Engineering success
Synthetic biology aims at creating and assembling novel biological components, mechanisms, and structures, along with modifying pre-existing natural biological systems for practical use (Nature, 2021). Just like in other branches of engineering, users of synthetic biology adhere to a cyclical approach known as the DBTL cycle (Design, Build, Test, Learn), which is essential for the effective development of novel or improved biological systems.
In CALIPSO, we made use of synthetic biology to engineer next-generation liposomes capable of specifically producing and delivering an anticancer drug next to cancer cells. Building the biosensing module has been at the core of our engineering approach, which we will present on this page. You can find the detailed design of CALIPSO on the design page, and the detailed experimental results on the Module 2: Biosensing systems page.
Cycle I: Constructing a HER2 biosensor with the split T7 RNA polymerase
We decided to engineer a heterodimer consisting of the split-T7 RNA polymerase (two subunits), coupled to anti-HER2 antibodies via a single transmembrane domain. The binding of the two antibodies to distinct epitopes of HER2 not only provides a means to anchor the liposome to the surface of cancer cells, it also enables complementation of the two split T7 RNA polymerase subunits that are facing the opposite side of the membrane. The functionally assembled T7 RNA polymerase can then trigger the expression of the thymidine phosphorylase gene internally.
DESIGN
When we decided to use the split-T7 RNA polymerase as a biosensor, the main challenge appeared to design a chimeric heterodimer harboring two functionalities: an HER2-binding element and a transcriptional activator. Importantly, the two anti-HER2 antibodies had to recognize different epitopes of the receptor, and binding to the target should enable complementation of a functional RNA polymerase, which imposes strong design constraints.
We opted to construct two plasmids, each containing a different antibody against HER2 fused to one of the two fragments of the split-T7 RNA polymerase. Using soluble linkers, Trastuzumab was linked to the C-terminal part of the polymerase (Trastuzumab-SL-T7Cterm), corresponding to the part BBa_K4768006, while Pertuzumab was linked to the N-terminal part (Pertuzumab-SL-T7Nterm), corresponding to the part BBa_K4768005.
part BBa_K4768006
part BBa_K4768005
BUILD & TEST
We first tried to construct them by In-Fusion molecular cloning using p0539_Trastuzumab-SL-T7Cterm and p0579_Pertuzumab-SL-T7Nterm with native p0539 and p0579 plasmids, and the two antibody sequences ordered from IDT.
- - For p0539_Trastuzumab-SL-T7Cterm, two inserts had to be cloned into plasmid p0539 by In-Fusion to obtain Trastuzumab linked to the C-term fragment of the T7 polymerase. A PCR was first performed to amplify the two gBlocks and linearize the plasmid p0539. The two amplified gBlocks were then cloned by In-fusion for simultaneous insertion into the linearised plasmid. After transformation and colony screening, no recombinant clones were obtained.
- - For p0579_Pertuzumab-SL-T7Nterm, PCR was performed to amplify the gBlock and linearise the native p0579 plasmid. In-Fusion was then carried out to assemble the insert and the linearized plasmid. After transformation and colony screening, no recombinant clones were obtained.
For both constructs, the In-Fusion reaction was attempted again, this time by considering the following aspects:
- - For Trastuzumab-SL-T7Cterm, we hypothesized that the primers used for the PCR screening did not work due to unspecified hybridation.
- - For Pertuzumab-SL-Nterm, as the homology regions respected all recommendations from Takara, we hypothesized that the size of the linearized vector was relatively small in comparison with that of the insert (respectively 2391 bp and 2087 bp). This could have led to a poor assembly efficiency.
Therefore, we switched the plasmid vector used for In-Fusion to pET21:
- - For pET21_Trastuzumab-SL-T7Cterm, the In-Fusion step with the two gBlocks failed again.
- - For pET21_Pertuzumab-SL-NtermT7, the In-Fusion assembly was successful.
New sets of primers were ordered to allow the In-Fusion assembly with pET21 for both constructs.
For production in PURE system, we do not necessarily need a plasmid. Linear DNA can be used as a template for gene expression. Hence, we directly amplified the gBlock for Pertuzumab-SL-T7Nterm and used the PCR product as a template for expression in PURE system. Protein gel analysis suggested that the full-length protein was successfully produced.
Figure 1: SDS-PAGE (10% polyacrylamide gel) of Pertuzumab-linker-T7Nterm reaction tube and visualization by Instant Blue coloration. The expected band is visible at 69 kDa. PUREfrex 2.0 was used. SP6 RNAP was added in the negative control (T-) and Reaction.
As an alternative to In-Fusion, we attempted to assemble the two inserts for Trastuzumab-Cterm by overlap-PCR. No complete PCR product was obtained.
LEARN
We conclude from these experiments that, despite the structural complexity of such chimeric proteins, it was possible to produce one of the two two-partite RNA polymerase-linked antibody proteins with PURE system. However, we repeatedly failed to generate the second construct, which prevented us from running functional assays. We therefore had to change our strategy to demonstrate the feasibility of using the split-T7 RNA polymerase for biosensing with PURE system.
Cycle II: Constructing a rapamycin biosensor based on the split T7 RNA polymerase
DESIGN
Inspired by a previous study, where the split-T7 RNA polymerase was fused to anti-rapamycin antibodies for rapamycin-induced gene expression (Bryan C. Dickinson et. al), we decided to modify the reported constructs for compatibility with PURE system.
