Peptide Assembly Directed and Quantified Using Megadalton DNA Nanostructures

In nature, co-assembly of polypeptides, nucleic acids, and polysaccharides is used to create functional supramolecular structures. Here, we show that DNA nanostructures can be used to template interactions between peptides and to enable the quantification of multivalent interactions that would otherwise not be observable. Our functional building blocks are peptide–oligonucleotide conjugates comprising de novo designed dimeric coiled-coil peptides covalently linked to oligonucleotide tags. These conjugates are incorporated in megadalton DNA origami nanostructures and direct nanostructure association through peptide–peptide interactions. Free and bound nanostructures can be counted directly from electron micrographs, allowing estimation of the dissociation constants of the peptides linking them. Results for a single peptide–peptide interaction are consistent with the measured solution-phase free energy; DNA nanostructures displaying multiple peptides allow the effects of polyvalency to be probed. This use of DNA nanostructures as identifiers allows the binding strengths of homo- and heterodimeric peptide combinations to be measured in a single experiment and gives access to dissociation constants that are too low to be quantified by conventional techniques. The work also demonstrates that hybrid biomolecules can be programmed to achieve spatial organization of complex synthetic biomolecular assemblies.

Mass spectrometry and analytical HPLC characterization of peptides  Figure S2 Additional CD spectra and thermal unfolding transitions for peptides  Figure S3 Additional AUC data for peptides  Figure S4 Determination of Kd for CC-Di-EK:KE dimer by CD spectroscopy  Figure S5 Heterodimer models  Figure S6 Peptide-DNA conjugation  Figure S7 LC-MS data for peptide-oligonucleotide conjugates  Figure S8 CD spectra of peptide-oligonucleotide conjugates  Figure S9 CaDNAno diagrams of DNA origami staple layouts  Figure S10 Diagrams of the ends of DNA origamis decorated with different numbers of peptides  Figure S11 Gel electrophoresis of DNA origamis, DNA origamis decorated with peptides and peptide-mediated origami assembly  Figure Tables   Table S1 Sequences of the designed heterodimeric coiled coils  Circular dichroism (CD) spectroscopy. Peptides and peptide-oligonucleotide conjugates were prepared at the desired concentration in phosphate-buffered saline (PBS) at pH 7.4 comprising Na2HPO4 (8.2 mM), KH2PO4 (1.8 mM), NaCl (137 mM) and KCl (2.7 mM) in quartz cuvettes (Starna Scientific) of appropriate path lengths. CD spectra were baseline corrected and recorded as the average of 8 scans (260 -190 nm) on a JASCO 815 spectropolarimeter fitted with a Peltier temperature controller at 5C, with a scanning speed of 100 nm/min, 1 nm bandwidth, 1 nm data pitch and 1 sec response time. Thermal unfolding profiles were measured by monitoring the signal at 222 nm (1 nm bandwidth) from 5 -95C (temperature ramp 40C/hr) in 1C increments. The midpoint of the transition (TM) was determined as the maximum of the first derivative of the thermal profile to the nearest C.
Determination of peptide Kd by thermodynamic analysis of CD data. The Kd of CC-Di-EK:KE was determined following methods described by Marky and Breslauer. 2 Thermal unfolding profiles were recorded in triplicate over a range of peptide concentrations (100 μM, 50 μM, 20 μM and 5 μM of each peptide) for CC-Di-EK with CC-Di-KE in a 1:1 ratio. A two-state transition was assumed (Equation 1), with a negligible change in heat capacity. A model describing equilibrium for a dimer formed from two non-self-interacting peptides was used (Equation 2) where: Ka = the association constant,  = fraction folded peptide, and ct is the total peptide concentration.  Analytical ultracentrifugation. AUC sedimentation equilibrium experiments were performed once for each peptide, or pair of peptides, at 55 M each in PBS (110 L) at 20C using a Beckman XL-A ultracentrifuge at speeds ranging 44-60 krpm in increments of 4 krpm. The reference channel contained 120 L buffer. Either a two-channel aluminium centre piece (with sapphire windows) or resin-filled epoxy centre piece (with quartz windows) was used. For the heterodimer linked by disulphide bonds formed between opposite ends of the constituent peptides, a six-channel resin-filled epoxy centre piece (with quartz windows) was used with 110 L in each sample channel and 120 L of buffer in the reference channels and scans recorded in 3 krpm increments from 24-39 krpm. Scans were performed across each cell at radial distances of 5.8-7.3 cm after 8 hours at each speed and then again after a further 1 hour to check the samples had reached equilibrium before moving onto the next speed. Data were fit to calculated profiles assuming single ideal species, and partial specific volumes determined using Ultrascan II (http://ultrascan2.uthscsa.edu/). 99% confidence limits were determined by Monte Carlo analyses of the fits.
