of 34
science.sciencemag.org/content/37
1/6531/
eabd6179
/suppl/DC1
Supplementary
Material
s for
Absolute and arbitrary orientation of single
-molecule shapes
Ashwin
Gopinath*, Chris
Thachuk, Anya
Mitskovets, Harry A.
Atwater
, David
Kirkpatrick
,
Paul W. K. Rothemund
*
*Corresponding author.
Email:
agopi@mit.edu (A.G.); pwkr@dna.caltech.edu (P.W.K.R.
)
Published
19 Febr
uary
2021,
Science
371
, eabd6179
(20
21)
DOI:
10.1126/science.
abd6179
This PDF file includes:
Materials and Methods
Figs. S1 to S17
References
and Notes
Contents
Materials and Methods
..........................................................................................
3
DNA origami designs, preparation and purification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3
Placement chip fabrication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5
Photonic crystal fabrication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5
FDTD simulations of PCCs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5
Origami placement experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6
Ethanol drying . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7
Troubleshooting placement experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
8
AFM characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10
TOTO-3 binding and optical experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10
Figs. S
1
through S
17
...........................................................................................
12
2
Materials and Methods
DNA origami designs, preparation and purification
Designs.
Here, all origami were designed with caDNAno (ref. 61,
http://cadnano.org/
) to position all staple ends on the
same face of the origami so that single-stranded 20T extensions to 5’ staple ends would all project from the same face of the
origami. All caDNAno design files and lists of staples are included as a supplementary zip archive:
AA-designs+scripts.zip
.
For right triangle designs, we list two versions of each staple: one is as designed from caDNAno and the other is with 20T
extension on the 5’ end. The three origami used in this work are as follows:
1.
Right-handed right triangle (RRT):
Staples on the right-hand face of this triangle were extended. The caDNAno design
and staple list files are
RRT.json
,
RRT-Staples.xls
and
RRT-T20-Staples.xls
.
2.
Left-handed right triangle (LRT):
This design is similar to that for the right-handed right triangle, except that staple ends
have been shifted by half a DNA turn so that they fall onto the left-hand face of the triangle. The caDNAno design and
staple list files are
LRT.json
,
LRT-Staples.xls
and
LRT-T20-Staples.xls
.
3.
Small moon:
CaDNAno design and staple list files are
small-moon.json
and
small-moon-staples.xslx
; staples
are extended with 20T on their 5’ ends.
Preparation.
Staple strands (Integrated DNA Technologies, 100
μ
M each in water) and the scaffold strand (single-stranded
M13mp18, 400 nM from Bayou Biolabs for right triangles; p8064, 100 nM from Tilibit for small moons) were mixed together to
target concentrations of 100 nM (each staple) and 40 nM, respectively (a 2.5:1 staple:scaffold ratio) in 10 mM Tris Base, 1 mM
EDTA buffer (adjusted to pH 8.35 with HCl) with 12.5 mM magnesium chloride (TE/Mg
2+
). 50
μ
L volumes of staple/scaffold
mixture were heated to 90
C for 5 min and annealed from 90
C to 20
C at -0.2
C/min in a PCR machine. We used 0.5 ml
DNA LoBind tubes (Eppendorf) to minimize loss of origami to the sides of the tube.
4
!
Do not use acetate in preparation of the formation buffer for DNA origami (
e.g.
using acetic acid to adjust
pH). For historical reasons acetate-containing TAE/Mg
2+
, a gel electrophoresis buffer, has been used for
preparing DNA origami. In the context of origami placement, acetate ions cause a high background of small
particles to appear, presumably insoluble acetate salts.
