Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
NUCLEIC ACID NANOTUBE LIQUID CRYSTALS AND USE FOR NMR STRUCTURE DETERMINATION OF MEMBRANE PROTEINS
Document Type and Number:
WIPO Patent Application WO/2007/127020
Kind Code:
A2
Abstract:
Compositions and methods for preparing nucleic acid nanotubes using DNA origami techniques are described, which provide for nanotubes of predictable and uniform length. The nucleic acid nanotubes thus formed are suitable as liquid crystal preparations enabling liquid-crystal NMR spectroscopy of proteins solubilized in detergent.

Inventors:
SHIH WILLIAM M (US)
DOUGLAS SHAWN M (US)
CHOU JAMES J (US)
Application Number:
PCT/US2007/008113
Publication Date:
November 08, 2007
Filing Date:
April 03, 2007
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
DANA FARBER CANCER INST INC (US)
SHIH WILLIAM M (US)
DOUGLAS SHAWN M (US)
CHOU JAMES J (US)
International Classes:
A61K48/00
Other References:
LIU H. ET AL.: 'Approaching the limit: Can one DNA oligonucleotide assemble into large nanostructures?' ANGEWANDTE, CHEMIE (INTERNATIONAL ED. IN ENGLISH) vol. 45, no. 12, March 2006, pages 1942 - 1945
LORIGAN G. ET AL.: 'Solid-state NMR spectroscopic studies of an integral membrane protein inserted into aligned phospholipid bilayer nanotube arrays' JOURNAL OF THE AMERICAN CHEMICAL SOCIETY vol. 126, no. 31, August 2004, pages 9504 - 9505
MATHIEU F. ET AL.: 'Six-helix bundles designed from DNA' NANO LETTERS vol. 5, no. 4, April 2005, pages 661 - 665
ROTHEMUND P.W.K.: 'Folding DNA to create nanoscale and patterns' NATURE vol. 440, no. 7082, March 2006, pages 297 - 302
Attorney, Agent or Firm:
WILLIAMS, Kathleen et al. (P.O. Box 55874Boston, MA, US)
Download PDF:
Claims:

CLAIMS

1. A composition comprising nucleic acid nanotubes having a uniform length, wherein each nanotube comprises a plurality of linked double-stranded nucleiacid helices, and wherein each nanotube is formed from at least one single-stranded scaffold nucleic acid molecule and a plurality of staple oligonucleotides.

2. The composition of claim 1, wherein the nucleic acid is DNA.

3. The composition of claim 1, wherein the average length of the nucleic acid nanotubes is given by the length of the single-stranded scaffold nucleic acid divided by the number of double stranded nucleic acid helices comprising each nanotube.

4. The composition of claim 1, wherein the distance between crossovers between any two adjacent helices is at least 10 base pairs.

5. The composition of claim 1, wherein there are at least two crossovers between any pair of adjacent helices.

6. The composition of claim 1, wherein each staple oligonucleotide comprises at least 7 base pairs, and links at least two non-contiguous segments of the single-stranded scaffold nucleic acid.

7. The composition of claim 6, wherein each staple oligonucleotide comprises 32 base pairs.

8. The composition of claim 6, wherein each staple oligonucleotide comprises 54 base pairs.

9. The composition of claim 1, wherein each staple oligonucleotide comprises 42 base pairs, and links three non-contiguous segments of the single-stranded scaffold nucleic acid .

10. The composition of claim 1, wherein the average length of the nucleic acid nanotubes is greater than 50 nanometers.

11. The composition of claim 1 , wherein the average length of the nucleic acid nanotubes is about 400 nanometers.

12. The composition of claim 1, wherein each nucleic acid nanotube comprises at least three double-stranded nucleic acid helices.

13. The composition of claim 1, wherein each nucleic acid nanotube comprises at least four double-stranded nucleic acid helices.

14. The composition of claim 1 , wherein each nucleic acid nanotube comprises five double-stranded nucleic acid helices.

15. The composition of claim 1, wherein each nucleic acid nanotube comprises six double-stranded nucleic acid helices.

16. The composition of claim 1, wherein each nucleic acid nanotube comprises seven double-stranded nucleic acid helices.

17. The composition of claim 1, wherein the standard deviation of the lengths of the nucleic acid nanotubes is no more than 20% of the average nanotube length.

