Inspired by immune system, our goal is to fablicate pore forming DNA nanostructure killing the cancer cell. The designed structure is shown below. Our design is characterized with three features: a broad plane part to anchor on the cell surface, bend of side edges to invade into the cell membrane, and connectable sites for oligomization.
It requires enough free energy to penetrate membranes. This is compensated by the free energy gained from biding of the DNA nanostructure bound cholresterols to the lipid bilayer. The more cholesterols are equipped to the structure, the more stabilized it stays near the membrane. A broad plane gives much cholesterol binding sites, so this feature might be suitable to penetrate membranes. Although previous reseach (Danilo D. Lasic, et al.) reported the non-specific binding of DNA to the liposome, which is usually an undesiable feature, we thought that we could utilized this feature positively by means of the broad plane. Integrating broad plane into our DNA nanostructure, we expect some free energy gain by the non-specific binding of the broad plane to the bilayer, which may stabilize the binding of our DNA structure to the bilayer.
To achieve an efficient pore forming system, one should balance between the cost of penetration of the nanostructure into the bilayer and the binding stability of the nanostructure to the bilayer. To overcome this dilemma, we introduced the "bend of side edges", which allow us to minimizing the penetration part, but also maximizing the anchoring part.
To oligomerize, the DNA nanostructure must have some binding site to each other. Therefore, we introduced "connectable sites" into our nanostructure. It should be noted that hybridization is used for oligomerization method in the figure, we examine other method as well.
Although we thought carefully, above design are mainly derived from speculation. We, therefore, need more information (e.g. interaction between DNA and liposome) to design our nanostructure more specifically. Thus, ss a first step toward our goal, we started with simple DNA origami structure: Cylinder in Barrel.
This barrel structure was designed to get some feedback for our final design of “Origomeric Cell Killer (hereafter “OCK")”. CaDNAno (version 2.2) was used to design the structure, and M13mp18 was chosen as the scaffold strand.
OCK needs to be oligomerized and be formed pore on the membrane, so we check following things by Cylinder.
1) Can DNA nanostructures penetrate lipid membranes?
2) Can DNA nanostructures bind each other and make dimer (or more complex structure ) in solution?
3) Can DNA nanostructures dimerize on membrane?
4) Can the direction of connection be controllable?
Therefore, the structure was equipped with following features.
To get reliable information, the design of cylinder needs to be simple and realistic. We designed our Cylinder in Barrel according to Martin Langecker et al.
The cylinder domain is about 65 nm long (195 bp) and consists of six dsDNA helixes, so its diameter is 6 nm. The barrel domain is approximately 44 nm long (128 bp) and 48 helixes build this domain The thickness of bilayer is 2 nm, therefore, our cylinder (about 20 nm penetration part) is enough long to stick into the bilayer. By connecting the cylinder with barrel, our cylinder in barrel structure can be integrated more cholesterols than that without barrels.
(Cholesterol is necessary to penetrate membranes, about which is written in next section.)
Because DNA has negative charge, the DNA nanorobots have to gain some energy to penetrate lipid membrane, which is composed of amphiphilic molecules. This problem is solved by binding cholesterols to the structures. The barrel domain has 26 staple strands complementary to cholesterol modified DNA oligos, and modified oligos are hybridized to these staples. The DNA structure anchors itself to membrane by cholesterols, and it gives stability for the structure to stay near membranes. Therefore, the structure can pierce lipid bilayer.
To hybridize cholesterol modified oligos, three different sequences of staples were prepared.
1:CCTCTCACCCACCATTCATC (Alexander johnson-Buck et al.)
2:TAACAGGATTAGCAGAGCGAGG (Martin Langecker et al.)
These are different in the length and in sequence. The sequence of cholesterol modified oligos are perfectly complementary to the above three sequences, and their 5' ends are modified with a cholesterol.
To achieve oligomerization, the OCK has binding site by means of hybridization. Two pairs of sequences were used for hybridization. Both sequences are according to previous work (Shawn M. Douglas, et al. ). These are derived from aptamer sequence, one is TE17, the other is sgc8c (and the complementary strands to these, so two pairs). These sequences were chosen because it is considered that these sequences don't prevent the folding of scaffold. The sequences are:
Designed structures have either the aptamer sequences (1, 2) or the two complementary strands (1, 2). When two structures with different pairs are mixed and hybridization happens, these structures hence bind each other through two binding sites. Two types of binding sites were designed to test whether the direction of connection can be controlled. One type of binding site uses staples sticking out from the ends of scaffolds, the other uses staples from side of OCK.
To detect the cylinder piercing membrane, biotin modified staple strands (biotin-oligo) are attached at the bottom of OCK and hybridized with complementally strands labeled with fluorescence. Strept-avidins (SA) encapsulated in the liposome could bind to the OCK, and only the biotin-oligo penetrating liposome could bind to the SA, which can be detected by the gel shift assay.
From the results of experiments with “Cylinder” we will decide our final design.