Programming chain-growth copolymerization of DNA hairpin tiles for in-vitro hierarchical supramolecular organization

Formation of biological filaments via intracellular supramolecular polymerization of proteins or protein/nucleic acid complexes is under programmable and spatiotemporal control to maintain cellular and genomic integrity. Here we devise a bioinspired, catassembly-like isothermal chain-growth approach to copolymerize DNA hairpin tiles (DHTs) into nanofilaments with desirable composition, chain length and function. By designing metastable DNA hairpins with shape-defining intramolecular hydrogen bonds, we generate two types of DHT monomers for copolymerization with high cooperativity and low dispersity indexes. Quantitative single-molecule dissection methods reveal that catalytic opening of a DHT motif harbouring a toehold triggers successive branch migration, which autonomously propagates to form copolymers with alternate tile units. We find that these shape-defined supramolecular nanostructures become substrates for efficient endocytosis by living mammalian cells in a stiffness-dependent manner. Hence, this catassembly-like in-vitro reconstruction approach provides clues for understanding structure-function relationship of biological filaments under physiological and pathological conditions.


Supplementary
were calculated from the AFM data. Length (L SAXS ) were calculated from the SAXS data.

I' CATGAGTGATGTAGTCTCGTCGGAGTCGTATAAGACTTACTCGTGACAGTCT
Supplementary In single-molecule fluorescence imaging experiment, we designed the positions of fluorescent dyes Cy3 and Cy5, and inter-dye distance in the copolymeric one-dimensional nanofilaments to minimize the occurrence of FRET between dyes. The efficiency of FRET is extremely sensitive to the separation distance between dyes. The range over which the energy transfer can take place is limited to approximately 10 nm. As shown in Supplementary Figure 9, fluorescent dyes Cy3 and Cy5 were labeled on the centres of monomers A and B through binding to the 5' ends of a1 and b1, respectively. In the copolymerized nanofilaments, the rigid strands between labeled dyes Cy3 and Cy5 is about 16.0 nm (47 bp) in length, therefore the FRET efficiency between Cy3 and Cy5 could not take place.

Supplementary Note 2 | Fluorescence analysis of the dependence of cellular uptake on cell lines Hela and MCF-7.
To check if there is any cell-type effect for our DNA structures, here we provide the internalization data of other two cell lines (Hela and MCF-7) incubated with DNA samples for 12 h (see Supplementary Figure 15 10 ) were used for endocytosis in Hela and MCF-7 cell lines, respectively. Although the internalization by Hela or MCF-7 was not so efficient as A549 cells, these results illustrated the similar increasing trends of uptake efficiency along with the shape and compliance of DNA structures. The results support our conclusion that the endocytosis of DNA nanofilaments we synthesized is structural and mechanical stiffness dependent, and no significant cell type effect found in the three cell lines we used.

Supplementary Note 3 | Nocodazole inhibition on cellular uptake of DHT nanofilaments.
Nocodazole, macropinocytosis inhibitor, was employed to verify if the internalization of one dimensional DHT nanofilaments was through the macropinocytosis pathway. The effects of nocodazole inhibition on cellular uptake efficiency of nanofilaments under FBS free incubation condition measured by flow cytometry and confocal images showing the distribution of ten pairs (I-[A-B] 10 ) DHT nanofilaments in cells pretreated with nocodazole or not. As shown in Supplementary Figure 17, results showed that all assembled DNA nanostructures presented a remarkable decrease (50-81%) in cellular internalization. Distribution of ten pairs (I-[A-B] 10 ) DHT nanofilaments before (in the plasma) and after (on the surface) nocodazole treatment was recorded by confocal microscopy, and results showed that nocodazole can significantly inhibit the internalization of the nanofilament.

Supplementary Note 4 | Calculation of the relative moment of inertia (I) for DHT nanofilament.
As shown in Supplementary Figure 18, we assume DHT nanofilament as a bundle of two rigidly linked DNA helices of radius r and no gaps between the helices. The moment of inertia of each helix with respect to its own center of mass is i, is displaced from the nanofilament's center of mass by a distance R. The Young's modulus of the nanofilament is the same as that of dsDNA. R = 1.5 nm, r = 1 nm.
Bend axis 1: Bend axis 2: Supplementary Note 5 | Calculation of the relative compliance of DNA structures.
The mechanical compliance of DNA structures is the inverse of stiffness. The bending stiffness (k b ) of beam-shaped structure: where E is the Young's modulus (elastic modulus), L is the length of the DNA nanostructure (calculated from theoretical values or AFM experimental values as shown in Table 2, L n ), I is the area moment of inertia.
To estimate the moment of inertia, we treat a nanofilament as a bundle of N rigidly linked cylindrical rods of radius r, I could be calculated in terms of i 1,2 . The moment of inertia of each dsDNA helix with respect to its own center of mass is i, is displaced from the nanofiament's center of mass by a distance R (Supplementary Figure 18 and Note 4). By the parallel axis theorem, the moment of inertia of the dsDNA helix with respect to the nanofiament's center of mass is i+MR 2 , where M is the mass of the helix. Assuming uniform density, = ! ! ! (4) and thus As shown in Supplementary Table 5, Cytochalasin D, which disrupts F-actin filaments via actin depolymerization, was used to identify whether phagocytosis or macropinocytosis pathway was included in the endocytosis process. Dansylcadaverine, which blocks the formation of coated pits by inhibiting transglutaminase in the cell membrane, was used to block clathrin-mediated endocytosis. Genistein, a tyrosine kinase inhibitor that inhibits actin recruitment in caveolae, was used to discern the role of caveolae-mediated endcytosis in internalization.