Module 2: Biosensing systems
This page presents the results about the two biosensing systems we designed for our project. One biosensor relies on the detection of the oncometabolite 2-hydroxyglutarate (2-HG) via the DhdR repressor system. The second biosensor involves a split T7 RNA polymerase linked to antibodies for the recognition of HER2 and subsequent transcriptional activation of a gene of interest. We attempted to implement both biosensing systems in PUREfrex, in bulk reactions, and within liposomes, in the case of the 2-HG-inducible system.
Our specific objectives were to purify the DhdR protein, validate its repressing role in PURE system, and quantify its susceptibility to physiological concentrations of 2-HG. Finally, we aimed to assess its functionality within liposomes.
Production and Purification of DhdR
This section presents our successful production and purification of DhdR. As a reminder, production in E. coli was conducted after cloning the his-tagged dhdR gene in a pET21a (+) plasmid (see Cloning results).
DhdR was expressed and purified as described in our Protocols page.
Pure fractions were analyzed by SDS-PAGE (Fig. 1). The expected size of DhdR is 28.27 kDa. Clear bands were observed for the four elution fractions E150, E1100, E1250, and E1500, but not in the negative control sample, as expected.
Figure 1: SDS-PAGE analysis of the different steps of the DhdR purification: Protein ladder (SE250 Mighty Small II Mini Vertical Protein Electrophoresis Unit), pellet DhdR, cell-free extract (CFE), flowthrough (FT), wash (W), elution with 50 mM imidazole, 100 mM imidazole, 250 mM imidazole and 500 mM imidazole (respectively E150, E1100, E1250 and E1500). The band corresponding to the protein was clearly identified by Coomassie Blue staining.
The E1250 fraction was dialysed leading to a concentration of 14.4 µM (> 95% pure protein). Fractions E1100 and E1500 were pooled and dialysed, resulting in a concentration of 7.63 µM (> 95% pure protein).
These data show that DhdR was efficiently purified and can be used for subsequent assays.
Functionality tests in PURE system
This section presents our results demonstrating the functionality of our biosensor detecting the oncometabolite 2-hydroxyglutarate (2-HG) with the repressor DhdR. Specifically, the aim of our experiments was to establish that the binding of the repressor DhdR to its operator site, dhdO, effectively inhibits transcription of a gene of interest regulated by dhdO. Then, we wanted to show that the presence of 2-HG leads to the de-repression of that gene in PURE system, as illustrated in Figure 2.
Figure 2: Operating principle of the biosensor.
The G-Block DhdR was cloned into pET21 (a+) plasmid (see Cloning results), and the repressor protein was expressed and successfully purified (see results below).
The sfGFP reporter gene harboring the dhdO site was used for assaying DhdR activity and the fluorescence signal of the GFP protein was measured with a spectrofluorometer.
1) Cell-free production of sfGFP
We used the PCR products of tymp, sfgfp, and anti-HER2 nanobody (anti-HER2 nb) as templates for expression with GeneFrontier PUREfrex2.0 kit (See the protocol here). Additionally, we supplemented the reaction with GreenLys reagent for the co-translational incorporation of fluorescent lysine residues, which facilitated the detection of synthesized proteins by SDS-PAGE.
Figure 3: SDS-PAGE analysis of the gene expression products (Mini-PROTEAN TGX Stain-free Gels). The overlay of the GreenLys and stain-free images are shown. Lanes from left to right: negative control without DNA, positive control with dhfr control plasmid, TYMP with an expected size of 52 kDa, sfGFP with an expected size of 26 kDa, anti-HER2-nb with an expected size of 17 kDa, protein ladder.
The presence of the GFP protein at the expected molecular weight was visible in lane 4 (Figure 3). This result confirms the successful production of GFP protein in PURE system under non-repressed conditions.
2) DhdR inhibits expression of a gene under control of dhdO
Our first goal was to validate that DhdR inhibits expression of our reporter gene, sfgfp, by binding to its operator site dhdO. sfGFP was synthesized in bulk either with 1.5 µM of DhdR or without DhdR (see PURE system protocol). Concentration of DhdR was determined from our biochemical network model (more details in our Global kinetic model page). These two conditions were repeated on different types of DNA templates, namely linear PCR products or circular plasmids, both of them being compatible with cell-free expression in PURE system.
Figure 4: Purified DhdR represses expression of a fluorescent reporter gene. Different DNA templates harboring the sfgfp gene were expressed in two PURE system kits, specifically PUREfrex2.0 or PUREfrex2.1. The intensity value of sfGFP without DhdR was used for normalization.