We designed two new parts with FKBP and FRB anti-rapamycin antibodies, respectively linked to the C-terminal (FKBP-SL-T7)Cterm and N-terminal (T7Nterm-SL-FRB) parts of the split-T7 RNA polymerase. Both constructs have optimized regulatory sequences for expression in PURE system. Moreover, the two genes are under control of an SP6 promoter to allow for orthogonal transcriptional regulation of the biosensing and reporting genes. We ordered the corresponding transcription cassettes as gBlocks. See the parts BBa_K4768009 and BBa_K4768010.
part BBa_K4768009
part BBa_K4768010
BUILD & TEST
We performed a two-step PCR on the two gBlocks, and used the products as templates for expressing the two recombinant proteins in PURE system. To enable functional expression of disulfide bond-containing proteins, as it is so for the majority of antibodies, we chose to use the PUREfrex2.1 kit. Co-translational labeling with Green-Lys, followed by SDS-PAGE analysis indicated successful biosynthesis of FKBP-SL-T7Cterm and T7Nterm-SL-FRB. Activity assays using a T7-promoter-based reporter gene suggested that, in the presence of the chaperone GroE during transcription-translation, our biosensor is responsive to rapamycin (see Module 2 : Biosensing systems).
Figure 2: SDS-page migration of FKBP-SL-T7Cterm and T7Nterm-SL-FRB on 13% acrylamide gel, stain-free and Pro-Q Emerald green revelations merged together. DhfR is visible at 18 kDa (lane 2), FKBP-SL-T7Cterm at 91 kDa (lane 3) and T7Nterm-SL-FRB at 32 kDa (lane 4).
Figure 3: Normalized fluorescence intensity of the anti-rapamycin biosensor when the biosensor was produced in presence of GroE. PUREfrex 2.1 custom kit was used. Fluorescence was measured at 513 nm. The fluorescence with rapamycin was normalized using the sfGFP signal for a native (full-length) T7 RNAP and with rapamycin. The fluorescence without rapamycin was normalized using the sfGFP signal for a native T7 RNAP and without rapamycin.
LEARN
We conclude that it is possible to produce a split-T7 RNA polymerase linked to an antibody for ligand-induced protein expression in PURE system. However, the inherent limitation of the rapamycin biosensor is that it does not respond to cancer biomarkers and can only serve as a proof-of-concept system. Therefore, we next focused on the development of a biosensor that responds to an oncometabolite that is found at higher concentrations in the microenvironment of cancer cells.
Cycle III: Implementation of a biosensor capable of detecting physiological concentrations of 2-hydroxyglutarate
Driven by the objective to trigger the biosynthesis of the thymidine phosphorylase only in the tumor’s vicinity, we engineered a biosensing strategy based upon the detection of the oncometabolite 2-hydroxyglutarate (2-HG), which is recognized by the transcriptional repressor DhdR from Achromobacter denitrificans.
DESIGN
We designed two plasmids: one for the expression and purification of DhdR, the other containing the DhdR binding site (dhdO) upstream of the gene encoding for the reporter sfGFP. The plasmid harboring the gene encoding for the Thymidine phosphorylase (tymp) in place of sfGFP was also prepared but is not described on this page (see Cloning results, Module 3: Anticancer drug in situ production and see the parts BBa_K4768000 and BBa_K4768001).
part BBa_K4768000
part BBa_K4768001
BUILD
DhdR was produced and purified after transformation of E. coli BL21 strain. Plasmids containing sfgfp under a T7-dhdO inducible promoter were successfully cloned and minipreped from Stellar competent cells.
TEST
- 1. We first proved that DhdR is capable of binding to dhdO, which inhibits expression of sfGFP in PURE system.
- 2. We determined the minimal concentration of DhdR required for strong repression. The tested concentrations were derived from the kinetical model, which was recalibrated afterward.
- 3. We showed that 2-HG de-represses transcription of DhdR-bound DNA in a concentration-dependent manner, and that our biosensor is sensitive to physiological concentrations of 2-HG.
- 4. We finally sought to demonstrate that liposomes loaded with the 2-HG biosensor could interact with 2-HG-producing cancer-cells cultured in the laboratory. Specifically, we wanted to know if our liposomes could anchor to tumor cells and express sfGFP at the same time. Liposomes functionalized with anti-HER2 nanobodies and folate molecules for specific anchoring were prepared. Preliminary results showed colocalization of GFP-producing liposomes with cultured Caco2 cells (Best Measurement page).
Figure 4: 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.
Figure 5: Fluorescence microscopy images of anti-HER2-nb- and folate-decorated liposomes on top of Caco2 cells (sample 2). a) Imaging of Caco2 cells in brightfield. b) Imaging of a liposome localized with a fluorescent membrane dye. c) Imaging of the same liposome in the GFP channel.
LEARN
This engineering cycle allowed us to create a functional 2-HG biosensor based on the DhdR transcriptional repressor in PURE system. This biosensor was capable of responding to physiological concentrations of the oncometabolite in bulk reactions and could be loaded in liposomes targeting cancer cells. In the course of these experiments, we also showed that repression was stronger when using a linear DNA template vs. a plasmid and that 2-HG can partly inhibit gene expression with PURE system.
These three engineering cycles allowed us to design, build and test two different biosensing systems. We learned that it is possible to produce a split-T7 RNA polymerase linked to an antibody for ligand-induced protein expression in PURE system. We obtained promising results concerning the functionality of the two-partite RNA polymerase-linked antibody anti-rapamycin we produced in PURE system. The next challenge is thus to produce a biosensing system based on the split-T7 RNA polymerase capable of responding to cancer biomarkers. The final cycle allowed us to obtain a 2-HG biosensor sensitive to physiological concentrations of 2-HG. We went as far as testing our 2-hydroxyglutarate biosensor in liposomes anchored to cancer cells.
Although further experiments need to be performed to validate the full spectrum of the liposomes' activity, the outcome of this engineering process constitutes a major step towards the development of intelligent anticancer liposomes.