Disulfide-linked dimers. The concentration of purified CC-Di-EK-C in 1 mL H2O was determined (280 = 10,340 M -1 cm -1 ) and the sample diluted 10-fold into 10 mL. Four equivalents of 2,2'-dipyridyldisulfide (DPDS) was added in 1 mL methanol and the reaction left for 1 hour under agitation. Unreacted DPDS was removed by 3 x 10 mL diethylether extractions. The aqueous fraction containing the DPDS-activated peptide was then freeze-dried before purification by RP-HPLC as above. The concentration of the purified DPDS-activated peptide was determined in H2O and reacted with 1 equivalent of CC-Di-KE-C and 1 equivalent of C-CC-Di-KE in PBS for 3 hours under agitation. The two disulphide-linked heterodimers were purified by HPLC and their identities confirmed by Nanospray-TOF mass spectrometry.
Peptide-oligonucleotide tag conjugation. Azide-functionalized peptides (CC-Di-EK-Z, Z-CC-Di-KE) and 5´ dibenzylcyloocytne (DBCO)-functionalized DNA oligos (,) were conjugated in 50 mM aqueous triethylammonium acetate (TEAA) overnight at 50C. The peptide-oligonucleotide conjugates, CC-Di-EK- and -CC-Di-KE, were isolated by HPLC purification using a JASCO system with a reverse-phase analytical Phenomenex Kinetex C18 column (5 M particle size; 100 Å pore size; 100 x 4.6 mm) running a linear gradient (10 -60%) of buffer B (100 mM TEAA in 80:20 MeCN:H2O) vs. buffer A (100 mM TEAA in H2O), while monitoring the absorbance at 280 and 260 nm. After freeze-drying, samples were dissolved in H2O and desalted into water using an ÄKTAprime plus system (GE Healthcare) with a HiTrap TM desalting column (GE Healthcare) and freeze dried. Conjugates were reconstituted into H2O, divided into aliquots and freeze dried. The stock concentration was measured with a Nanodrop 2000 spectrophotometer (Thermo Scientific) using the sum of the extinction coefficients of the peptides and their ligated oligos: CC-Di-EK- 260 = 292,955 M cm -1 , -CC-Di-KE 260 = 293,714 M cm -1 . The identities of CC-Di-EK- and -CC-Di-KE were confirmed by submitting samples to the LC-MS service at ATDBio Ltd (Oxford) and their purities verified by in-house analytical HPLC using the same conditions as for purification of the reaction mixtures above.
Scaffold DNA preparation. Commercial pUC19 (NEB) was transformed into NEB® 5-alpha Competent E. coli and double-stranded (ds) pUC19 was collected and purified with QIAGEN Plasmid Kit. Singlestranded (ss) pUC19 was prepared by sequential reaction with Nt.BspQI at 50°C and Exonuclease III at 37°C to digest the non-template strand. 3 Enzymes were removed with QIAGEN Plasmid Kit and the DNA then recovered by ethanol precipitation. DNA origami folding. The DNA origami was assembled from a scaffold strand (ss pUC19, 60 nM) and a 5-fold excess of staple strands in 1×TE buffer (10 mM Tris, 1 mM EDTA, pH8.0) with 16 mM MgCl2. The folding mixture was heated to 80C for 5 min then cooled to 60C at a rate of 1C / 4 min, to 25C at a rate of 0.5C / 45 min, to 15C at a rate of 1C / 2 min and held at 15C. DNA origami-peptide preparation. Assembled, unpurified DNA origami with an appropriate number of handles (approx. 60 nM) was mixed with a 3× excess of corresponding peptide-oligonucleotide conjugate and incubated at room temperature (approx. 20C) for 2 hours.
Gel purification. The assembled origami-peptide nanostructures were purified by agarose gel electrophoresis using 0.8% agarose gel containing 0.5xTBE buffer (44.5 mM Tris, 44.5 mM boric acid, 1 mM EDTA, pH 8.3), 11 mM MgCl2 at 60 V for 2-3 hours in a gel tank incubated in an ice-water bath. Target bands were cut out and squeezed on Parafilm® M. The concentration of the collected solution was determined by measuring ultraviolet absorbance at 260 nm (Cary 50 Probe UV-Visible Spectrophotometer, Varian). We assumed the extinction coefficient of the origami to be equal to that of double-stranded pUC19 (4.24×10 7 M -1 cm -1 ). 4 DNA origami-peptide assembly. Gel-purified, peptide-functionalized DNA origamis of different initial concentrations (Table S4) were mixed at room temperature (approx. 20C) and incubated for at least 2 hours to equilibrate.
Transmission electron microscopy. DNA origami-peptide assemblies prepared as described above were adsorbed onto glow-discharged Formvar & heavier carbon film TEM grids (Agar Scientific) and then stained using 2% aqueous uranyl acetate. Imaging was performed using an FEI Tecnai T12 Transmission Electron Microscope operated at 120 kV.