Repression of sfGFP production by DhdR was confirmed with both DNA templates (Figure 4). The strength of repression was slightly higher with PCR products as a template than with a plasmid. Therefore, the next experiments were conducted using the PCR product of clone 8 as a template. Moreover, no significant differences were observed between the two PUREfrex kits, indicating that DhdR is active in both reducing and nonreducing conditions.
Then, we wanted to determine the minimal concentration of DhdR required to obtain strong repression. sfGFP was synthesized in the presence of different concentrations of DhdR. Our biochemical network model predicted a range of DhdR concentrations expected to lead to different sfGFP levels (see Global kinetic model page), which we experimentally tested.
Figure 5: Effect of different concentrations of DhdR on the expression of a fluorescent reporter gene. Experiments 1 and 2 were performed with the same batch of PCR product from clone 8, while a new batch of PCR product from the same clone was used in experiment 3. PUREfrex2.0 was used in all conditions. The intensity value of sfGFP without DhdR was used for normalization.
As expected, the higher the concentration of DhdR, the stronger the repression in all three experiments (Figure 5). With the new batch of linear DNA, repression was consistently stronger. We deduced from these results that the optimal concentration of DhdR to efficiently repress expression of a gene under transcriptional control of a dhdO operator sequence was 1.5 µM, validating the predictions of the biochemical network model.
In turn, the obtained experimental data allowed us to recalibrate our biochemical network model for better fitting the observations (Global kinetic model page).
3) 2-HG relieves the repression by DhdR
Before testing the inducible role of 2-HG, we wanted to check if its presence in PURE system could influence the synthesis of sfGFP. We ran PURE system reactions in the presence of different concentrations of 2-HG without DhdR. We chose physiological concentrations of 2-HG found around tumor cells, i.e., between 10 and 100 µM. A higher concentration was also tested corresponding to full saturation of the DhdR repressor.
Figure 6: Effect of different concentrations of 2-HG on the expression of sfGFP. PUREfrex2.1 was used.
Figure 6 shows that the signal of sfGFP decreases as the concentration of 2-HG increases, indicating that 2-HG partly inhibits protein synthesis in PUREfrex2.1. This inactivation has been taken into account in our next experiments.
Induction of gene expression that was repressed by 1.5 µM of DhdR was then assayed using the same range of 2-HG concentrations as above. The results demonstrate that 2-HG de-represses transcription of DhdR-bound DNA in a concentration-dependent manner (Figure 7). Up to 48% of sfGFP signal was recovered at a saturating concentration of 2-HG. The reason why protein production is not fully restored remains to be investigated.
Figure 7: Effect of different concentrations of 2-HG on DhdR repression. PUREfrex2.1 was used. The intensity value of sfGFP without DhdR and with 10 µM 2-HG was used for normalization. Each concentration was corrected taking into account the inactivation effect of 2-HG on PURE system.
In-liposome expression of the sfgfp gene
This section presents our results demonstrating successful expression of sfGFP inside liposomes.
We conducted two experiments in which we encapsulated the PURE system solution along with the sfgfp gene, either with 1.5 µM of DhdR or without DhdR (see In-liposome gfp expression protocol). We visualized the liposomes by optical microscopy.
Figure 8a displays a population of liposomes localized by the membrane dye Topfluor594. A zoom-in image of liposomes showed the fluorescent rim characteristic of membrane-labeled vesicles (Fig. 8b). The line intensity profile generated with ImageJ confirmed that the intensity was highest at the membrane and lower inside the liposome (Fig. 8c).
Figure 8: a) Liposomes were localized with a fluorescent membrane dye. b) Zoom-in image of the liposome area depicted with a red arrow in a. A yellow line crossing the liposome has been appended. c) Fluorescence intensity profile along the line appended in b. The two peaks correspond to the two regions of membrane crossing.
We were curious to image our liposomes with a high-resolution microscope (Nikon Eclipse Ti microscope at the photonic platform LITC at CBI: see In-liposome gfp expression protocol). Figure 9 shows a 72-ms stream of a single liposome with its fluorescent membrane rim.
Figure 9: 72 ms streaming of liposomes localized with a fluorescent membrane dye.
