2009.72 supplementary information - media.nature.com · procedure by Damha and co-workers.S2 Typically, 10 µL of cyanogen bromide (5 M in acetonitrile

nature nanotechnology | www.nature.com/naturenanotechnology 1

supplementary informationdoi: 10.1038/nnano.2009.72

Modular construction of DNA nanotubes of tunable geometry

and single- or double-stranded character

Faisal A. Aldaye, Pik Kwan Lo, Pierre Karam, Christopher K. McLaughlin, Gonzalo Cosa & Hanadi F. Sleiman*

Department of Chemistry, McGill University, 801 Sherbrooke Street West, Montreal, QC H3A 2K6, Canada.

Here we present an approach to DNA nanotube construction that provides control over their size and geometry, one rung at a time. Specifically, we constructed the first triangular and square-shaped DNA nanotubes that can be assembled in fully double-stranded or partially single-stranded forms.

© 2009 Macmillan Publishers Limited. All rights reserved.

2 nature nanotechnology | www.nature.com/naturenanotechnology

supplementary information doi: 10.1038/nnano.2009.72

S1

Contents I. General S2

II. Instrumentation S2

III. Synthesis of triangle 3 and square 4 S2

IV. Construction of triangular rung 3' and square rung 4' S6

V. Assembly of triangular and square-shaped DNA nanotubes 3nt and 4nt S8

VI. Assembly of triangular DNA nanotubes with radially protruding hairpins 3nt-hp S12

VII. Assembly of partially single-stranded square-shaped DNA nanotubes 4nt-ss S14

VIII. Confocal fluorescence microscopy imaging of 3nt S15

IX. Atomic force microscopy S18

X. References S18




S1

Contents I. General S2

II. Instrumentation S2

III. Synthesis of triangle 3 and square 4 S2

IV. Construction of triangular rung 3' and square rung 4' S6

V. Assembly of triangular and square-shaped DNA nanotubes 3nt and 4nt S8

VI. Assembly of triangular DNA nanotubes with radially protruding hairpins 3nt-hp S12

VII. Assembly of partially single-stranded square-shaped DNA nanotubes 4nt-ss S14

VIII. Confocal fluorescence microscopy imaging of 3nt S15

IX. Atomic force microscopy S18

X. References S18

S2

I. General

Acetic acid, boric acid, cyanogen bromide (5M in acetonitrile), formamide, 4-

morpholineethanesulfonic acid (MES), MgCl2·6H2O, StainsAll®, and tris(hydroxymethyl)-

aminomethane (Tris) are used as purchased from Aldrich. 5-Ethylthiotetrazole, 2000Å

phosphate-CPG with a loading density of 5.4 µmol/g, and reagents used for automated DNA

synthesis are purchased from ChemGenes. Exonuclease VII (ExoVII; source: recombinant), and

sephadex G-25 (super fine DNA grade) are used as purchased from Amersham Biosciences.

Microcon® size-exclusion centrifugal filter devices (YM10) are purchased from Millipore.

RubyRed mica sheets for AFM are purchased from Electron Microscopy Sciences. Etched

silicon cantilevers (OMCL-AC160TS) for AFM imaging are used as purchased from Olympus.

II. Instrumentation

Standard automated oligonucleotide solid-phase syntheses are performed on a Perspective

Biosystems Expedite 8900 DNA synthesizer. UV/vis quantifications are conducted on a Varian

Cary 300 biospectrophotometer. Gel electrophoresis experiments are carried out on an

acrylamide 20 X 20 cm vertical Hoefer 600 electrophoresis unit. Electroelutions are performed

using a Centrilutor® electroeluter from Millipore. Temperature controlled hybridizations are

conducted using a Flexigene Techne 60 well thermocycler. AFM images are either acquired on a

Digital Instruments “Dimension 3100” or on an E-scope microscope (Santa Barbara, CA).

III. Synthesis of triangle 3 and square 4

The construction of templates 3 and 4 involves the initial synthesis of the linear analogues of

triangle 3 and square 4, followed by their subsequent templated cyclization and chemical ligation

to generate the fully cyclic and single-stranded building blocks 3 and 4 (Scheme S1). The linear

analogous of 3 and 4 are synthesized on 2000Å phosphate-CPG with a loading density of 15

µmol/g using standard oligonucleotide synthetic protocols. The coupling of vertex 1 is conducted

using a trityl protected amidite derivative (i.e. 2), with an extended coupling and deprotection

times of 15 and 2 minutes. 2 is prepared according to a previously reported method by Sleiman

and co-workers.S1 In the case of triangle 3, for example, sixty bases of the appropriate sequence

are synthesized and are embedded with three units of vertex 2 at positions 10, 30, and 50. A 5′

phosphate group is then synthetically incorporated to facilitate subsequent chemical ligation.