Recipes of all origami used in this paper:
Scaffold
Staple
10x Buffer Water
RRT, 0% T
5
μ
L
(M13mp18) 16
μ
L
(RRT)
5
μ
L
24
μ
L
RRT, 12.5% T 5
μ
L
2
μ
L
(RRT-20T) + 14
μ
L
(RRT) 5
μ
L
24
μ
L
RRT, 25% T 5
μ
L
4
μ
L
(RRT-20T) + 12
μ
L
(RRT) 5
μ
L
24
μ
L
RRT, 37.5% T 5
μ
L
6
μ
L
(RRT-20T) + 10
μ
L
(RRT) 5
μ
L
24
μ
L
RRT, 50% T 5
μ
L
8
μ
L
(RRT-20T) + 8
μ
L
(RRT) 5
μ
L
24
μ
L
RRT, 62.5% T 5
μ
L
10
μ
L
(RRT-20T) + 6
μ
L
(RRT) 5
μ
L
24
μ
L
RRT, 75% T 5
μ
L
12
μ
L
(RRT-20T) + 4
μ
L
(RRT) 5
μ
L
24
μ
L
RRT, 87.5% T 5
μ
L
14
μ
L
(RRT-20T) + 2
μ
L
(RRT) 5
μ
L
24
μ
L
RRT, 100% T 5
μ
L
16
μ
L
(RRT-20T)
5
μ
L
24
μ
L
LRT, 0% T
5
μ
L
16
μ
L
(LRT)
5
μ
L
24
μ
L
LRT, 12.5% T 5
μ
L
2
μ
L
(LRT-20T) + 14
μ
L
(LRT) 5
μ
L
24
μ
L
LRT, 25% T
5
μ
L
4
μ
L
(LRT-20T) + 12
μ
L
(LRT) 5
μ
L
24
μ
L
LRT, 37.5% T 5
μ
L
6
μ
L
(LRT-20T) + 10
μ
L
(LRT) 5
μ
L
24
μ
L
LRT, 50% T
5
μ
L
8
μ
L
(LRT-20T) + 8
μ
L
(LRT) 5
μ
L
24
μ
L
LRT, 62.5% T 5
μ
L
10
μ
L
(LRT-20T) + 6
μ
L
(LRT) 5
μ
L
24
μ
L
LRT, 75% T
5
μ
L
12
μ
L
(LRT-20T) + 4
μ
L
(LRT) 5
μ
L
24
μ
L
LRT, 87.5% T 5
μ
L
14
μ
L
(LRT-20T) + 2
μ
L
(LRT) 5
μ
L
24
μ
L
LRT, 100% T 5
μ
L
16
μ
L
(LRT-20T)
5
μ
L
24
μ
L
Small moon
20
μ
L
(p8064)
10
μ
L
(20T modified)
5
μ
L
15
μ
L
3
Purification.
A high concentration of excess staples will prevent origami placement. Thus origami were purified away from
excess staples using 100 kD molecular weight cut-off filters spin filters (Amicon Ultra-0.5 Centrifugal Filter Units with
Ultracel-100 membranes, Millipore, UFC510024). By the protocol below, recovery is generally 40–50% and staples are no
longer visible by agarose gel:
1.
Wet the filter by adding 500
μ
L
TE/Mg
2+
.
2.
Spin filter at 2000 rcf for 6 min at 4
C , until the volume in the filter is 50
μ
L
. Discard the filtrate.
3.
Add 50
μ
L
of unpurified origami and 400
μ
L
TE/Mg
2+
. Spin at 2000 rcf for 6 min at 4
C.
4.
Discard the filtrate. Add 450
μ
L
TE/Mg
2+
and spin at 2000 rcf for 6 min at 4
C.
5.
Repeat step (4) three more times.
6.
Invert the filter onto a clean tube and spin at 2000 rcf for 6 min at
C to collect purified origami (
50
μ
L
).
Total time for this purification is roughly 40 minutes. Post-purification, origami are quantified using a NanoDrop spectrophoto-
meter (Thermo Scientific), estimating the molar extinction coefficient of the DNA origami as that of a fully double-stranded
M13mp18 molecule (
=
123,735,380/M/cm; we do not correct for small single-stranded loops which are present on the edges
of some designs). We typically work with stock solutions of 15–20 nM DNA origami (2–2.5 OD). The working concentration
for origami during placement is 100 pM, which is too small to be measured with the NanoDrop, so serial dilutions must be
performed. High quality placement is very sensitive to origami concentration. To maintain consistency for each series of
experiments for a particular shape, a single high concentration stock solution (from a single purification) was maintained and
diluted to a nominal concentration of 100 pM as needed.
Note:
All of the work reported in this paper was performed with spin-column purified origami, which is suitable for
small amounts of origami. Larger-scale purification can be achieved using PEG precipitation; we have performed placement
experiments using PEG-purified origami, and achieved good results. See ref. (
63
) for other large-scale purification techniques
and a comparison of their efficiency.
4
!
After purification and quantification, it is especially important to use DNA LoBind tubes (Eppendorf) for
storage and dilution of low concentration DNA origami solutions. Low dilutions,
e.g.
100 pM, must be made
fresh from more concentrated solutions and used immediately—even overnight storage can result in total loss
of origami to the sides of the tube. Addition of significant amounts of carrier DNA to prevent origami loss
may prevent origami placement, just as excess staples do. We have not yet determined whether other blocking
agents such as BSA might both prevent origami loss and preserve placement.