18. The composition of claim 1 , wherein the standard deviation of the lengths of the nucleic acid nanotubes is no more than 10% of the average nanotube length.

19. The composition of claim 1, wherein the nucleic acid nanotubes form a liquid crystalline phase.

20. The composition of claim 19, wherein the nucleic acid nanotubes in liquid crystal phase are birefringent.

21. The composition of claim 19, wherein the nucleic acid nanotubes in liquid crystal phase give rise to measurable residual dipolar couplings in a solution with detergent-solubilized membrane protein.

22. The composition of claim 1, wherein the scaffold nucleic acid is derived from a natural source.

23. The composition of claim 1, wherein the scaffold nucleic acid is derived from Ml 3 DNA.

24. The composition of claim 1, wherein the scaffold nucleic acid is artificial.

25. The composition of claim 1 , further comprising a protein.

26. The composition of claim 25, further comprising a detergent.

27. The composition of claim 26, wherein the protein is a membrane protein.

28. The composition of claim 25, wherein the protein is present at a concentration of at least 0.1 mM.

29. A method of making nucleic acid nanotubes comprising the steps of:

(a) preparing a solution comprising a single-stranded scaffold nucleic acid and a plurality of staple oligonucleotides;

(b) heating the solution to denature the scaffold nucleic acid and staple oligonucleotides; and

(c) cooling the solution to assemble the nanotubes;

wherein the nucleic acid nanotubes have a uniform length.

30. The method of claim 29, wherein the solution is heated to about 90 degrees C.

31. The method of claim 29, wherein the solution is cooled by allowing it to stand at room temperature for at least 2 hours.

32. The method of claim 29, wherein the solution is cooled by allowing it to stand at room temperature for about 20 hours.

33. A method of performing NMR spectroscopy of proteins using nucleic acid nanotubes comprising the steps of:

(a) suspending nucleic acid nanotubes in a solution;

(b) forming a liquid crystalline phase comprising the nucleic acid nanotubes;

(c) adding a protein to the solution; and

(d) performing NMR spectroscopy on the protein and nucleic acid nanotube mixture.

34. The method of claim 33, wherein the protein is a membrane protein.

35. The method of claim33, wherein the solution comprises a detergent.

36. The method of claim 33, wherein the protein is present at a concentration of at least 0.1 mM.

37. A plurality of nucleic acid nanotubes having a uniform length, wherein each nanotube comprises a plurality of linked double-stranded nucleic acid helices, and wherein each nanotube is formed from at least one single-stranded scaffold nucleic acid molecule and a plurality of staple oligonucleotides.

38. The composition of claim 1, wherein the nucleic acid nanotubes comprise at least two basepair sequences; the first basepair sequence nanotube having a first head end and a first tail end, said first head end comprising paired complementary strands of nucleic acids, and said first tail end comprising a plurality of first unpaired strands of nucleic acids; and the second basepair sequence nanotube having a second head end and a second tail end, said second head end comprising a plurality of second unpaired strands of nucleic acids, and said second tail end comprising paired complementary strands of nucleic acids; wherein said first unpaired strands of nucleic acids are substantially complementary to said second unpaired strands of nucleic acids.

39. The composition of claim 1, wherein the average length of the nucleic acid nanotubes is about 800 nanometers.

Description:

NUCLEIC ACID NANOTUBE LIQUID CRYSTALS AND USE FOR NMR STRUCTURE DETERMINATION OF MEMBRANE PROTEINS

RELATED APPLICATIONS This application claims the benefit of priority to U.S. Provisional Application

Ser. No. 60/793,788, filed April 21, 2006, and U.S. Provisional Application Ser. No. 60/904,266, filed February 28, 2007.

TECHNICAL FIELD The invention generally relates to nucleic acid nanotubes. More particularly the invention relates to compositions and methods for making nucleic acid nanotubes that are suitable for performing liquid-crystal NMR spectroscopy of detergent- solubilized membrane proteins.