S3

The DNA strands are cleaved and deprotected from the solid-support in a concentrated

solution of ammonium hydroxide (55 °C, 12 hrs), purified using 24% 7 M urea polyacrylamide

gel electrophoresis, extracted into 3 mL of water (16 hours, 37 °C), and desalted using Sephadex

G-25 column chromatography. Quantification is carried by UV/vis analysis using Beer’s law

(Atotal = Avertex + ADNA), in which the extinction coefficient of the vertex at 260 nm is calculated to

be 2.30 X 105 L mol-1 cm-1. Table S1 summarizes the sequences of the linear templates 3 and 4,

and of the template strand used to cyclize them.

p p p

5' OH 3' pp

i ii iii

(top) Single-stranded and cyclic DNA templates triangle 3 and square 4, embedded with vertex 1. (bottom) DNA of the appropriate length, sequence, and number of 1 molecules is (i) synthesized on phosphate-CPG to generate a linear analogue of 3, (ii) which is then cyclized using a complementary DNA template, and (iii) chemically ligated to yield the single-stranded and cyclic DNA triangle 3. The black, green and red strands denote different sequences.

13 4 O O

PPOO O

O

DNADNA

OO 1

Scheme S1 Templates 3 and 4.

Sequences (5' - 3')

Linear 3TATTGGTTTG-1-TGACCAATAACACAAATCGG-1-

TCAGTAATCTCTTGAAGGTA-1-GGAAACGACA-p hosp hate

Linear 4TATTGGTTTG-1-TGACCAATAACACAAATCGG-1-AAAGCTTGAAGG

GGGGAATC-1-TCAGTAATCTCTTGAAGGTA-1-GGAAACGACA-p hosp hate

Tem plate CAAACCAATATGTCGTTTCC

Table S1 Sequences of 3, 4, and template strand.




S3

The DNA strands are cleaved and deprotected from the solid-support in a concentrated

solution of ammonium hydroxide (55 °C, 12 hrs), purified using 24% 7 M urea polyacrylamide

gel electrophoresis, extracted into 3 mL of water (16 hours, 37 °C), and desalted using Sephadex

G-25 column chromatography. Quantification is carried by UV/vis analysis using Beer’s law

(Atotal = Avertex + ADNA), in which the extinction coefficient of the vertex at 260 nm is calculated to

be 2.30 X 105 L mol-1 cm-1. Table S1 summarizes the sequences of the linear templates 3 and 4,

and of the template strand used to cyclize them.

p p p

5' OH 3' pp

i ii iii

(top) Single-stranded and cyclic DNA templates triangle 3 and square 4, embedded with vertex 1. (bottom) DNA of the appropriate length, sequence, and number of 1 molecules is (i) synthesized on phosphate-CPG to generate a linear analogue of 3, (ii) which is then cyclized using a complementary DNA template, and (iii) chemically ligated to yield the single-stranded and cyclic DNA triangle 3. The black, green and red strands denote different sequences.

13 4 O O

PPOO O

O

DNADNA

OO 1

Scheme S1 Templates 3 and 4.

Sequences (5' - 3')

Linear 3TATTGGTTTG-1-TGACCAATAACACAAATCGG-1-

TCAGTAATCTCTTGAAGGTA-1-GGAAACGACA-p hosp hate

Linear 4TATTGGTTTG-1-TGACCAATAACACAAATCGG-1-AAAGCTTGAAGG

GGGGAATC-1-TCAGTAATCTCTTGAAGGTA-1-GGAAACGACA-p hosp hate

Tem plate CAAACCAATATGTCGTTTCC

Table S1 Sequences of 3, 4, and template strand.

S4

The clean isolation of the linear analogues of

triangle 3, square 4 and the template strand is

demonstrated using 24% polyacrylamide gel

electrophoresis (Fig. S1). The gel is visualized

following staining in a solution of StainsAll® for

two hours (12.5 mg StainsAll® in 125 mL of

distilled water and 125 mL of formamide).

The templated cyclization of the linear triangle

3 and the linear square 4 is monitored using 10%

native PAGE, and is found to occur quantitatively

for both 3 and 4. Generally, 1.2 X 10-10 moles of

either of the linear strands, and 1.2 X 10-10 moles

of the template strand are mixed in 10 μL of

TAmg buffer (40 mM Tris, 20 mM acetic acid,

12.5 mM MgCl2·6H20; pH 7.8), and are left

incubating at 0 °C for 10 minutes. As seen in Fig.

S2, the templated closure of the linear analogue of

3 (lane 2), and of the linear analogue of 4 (lane 4) occurs quantitatively in both cases (lanes 3 and

5 respectively).

Figure S2 Assembly of cyclic 3 and 4. Native PAGE analysis reveals the clean templated closure of 3(lane 3) and 4 (lanes 5), using their respective linear analogues (lanes 2 and 4, respectively) and the complementary template strand (lane 1).

p

2 3 4 5

OH 5' p3'

p3'

p

1

OH 5'

Figure S1 Template strand, and the linear analogues of 3 and 4. Denaturing PAGE analysis of the linear analogues of 3 (lane 2) and 4 (lanes 3), and of the template strand used to cyclize them (lane 1).