4
Fabrication of binding sites
Fabrication of binding sites is very similar to that found in (
17
) and (
18
) here we give an overview of the process and a couple
places where it departs from previous work. All steps were caried out in Caltech’s Kavli Nanoscience Institute cleanroom.
For non-PCC exmperiments, fabrication begins with a thermally-grown SiO
2
layer (on a silicon wafer) which is cleaned and
silanized with a trimethylsilyl passivation layer by vapor deposition of HMDS (hexamethyldisilazane). A thin (80 nm) layer of
PMMA 950 A2 (MicroChem Corp.; our previous work used a thicker layer of PMMA 950 A3) is spun-coat on the substrate as
a resist. Binding sites in the shape of a DNA origami are defined in the resist with e-beam lithography and developed. After
the binding sites are defined, the trimethylsilyl passivation layer is selectively removed at the binding sites using an anisotropic
O
2
-plasma etch, in a process we term ‘activation’. Finally, the residual PMMA resist is removed to reveal a substrate that is
composed of two chemically distinct regions: (i) origami-shaped features covered with ionizable surface silanols (-OH) and (ii)
a neutrally-charged background covered with trimethylsilyl groups. This procedure enables good placement in 35 mM Mg
2+
.
For the photonic crystal experiments on silicon nitride, the complex geometry of the holes and membranes means that we
cannot add an HMDS passivation layer to some surfaces. To avoid nonspecific binding of origami to these surfaces, we perform
DOP at a lower Mg
2+
concentration of 12.5 mM. To achieve strong adhesion to binding sites under this condition, we silanize
activated sites with 0.1% CTES (carboxyethylsilanetriol from Gelest, 25% w/v Catalog # SIC2263.0) in 10 mM Tris, pH 8.0
for 30 minutes
before
the resist is stripped. In our previous work (
18
), silanization was performed with lower concentration
CTES (0.01% for 10 minutes)
after
the resist was stripped but the new protocol results in lower background binding since the
HMDS passivation layer is protected beneath the resist during silanization.
Fabrication of PCC arrays
Here, fabrication of PCC arrays is very similar to the process found in (
18
) for “isolated PCCs”, rather than the process for
“close-packed arrays”; this is because the PCC arrays described here are smaller and do not justify the more complex process
used to fabricate very large, suspended arrays of PCCs. All steps were carried out in Caltech’s Kavli Nanoscience Institute
cleanroom.
A schematic of the fabrication process is shown in Fig. S13 and SEM of the result in Fig. S14. Fabrication began with
double-side polished silicon wafers (DSP,
h
100
i
, 380
±
10
μ
m thick, University Wafers, Rogue Valley Microdevices) with
275 nm layers of LPCVD-grown SiN on both sides of each wafer. The wafer was cleaned and alignment markers were
defined in the SiN layer by e-beam lithography and modified-Bosch ICP etching. The substrate was then cleaned and silanized
with a trimethylsilyl passivation layer using vapor deposition of HMDS. Next, binding sites in the shape of a DNA origami
were defined using e-beam lithography at specific locations on the front face using the previously-defined alignment markers.
Binding sites were then activated with a short O
2
plasma etch to create silanols, the silanols were converted to carboxyl groups
(see “Fabrication of binding sites”), and the resist was stripped. New resist was spun on, and PCCs were defined around
binding site by e-beam lithography and modified-Bosch ICP etching of the SiN layer. Finally, PCCs were suspended using a
XeF
2
isotropic etch of the underlying Si layer.
FDTD simulations of PCCs
Three dimensional (3D) finite difference time domain (FDTD) simulation was used both for PCC design and to generate
simulated LDOS for comparison with experimental maps of the resonant cavity modes. All simulations were performed
using
FDTD Solutions
from Lumerical Solutions, Inc (
https://www.lumerical.com/
). Lumerical simulation files can be
found in the directory
LumericalScripts
in the zip archive
AA-designs+scripts.zip
. Matlab files for creating Autocad
versions of optimized resonators can be found in the directory
AutocadScriptGenerator
in the same zip archive.
To design the photonic crystal we fixed the refractive index of SiN at 2.05, the thickness of the SiN membrane at 275 nm,
and adjusted
r
,
r/a
,
r
1
,
r
2
and
s
(inset, Fig. S14A) to maximize quality factor within the wavelength range of 655–660 nm.
Photonic crystal size was set to
20
a
in the
x
direction and
34
.