BACKGROUND

Structure determination of membrane proteins is an important challenge for biomedical science. About thirty percent of expressed proteins span lipid bilayers, yet structures of only about one hundred membrane proteins have been resolved. Membrane proteins are encoded by 20-35% of genes but represent fewer than one percent of known protein structures to date. Knowledge of their structures will be enormously insightful for cell biology. Furthermore, membrane proteins are important as drug targets. The slow rate of membrane-protein structure determination represents a significant bottleneck for both basic and applied bioscience discovery. This bottleneck largely derives from difficulties in forming well-ordered three-dimensional crystals of membrane proteins. Solution NMR presents an attractive alternative for the study of membrane proteins, as high-resolution structural information can be obtained for proteins up to 80 kD in size without the need for crystallization. Residual dipolar couplings (RDCs), commonly measured for biological macromolecules weakly aligned by liquid-crystalline media, are important global angular restraints for NMR structure determination. For membrane proteins greater than 15-kDa in size, Nuclear-O verhauser-effect (NOE)-derived distance restraints are difficult to obtain, and RDCs could serve as the main reliable source of NMR

structural information. In many of these cases, RDCs would enable full structure determination that otherwise would be impossible. However, none of the existing liquid-crystalline media used to align water-soluble proteins are compatible with the detergents required to solubilize membrane proteins. For solution NMR, macromolecules must be solubilized in water to facilitate fast tumbling; the faster the tumbling, the better the spectra. To promote water solubility, membrane proteins must be complexed with detergent micelles. The micelle-protein complex is considerably larger than the protein alone, and tumbling is relatively slow as a result. This increase in effective size is especially problematic for α-helical membrane proteins greater than 15 kD in size, where resonance peaks are closely spaced and become irresolvable with the fast coherence relaxation of slowly tumbling macromolecules. In order to obtain information about the internuclear angles, each protein must be made to tumble in a weakly ordered regime. The appropriate weak ordering, about 0.1%, can be achieved by dissolving the protein in an appropriate.concentration of a suitable alignment material. For example, water- soluble proteins can be aligned weakly by a suitable amount with ~1.5—2% PfI filamentous phage, which forms a liquid crystal at that concentration. The easiest method for weak alignment of proteins is through mixing the protein with a liquid- crystalline medium, such as PfI filamentous phage, DMPC/DHPC bicelles, C12E5 polyethylene glycol, or cellulose crystallites. However, none of these media are compatible with detergent-solubilized membrane proteins.

The general applicability of solution NMR spectroscopy to structural characterization of intact α-helical membrane proteins has been demonstrated by the structure determination of the 15-kDa Mistic protein and the 30-kDa pentameric phospholamban, as well as the complete assignment of backbone resonances and secondary structures of the 44-kDa trimeric diacylglycerol kinase and the 68-kDa tetrameric KcsA potassium channel. Despite such progress, full-scale structure determination of α-helical membrane proteins remains challenging and rare. Due to the large fraction of methyl-bearing residues in membrane proteins and to the added molecular weight of detergent micelles, the low chemical-shift dispersion of α-helical proteins is obscured by resonance overlap and line broadening, making assignment of side-chain methyl resonances extremely difficult. Without side-chain chemical shifts, it is impossible to obtain a sufficient number of long-range NOE-derived distance

restraints for folding secondary segments into the correct tertiary structure. Therefore, development of alignment media for accurate RDC measurements from α-helical membrane proteins would enhance significantly the capability of solution NMR in structure determination of this important class of targets. The most effective method for weak alignment involves mixing the protein of interest with large particles that form stable liquid crystals at low concentration (~1.5- 5% w/v). Liquid crystals that have been used to align water-soluble proteins include DMPC/DHPC-bicelle liquid crystals, filamentous phage particles, ternary mixtures of cetylpyridinium CI/Br, hexanol, and sodium CI/Br, binary mixtures of polyethylene glycol and hexanol, and cellulose crystallites. However, none could be applied to membrane proteins due to incompatibility with the zwitterionic or anionic detergents typically used to solubilize membrane proteins for structural study. The only method currently available for weak alignment of membrane proteins involves the use of strained (radially or axially compressed) polyacrylamide gels. However, dissolving protein-micelle complexes to high concentration in gels is notoriously difficult due to the inhomogeneous pore size of randomly cross-linked gel matrices. Thus the measured RDCs are of limited accuracy.

