed at the position between 3000 bp and 2000 bp. b BBa_K4195126 in E. coli BL21(DE3). Target bands (2505 bp) can be observed at the position between 3000 bp and 2000 bp. c BBa_K4195131 in E. coli BL21(DE3). Target bands (3879 bp) can be observed at the position between 5000 bp and 3000 bp. d BBa_K4195133 in E. coli BL21(DE3). Target bands (3900 bp) can be observed at the position between 5000 bp and 3000 bp.

Immunofluorescence (IF)

The bacterial culture was cultivated overnight after L-arabinose induction. Then the FITC-labeled anti-His-Tag antibody was used to target the fused His-tag (6×His) displayed via ClyA and INPNC, followed by measuring the fluorescence intensity and OD₆₀₀ of the culture.

Fig. 40 The results of immunofluorescence to characterize the function of the two surface display systems. a Fluorescence intensity/OD₆₀₀ of E. coli whether rFET is fused to anchors or not. The left is for BBa_K4195127 (p = 0.0024) and the right is for BBa_K4195126 (p = 0.0199). b Fluorescence intensity/OD₆₀₀ of E. coli whether rLvAPN1 is fused to anchors or not. The left is for BBa_K4195131 (p = 0.0092) and the right is for BBa_K4195133 (p = 0.0406).

The results showed that the ratio of fluorescence intensity to OD₆₀₀ of positive control (bacteria harboring surface display system) is higher than that of negative control (bacteria without surface display system) (Fig. 40), which indicates that both of the two receptors can be successfully located on the surface of engineered bacteria.

Receptors can bind to the toxins with different affinities in vitro

Show Details

Molecular cloning and protein purification of the toxins

In order to purify PirA and PirB, a His-tag (6×His) was fused to the C-terminal of PirA and the N-terminal of PirB. Likewise, L-arabinose-inducible expression system was built to obtain the composite part BBa_K4195137 (for PirA-his) and BBa_K4195138 (for his-PirB) at pSB1C3. The constructed plasmid was transformed into E. coli BL21(DE3), then the positive transformants were selected by chloramphenicol and confirmed by colony PCR (Fig. 41) and sequencing.

Fig. 41 DNA gel electrophoresis of the colony PCR products. a BBa_K4195137 in E. coli BL21(DE3). Target bands (2086 bp) can be observed at the position between 3000 bp and 2000 bp. b BBa_K4195138 in E. coli BL21(DE3). Target bands (3070 bp) can be observed at the position around 3000 bp.

After being cultivated and induced by arabinose, GE AKTA Prime Plus FPLC System was employed to get purified protein from the lysate supernatant. Purified protein was verified by sodium dodecyl sulfate (SDS)-12% (wt/vol) polyacrylamide gel electrophoresis (SDS-PAGE) and Coomassie blue staining.

After the first time of purification, we got PirA-his successfully. As shown in the gel image of PirA-his (Fig. 42), the target protein (14.4 kDa) can be observed at the position between 25 kDa and 14 kDa on the purified protein lanes (FR).

Fig. 42 SDS-PAGE analysis of PirA-his protein in lysate of E. coli BL21(DE3) and the elution samples. Target bands (14.4 kDa) can be observed at the position between 25 kDa and 14 kDa.

However, we found that PirB was expressed so poorly (data not shown) that we did not obtain the target protein. Therefore, we turned to construct this part on the common expression vector pET-28a(+) by Gibson assembly, then transformed the plasmid into E. coli BL21(DE3). The positive transformants were selected by kanamycin and confirmed by colony PCR (Fig. 43) and sequencing.

Fig. 43 DNA gel electrophoresis of the colony PCR products of BBa_K4195036_pET-28a(+). Target bands (1539 bp) can be observed at the position between 2000 bp and 1500 bp.

Unfortunately, after the second round of purification, we found that the expression of PirB remained low as the first time (data not shown). Under this circumstance, we co-transformed another plasmid that could express molecular chaperones GroES and GroEL with his-PirB_pET-28a(+) into E. coli BL21(DE3) to try to improve the expression efficiency. As shown in the gel image of his-PirB (Fig. 44), the target protein (51.9 kDa) can be observed at the position between 60 kDa and 50 kDa on the purified protein lanes (FR), although there were some unwanted bands shown together.