64
a
in the
y
direction. Boundary conditions were implemented
by introducing a perfect matching layer around the structure. The simulation discretization was set to
a/R
in the
x
-direction,
0
.
866
a/R
in the
y
-direction, and
a/R
in the
z
-direction, where the variable
R
was set to 10 for PCC design (so that PCC
parameter could be quickly optimized), and set to 20 to generated simulated LDOS of higher resolution for comparison with
experimental mode maps. The simulation modeled emission from a single dipole with polarization
P
(
x,y,z
)=(1
,
1
,
0)
,
located at a weak symmetry point close to the cavity surface.
5
Origami placement experiments
Below we describe the placement protocol in four steps. See troubleshooting guide on page
8
for an enumeration of problems
and suggestions. See our previous work (
17
) for a greater discussion of origami placement; the supplemental material for that
work provides a figure (Fig. S3) showing how substrates should look during the placement process.
1.
Binding.
A 50 mm petri dish was prepared with a moistened lint-free wipe (Chemwipe) to limit evaporation. For
non-PCC samples, solution with 100 pM origami was prepared in
placement buffer
(10 mM Tris, 35 mM Mg
2+
, pH 8.3)
and a 20
μ
L
drop was deposited in the middle of the chip on top of the patterned region. For PCC arrays, 12.5 mM Mg
2+
was used in the placement buffer (see note below). The chip was placed in a closed, humid petri dish and the origami
solution was allowed to incubate on the chip for 1 hour.
2.
Initial wash.
After the 1 hour incubation, excess origami (in solution) were washed away with at least 8 buffer washes
by pipetting 60
μ
L
of fresh
placement buffer
onto the chip, and pipetting 60
μ
L
off of the chip. Each of the 8 washes
consisted of pipetting the 60
μ
L
volume up and down 2–3 times to
mix
the fresh buffer with existing buffer on the chip.
This initial wash took about 2 minutes.
3.
Tween wash.
Next, in order to remove origami that were non-specifically bound to the passivated background, the chip
was buffer-washed 5 times using a
Tween washing buffer
made by adding 0.1% Tween 20 (v/v) to placement buffer.
This took about 1 minute. Because of the low surface tension of the Tween washing buffer, these washes were somewhat
tricky: they involve adding 20–40
μ
L
of tween wash buffer, just enough to cover most of the chip, but not enough to spill
over the chip and wet the back side of the chip (this may introduce dust contamination from the petri dish). After the 5th
wash, the chip was left to incubate for 30 minutes.
4.
Final wash.
Lastly, the chip was buffer-washed 8 times back into either a higher pH
stabilizing buffer
for wet AFM
imaging (10 mM Tris, 35 mM Mg
2+
, pH 8.9; this prevents movement during AFM) or placement buffer for subsequent
drying. This took about 2 minutes. These washes were high volume (60
μ
L
) and were intended to completely remove the
Tween 20. The amount of Tween 20 left was monitored qualitatively by the surface tension of the drop (roughly, by eye).
When a 20
μ
L
drop covered roughly the same area as the initially deposited drop, it was assumed that the Tween 20 had
been sufficiently removed. After the last wash, the chip was left with roughly 20
μ
L
of buffer and was ready for AFM
imaging or drying.
4
!
Do not use EDTA in placement, Tween washing, or imaging buffers. It is unnecessary in this context, and
will slightly change the effective Mg
2+
concentration available for placement.
4
!
Do not allow the patterned region with binding sites to dry at any point during the binding step or subsequent
buffer washes. Inadvertent dewetting of the binding sites leads to distortion of the origami (causing them to
ball up) as well as the formation of salt crystals on the binding sites. If the substrate needs to be dried follow
the ethanol drying procedure presented in the next section.
4
!
Use Tween 20, rather than other surfactants. Tween 80 and SDS, which are two other common surfactants,
lead to very different results. Tween 80 leads to the total removal of placed origami from the substrate. SDS
does not remove excess origami from the trimethylsilyl background.
4
!
Make sure that chips are not exposed to Tween 20 until
after
the origami have been deposited. Tween 20
applied before binding significantly reduces binding to activated sites.
4
!
Make fresh buffer solutions every week. Here and elsewhere in this work, we use buffers at low strength
(typically 10 mM) to minimize background binding and to make complete washing into different buffers
easier. This means the buffers have low buffering capacity and the pH will decrease with time (and placement
may cease to work).