Nucleic acid nanotube liquid crystals can extend the advantages of weak alignment to NMR structure determination of a broad range of detergent-solubilized membrane proteins. Alignment media comprised of 800 nm heterodimer DNA nanotubes should be broadly useful for providing global structural restraints in solution NMR studies of membrane proteins. As a large number of helical membrane proteins of great biomedical interest are between 20-30 kDa in size — well below the current size limitation of solution NMR spectroscopy — new experimental systems for obtaining NMR structural information in the presence of detergents are of fundamental importance. DNA nanotechnology, which affords versatile molecular design and sub- nanometer-scale precision, has been pursued as a route towards building host lattices to position guest macromolecules for crystallographic structural studies. The present invention employs solution NMR instead of crystallographic methods, and validates the potential of DNA nanotechnology for imposing order on target macromolecules to acquire atomic-resolution structural information.

SUMMARY OF THE INVENTION

The invention is related to novel compositions and methods for preparing liquid crystalline solutions of nucleic acid nanotubes suitable for performing liquid- crystal NMR spectroscopy of proteins, including detergent-solubilized membrane proteins. By virtue of being constructed from nucleic acids, these nanotubes generally are resistant to detergents, and can be constructed, for example, to mimic the shape and size of filamentous phage particles.

It is an object of the invention to provide a composition comprising nucleic acid nanotubes having uniform length. Each nanotube comprises a plurality of linked double-stranded nucleic acid helices, and each nanotube is formed from at least one single-stranded scaffold nucleic acid molecule and a plurality of staple oligonucleotides. The nucleic acid can be DNA. The nucleic acid nanotubes can form a liquid-crystalline phase in solution, and proteins solubilized in detergent can be aligned weakly using the nanotube liquid crystals.

A further object of the invention is to provide nucleic acid nanotubes in which the average length of the nucleic acid nanotubes is given by the length of the single- stranded scaffold nucleic acid divided by the number of double stranded nucleic acid helices comprising each nanotube. In some embodiments the length of the helices comprising the nucleic acid nanotubes varies by no more than 20% of the average length of the helices, and in certain embodiments by no more than 10 % of the average length of the helices. The length of the nucleic acid nanotubes is greater than about 50 nanometers, and about 400 nanometers. The nanotubes comprise at least 3 adjacent double-stranded helices. In certain embodiments the nanotubes consist of 5, 6, or 7 adjacent helices. In other embodiments, DNA heterodimer nanotubes SOO nanometers in length can be constructed from 400 nanometer monomers of two types, one type of monomer self-assembling with the second type of monomer.

A further object of this invention is to provide a method of preparing nucleic acid nanotubes. The method comprises preparing a solution comprising a single- stranded scaffold nucleic acid and a plurality of staple oligonucleotides, heating the solution to denature the scaffold nucleic acid and staple oligonucleotides, and cooling

the solution to room temperature. The nucleic acid nanotubes thus formed will have a uniform length.

A further object of this invention is to provide a method to perform liquid- crystal NMR spectroscopy of proteins using nucleic acid nanotubes. The method comprises suspending nucleic acid nanotubes in a solution, forming a liquid crystalline phase comprising the nucleic acid nanotubes, adding a protein to the solution, performing NMR spectroscopy on the protein and nucleic acid nanotube mixture. In some embodiments, the protein added to the solution is solubilized in detergent. In other embodiments, the protein is a membrane protein solubilized in detergent.

Further features and advantages of the invention and further embodiments will become more fully apparent in the following description of the embodiments and drawings thereof, and from the claims.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 shows various representations of a 800 nm DNA nanotube heterodimer.

Figure Ia is a stylized 3-dimensional representation highlighting the 14nm segment that forms the junction between the two 400 nm monomers. Figure Ib is a segment diagram in which each monomer consists of 28 segments of length 42 base pairs, as well as a head and tail segment on each end.

Figure Ic is a schematic of the two scaffold strands of each monomer (without the complementary staple oligonucleotides), each monomer consisting of a modified M13 bacteriophage single-stranded DNA genome of length 7308 bases. Figure Id is a cross-sectional schematic view of the DNA nanotube shown in

Figure Ia.