1 2 3

p

p

Template




S5

Ligations using cyanogen bromide are conducted according to a previously reported

procedure by Damha and co-workers.S2 Typically, 10 µL of cyanogen bromide (5 M in

acetonitrile) is added to the pre-cyclized template in 30 μL of the ligation buffer MES (250 mM

MES, 20 mM MgCl2, pH 7.6), and is left incubating at 0°C for 15 minutes. The resulting mixture

is then recovered using Microcon® size-exclusion centrifugal filter devices (YM10), and analyzed

using denaturing PAGE. As seen in Fig. S3, the cyclization of linear 3 and 4 results in a single

other band of relatively slower mobility assigned to the fully cyclic triangle 3 and square 4,

respectively.

The cyclic nature of triangle 3 and square 4 is confirmed using enzymatic digestion assays

with ExoVII. This enzyme is selective for the digestion of open single-stranded DNA, and will

not digest single-stranded DNA that is cyclized. The digestion of the mixtures generated from the

ligation of 3 and 4 (5 units, 37 °C, 22 mins) results in complete degradation of the linaer

analogues of 3 and 4, and does not affect the bands assigned to cyclic 3 and 4 (Fig. S4). This

confirms the cyclic assignment of triangle 3 and square 4. Finally, 3 and 4 are purified via

electroelution.

p

2 3 4

p

1

4

3

Figure S3 Chemical ligation of 3 and 4. Native PAGE analysis reveals the generation of a single other band of relatively retarded electrophoretic mobility when the cyclic assemblies of 3 (lane 1) and 4 (lane 3) are chemically ligated using cyanogen bromide (lanes 2 and 4, respectively). These single other bands are respectively assigned to the cyclic templates 3 and 4.




S5

Ligations using cyanogen bromide are conducted according to a previously reported

procedure by Damha and co-workers.S2 Typically, 10 µL of cyanogen bromide (5 M in

acetonitrile) is added to the pre-cyclized template in 30 μL of the ligation buffer MES (250 mM

MES, 20 mM MgCl2, pH 7.6), and is left incubating at 0°C for 15 minutes. The resulting mixture

is then recovered using Microcon® size-exclusion centrifugal filter devices (YM10), and analyzed

using denaturing PAGE. As seen in Fig. S3, the cyclization of linear 3 and 4 results in a single

other band of relatively slower mobility assigned to the fully cyclic triangle 3 and square 4,

respectively.

The cyclic nature of triangle 3 and square 4 is confirmed using enzymatic digestion assays

with ExoVII. This enzyme is selective for the digestion of open single-stranded DNA, and will

not digest single-stranded DNA that is cyclized. The digestion of the mixtures generated from the

ligation of 3 and 4 (5 units, 37 °C, 22 mins) results in complete degradation of the linaer

analogues of 3 and 4, and does not affect the bands assigned to cyclic 3 and 4 (Fig. S4). This

confirms the cyclic assignment of triangle 3 and square 4. Finally, 3 and 4 are purified via

electroelution.

p

2 3 4

p

1

4

3

Figure S3 Chemical ligation of 3 and 4. Native PAGE analysis reveals the generation of a single other band of relatively retarded electrophoretic mobility when the cyclic assemblies of 3 (lane 1) and 4 (lane 3) are chemically ligated using cyanogen bromide (lanes 2 and 4, respectively). These single other bands are respectively assigned to the cyclic templates 3 and 4.

S6

IV. Construction of triangular rung 3' and square rung 4'

The construction of the triangular rung 3' and the square rung 4' is conducted using a number of

complementary CS and rigidifying strands RS (Scheme S2). 3', for example, is constructed from

one unit of the triangular template 3, three complementary strands containing sticky-end

overhang cohesions CS1-CS3, and from three rigidifying strands that serve to orient each of these

sticky-ends into one of two lateral directions RS1-RS3.

Assemblies are typically conducted by combining all DNA strands in the correct molar ratios

(final assembly 1.2 X 10-10 moles; 30 μL TAmg buffer), and by incubating at 95 °C for 10

Figure S4 Enzymatic digestions of 3 and 4. The crude mixtures generated from the cyclization of 3 (lane 1) and 4 (lane 3) are enzymatically digested using ExoVII (lanes 2 and 4, respectively), which confirm the cyclic nature of the bands assigned to the single-stranded and cyclic DNA templates 3 and 4.

p

2 3 4

p

1

4

3

Scheme S2 Construction of 3' and 4'.




S7

minutes followed by slowly cooling to 5 °C over a period of 16 hours. Table S2 summarizes the

sequences of the strands used to construct 3' and 4' from 3 and 4, respectively. This process is

monitored sequentially and is found to occur quantitatively at each step leading to, and including,

the construction of the well-defined triangular and square rungs 3' and 4' (Fig. S5a and S5b).