Figure 10a displays a population of liposomes expressing the sfgfp gene. In the liposome shown in Fig. 10b, one can clearly see the distribution of GFP fluorescence inside the lumen of the liposome. A quantitative analysis is represented in Fig. 10c. Analysis of the two samples with or without DhdR did not reveal notable differences neither in the occurrence of liposomes exhibiting GFP nor in the intensity level of GFP inside individual liposomes. Follow-up experiments will be necessary to optimize the relative and absolute amounts of DNA and DhdR in liposomes, which should allow for a better discrimination between repressing and non-repressing conditions.
Figure 10: Fluorescence microscopy image of liposomes in the GFP channel. Expressed sfGFP signal was localized in the liposome lumen. a) Large field-of-view. b) Blow-up of the liposome depicted with the red arrow in a. c) Intensity profile along the yellow line appended in b.
Conclusion
These experiments provide evidence that our 2-HG biosensor is functional in bulk reactions at concentrations of oncometabolite that are physiologically relevant. Our biochemical network model was used to predict DhdR concentrations, and then optimized according to experimental results. Moreover, we established a protocol for encapsulating the biosensor inside gene-expressing liposomes. Although DhdR-based repression was not clearly demonstrated in vesicles, we gave recommendations and provided image analysis tools for future investigations.
Herein, we describe the results about the two-partite RNA polymerase-linked antibody biosensor. The main challenge was to obtain a functional biosensor entirely produced with PURE system. We faced cloning difficulties with one of the two biosensor subunits. As a proof-of-concept of the split T7 RNAP recombination assay in PURE system, we decided to produce an anti-rapamycin biosensor that led to promising results.
Cell-free production of Pertuzumab-SL-T7Nterm
The objective of this experiment was to assess whether the protein Pertuzumab-SL-T7Nterm, consisting of the N-terminal fragment of the split T7 RNA polymerase fused to the anti-HER2 antibody pertuzumab through a soluble linker (SL), could be expressed from its DNA template in PURE system. The challenge resides in the complexity of this chimeric protein that contains structurally different domains.
1) Successful production of Pertuzumab-SL-T7Nterm with PUREfrex2.0
We first expressed Pertuzumab-SL-T7Nterm from its DNA template (see Cloning results) using the PUREfrex2.0 kit supplemented with SP6 RNAP (see PURE system protocol). The reaction products were analyzed by SDS-PAGE. Because the theoretical molecular weight is 69 kDa, no other band from PURE system proteins was expected to migrate at this size. The protein pattern shown in Figure 11 exhibits an additional band around 69 kDa compared to the negative controls.This result indicates successful production of the full-length Pertuzumab-SL-T7Nterm in PUREfrex2.0.
Figure 11: SDS-PAGE (10% polyacrylamide) analysis of Pertuzumab-SL-T7Nterm visualized by Instant Blue staining. The arrowhead indicates the additional band corresponding to Pertuzumab-SL-T7Nterm (69 kDa). PUREfrex2.0 was used. SP6 RNAP was added in the negative control (T-). In the positive control (T+), DHFR was expressed but the band ran out of the gel due to the low molecular weight.
2) Successful production of Pertuzumab-SL-T7Nterm with PUREfrex2.1
Next, we produced Pertuzumab-SL-T7Nterm using the PUREfrex2.1 kit to promote disulfide bond formation due to non reducing conditions (see PURE system protocol). SDS-PAGE analysis showed the expected band of the protein at 69 kDa (Figure 12).
Figure 12: SDS-PAGE (10% polyacrylamide) analysis of Pertuzumab-SL-T7Nterm visualized by co-translational labeling with GreenLys. The arrowhead indicates the additional band corresponding to Pertuzumab-SL-T7Nterm (69 kDa). PUREfrex2.1 was used. SP6 RNAP was added in the negative control (T-). In the positive control (T+), DHFR was expressed but the band ran out of the gel due to the low molecular weight.
The experiment was repeated a second time with the GreenLys reagent for co-translation labeling of synthesized proteins. The same band was obtained at the expected size (Fig. 13), confirming successful production of Pertuzumab-SL-T7Nterm in PUREfrex2.1.
Figure 13: SDS-PAGE analysis and visualization of expressed proteins by co-translational labeling with GreenLys. Expected band is visible at 69 kDa. PUREfrex2.1 was used. SP6 RNAP was added in the negative control (T-).
Conclusion
We can conclude that the recombinant protein Pertuzumab-SL-T7Nterm can be cell-free expressed in both reducing PUREfrex2.0 and non reducing (disulfide bond promoting) PUREfrex2.1. It is worth mentioning that the sequencing results obtained after we performed these experiments showed a 675-nucleotide deletion within the region of the heavy chain of Pertuzumab. It remains to be investigated whether this deletion originated during the PCR that preceded the cloning or at the sequencing stage.
