INTRODUCTION
Throughout the development of our project, we have faced multiple obstacles that have forced us to look for other ways to achieve our objectives. We have had to go through various troubleshooting processes that, satisfyingly, have boosted our learning in several concepts and techniques. While doing so, we have performed a series of engineering cycles.
Assembly strategy
The first strategy that we conceived for the assembly of our constructs was based on the Golden Gate technique. For that, we designed our constructs as multiple fragments with sticky ends generated within a Golden Gate reaction to be assembled into the recipient pJUMP29-1A KanR Level 1 backbone. In order to save time and simplify the reactions, all the constructs were designed to be assembled in a level 1 backbone, avoiding the need of cloning them on level 0 backbones or further cloning them into level 2 ones (for more information see Contribution wiki). Furthermore, we designed the fragments in a modular fashion to take advantage and use the same common fragment for multiple constructions (CMV promoter, SV40 polyA). Once we tested that the assembly of each of our constructs was successful in silico (analysis performed with benchling.com), we sent the fragments to be built by iGEM sponsors IDT and Twist Bioscience. Once in the wet lab, the experimental tests did not work as we expected: we only achieved the assembly of the constructs formed by a low number of fragments (1-2 fragments), whereas the assemblies of the ones formed by more fragments (4-6 fragments) were not achieved.
In that context, we have performed several engineering cycles to determine the proper Golden Gate conditions for each assembly. We tried out several conditions for the reaction, modifying both the number of cycles, the duration of each temperature, the moment of introducing each enzyme and the duration of the last ligation step. We even tried to cut and purify the open backbone beforehand to increase the ligation of the fragments. With that many experiments performed we learned that the proper conditions for our constructs depend on the number of fragments to be assembled in a single Golden Gate reaction.
However, after we achieved the ligation of most of our constructs, we made an awful discovery that explained the bad efficiency of our Golden Gate assemblies. We were performing the assemblies using BsmBI, which is in the reverse orientation! Instead, we should have used the BsaI sites, as they have a correct orientation that allows the ligation of the fragments while hindering the entry of the removed part of the backbone. The source of this problem in the design, apart from our inexpertise and lack of contact with experts, resides in the backbone's map sequence in the Part Registry. This sequence is incomplete! It lacks the BsaI sites and the GFP in between the main assembly site. Sadly, we did not realize this until when, upon request, we obtained the correct map sequence of the backbones.
It was too late, but we re-designed our unsuccessfully assembled constructs to be divided into fragments for a BsaI assembly (for example BBa_K4501021).
In parallel, given the technical difficulties for the assembly we resolved to re-design our assembly strategy and try the fusion PCR technique. For that, we asked Dr. Fernando Setién, an expert on the field, for guidance and general guidelines. Once we came up with the desired strategy, we designed several pairs of primers that we used to amplify and add the adapters required to the gene fragments. Once more, the test did not work and we tried several conditions for the fusion PCR, but none of them worked (see more on Results).
Curiously, at the end the constructs were successfully assembled by Golden Gate technique, so the fusion PCR strategy was discontinued.
pJUMP-based Lentiviral Transfer Plasmid
Our first design of the pJUMP-based transfer plasmid had some important limitations that restricted the potential versatility and amenability of the device. Namely, the 5’ module had some BsmBI and BsaI recognition sites that impeded the assembly on the main site. For that reason, the device was first conceived to be cloned sequentially: the 5’ module was to be cloned after the main assembly, which extended the procedure and made it more complicated and expensive. When we built and tested the device, we faced assembly difficulties and we realized that the pJUMP-based transfer plasmid was not optimal at all.
For that reason, we decided to re-design the 5’ and 3’ modules so they didn't incorporate BsmBI nor BsaI recognition sites. That way, we would be able to clone both of them previous to the main assembly reaction, shortening the procedure and reducing costs. Furthermore, we decided to incorporate a SV40 ori of replication downstream of the 3’ module, supplying the backbone with a powerful ori that could increase the transfection efficiency onto the HEK293T cell line.
Validation of anti-myc shRNAs
One of our greatest engineering successes is the design of our own shRNA against myc. Using the VectorBuilder Wizard for shRNA generation, we designed a shRNA against the 1741-1761 region of myc, for which we constructed an expression cassette coupled to a GFP reporter. To that shRNA (BBa_K4501018), we made two different variants that incorporated the C/D-box label for exosome loading: one containing the C/D-box after the 3’ end of shRNA (BBa_K4501019) and another incorporating the C/D-box in substitution of the traditional loop (BBa_K4501020) (Figure 3).
We managed to assemble all the constructs into a pJUMP backbone by Golden Gate assembly, and we performed a qPCR to assess myc expression (see Results webpage). In that, we obtained an almost 2-fold reduction of myc expression in Ramos cells treated with the shRNA, indicating its functionality and the success of the design.
His-tag-based affinity chromatography purification of exosomes
We also succeeded in developing our own way of purifying exosomes with only instruments that can be found in a normal laboratory. By using homemade column affinity chromatography with HisPurTM Ni-NTA resin (ThermoFisher, Batch protocol), we achieved purification (though with a low yield) of His-tagged exosomes (see Results webpage). However, further improvement of the system (p.e. stably transfecting with lentiviral vectors) could increase the yield of exosome purification.