Figure Ie is a schematic of the 14 nm junction between the linked nanotube monomers, showing the 42 base pair link between the two scaffold strands and their complementary staple strands. Figure If is a schematic of a typical 42 base pair segment of the completed nanotube, showing that a crossover pattern consisting of six staple strands repeats itself every 42 base pair segment along the length of the nanotube.

Figure 2 A is a schematic representation of a six-helix bundle DNA nanotube. Figure 2B is a representation of the arrangement of scaffold strand and staple oligonucleotides of the six-helix DNA nanotube of Figure 2A. The dark line represents the uninterrupted scaffold DNA, folded into a six-strand arrangement, with a seam formed between helix 2 and helix 3, and another seam between helix 4 and helix 5. The staple oligonucleotides are generally 42 bases long, each contributing three 14-base segments, one to each of three adjacent portions of the scaffold DNA. Figure 3 shows schematic representations of the front and rear overhangs of the pre-dimerization monomers that combine to form DNA heterodimer nanotubes Figure 3a and 3b show, respectively the "capped" scaffold-plus-staples head, and the unpaired scaffold and staple strands of the tail of the front monomer.

Figure 3c and 3d show, respectively, the unpaired scaffold and staples of the head, and the "capped" scaffold-plus-staples tail of the rear monomer.

Figure 3e shows the inter-monomer junction of a DNA heterodimer nanotube, in which the unpaired scaffold and staples of the front monomer are complementary to and join with the unpaired staples and scaffold of the rear monomer.

Figure 4A-4D shows a Python computer program script for the sequence determination of the components of six-helix bundle DNA nanotubes.

Figure 5A-5B is the Python computer program script used to generate front monomer core oligonucleotides and head caps.

Figure 6A-6B is the Python computer program script used to generate rear monomer core oligonucleotides and tail caps.

Figure 7A-7B is the Python computer program script used to generate front monomer tail connector oligonucleotides and rear head connector oligonucleotides. Figure 8 shows the sequences used in the example for front monomer head cap staples of a six-helix bundle DNA nanotube.

Figure 9 shows the sequences .used in the example for front monomer tail connector staples of a six-helix bundle DNA nanotube.

Figure 10 shows the sequences used in the example for rear monomer head connector staples of a six-helix bundle DNA nanotube.

Figure 11 shows the sequences used in the example for rear monomer tail cap staples of a six-helix bundle DNA nanotube.

Figure 12A-C is the sequence of the M13mpl8 derived single stranded DNA scaffold used in the DNA nanotube of Figure 2.

Figure 13A-D lists the sequences of each of the staple oligonucleotides that build to the scaffold DNA used in the DNA nanotube of Figure 2. Figure 14 shows a computer-generated random 59-base sequence inserted into

M13mpl8 at insert position 6258 to generate recombinant M13 filamentous bacteriophage.

Figure IS shows the sequences that were used in the example to construct the M13mpl8 insert fragment of Figure 14, together with flanking regions (109 base pairs total).

Figures 16A-C show the recombinant Ml 3 filamentous bacteriophage genome sequence used in the example that serves as the input to all Python scripts to generate the scaffold strand of the DNA nanotubes.

Figures 17A-D list the sequences used in the example for front monomer core staples of a six-helix bundle DNA nanotube.

Figure 18A-D list the sequences used in the example for rear monomer core staples of a six-helix bundle DNA nanotube.

Figure 19a shows a gel-shift analysis of folding and heterodimerization of DNA nanotubes. Figure 19b shows a negative-stain electron micrograph of DNA nanotube heterodimers.

Figure 19c shows a photograph of the birefringence exhibited between crossed polarizers by DNA nanotube dimmers at 28 mg/mL in a glass NMR tube.

Figure 19d shows the NMR spectrum of a 90% H 2 O/10%D 2 O sample containing 28 mg/mL DNA nanotube heterodimers.

Figure 20 shows the 2 H NMR spectra of D 2 O in liquid crystalline DNA nanotubes recorded at 30 degrees C and 1 H frequency of 500 MHz. In panel (A), the 2 H quadrupolar coupling constant was 6.7Hz for liquid-crystalline DNA nanotube at a concentration of 30 mg/ml in 50 mM HEPES, 50 mM NaCl, 10 mM MgCl 2 , pH 7.5. In panel (B), lOOrnM LMPG detergent was added, yielding a coupling constant of 5.1 Hz (consistent with mere dilution of the D 2 O). Panel (C) shows that no change in the coupling constant was observed 24 hours after addition of the detergent.