Note:
For non-PCC samples the binding of DNA origami to SiO
2
is mediated by Mg
2+
binding to surface silanols. For
PCC samples, the origami binding is mediated by Mg
2+
binding to carboxyl groups generated by CTES silanization. The use
of carboxylated binding sites allows high-quality origami placement and orientation on SiN PCC membranes at a much lower
Mg
2+
concentration (12.5 mM) than that required (35 mM) for O
2
plasma-activated binding sites on SiO
2
. We suggest that the
effect is due to the difference in pK
a
between these two functional groups: similar surface carboxyl groups have a pK
a
6, while
6
silanol groups have a pK
a
of 8.3. Thus binding sites with carboxyl groups should carry a higher negative charge at our working
pH of 8.3, they should bind more Mg
2+
, and should enable the observed binding of origami at lower Mg
2+
concentration. In
addition to decreasing the potential for salt artifacts during drying, the use of carboxyl groups has a further very important added
benefit. During the extensive PCC fabrication process, different surface types as identified by a specific series of treatments, are
created. Some of these, for example the inside of the PCC holes or the back side of the PCC membranes, are not passivated with
trimethylsilyl groups, and appear to bind some origami at higher Mg
2+
concentrations. Thus the use of carboxylated binding
sites (and hence a lower Mg
2+
concentraton for placement) decreases nonspecific origami binding and ensures that under our
buffer conditions the only locations at which origami can stably bind are the intended binding sites.
Ethanol drying
After DNA origami were immobilized on chips (and potentially labeled with TOTO-3), they were dried by exposure to an
ethanol dilution series: 10 seconds in 50% ethanol, 30 seconds in 75% ethanol, and 120 seconds in 90% ethanol. To remove
remaining 90% ethanol, chips were air dried.
4
!
If arrays of placed origami are subjected to solutions with less than 80% ethanol for an extended period
(
>
2
minutes), a significant reduction in binding is observed.
4
!
Drying with stream of N
2
can lead to drying artifacts (e.g. micron-scale streaks visible via AFM).
7
Troubleshooting origami placement
Problem
Likely cause
Solution
Site occupancy below 90%.
Old chip with inactive sites.
Low origami concentration.
Short incubation time.
Low Mg
2+
or pH, esp.
if site occupancy
<
30%.
Chips work best
24 hours after activation.
Use higher origami concentration,
100
pM.
Prepare dilution fresh. Use Lo-Bind tubes.
Incubate origami for an hour.
If using silanol surface, use
35 mM Mg
2+
.
If using carboxyl surface, test carboxylation
by placing on an unpatterned activated chip.
Use pH 8.3–8.5.
High multiple binding.
Primarily:
High origami concentration.
Long incubation time.
Oversized features.
Secondarily:
High pH.
High Mg
2+
.
First try:
Use
100 pM origami.
Keep incubation between 30 and 90 min.
Look at features in resist by SEM and
adjust e-beam write (feature size, dose)
and/or minimize O
2
activation time.
Second try:
Keep pH in the range 8.3–8.5.
Use 35 mM Mg
2+
.
Poor alignment of origami
with few multiple bindings.
High pH.
High Mg
2+
.
Keep pH in the range 8.3–8.5.
Use
35 mM Mg
2+
(if using silanols).
Symmetry breaking non-sticky patch is
absent, e.g. poorly written.
High background binding.
Whole or partial origami
on background in AFM.
Unstable AFM,
e.g.
whole scanlines of
identical value (“scars”).
For fluorescent origami,
high background under
optical imaging.
Poor initial TMS quality.
TMS hydrolyzed by high pH.
TMS hydrolyzed by long
incubation.
Failure to wash weakly
bound origami from TMS.
Dehydrate the wafer by baking before
and
after TMS formation.
Keep pH
<
9 preferably in the range 8.3–8.5.
Keep incubation between 30 and 90 minutes.
Remove weakly bound origami with
8
Tween 20 washes.
Large particulates on sites
but few or no origami.
Sample dewetted or dried.
Salts and origami aggregates
occupy the site.
Do not let chip dewet during origami
deposition or subsequent buffer washes.
Small particles on
background.
Overbaked PMMA.
Acetate causes fine precipitate.
Bake PMMA for 30 s at 180
C.
Use non-acetate salts/acids when preparing
buffers,
e.g.
use MgCl
2
, and HCl to adjust.
Placement requires more
than 35 mM Mg
2+
.
Surface is too rough
or improperly cleaned.
Include HF and NH
4
F cleaning steps.
Continues on next page...
8
Problem
Likely cause
Solution
AFM unstable; false engages.
Tween 20 still present.