Due to unsuccessful cloning of the second part of the biosensor, Trastuzumab-SL-T7Cterm, the functionality assays could not be performed. We encourage future iGEM teams to attempt other cloning strategies, pursue the characterization of the biosensor, and to contact us for further details. As a proof-of-concept for biosensing based on the split T7 RNA polymerase in PURE system, we decided to study a rapamycin-induced RNA polymerase. This will be the topic of the section below.
Production and activity test of a rapamycin biosensor based on a split T7 RNA polymerase.
This section presents our results for the production and cell-free characterization of a rapamycin biosensor relying on the inducible complementation of a split T7 RNA polymerase. Specifically, the aim of our experiments was to express in PURE system the two proteins of the biosensor and to demonstrate the functional complementation of the split T7 RNAP induced by a ligand (here rapamycin) as a new tool in cell-free synthetic biology.
![Figure 14: Principle of a rapamycin biosensor based on inducible complementation of the split T7 RNAP. Image from [1].](https://static.igem.wiki/teams/4768/wiki/module-2/rapamycine-article.png "Figure 14: Principle of a rapamycin biosensor based on inducible complementation of the split T7 RNAP. Image from [1].")
Figure 14: Principle of a rapamycin biosensor based on inducible complementation of the split T7 RNAP. Image from [1].
1) Replacement of the promoter and terminator sequences in the initial constructs
The C-term and N-term fragments of the split-T7 RNA polymerase are fused to two antibodies, FKBP and FRB, that recognize two different epitopes of rapamycin through a soluble linker [1] These two genes are carried by two plasmids p0539 and p0579 that were a kind gift from Jerôme Bonnet (CBS, Montpellier, France), see Materials page. For general applicability in PURE system, it is necessary to express the two genes using an orthogonal transcription system to T7, as the latter is by design employed to respond to the presence of rapamycin. We opted for the SP6 RNA polymerase and its SP6 promoter as it showed strong orthogonality in PURE system [2].
First, we performed two successive PCR reactions to replace in both constructs the original promoter and terminator from E. coli by an SP6 promoter and T7 terminator [2]. For the first PCR we used the pair of primers RBS-FRB-F1/T7Nterm-FRB-R1 and RBS-FKBP-F1/R1-rap-FKBP to amplify FRBP- and FRB- containing genes, respectively. The amplification products were checked on agarose gel by EtBr staining (see Protocol page). Figures 15A and B show amplification products at the expected sizes. For the second PCR, we used the primers SP6-F and T7term-rap-R2 for both constructs. The amplification products were checked on agarose gel. Figures 15C and D show that both PCR products had the expected sizes. The linear DNA fragments were used as templates for expression in PURE system.
Figure 15: Analysis of the two-step PCR reaction products with modified transcription elements for expression of the rapamycin biosensor in PURE system. PCR products were analyzed on 0.8% agarose gel and EtBr staining. (A, B) First-step PCR products of the fragments RBS + FRB-T7Nterm + ½ T7 Terminator (A) and RBS + FKBP-T7Cterm + ½ T7 Terminator (B) appear at the expected size of 913 bp and 2573 bp, respectively. (C, D) Fully assembled DNA templates SP6 + FRB-T7Nterm + T7 Terminator (C) and SP6 + FKBP-T7Cterm + T7 Terminator (D) appear at the expected size of 983 bp and 2596 bp, respectively.
2) The two protein parts of the rapamycin biosensor can be expressed in PURE system
Both linear DNA templates were separately expressed with PUREfrex2.1 (see PURE system protocol). This kit was used as it promotes formation of disulfide bonds, which is of general relevance when expressing antibodies. SP6 RNA polymerase was supplied to the reaction mixture to enable constitutive transcription of the two genes. Moreover, GreenLys reagent was supplemented for co-translational incorporation of fluorescent lysine residues, which facilitated the detection of synthesized proteins by SDS-PAGE.
Clear bands corresponding to FRB-T7Nterm (32 kDa) and FKBP-T7Cterm (91 kDa) were obtained (Fig. 16). We can therefore conclude that it is possible to produce the split-T7 RNA polymerase coupled with anti-rapamycin antibodies in PURE system.