Figure 21 shows analysis of DNA nanotubes. Figure 21a shows native

agarose gel electrophoresis.1, 1 kb ladder; II, naked 7308 base scaffold; III, folded DNA nanotube. Figure 21b shows negative-stain electron micrograph of 200 nm DNA nanotube; scale bar is 50 ran. Figure 21c show negative-stain electron micrograph of 400 nm DNA nanotube, scale bar is 200 nm. Figures 22 a-f show negative-stain electron micrographs of DNA monomers and heterodimers. Figure 22a shows a DNA front monomer at 68000X magnification. Figure 22b shows a DNA nanotube heterodimer at 49000X magnification. Figure 22c shows DNA nanotube front monomers at 23000X magnification. Figure 22d shows DNA nanotube heterodimers at 185OOX magnification. Figure 22e shows DNA nanotube front monomers at 6800X magnification. Figure 22f shows DNA nanotube heterodimers at 6800X magnification.

Figure 23a shows an analysis of the residual dipolar couplings (RDCs) measured for the detergent-reconstituted transmembrane domain of the ζ chain of the T-cell receptor complex, weakly aligned in a 28 mg/mL DNA nanotube mixture.

Shown is a 0.98 correlation coefficient between the observed backbone RDCs and the RDCs predicted for the known NMR structure of the ζ-ζ transmembrane domain (2HAC) obtained from the Protein Data Bank.

Figure 23b shows the principal axes of the alignment tensor relative to 2HAC, the ζ-ζ dimer of the T cell receptor.

DETAILED DESCRIPTION Definitions

The term "nanotube" as used herein refers to a cylindrical arrangement of nucleic acid helices aligned in parallel and linked to one another, forming a tubular structure with approximate radial symmetry around a central axis.

The term "scaffold nucleic acid" as used herein refers to a single-stranded nucleic acid that is able to fold into various conformations through the complementary binding of shorter single-stranded nucleic acids (staple oligonucleotides) to non- contiguous segments of the longer nucleic acid.

The term "staple oligonucleotide" as used herein refers to a single-stranded oligonucleotide with successive segments that are complementary to non-contiguous segments of a scaffold, each scaffold segment forming part of a different helix in a

nucleic acid nanotube. As used herein, the term "staple" refers to staple oligonucleotide.

The terra "crossover" as used herein refers to the point at which a staple oligonucleotide crosses over from a binding site on one helix to a binding site on an adjacent helix in a nucleic acid nanotube. A crossover comprises either a covalent bond joining atoms in adjacent helices or a chemical group which is covalently linked to atoms in adjacent helices. The chemical group can be, for example, a phosphate group which forms part of the nucleic acid backbone of a staple oligonucleotide.

The term "seam" as used herein refers to the point at which a scaffold nucleic acid crosses from one helix to an adjacent helix. A seam comprises either a covalent bond joining atoms in adjacent helices or a chemical group which is covalently linked to atoms in adjacent helices. The chemical group can be, for example, a phosphate group which forms part of the nucleic acid backbone of a scaffold nucleic acid.

Description

The inventors have discovered how to make nucleic acid nanotubes of a uniform length that will self-assemble into liquid crystals. A solution comprising liquid crystalline nucleic acid nanotubes is resistant to detergent and enables liquid- crystal NMR spectroscopy of membrane proteins solubilized in detergent. Rod-like molecules are more likely to self-assemble into liquid crystals if they have large aspect ratios (length-to-cross-section diameter) and if they are homogeneous in length.

Nucleic acid nanotubes have been prepared using DNA origami techniques. These nanotubes can form detergent-resistant liquid crystals that make possible the accurate measurement of NMR residual dipolar couplings (RDCs) for a wide array of detergent-solubilized proteins. Acquisition of RDCs, which encode global orientation constraints, facilitates the de novo NMR structure determination of polytopic alpha- helical membrane protein monomers larger than 15 kDa in size. The previous size limit for solution-NMR-based de novo structure determination of membrane proteins can be extended by employing liquid-crystalline nucleic acid nanotubes to facilitate the accurate measurement of residual dipolar couplings, from which global orientation information can be derived. Thus these nanotube liquid crystals have made feasible

the structure determination of a wide range of biomedically important targets that currently are very difficult to characterize.