Increase buffer washes until surface tension
is restored.
Origami fall off during
ethanol drying.
Too much time spent in
dilute ethanol
<
80%.
Move quickly from low to high % ethanol.
Origami ball up into site
during ethanol drying and
corners are double height.
Origami project onto
non-sticky TMS surface.
Hydrolyze TMS surface before drying
by incubating in pH 9 buffer.
9
AFM characterization
All AFM images were aquired using a Dimension Icon AFM/Nanoscope V Scanner (Bruker) using the “short and fat” cantilever
from an SNL probe (“sharp nitride lever”, 2 nm tip radius, Bruker). Non-PCC samples were imaged in fluid tapping mode, using
a cantilever resonance between 8 and 10 kHz. The use of phase imaging allowed us to minimize the tip-sample interaction and
still achieve high enough contrast for image analysis. (High-contrast height imaging required large enough tip-sample forces
that origami would occasionally detach from the surface.) PCC samples were imaged in air in contact mode. AFM images
were processed using Gwyddion (
http://gwyddion.net/
). Single and multiple binding events for placed origami were
hand-annotated and measurements of right triangle and small moon orientation were made by hand using Matlab scripts as
an aid. All scripts for measurement of the right triangle orientation are available in the Auxiliary Supplementary Materials
file
Left+Right-Handed-Placement-Combined.zip
. For example, in the case of right triangles, an overlay of a green
reference image on top of a red test image was used to allow an operator to translate and rotate the reference image relative to
the test image, until a maximum overlap was achieved. The script automatically recorded
x
,
y
, and
for that test image and
presented the next test image. To prevent operator bias, the orientation of the reference image used to measure the angle of
right triangles (to within the nearest 0.5
) was randomized for each test image, and the orientation of the reference image was
obscured from the script operator. This is important to ensure that all angles are accurately measured. Without randomization,
if the initial configuration of 0
was “close enough” to the observed orientation, an operator might be tempted to simply register
0
as the orientation. Similarly, if a particular orientation appears regularly, (say a common rotation such as -150
) an operator
might be tempted to rotate a reference image to exactly -150
and call this “close enough”. Randomization of the reference
image angle and obscuration of the numerical measurement angle prevented bias from arising in these situations.
TOTO-3 binding and optical experiments
After placement, small moon origami were labeled with TOTO-3 (Invitrogen/ThermoFisher) and dried via ethanol drying.
TOTO-3 labeling was performed by incubating placed origami in a buffer (10 mM Tris, 35 mM Mg
2+
at pH 8.3) containing
1 nM TOTO-3 for 10 minutes at room temperature.
All fluorescence imaging was performed with an Olympus BX-61 microscope with a xenon excitation source and Hamamatsu
EMCCD cooled to -75
C. For fluorescence imaging of simple placed samples (without PCCs), excitation light was filtered
with a 640 nm shortpass filter and emission light was longpass-filtered via a 645 nm dichroic. For the PCC array, an additional
655
±
5 nm bandpass filter was used to select the PCC’s fundamental wavelength of 657.2 nm. For non-PCC samples, excitation
light was filtered with an additional linear polarizer, mounted on a rotatable adaptor to allow selection of the desired excitation
polarization
relative to the sample axis. For non-PCC samples, fluorescence emission was collected using a 50
objective
(1.0 NA oil, optimized for polarized light); for the PCC array, a 50
(0.8 NA air) objective was used.
Photoexposure was limited to prevent photobleaching, which could influence data for which multiple serial images were
taken. For both PCC and non-PCC samples, we observed that complete bleaching took approximately 45 seconds under
constant illumination; we took care to limit exposure to less than 10% of this time. For non-PCC samples, the integration time
for each polarization angle was 100 milliseconds. For orientation measurements this meant a total of 3.6 seconds of exposure,
for the polarimeter this meant a total of 1.2 seconds of exposure. The final image of the PCC arry (Fig. 4E) was created by
averaging images from three separate samples; each sample was individually imaged with an integration time of 1 second.
4
!
Do not label origami with TOTO-3 prior to placement. Our attempts to label origami with TOTO-3 in
solution, prior to placement, resulted in no origami binding. This is likely due to distortion of the origami’s
3D shape upon TOTO-3 intercalation (which changes DNA twist); profound distortions of DNA origami
have been observed upon the binding of other intercalators. By intentionally designing DNA origami with
underwinding so that intercalated origami have the desired (flat) 3D shape (
64
) it should be possible to achieve
placement with origami labelled with TOTO-3 or other intercalators.
10