Experimental approach

We have designed an universal membrane platform which is inherently modular and which can elicit both slow as well as fast responses. To verify the viability of the platform, we devised characterizing the fast cellular responses (see Figure 1). In the construction of our device, we planned on conducting a wide range of experiments to verify whether the individual elements of our system worked. These elements include whether the click reaction occurs, verifying whether the signaling components work, whether proximity invokes a response and whether the DBCO-modified aptamers work. An overview of the experiments is given below.

Figure 1: As a proof of concept, we will construct and express our device within E.coli BL21DE3. The device we test features the fast cellular response signaling components. We will verify the construction of our device by conducting numerous experiments. These experiments include verification of the click reaction, of the individual signaling components, whether proximity invokes a response and if the aptamers work.
The BL21DE3 cells will be cotransformed with two plasmids. The first plasmid, pET-Duet1 (blue), carries the genes for the outer membrane proteins and expression is triggered by the addition of IPTG. The second plasmid, pEVOL pAzF (red), is necessary for the incorporation of the non-natural amino acid within the outer membrane proteins.

The Vectors

As shown in Figure 1, our system relies on the presence of two vectors within the host cell. The first of these vectors is pEVOL-pAzF. The second one is the pETDuet-1 expression vector.

pEVOL-pAzF

pEVOL is a small vector which has been designed and optimized for the incorporation of unnatural amino acids into proteins in E.coli. The coding sequence of pEVOL encodes tRNA synthetases, which can translate the amber stop codon sequence into the incorporation of the non-natural amino acid. Optimization of the vector has enabled higher yields of mutant proteins in comparison to previous vectors: pEVOL showed roughly 250% greater yields in comparison with vectors previously used for the incorporation of non-natural amino acids [1]. One of the first amino acids which has been incorporated into proteins using the relatively novel pEVOL vector was pAzF and we will use this exact vector to construct our mutant protein in vivo.

Features of the pEVOL vector:

The pEVOL expression vector features the p15A origin of replication which makes the pEVOL plasmid compatible
Plasmid compatibility is generally defined as the failure of two coresident plasmids to be stably inherited in the absence of external selection [2]. The cause of the failure to be stably co-inherited lies in the fact that the origins of replication are too analogous. In that case, the bacteria cannot distinguish between the plasmids and can eventually lose either one of the plasmids as the amount of both plasmids is limited by a single copy number.
with many other frequently used plasmids.
The chloramphenicol resistance gene
The tRNA synthetase is under the control of the arabinose-inducible AraBAD promotor

Figure 2: Simplified vectormap of the pAzF

pETDuet-1

The designed system relies on the two membrane proteins which come into close proximity as a result of ligand binding. Often, when such protein assemblies are to be obtained, one can isolate endogenous complexes and reconstitute those components in vitro to analyze whether the assembly takes place [3]. As we have designed the system to be used in vivo, however, we rely on the co-expression of all components within the same host cell.
Generally, this heterologous expression can be reached into two different ways, firstly by transforming multiple constructs and secondly by introducing a single plasmid carrying multiple genes in E.coli. As our device already featured two plasmids, we devised to use a plasmid which could co-express multiple genes, preferably in an equimolar ratio. Since the pETDuet-1^TM Expression System from Novagen has been developed for this particular purpose, we have chosen to use this system as our vector of choice.

Features of the pETDuet-1 expression vector:

The pET-Duet1 vector features two multiple cloning sites, each carrying a dozen cloning sites. This enables insertion of multiple fragments.
Each of the multiple cloning sites contains its own T7 lac promotor and a ribosome binding site. A single terminator is located after MCS2, such that transcription yields two different mRNAs.
The pET-Duet1 vector not only carries multiple genes but is bicistronic. Therefore, it allows simultaneous expression of two proteins separately from the same RNA transcript. Hence, both the MCS1 insert and MCS2 insert are expressed from the longer mRNA strand. Only the MCS2 insert is expressed from the shorter mRNA strand.
Figure A: Transcription of pETDuet-1 yields two different messenger RNA strands. The longer strain is bicistronic and encodes both the MCS1 as well as the MCS2 insert. The smaller strain only encodes MCS2. The smaller strain compensates for the reduced expression of MCS2 on the longer strand.

Figure 3: General overview of the pETDuet-1 expression vector. The vector carries the Ampicillin resistance gene and carries the pBR332 origin of replication

Usually, bicistronic vectors containing two target genes under the control of a single promotor preceding the two genes show strongly reduced expression of the gene located more distant from the promotor site [4]. The second promotor which initiates the translation of the second mRNA aims to correct for this reduced expression.