Fig. 44 SDS-PAGE analysis of his-PirB protein in lysate of E. coli BL21(DE3) and the elution samples. Target bands (51.9 kDa) can be observed at the position between 60 kDa and 50 kDa.

The interaction between rFET and toxin PirA and PirB

Molecular cloning and protein purification of rFET

In order to purify rFET, a His-tag (6×His) was fused to the C-terminal of it. L-arabinose-inducible expression system was also constructed and we obtained the composite part BBa_K4195112 at pSB1C3. The plasmid was transformed into E. coli Shuffle T7, then the positive transformants were selected by chloramphenicol and confirmed by colony PCR (Fig. 45a) and sequencing.

Fig. 45 DNA gel electrophoresis of the colony PCR products of BBa_K4195112 (a) and BBa_K4195009(+) (b). a Target bands (1801 bp) can be observed at the position around 2000 bp. b Target bands (215 bp) can be observed at the position between 250 bp and 100 bp.

After the first time of purification, we found that rFET was expressed so poorly (data not shown) to obtain the target protein. Like his-PirB, we constructed this part on the expression vector pET-28a(+) by Gibson assembly, then transformed the plasmid into E. coli BL21(DE3). The positive transformants were selected by kanamycin and confirmed by colony PCR (Fig. 45b) and sequencing.

It was so regret that we could not confirmed that we obtained the purified rFET during the whole season, since the rFET-his is so small (37 aa, 4.1 kDa) and the target bands cannot be observed obviously even though the concentration of separating gel has been increased to 20% for SDS-PAGE. Thus, we were not obsessed with purifying this protein for in vitro bio-chemical test and turned to the characterization on the surface of bacteria.

The interaction between rLvAPN1 and toxin PirA and PirB

Molecular cloning and protein purification of rLvAPN1

Fig. 46 DNA gel electrophoresis of the colony PCR products of BBa_K4195134 (a) and BBa_K4195038(+) (b). a Target bands (2925 bp) can be observed at the position around 3000 bp. b Target bands (1386 bp) can be observed at the position between 1500 bp and 1000 bp.

rLvAPN1-his (BBa_K4195038) was designed for protein purification and we obtained the composite part BBa_K4195134 at pSB1C3, which was also induced by L-arabinose (Fig. 46a). However, this expression system did not work as expected and we reconstructed this part at pET-28a(+) like before (Fig. 46b). Finally, as shown in the gel image of rLvAPN1-his (Fig. 47), the bands of target protein (45.5 kDa) can be observed at the position between 50 kDa and 35 kDa on the purified protein lanes (FR). Besides, with the help of CUG-China for offering the operation of western blot, we could confirm that the rLvAPN1-his was expressed indeed (Fig. 47b), after struggling to obtain the purified protein for several rounds mentioned above (Learn more from our Partnership page).

Fig. 47 Confirmation of the purification for rLvAPN1-his. a SDS-PAGE analysis of rLvAPN1-his in lysate of E. coli BL21(DE3) and the elution samples, in which target bands (45.5 kDa) can be observed at the position between 50 kDa and 35 kDa. b Western blot analysis of rLvAPN1-his in lysate supernatant, in which target bands (45.5 kDa) can be observed at the position between 43 kDa and 55 kDa.

Isothermal titration calorimetry (ITC)

ITC was used to detect the interaction of rLvAPN1 to PirA. Assuming a one-site binding model, the apparent dissociation constants (K_d) of rLvAPN1 to PirA, as calculated from the ITC curves, were 5.9 μM (Fig. 48), which is similar to the reported binding affinity (3.2 μM) that determined through ELISA. The result indicates that rLvAPN1 can bind directly to toxin PirA. However, due to the need of large sample amount of ITC, we could not obtain a satisfying titration curve of rLvAPN1 to PirB because of the few amount of soluble components of this toxin.

Fig. 48 The binding affinity of rLvAPN1 to PirA determined by ITC. One-site binding model was applied to curve fitting.