Figure S5 Construction of 3' and 4'. (a) The single-stranded triangle 3 (lane 1) is sequentially titrated with the complmentary strands CS1-CS3 (lanes 2-4, respectively), and with the RS1-RS3 strands to quantitatively generate a fully assembled triangular rung 3' (lane 5). (b) 4 (lane 1) is similarly titrated with CS1'-CS4' (lanes 2-5, respectively), and RS1'-RS4' to quantitatively generate 4' (lane 6). Inset: An expanded view of the gel to better illustrate the shifts in band mobility for the intermediate assemblies.

Table S2 Sequences of CS and RS.

CS1

CS2

CS3

RS1 TTCAACCTAACAGCAAACCT

RS2 GCAATACTATCAAGAGTTCC

CS1'

CS2'

CS3'

CS4'

RS1' Same as RS1 RS2' Same as RS2

RS3' TTCGATCTGGCCAGCCTTTC RS4' Same as RS3

RS3 TTCCTACCTTGAGATGTCGT

Same as CS1

AAAAAGGAACTCTTGACTGGTTATTGTGTTTAGCCCCAGATCGAAACGAC

ATCTCGAAAGGCTGGTTTCGAACTTCCCCCCTTAGAAGGTAGGAATAGGA

Same as CS3

Sequences (5' - 3')

GCTGGGAAGGTTTGCTGCCTTTGCTGTATAACCAAACATAGTATTGCACCCA

AAAAAGGAACTCTTGACTGGTTATTGTGTTTAGCCAAGGTAGGAATAGGA

TGAAGACGACATCTCAGTCATTAGAGAACTTCCATTTAGGTTGAAAACTCTG




S7

minutes followed by slowly cooling to 5 °C over a period of 16 hours. Table S2 summarizes the

sequences of the strands used to construct 3' and 4' from 3 and 4, respectively. This process is

monitored sequentially and is found to occur quantitatively at each step leading to, and including,

the construction of the well-defined triangular and square rungs 3' and 4' (Fig. S5a and S5b).

Figure S5 Construction of 3' and 4'. (a) The single-stranded triangle 3 (lane 1) is sequentially titrated with the complmentary strands CS1-CS3 (lanes 2-4, respectively), and with the RS1-RS3 strands to quantitatively generate a fully assembled triangular rung 3' (lane 5). (b) 4 (lane 1) is similarly titrated with CS1'-CS4' (lanes 2-5, respectively), and RS1'-RS4' to quantitatively generate 4' (lane 6). Inset: An expanded view of the gel to better illustrate the shifts in band mobility for the intermediate assemblies.

Table S2 Sequences of CS and RS.

CS1

CS2

CS3

RS1 TTCAACCTAACAGCAAACCT

RS2 GCAATACTATCAAGAGTTCC

CS1'

CS2'

CS3'

CS4'

RS1' Same as RS1 RS2' Same as RS2

RS3' TTCGATCTGGCCAGCCTTTC RS4' Same as RS3

RS3 TTCCTACCTTGAGATGTCGT

Same as CS1

AAAAAGGAACTCTTGACTGGTTATTGTGTTTAGCCCCAGATCGAAACGAC

ATCTCGAAAGGCTGGTTTCGAACTTCCCCCCTTAGAAGGTAGGAATAGGA

Same as CS3

Sequences (5' - 3')

GCTGGGAAGGTTTGCTGCCTTTGCTGTATAACCAAACATAGTATTGCACCCA

AAAAAGGAACTCTTGACTGGTTATTGTGTTTAGCCAAGGTAGGAATAGGA

TGAAGACGACATCTCAGTCATTAGAGAACTTCCATTTAGGTTGAAAACTCTG

S8

V. Assembly of triangular and square-shaped DNA nanotubes 3nt and 4nt

Assemblies are typically conducted in 30 μL of TAmg buffer, and involve addition of the double-

stranded linking strands to the already assembled rungs 3' or 4' at 40 °C, for 10 minutes, followed

by the slow cooling to 5 °C over a period of 16 hours. The linking strands are mixed with their

respective rungs, in the correct molar ratio, to generate an assembly with a final concentration of

4.0 X 10-6 mol L-1. Table S3 summarizes the sequences used to assemble 3nt and 4nt. AFM

analysis reveals the clean formation of the triangular and square-shaped DNA nanotubes in high

yields (Fig. S6a), while cross-

sectional analysis show them

all to be of the same size (Fig.

S6b). The square-shaped DNA

nanotubes are similarly

assembled in high yields, and

are of the uniform size (Fig.

S7). It is of interest to note

that sample preparation of the

DNA nanotubes for imaging

using AFM resulted in

nanotube assemblies that are

always somewhat embedded

on the mica surface. Cross-

sectional analysis of 3nt and

4nt is therefore conducted on the phase images, and can only be used to confirm that all of the

DNA nanotube assemblies are of the same size. Given that our AFM images clearly show a

large number of individual, unassociated nanotubes, and that cross-sectional height

analysis reveals all DNA nanotubes within a given assembly to essentially be of the same

height, the observed aggregation in some of the AFM images is most likely due to

association of pre-formed DNA nanotubes during sample drying.

