Figure 1

DNA tile used in this study. (a) Structural schematic showing the intertwining of the five DNA strands. Five unpaired bases at both ends of the rightmost (orange) and leftmost (blue) strands are sticky ends, responsible for directing tile assembly. (b) Cartoon clarifying interstrand base pairing. (c) Local arrangement of tiles into crystals with paired sticky end sequences explicit. (d) Global geometry of the intrinsically curved tile lattice that forms hollow tubes with between four and ten tiles around [13]. (e) Complete sequences of the five DNA oligonucleotides. Bases that participate in sticky ends are boldface.Reuse & Permissions

Figure 2

Nanotube assembly from preformed tiles following a quench ($45°\mathrm{C}\to 22°\mathrm{C}$). At various times after the quench, a small amount of solution was diluted to 40 nM and deposited on bare glass [17]. Adhesion to the glass stopped the assembly reaction and facilitated measurement of short nanotubes. (a) After 2 min, small aggregates appear. At 10 min, tubes are clearly visible. After 100 min, long tubes ($>20\mu \mathrm{m}$) are present. More than a day later, tubes $>50\mu \mathrm{m}$ can be found. Scale bar: $20\mu \mathrm{m}$. (b) The distribution of nanotube lengths is exponential at all times. (c) Average length (●) increases while nanotube density (□) decreases over time. Steady state is reached after $\sim 1000\mathrm{m}\mathrm{i}\mathrm{n}$.Reuse & Permissions

Figure 3

Top panels: Stills from a movie in which two short nanotubes join end to end, forming a single nanotube. Bottom panels: End-to-end joining results in patched nanotubes. Tiles labeled with green and red fluorescent dye were assembled separately, mixed, sonicated briefly and left to equilibrate for 48 h. Two examples are shown in which the same field of view is imaged with both green (left) and red (right) filter sets. Scale bars: $5\mu \mathrm{m}$.Reuse & Permissions

Figure 4

Nanotubes disassemble upon heating via scission. (a) Representative fields at increasing temperature. Scale bar: $20\mu \mathrm{m}$. (b) Top: Total polymer length (per field of view) exhibits a typical melting curve shape (solid line). The derivative (dashed line) peaks at $\sim 35°\mathrm{C}$. Bottom: As the average length (dashed line) steadily decreases, the number of nanotubes per field of view (solid line) increases, peaking when $⟨L⟩\sim 4\mu \mathrm{m}$, before decreasing to zero. (c) $⟨L⟩$ versus $1/T$ for nanotube solutions prepared at three different cooling rates: $5°\mathrm{C}$/$5^{\prime }$ (▽), $5°\mathrm{C}$/${20}^{\prime }$ (□), $5°\mathrm{C}$/${60}^{\prime }$ (△). The slope is proportional to $E_{\mathrm{s}\mathrm{c}}$, the scission energy required to create two new ends. Linear fits to the three data sets yield an average value $E_{\mathrm{s}\mathrm{c}}=110\mathrm{k}\mathrm{c}\mathrm{a}\mathrm{l}/\mathrm{m}\mathrm{o}\mathrm{l}\sim 180k_{B}T$.Reuse & Permissions