The function of receptors can be retained after displayed on the bacterial surface

Show Details

Molecular cloning

In order to verify whether two receptors on the surface of the engineered bacteria can bind toxins, we constructed four composite parts BBa_K4195125 (for INPNC-rFET), BBa_K4195128 (for ClyA-rFET), BBa_K4195130 (for ClyA-rLvAPN1), and BBa_K4195132 (for INPNC-rLvAPN1) at pSB1C3, which are all under the control of L-arabinose-inducible system. The plasmid was transformed into E. coli BL21(DE3), then the positive transformants were selected by chloramphenicol and confirmed by colony PCR (Fig. 49) and sequencing.

Fig. 49 DNA gel electrophoresis of the colony PCR products.a BBa_K4195125 in E. coli BL21(DE3). Target bands (2484 bp) can be observed at the position between 3000 bp and 2000 bp. b BBa_K4195128 in E. coli BL21(DE3). Target bands (2466 bp) can be observed at the position between 3000 bp and 2000 bp. c BBa_K4195130 in E. coli BL21(DE3). Target bands (3816 bp) can be observed at the position between 5000 bp and 3000 bp. d BBa_K4195132 in E. coli BL21(DE3). Target bands (3834 bp) can be observed at the position between 5000 bp and 3000 bp.

The interaction between displayed rFET and toxin PirA and PirB

Immunofluorescence (IF)

After induction, E. coli harboring BBa_K4195125 and BBa_K4195128 were cultivated overnight. The bacterial culture with his-PirB added was set as positive control while which had no toxins added was set as negative control. Then IF experiment was performed again to test whether the receptor activity is retained or not after displayed on the bacterial surface.

The ratio of fluorescence intensity to OD₆₀₀ of the positive control (culture was incubated with toxins) is higher than that of negative control (culture was incubated with 1×TBST) (Fig. 50). The results proved that our surface display system works well and the rFET can still bind PirA-his and his-PirB, even on the surface of bacteria.

Fig. 50 Ability of binding his-PirB on the surface of E. coli BL21(DE3). a Fluorescence intensity/OD₆₀₀ of E. coli harboring BBa_K4195128 which was incubated with PirB or not (p = 0.0002). b Fluorescence intensity/OD₆₀₀ of E. coli harboring BBa_K4195125 which was incubated with PirB or not (p = 0.0426).

The interaction between displayed rLvAPN1 and toxin PirA and PirB

Immunofluorescence (IF)

After induction, E. coli harboring BBa_K4195130 and BBa_K4195132 were cultivated overnight. The bacteria culture with PirA-his or his-PirB added was set as positive control while which had no toxins added was set as negative control. IF was implemented once again to test whether the receptor activity was retained or not after displayed on the bacterial surface.

The ratio of fluorescence intensity to OD₆₀₀ of positive control (culture was incubated with toxins) is higher than that of negative control (culture was incubated with 1×TBST) (Fig. 51), which indicates that our surface display system works well and the displayed rLvAPN1 can still bind PirA-his and his-PirB.

Fig. 51 Ability of binding PirA-his and his-PirB on the surface of E. coli BL21(DE3). a Fluorescence intensity/OD₆₀₀ of E. coli harboring BBa_K4195130 which was incubated with toxins or not (p = 0.0082 for PirA-his, p = 0.0213 for his-PirB). b Fluorescence intensity/OD₆₀₀ of E. coli harboring BBa_K4195132 which was incubated with toxins or not (p = 0.0014 for PirA-his, p = 0.0329 for his-PirB).

Conclusion

We successfully proved that the two receptors (rFET and rLvAPN1) can be displayed by two surface display systems (ClyA and INPNC) on the surface of engineered bacteria. Together with the results of Treatment part, it is convincing that the surface display system we chose can function as expected to locate the cargo proteins onto the cell surface. Besides, the function of rFET to bind the toxin PirB and rLvAPN1 to bind both PirA and PirB was verified after the receptor was displayed. As for the binding affinity of displayed rFET to PirA, we attributed this to the potential weak interaction between the two proteins, which may be similar to the situation that the interaction between rLvAPN1 and PirA (K_d = 3.2 μM) is weaker than that between rLvAPN1 and PirB (K_d = 0.5 μM). Because the concentration of some protein purified can hardly meet the need of ITC, we only obtained the binding affinity of rLvAPN1 to PirA (K_d = 5.9 μM) in this in vitro characterization. In summary, the outcome we have gotten set the fundamental for the verification at the level of OMVs (more details shown in our Proof of Concept page), since OMVs are originated from the bacterial surface (3).