Although it is difficult to obtain an accurate value for the diameter of either 3nt or 4nt using

height images in which the DNA assemblies are somewhat embedded, the relative ratio of the

observed dimensions can be used to better ascertain the formation of triangular and square-shaped

DNA nanotube assemblies of the expected size. The theoretically calculated diameter of 3nt is

Sequences (5' - 3')

LS1 TCCCAGCACATCACCTTGGTTGGCTGCTCATACCAGAGTT

LS2 TTTTTTGACATCACCTTGGTTGGCTGCTCATACTCTGGGT

LS3 CTTCATGACATCACCTTGGTTGGCTGCTCATACTCTCCTA

dsLS1 GTATGAGCAGCCAACCAAGGTGATGT

dsLS23 GAGTATGAGCAGCCAACCAAGGTGATGTCA

LS1' Same as LS1

LS2' Same as LS2

LS3' GAGATTGACATCACCTTGGTTGGCTGCTCATACTCGTCGT

LS4' Same as LS3

dsLS1' Same as dsLS1

dsLS234' Same as dsLS23

Table S3.Sequences of LS and dsLS. LS1-3, dsLS1 and dsLS23 are used

to generate 3nt from 3´, and LS1´-4´, dsLS1´ and dsLS234´ are used to

generate 4nt from 4´.




S9

8.9 nm, while that of 4nt is 11 nm. Therefore, the relative diameter ratio of 3nt to 4nt is

theoretically expected to be 0.81. The experimentally obtained cross-sectional height analysis of

b

a

Figure S6 AFM characterization of 3nt. (a) AFM analysis of 3nt reveals the formation of extended one-dimensional DNA nanotube assemblies. Bar is 2.5 μm. (b) Cross-sectional analysis shows them all to be of the same size.




S9

8.9 nm, while that of 4nt is 11 nm. Therefore, the relative diameter ratio of 3nt to 4nt is

theoretically expected to be 0.81. The experimentally obtained cross-sectional height analysis of

b

a

Figure S6 AFM characterization of 3nt. (a) AFM analysis of 3nt reveals the formation of extended one-dimensional DNA nanotube assemblies. Bar is 2.5 μm. (b) Cross-sectional analysis shows them all to be of the same size.

S10

Figure S7 AFM characterization of 4nt. (a) AFM analysis of 4nt reveals the formation of extended one-dimensional DNA nanotube assemblies. Bar is 5 μm. (b) Cross-sectional analysis shows them all to be of the same size.

b

a




S11

the 3nt and 4nt using the height images is consistently found to be 2.95 nm and 3.56 nm for all

nanotubes measured, which translates into a diameter ratio of 3nt to 4nt of 0.83 (Figure S8). This

value is in good agreement with the theoretically calculated value of 0.81, and can be used to

indirectly confirm the formation of triangular and square DNA nanotubes of the expected size.

To determine the thermal stability of our DNA nanotubes, thermal denaturation

experiments are conducted on the assembled triangular DNA nanotube 3nt, on the triangular rung

3', and on the double-stranded linking strands dsLS used to generate the final assembly. As seen

in Fig. S9, 3nt possesses three distinct melting ranges. Individual thermal denaturation

experiments on 3' and dsLS permit us to assign the observed Tm at 80 °C to the dissociation of

the double-stranded portion of dsLS, the Tm at 50 °C to 3', and the Tm at 15 °C to the relatively

shorter sticky-end overhangs connecting the linking strands to their respective rungs. DNA

vert distance 2.985 nm




a

b

Figure S8 Height analysis of 3nt and 4nt. Cross-sectional analysis of 3nt and 4nt reveals a consistent height of 2.95 nm and 3.56 nm for each respective DNA nanotube. The relative height ratio between 3ntand 4nt is thus 0.83.




S11

the 3nt and 4nt using the height images is consistently found to be 2.95 nm and 3.56 nm for all

nanotubes measured, which translates into a diameter ratio of 3nt to 4nt of 0.83 (Figure S8). This

value is in good agreement with the theoretically calculated value of 0.81, and can be used to

indirectly confirm the formation of triangular and square DNA nanotubes of the expected size.

To determine the thermal stability of our DNA nanotubes, thermal denaturation

experiments are conducted on the assembled triangular DNA nanotube 3nt, on the triangular rung

3', and on the double-stranded linking strands dsLS used to generate the final assembly. As seen

in Fig. S9, 3nt possesses three distinct melting ranges. Individual thermal denaturation

experiments on 3' and dsLS permit us to assign the observed Tm at 80 °C to the dissociation of

the double-stranded portion of dsLS, the Tm at 50 °C to 3', and the Tm at 15 °C to the relatively

shorter sticky-end overhangs connecting the linking strands to their respective rungs. DNA





a

b

Figure S8 Height analysis of 3nt and 4nt. Cross-sectional analysis of 3nt and 4nt reveals a consistent height of 2.95 nm and 3.56 nm for each respective DNA nanotube. The relative height ratio between 3ntand 4nt is thus 0.83.