Verifying the click reaction

A vital aspect of our device is clicking the aptamers to the membrane proteins. For this click, we made use of the exact same click chemistry used by iGEM TU Eindhoven 2014. The click reaction was used N-terminally by iGEM TU Eindhoven 2014, in order to minimize sterical strain. To analyze whether the localization of the azide-functionalized amino acid within the loops of OmpX impedes the click reaction, we clicked a DBCO-functionalized fluorophore (TAMRA) to the outer membrane proteins. After some washing steps and spinning down, we expected the cells to remain fluorescent.

Figure 4: To verify whether the click reaction has occured, we incubate the cells with DBCO-functionalized TAMRA. If the outer membrane protein is functionalized with the non-natural amino acid, this TAMRA dye binds to the membrane proteins covalently. In that case, the cells will remain fluorescent after a few washing steps.

Since the fluorescence of the TAMRA-dye falls within the visible spectrum, we expected to see whether the dye clicked on the outer membrane proteins with the naked eye. To analyze the fluorescence at the single-cell level, we measured cells using the Fluorescence-Activated Cell Sorter (FACS)

.

A Fluorescence-Activated Cell Sorter (FACS) is a specialized flow cytometer (see Figure B). The FACS can provide information about cell size, complexity and fluorescence. The relative cell complexity is measured using size scatter (SSC). The relative cell size is measured using forward scatter (FSC). The fluorescence can be measured by using a wide range of filters. These cell characteristics can be combined to sort cells.

Measuring bioluminiscence & fluorescence

DNA Strand displacement

As we have already touched upon lightly, it has long been thought that nucleic acids had only a single role: carrying hereditary information. As discussed, the discovery that DNA could fold into higher-order structures gave way to SELEX, an evolutionary method of discovering aptamers. The construction of DNA nanostructures, however, was not only carried out through a combinatorial approach: the specificity and predictability of Watson-Crick basepairing enabled rational design for engineering at the nanoscale [5]. Initially, the designed DNA nanostructures were mostly static, but dynamic nanostructures have become available over the years. Most of these dynamic nanostructures share a common feature: they exploit a biophysical phenomenon known as DNA strand-displacement.

DNA Strand displacement is the workhorse of dynamic DNA technology, the field which uses DNA’s non-covalent interactions to assemble higher-order structures. DNA strand displacement is a process where two strands with partial complementarity are hybridized. These pre-assembled DNA strands have only partial complementary, leaving room for a toehold region. When a DNA sequence with full complementary binds to this region (the input), branch migration takes place (see the figure below). In the end, the sequence with full complementary binds to the sequence, yielding the output.

Figure C: Schematic overview of DNA strand displacement. DNA Strand displacement is initiated when the input binds to the toehold region of a partially hybridized DNA strand. After the initiation, branch migration takes place and the input fully hybridizes with its perfect match, yielding the output.

DNA strand-displacement is a very robust technology: the sequences of the used strands often go unreported as they play a minor role. The robustness of the technology enabled rational design of numerous nanostructures. These nanostructures include DNA walkers, strand displacement cascades and self-assembling dendrimers [5].

In our project, we want to bring two membrane proteins in close proximity through aptamers. To test whether a close proximity indeed triggers an intracellular signal, we used DNA to bring the membrane proteins in close proximity. DNA is the ideal probe for this purpose, as its high specificity and predictability allowed us to bring the membrane proteins in vicinity. To make the system with DNA reversible, we designed a simple system exploiting the strand displacement (see Figure 3).

Figure 5: The membrane proteins can be brought in close proximity by clicking oligonucleotides on the loops and adding a long strand complementary to both oligonucleotides. It is expected that this results in a measurable signal. The longer strand is functionalized with a toehold region. Upon addition of a strand perfectly complementary to the longer strand, the system disassembles and the signal will fade out.

[1] T. S. Young, I. Ahmad, J. A. Yin, and P. G. Schultz, “An enhanced system for unnatural amino acid mutagenesis in E. coli.,” J. Mol. Biol., vol. 395, no. 2, pp. 361–74, Jan. 2010.
[2]R. P. Novick, “Plasmid incompatibility.,” Microbiol. Rev., vol. 51, no. 4, pp. 381–95, Dec. 1987.
[3]D. Busso et al., “Expression of protein complexes using multiple Escherichia coli protein co-expression systems: a benchmarking study.,” J. Struct. Biol., vol. 175, no. 2, pp. 159–70, Aug. 2011.
[4]J. M. Glück, S. Hoffmann, B. W. Koenig, and D. Willbold, “Single vector system for efficient N-myristoylation of recombinant proteins in E. coli.,” PLoS One, vol. 5, no. 4, p. e10081, Jan. 2010.
[5]D. Y. Zhang and G. Seelig, “Dynamic DNA nanotechnology using strand-displacement reactions.,” Nat. Chem., vol. 3, no. 2, pp. 103–13, Mar. 2011.