Supporting Discussion
The following is an analysis of random and systematic errors in the calculation of dissociation constants based on analysis of counted numbers of origami monomers, dimers and higher multimers in transmission electron micrographs.

Concentration measurement errors
Stock origami concentrations were calculated from their absorbance at 260 nm measured using a Cary 50 Probe uv-visible spectrophotometer (Varian). The molar extinction coefficients of the two origami nanostructures were assumed to be equal to that of double-stranded pUC19 (4.24×10 7 M -1 cm -1 ). The standard deviation (SD) of absorbance measurements performed in triplicate was typically <8%. The inferred concentrations of all species were scaled from these measurements: we assume a corresponding random fractional error c = 0.08 in the concentrations of all species.
All concentrations in a given experiment were scaled from the measured total concentration of origami A (see note on equilibrium concentration calculation in below). The fractional random error in [A]×[B]/[A:B] due to uncertainties in the measurements of origami concentrations is therefore equal to the fractional random error in [A]tot, c. Correlation between concentration errors in a given experiment was ignored in the error analysis presented below: this slightly increases the inferred error in Kd.

Sampling bias, relative abundances
Longer staining times were used for lower-concentration samples to ensure adequate densities of imaged particles. We found no systematic variation of calculated Kd with staining time.
Particle densities vary on an EM grid, therefore several locations were sampled in each experiment to maximize consistency of sampling.
Although we used nominally equal concentrations of the two peptide-decorated origamis An and Bn, the total numbers of Bn counted are systematically 20%±13% higher than An (the quoted uncertainty is the standard deviation of 24 pairs of measurements). This may reflect a sampling bias -Bn sticks to the TEM grid more readily than Anor a hypochromic correction to DNA absorbance at 260 nm that depends on the sub-wavelength-scale origami structure. We have assumed that the total counted number of each species (An or Bn) is proportional to its total concentration and have scaled the concentrations of all species in a given experiment to the measured total concentration of origami A. Uncertainty in this correction leads to a systematic error of approximately 20% in all Kds.

Statistical sampling errors
For each system (0/1/2/3 peptides per origami) and each experiment (different initial concentrations), the total numbers of origamis An and Bn counted were similar (in the range 538-2137) while the counted numbers of homo-and heterodimers varied greatly. We estimate the standard error of the total number of each species 'x' in a sample as the square root of the number Nx counted. The corresponding fractional random error in the inferred concentration is (Nx) -0.5 .