S12

nanotubes 3nt are thus stable at temperature below 15 °C. In principle, the thermal stability of our

DNA nanotubes can be increased to values that are well above room temperature by incorporating

longer sticky-end overhangs connecting each rung to its respective linking strands.

VI. Assembly of triangular DNA nanotubes with radially protruding hairpins 3nt-hp

Hairpins are incorporated into the triangular-shaped DNA nanotubes to aid in their

characterization. In this case, the well-defined triangular rung 3'-hp is constructed using

rigidifying strands with protruding hairpins RS-hp to generate an assembly with three hairpins at

each of its corner units (Scheme S3). The sequences of the modified RS-hp strands are

summarized in Scheme S3; the hairpins are colored in dark purple for clarity. The assembly of a

triangular-shaped DNA nanotube 3nt-hp from this set of modified rungs results in one-

Scheme S3 Construction of 3'-hp using RS-hp.

RS-hp1 - TTCAACCTAAGACATCACCTTTGGGTGATGTCCAGCAAACCT RS-hp2 - GCAATACTATGACATCACCTTTTGGTGATGTCCAAGAGTTCC RS-hp3 - TTCCTACCTTGACATCACCTTTTGGTGATGTCGAGATGTCGT

a b3nt

3' dsLS

Sticky-end overhangs

3'

dsLS

Figure S9 (a) Thermal denaturation profile of 3nt reveals three distinct melting regions: the first is assigned to sticky-end overhangs connecting dsLS to their respective rungs 3', the second is assigned to 3', and the third is assigned to dsLS. (b) Thermal denaturation profile of 3' and dsLS.




S13

dimensional DNA nanotubes extending over tens of microns (Fig. S10a), with a protruding

periodicity at the nanoscale (Fig. S10b). Fourier analysis of the height-trace shows these

assemblies to possess a periodicity of 45 nm (Fig. S10c), which exactly corresponds to the

distance between three rungs within the nanotube assembly.

Figure S10 AFM characterization of 3nt-hp. (a) AFM analysis of 3nt-hp reveals the formation of extended 1D DNA nanotube assemblies (b) that possess detectable protrusions. Bar is 1 μm. (c) Fourier analysis of the height trace reveals a fundamental frequency (i.e. first harmonic) of 9.15. Given that each 15 pixels correspond to 45 nm, this translates into an experimentally calculated periodicity of 43 nm.

10 20 30 40 50 60

First harmonic

Second harmonic

50 100 150 0Distance (pixels)

a

b

c

45 nm




S13

dimensional DNA nanotubes extending over tens of microns (Fig. S10a), with a protruding

periodicity at the nanoscale (Fig. S10b). Fourier analysis of the height-trace shows these

assemblies to possess a periodicity of 45 nm (Fig. S10c), which exactly corresponds to the

distance between three rungs within the nanotube assembly.

Figure S10 AFM characterization of 3nt-hp. (a) AFM analysis of 3nt-hp reveals the formation of extended 1D DNA nanotube assemblies (b) that possess detectable protrusions. Bar is 1 μm. (c) Fourier analysis of the height trace reveals a fundamental frequency (i.e. first harmonic) of 9.15. Given that each 15 pixels correspond to 45 nm, this translates into an experimentally calculated periodicity of 43 nm.

10 20 30 40 50 60

First harmonic

Second harmonic

50 100 150 0Distance (pixels)

a

b

c

45 nm

S14

Although cross-sectional height analysis of the AFM images obtained from 3nt-hp reveal a

periodicity of 45 nm, no periodicity is observed when images from DNA nanotubes 3nt or 4nt

are analyzed. Furthermore, the effect of inward bowing on the morphology of our DNA

architectures, as determined via simple molecular modeling experiments on a three-tier triangular

DNA nanotube, reveals an inward dip of 2.8 nm upon a twist of 40° per rung (Fig. S11),

which in principle should be detectable by AFM. Given that no periodicity is observed for either

3nt or 4nt, we therefore cannot preclude the possibility that the incorporation of hairpins induces

twisting within 3nt-hp.

VII. Assembly of partially single-stranded square-shaped DNA nanotubes 4nt-ss

The approach to DNA nanotube construction presented in this contribution is amenable to the

generation of DNA nanotubes that can also exist in a partially single-stranded and open state. An

open square-shaped DNA nanotube is constructed using well-defined square rungs 4', one

double-stranded liking strands, and three single-stranded linking strands. As expected, AFM

analysis of these nanotubes shows them all to be more flexible when compared to their fully

double stranded analogue (Fig. S12). The fully double-stranded square DNA nanotubes 4nt are

rigid over distances that are greater than those obtained using the partially single-stranded open

square nanotubes.