Figure 16: SDS-PAGE of the two rapamycin biosensor proteins FKBP-linkerT7Cterm and T7Nterm-linker-FRB expressed in PURE system. 13% acrylamide was used. (A) Pattern of total proteins, i.e., all proteins in PURE system and the additional bands (indicated by arrowheads) are synthetized proteins. (B) Fluorescence image showing only the synthesized proteins that contain fluorescent lysine residues. (C) Overlay of the images in A and B. DHFR used as positive control is visible at 18 kDa (lane 3), FKBP-T7Cterm at 91 kDa (lane 4) and FRB-T7Nterm at 32 kDa (lane 5). The protein marker is in lane 1 and the negative control (no DNA) in lane 2.
3) Preliminary data suggest rapamycin-induced expression of a gene reporter controlled by a T7 promoter
After validating the production of the two subunits of the rapamycin biosensor in PURE system, we decided to test its activity. In presence of rapamycin, the recombined split T7 RNA polymerase should promote expression of the sfgfp gene that is under control of a T7 promoter. Importantly, we used a PUREfrex2.1 custom kit devoid of T7 RNAP (present in Solution II of the regular kit) that would otherwise bypass the effect of the synthesized polymerase (see PURE system Protocol).
Two consecutive PURE system experiments were performed. The two proteins of the biosensor were produced in the same PUREfrex reaction mixture using the SP6 RNA polymerase and no rapamycin. In a second step, a small volume of the two pre-ran reactions supposedly containing the biosensor proteins was added to a fresh PURE solution along with the sfGFP DNA template and 100 μM of rapamycin. A positive control consisting of the expression of sfGFP with a purified, active T7 RNA polymerase (constitutive expression), was used for normalization of the sfGFP signal in the two samples of interest.
Figure 17: Normalized fluorescence intensity of the sfGFP reporter following expression of the rapamycin biosensor. Intensity was normalized using the fluorescence signal of constitutively expressed sfGFP.
Figure 17 shows that sfGFP expression failed, both with or without rapamycin. We suspected that the problem may come from impaired protein folding of either the T7 RNA polymerase subunits or the antibodies, and decided to add the GroE chaperone to enhance functional folding. As additional modifications of the previous protocol, we added rapamycin in the first PURE reaction and increased the production time from 6 hours to overnight incubation. The idea was to leave more time to start forming rapamycin-induced RNA polymerase complexes prior to the second PURE reaction. Figure 18 shows that the normalized sfGFP intensity was higher in the presence of rapamycin. This result suggests that GroE could enhance protein folding, which enables formation of an active rapamycin-responsive biosensor. More experiments will have to be performed with other chaperones and different concentrations of rapamycin.
Figure 18: Normalized fluorescence intensity of the sfGFP reporter following expression of the rapamycin biosensor with an optimized protocol. Intensity values were normalized using the fluorescence signal of constitutively expressed sfGFP with (green) or without (red) rapamycin.
Conclusion
We have designed a rapamycin biosensor with transcriptional elements that are compatible with expression in PURE system. Cell-free production of the two complementary biosensor proteins was demonstrated. Preliminary experiments suggest that rapamycin-induced formation of an active RNA polymerase from two split fragments is possible when the chaperone GroE assists protein folding. It should be noted that the increase of the sfGFP reporter signal with addition of rapamycin was modest (from 3 to 7%) and further experiments will have to be repeated to confirm these results. If the benefit of adding chaperones is supported, this implies that GroE or other chaperones (e.g., DnaK) might be included in the molecular composition of our therapeutic liposomes for enhanced activity.
We have successfully demonstrated the functionality of our 2-HG biosensor system in bulk, both the repression step by DhdR and the inducible role of 2-HG. This result is invaluable toward the development of 2-HG-responsive therapeutic liposomes.
We also proved that it is possible to cell-free express in both reducing and non reducing (disulfide bond promoting) conditions a split-T7 RNA polymerase linked to an antibody for ligand-induced protein expression. The inherent limitation of the functional rapamycin biosensor is that it does not respond to cancer biomarkers and can only serve as a proof-of-concept system. Nonetheless, we learnt from this system that adding chaperones during gene expression may enhance functionality, something we should implement with our HER biosensing module. We encourage future iGEM teams to use our experimental results to address the difficulties encountered, especially during cloning, and succeed in producing the anti-HER2 split T7 RNAP biosensor.
References
- [1] Pu, J., Zinkus-Boltz, J., & Dickinson, B. B. 2017, Evolution of a split RNA polymerase as a versatile biosensor platform. Nat Chem Biol, 13(4), 432-438.
- [2] Duco Blanken, David Foschepoth, Adriana Calaça Serrão and Christophe Danelon. 2020. Genetically controlled membrane synthesis in liposomes. Nature Communications 11, 4317.