Identification errors
Species were identified and counted by eye. We estimate that errors due to misidentification of origami A as B and vice versa were small compared to statistical sampling errors. Some species identified as peptide-linked dimers are, in fact, unlinked pairs of monomers whose functionalized ends lie, by chance, close to each other. Dimers are thus overcounted. Each electron micrograph is square of side L = 2 μm. Let t be the tolerance for judging whether two origamis are actually connected, i.e., pairs of origamis whose peptide-functionalized ends are separated by a distance less than t are identified as dimers whether or not they are physically linked (we estimate that t is approximately 10 nm). The number of false positives (FP) per micrograph is therefore NANB/nm 2 ×πt 2 /L 2 (heterodimer) or Nx 2 /(2nm 2 )×πt 2 /L 2 (homodimer), where NX is the total number of origamis of type X counted in nm micrographs. The corresponding total numbers of false positives in each experiment is NFP = NANB/nm×πt 2 /L 2 (heterodimer) or Nx 2 /(2nm)×πt 2 /L 2 (homodimer).
If we assume that there is no significant association between origamis with no linking peptides attached, i.e., all dimers counted in such experiments are 'false positives', then we obtain from these data an improved estimate of the range parameter t = 8±1 nm. (There is a small but significant systematic difference between values of t calculated for hetero-and homo-dimeric complexes which may reflect a real attractive interaction between the ends of linked origamis.) We use this estimate to calculate the 'false positive' background for each species of dimer in each experiment and subtract it from the total number counted. The uncertainty in this background correction is (NFP) 1/2 .

Curve fitting
Relative uncertainties in data were estimated as follows (see above): [ : ] Uncertainties (Err) in data used as input parameters for the linear fits used to estimate Kds are listed in Table S5.

Assessments of fits, errors Linear fits to graphs of [An] × [Bn] vs. [An:
Bn] etc. were obtained using OriginPro, with uncertainties in both quantities calculated as described above and listed in Table S5. The gradient of the fitted line was accepted as a useful measure of Kd if the change in fitted gradient on constraining the line to pass through the origin was <30%. Errors quoted in Table 1 are standard errors on the gradient of the constrained fits and do not include systematic errors discussed above.     Table S2.

Figure S3
Additional AUC data for peptides. Sedimentation equilibrium AUC data (top, circles) and fits to single ideal species models (top, black lines) with residuals for the fits (circles, bottom) plotted against r 2 -r0 2 (r, radius of the sample from the centre of the rotor and a reference radius, r0. The masses returned from the fits are given in Table S3.         Table S6-1) and one staple strand in Origami B is labelled with Cy5 (red; for details see Table S6       Midpoints of thermal unfolding transitions were determined from the maxima of first order derivatives of the thermal profiles recorded at 222 nm by CD spectroscopy. Experimental conditions and unfolding data are in Figure S2.  Table S3. Peptide oligomeric states as determination by AUC. Calculated partial specific volumes (vbar) and molecular weights vs. AUC-determined molecular weights and consequent deduced oligomeric states for peptides in this study and for which the data are shown in Figure S3.  Each experiment corresponds to a mixture prepared with different initial concentrations (Conc(0)) of origamis A0 and B0. After equilibration, numbers of each assembly/complex (N) in TEM images were counted and corresponding equilibrium concentrations were calculated. The numbers of "dimers" observed are consistent with a random distribution of origami nanostructures in each micrograph with a minimum detectable separation between the ends of adjacent origamis of 8 nm: pairs of origamis with a closer separation are "false positives", counted as dimers but not linked by peptide-peptide interactions. Equilibrium concentrations shown here were calculated without deduction of this false positive background.

Table S4-2 Equilibrium concentrations of different assemblies in mixtures of origamis A1 and B1.
Each experiment corresponds to a mixture prepared with different initial concentrations (Conc(0)) of origamis A1 and B1. After equilibration, numbers of each assembly/complex (N) in TEM images were counted and corresponding equilibrium concentrations were calculated before (Conc(etot)) and after (Conc(e)) correction for a background corresponding to random coincidences (NFP   (0)) of origamis A2 and B2. After equilibration, numbers of each assembly/complex (N) in TEM images were counted and corresponding equilibrium concentrations were calculated before (Conc(etot)) and after (Conc(e)) correction for a background corresponding to random coincidences (NFP