2.8 nm

Figure S11 Molecular modeling on a three-tiered nanotube assembly reveals a central dip of 2.8 nm upon twisting. (AMBER force field calculations)




S15

VIII. Confocal fluorescence microscopy imaging of 3nt

Cover-slip and chamber preparation: In order to prevent any unwanted DNA-surface interactions

microscope glass-coverslips were functionalized using poly(vinylpyrrolidone) according to a

previously reported procedure (Rothemund et alS3). Glass-coverslips were soaked in a 1 M

sodium hydroxide solution (1 hour), rinsed with de-ionized (DI) water, immersed in a 1% v/v

acetic acid solution (2 hours), rinsed with DI water, and finally silanized using 3-

(trimethoxysilyl)propylmethacrylate (1% v/v) in acetic acid (1% v/v) for 36 hours. In a

Figure S12 AFM characterization of 4nt-ss. (a) AFM analysis of 4nt-ss reveals the formation of extended DNA nanotubes that are relatively more flexible than their fully double-stranded analogous 4nt.(b) Cross-sectional analysis of these assemblies shows them all to be of the same size. Bar is 5 μm.

b

a

S16

subsequent step glass-coverslips were functionalized with poly(vinylpyrrolidone) (PVP, Mw =

360,000). Herein, 500 mL of a 4% w/v poly(vinylpyrrolidone) (PVP) solution were mixed with

2.5 mL of 10% w/v ammonium persulfate solution and 250 μL of N,N,N0,N0

tetramethylethylenediamine (TEMED, Acros). Glass-coverslips were incubated in the PVP

solution at 80 °C for 18 hours, then rinsed and stored in DI water until further use. Immediately

prior to use, two PVP-coated glass-coverslips were rinsed with ethanol and dried. 9 μL of a

solution containing 260 nM DNA (DNA in strand concentration) and 65 μM Picogreen

(Molecular probes) in 40 mM Tris, 20 mM acetic acid, 2 mM EDTA and 12.5 mM MgCl2

(pH=7.8 buffer) were deposited onto one slide and covered by a second one. The DNA solution

was drawn under the capillary forces created by the two slides.

Experimental setup: The imaging setup has been described previouslyS4. The samples were

imaged using a stage-scanning confocal microscope setup. It consisted of a closed-loop sample

scanning stage model Nano LP100, Mad City Labs, (Madison, WI). Continuous circularly

polarized wave excitation at 488 nm (0.407 μW/μm2) from an Ar+ laser model 35 LAL 030 (from

Melles Griot) was introduced via an optical fiber and directed by a dichroic beamsplitter (z488rdc

DCLP, Chroma, Rockingham, VT) to the sample via a high numerical aperture (N.A.) = 1.40 oil

immersion microscope objective (Olympus U PLAN SAPO 100X). Fluorescence emission from

the sample was collected by the objective and directed, through an HQ535/50 emission filter, to

an avalanche photodiode detector (Perkin Elmer Optoelectronics SPCM AQR-14, Vaudreuil,

Quebec, Canada). Images consisting of 256 x 256 pixels were acquired by collecting the intensity

during 1 ms at each pixel. A home built LabView (National Instruments, Austin, TX) routine was

used for the data acquisition and the stage positioning. A National Instruments NI-PCI- 6602

board was used as a counter board.

Images (Fig. S13 and S14): Picogreen-stained DNA nanotubes were imaged using a stage

scanning confocal microscope. Stiff DNA nanotubes approximately 10 μm in length and 350 nm

in diameter (diffraction-limited resolution) were observed. Some nanotubes showed discontinuity

which could be attributed to an uneven staining. Picogreen solution without DNA-nanotubes

showed no longitudinal patterns but bright spots ca. 400 nm in diameter (i.e., diffraction-limited

spots) were however observed which may be attributed to impurities in the sample.




S15

VIII. Confocal fluorescence microscopy imaging of 3nt

Cover-slip and chamber preparation: In order to prevent any unwanted DNA-surface interactions

microscope glass-coverslips were functionalized using poly(vinylpyrrolidone) according to a

previously reported procedure (Rothemund et alS3). Glass-coverslips were soaked in a 1 M

sodium hydroxide solution (1 hour), rinsed with de-ionized (DI) water, immersed in a 1% v/v

acetic acid solution (2 hours), rinsed with DI water, and finally silanized using 3-

(trimethoxysilyl)propylmethacrylate (1% v/v) in acetic acid (1% v/v) for 36 hours. In a

Figure S12 AFM characterization of 4nt-ss. (a) AFM analysis of 4nt-ss reveals the formation of extended DNA nanotubes that are relatively more flexible than their fully double-stranded analogous 4nt.(b) Cross-sectional analysis of these assemblies shows them all to be of the same size. Bar is 5 μm.

b

a

S16

subsequent step glass-coverslips were functionalized with poly(vinylpyrrolidone) (PVP, Mw =

360,000). Herein, 500 mL of a 4% w/v poly(vinylpyrrolidone) (PVP) solution were mixed with