Future prospects

Given the length of rFET (31 aa), such protein should be considered as polypeptide and obtained through solid-phase peptide synthesis (SPPS). Alternative characterization method, such as ESI-MS or MALDI-MS, can be applied to verification rather than SDS-PAGE. We hope this will be realized thus the in vitro protein-protein interaction could be probed.

To achieve the virulence loss of V. parahaemolyticus for further prevention of AHPND, we also designed a CRISPR/Cas system to destroy both pirA and pirB gene (Learn more in our Design page). However, due to the time limit, we have not cloned the type I-F cas cluster from the V. parahaemolyticus-alternative strain V. alginolyticus (BSL1) and constructed the complete circuits as designed. We are looking forward to proving its effectiveness in the future.

Kill Switch

The kill switch system is mainly composed of an inducible promoter (pBAD/araC) and a mazF gene (express toxin protein of MazF). Both promoters could be induced by L-arabinose to express toxin protein. Our GMOs will be mixed in the fodder and put into the water environment of prawn ponds, then colonize directly in prawns’ intestines. The kill switch is essential for preventing the leakage of GMOs. Before harvesting and marketing mature prawns, we must ensure that the GMOs colonized in prawn intestines are killed. Besides, GMOs, that remain in the tailwater of prawn ponds, must be killed before the tailwater is drained into the environment out of the prawn ponds. We planned to throw L-arabinose into the tailwater, so that MazF can be expressed and kill them, guaranteeing biosafety by cleaving the gnome of GMOs.

We use pBAD/araC (BBa_I0500), RBS (BBa_B0034), terminator (BBa_B0015), mazF (BBa_K1096002) to construct the composite part BBa_K3332083, which were assembled on pSB1C3 backbone by standard assembly. The constructed plasmids were transformed into E. coli DH5α and BL21(DE3), then the positive transformants were selected by chloramphenicol and confirmed by colony PCR (Fig. 52) and sequencing.

Fig. 52 DNA gel electrophoresis of the colony PCR products (BBa_K3332083).

CFU assay was implemented to characterize the effect of toxin MazF (Fig. 53a). We observed that the number of colonies on the plate decreased significantly after induction, while there was no significant change in the non-induced group (Fig. 53b). This demonstrated that toxin MazF could be lethal to the bacteria.

Fig. 53 CFU assay for characterizing the function of MazF. A Solid medium plate of CFU assay. b survival Ratio of different groups against time (h).