2.5 mL of 10% w/v ammonium persulfate solution and 250 μL of N,N,N0,N0

tetramethylethylenediamine (TEMED, Acros). Glass-coverslips were incubated in the PVP

solution at 80 °C for 18 hours, then rinsed and stored in DI water until further use. Immediately

prior to use, two PVP-coated glass-coverslips were rinsed with ethanol and dried. 9 μL of a

solution containing 260 nM DNA (DNA in strand concentration) and 65 μM Picogreen

(Molecular probes) in 40 mM Tris, 20 mM acetic acid, 2 mM EDTA and 12.5 mM MgCl2

(pH=7.8 buffer) were deposited onto one slide and covered by a second one. The DNA solution

was drawn under the capillary forces created by the two slides.

Experimental setup: The imaging setup has been described previouslyS4. The samples were

imaged using a stage-scanning confocal microscope setup. It consisted of a closed-loop sample

scanning stage model Nano LP100, Mad City Labs, (Madison, WI). Continuous circularly

polarized wave excitation at 488 nm (0.407 μW/μm2) from an Ar+ laser model 35 LAL 030 (from

Melles Griot) was introduced via an optical fiber and directed by a dichroic beamsplitter (z488rdc

DCLP, Chroma, Rockingham, VT) to the sample via a high numerical aperture (N.A.) = 1.40 oil

immersion microscope objective (Olympus U PLAN SAPO 100X). Fluorescence emission from

the sample was collected by the objective and directed, through an HQ535/50 emission filter, to

an avalanche photodiode detector (Perkin Elmer Optoelectronics SPCM AQR-14, Vaudreuil,

Quebec, Canada). Images consisting of 256 x 256 pixels were acquired by collecting the intensity

during 1 ms at each pixel. A home built LabView (National Instruments, Austin, TX) routine was

used for the data acquisition and the stage positioning. A National Instruments NI-PCI- 6602

board was used as a counter board.

Images (Fig. S13 and S14): Picogreen-stained DNA nanotubes were imaged using a stage

scanning confocal microscope. Stiff DNA nanotubes approximately 10 μm in length and 350 nm

in diameter (diffraction-limited resolution) were observed. Some nanotubes showed discontinuity

which could be attributed to an uneven staining. Picogreen solution without DNA-nanotubes

showed no longitudinal patterns but bright spots ca. 400 nm in diameter (i.e., diffraction-limited

spots) were however observed which may be attributed to impurities in the sample.




S17

Figure S13 Fluorescence intensity images obtained for picogreen-stained DNA-nanotubes upon 488 nm excitation. 9 μL of a solution containing 260 nM DNA (DNA in strand concentration) and 65 μM Picogreen was deposited onto one slide and covered by a second one.

Figure S14 Control fluorescence intensity images obtained with picogreen upon 488 nm excitation. 9 μL of a solution containing 65 μM Picogreen was deposited onto one slide and covered by a second one.




S17

Figure S13 Fluorescence intensity images obtained for picogreen-stained DNA-nanotubes upon 488 nm excitation. 9 μL of a solution containing 260 nM DNA (DNA in strand concentration) and 65 μM Picogreen was deposited onto one slide and covered by a second one.

Figure S14 Control fluorescence intensity images obtained with picogreen upon 488 nm excitation. 9 μL of a solution containing 65 μM Picogreen was deposited onto one slide and covered by a second one.

S18

IX. Atomic force microscopy

AFM sample preparation typically involves the deposition of 10 μL of the self-assembled mixture

(concentration of 10 pM) onto freshly cleaved mica (dimensions 2 X 2 cm), followed by adequate

evaporation to achieve complete dryness (typically 30 mins in a fumehood). Whenever possible,

imaging is conducted within 24 hours to minimize time-dependant sample degradation. AFM

images are acquired in air, and at room temperature. “Tapping mode” (i.e. intermittent contact

imaging) is performed at a scan rate of 1 Hz using etched silicon cantilevers with a resonance

frequency of ~ 300 kHz, a spring constant of ~ 42 N/m, and a tip radius of < 10 nm. All images

are acquired with medium tip oscillation damping (20-30%).

X. References

S1. See references 12, 13 and 14 in manuscript.

S2. Carriero, S. & Damha, M. J. Synthesis of lariat-DNA via the chemical ligation of a

dumbbell complex. Org. Lett. 5, 273-276 (2003).

S3 Rothemund, P. W. K. et al. Design and characterization of programmable DNA

nanotubes. J. Am. Chem. Soc. 126, 16344-16352 (2004).

S4 Ngo, A. T., Karam, P., Fuller, E., Burger, M. & Cosa, G. Liposome encapsulation

of conjugated polyelectrolytes: toward a liposome beacon. J. Am. Chem. Soc. 130,

457-459 (2008).


Documents

2009.72 supplementary information - media.nature.com · procedure by Damha and co-workers.S2 Typically, 10 µL of cyanogen bromide (5 M in acetonitrile