Reference

1. M. B. Holowko, H. Wang, P. Jayaraman, C. L. Poh, Biosensing Vibrio cholerae with Genetically Engineered Escherichia coli. ACS Synth. Biol. 5, 1275-1283 (2016).
2. S. Gupta, E. E. Bram, R. Weiss, Genetically programmable pathogen sense and destroy. ACS Synth. Biol. 2, 715-723 (2013).
3. P. Jayaraman, M. B. Holowko, J. W. Yeoh, S. Lim, C. L. Poh, Repurposing a Two-Component System-Based Biosensor for the Killing of Vibrio cholerae. ACS Synth. Biol. 6, 1403-1415 (2017).
4. Y. Wu, C. W. Wang, D. Wang, N. Wei, A Whole-Cell Biosensor for Point-of-Care Detection of Waterborne Bacterial Pathogens. ACS Synth. Biol. 10, 333-344 (2021).
5. K. A. Egland, E. P. Greenberg, Quorum sensing in Vibrio fischeri: elements of the luxl promoter. Mol. Microbiol. 31, 1197-1204 (1999).
6. A. Boo, R. Ledesma Amaro, G.-B. Stan, Quorum sensing in synthetic biology: A review. Curr. Opin. Syst. Biol. 28, 100378 (2021).
7. L. Gambill et al.,https://doi.org/10.1101/2022.01.12.476080 (2022)
8. J. N. Zadeh et al., NUPACK: Analysis and design of nucleic acid systems. J. Comput. Chem. 32, 170-173 (2011).
9. A. Lopreside, X. Wan, E. Michelini, A. Roda, B. Wang, Comprehensive Profiling of Diverse Genetic Reporters with Application to Whole-Cell and Cell-Free Biosensors. Anal. Chem. 91, 15284-15292 (2019).
10. W. Wang, M. Li, H. Lin, J. Wang, X. Mao, The Vibrio parahaemolyticus-infecting bacteriophage qdvp001: genome sequence and endolysin with a modular structure. Arch. Virol. 161, 2645-2652 (2016).
11. Q. Li, Z. Yu, X. Shao, J. He, L. Li, Improved phosphate biosorption by bacterial surface display of phosphate-binding protein utilizing ice nucleation protein. FEMS Microbiol. Lett. 299, 44-52 (2009).
12. K. Murase, Cytolysin A (ClyA): A Bacterial Virulence Factor with Potential Applications in Nanopore Technology, Vaccine Development, and Tumor Therapy. Toxins (Basel) 14, 78 (2022).
13. E. van Bloois, R. T. Winter, H. Kolmar, M. W. Fraaije, Decorating microbes: surface display of proteins on Escherichia coli. Trends Biotechnol. 29, 79-86 (2011).
14. A. Hjelm et al., Autotransporter-based antigen display in bacterial ghosts. Appl. Environ. Microbiol. 81, 726-735 (2015).
15. J. Jarmander, M. Gustavsson, T. H. Do, P. Samuelson, G. Larsson, A dual tag system for facilitated detection of surface expressed proteins in Escherichia coli. Microb. Cell Fact. 11, 118 (2012).
16. M. Gustavsson, E. Backlund, G. Larsson, Optimisation of surface expression using the AIDA autotransporter. Microb. Cell Fact. 10, 72 (2011).
17. Z. Zhang et al., Surface Immobilization of Human Arginase-1 with an Engineered Ice Nucleation Protein Display System in E. coli. PLoS One 11, e0160367 (2016).
18. Z. Yang, Q. Liu, Q. Wang, Y. Zhang, Novel bacterial surface display systems based on outer membrane anchoring elements from the marine bacterium Vibrio anguillarum. Appl. Environ. Microbiol. 74, 4359-4365 (2008).
19. Y. Zhu, Y. Liu, M. Ai, X. Jia, Surface display of carbonic anhydrase on Escherichia coli for CO₂ capture and mineralization. Synth. Syst. Biotechnol. 7, 460-473 (2022).
20. E. van Bloois, R. T. Winter, D. B. Janssen, M. W. Fraaije, Export of functional Streptomyces coelicolor alditol oxidase to the periplasm or cell surface of Escherichia coli and its application in whole-cell biocatalysis. Appl. Microbiol. Biotechnol. 83, 679-687 (2009).
21. Y. Liu et al., Display of a novel carboxylesterase CarCby on Escherichia coli cell surface for carbaryl pesticide bioremediation. Microb. Cell Fact. 21, 97 (2022).
22. C. L. Kelly et al., Synthetic Chemical Inducers and Genetic Decoupling Enable Orthogonal Control of the rhaBAD Promoter. ACS Synth. Biol. 5, 1136-1145 (2016).
23. M. Zhou, Q. Li, R. Wang, Current Experimental Methods for Characterizing Protein-Protein Interactions. ChemMedChem 11, 738-756 (2016).
24. M. Harish, P. Venkatraman, Evolution of biophysical tools for quantitative protein interactions and drug discovery. Emerging Top. Life Sci. 5, 1-12 (2021).
25. W. Luangtrakul et al., Cytotoxicity of Vibrio parahaemolyticus AHPND toxin on shrimp hemocytes, a newly identified target tissue, involves binding of toxin to aminopeptidase N1 receptor. PLoS Pathog. 17, e1009463 (2021).
26. M. Toyofuku, N. Nomura, L. Eberl, Types and origins of bacterial membrane vesicles. Nat. Rev. Microbiol. 17, 13-24 (2019).