Multi-helix bundle nucleic acid nanotubes were prepared by adapting the scaffolded DNA origami technique described by Rothemund (Rothemund, P. W., J. Biomol. Struct. Dyns. 22, addendum, 2005; and Rothemund, P.W., Nature 440, 297- 302, 2006; both hereby incorporated by reference in their entireties). In one embodiment, the origami technique was used to create a six-helix bundle DNA- nanotube architecture similar to that described by Mathieu et. al (Nano Lett. 5, 661-5, 2005). The scaffold DNA used to construct the multi-bundle DNA nanotubes can be one or more long single-stranded DNA molecules of known sequence. To the scaffold are added many short staple oligonucleotides with complementarity to at least two sections of the scaffold DNA, the staple oligonucleotides force the scaffold into the shape of an array of parallel double helices. This construction technique permits the construction of bundles comprising various numbers of helices, and of predictable lengths. The length of a bundle is determined by the length of the scaffold DNA, which is folded using the staple oligonucleotides into approximately equal smaller lengths of DNA helices that are linked to one another. The average length of the nanotubes is given by the length of the single-stranded scaffold strand divided by the number of double-stranded helices present in each nanotube. Each helix is linked to an adjacent helix by at least two of the staple oligonucleotides which cross over to an adjacent helix. The bundles are linked together in this manner to form a closed tube-like structure. In a preferred embodiment of this invention, six-helix bundle DNA nanotubes were assembled by combining a single-stranded scaffold DNA with a plurality of oligonucleotides complementary to segments of the scaffold, in a manner that causes the length of the DNA nanotube to be one-sixth of the length of a DNA double helix comprising the scaffold as one strand. Thus, adapting the DNA origami technique to the preparation of DNA nanotubes results in nanotubes of predictable and uniform length and aspect ratio.

The nanotubes of the present invention are particularly well-suited to forming liquid crystals useful in a variety of applications. Moreover, the efficiency of producing a liquid crystal nanotube solution is substantially improved with the present methods. Competition for binding to the scaffold is likely to select for those oligonucleotides with fewer defects, thus mitigating somewhat complications from

using chemically-synthesized oligonucleotides. This strategy allows for flexibility in the length of each double helix in the array, as well as in the angle of curvature between any three parallel helices. The result is a robust and facile method. This method does not require any sequence design for the scaffold, nor does it require purification of the oligonucleotides. Thus the amount of labor required for assembly of such structures is reduced greatly compared to previous methods, and the material costs are relatively low.

Sequence structure of the nucleic acid nanotubes The basic strategy of preparing DNA origami structures is described by

Rothemund (Rothemund, P. W., J. Biomol. Struct. Dyns. 22, addendum, 2005; and Rothemund, P.W., Nature 440, 297-302, 2006; both hereby incorporated by reference in their entireties). According to the invention, one or more scaffold nucleic acids are combined with a plurality of staple oligonucleotides whose sequences are chosen to form complementary base pairings with the scaffold strand(s), thereby causing the scaffold to fold into a framework which, together with the base-paired staple oligonucleotides, forms three or more double helices linked side to side (i.e., a multi- helix bundle) to form a nanotube.

The double helices comprising each component of a multi-helix bundle can be rendered as a sketch drawing, followed by conversion of the general structure into an Adobe Illustrator file that indicates the details of the spacing between scaffold crossovers and oligonucleotide crossovers. The minimum distance between scaffold crossovers and oligonucleotide crossovers on adjacent lines is about 10 base pairs. In a preferred embodiment, the DNA nanotube structure mimics the shape and size of PfI , a rod-like viral particle that is 6nm in diameter and 2 μm in length. Its structural rigidity and negative-charge surface density allow it to form a stable and useful liquid crystal at low concentrations. To achieve a PfI -like DNA structure, a six-helix bundle DNA-nanotube architecture can be adopted. This design resembles a parallel array of six double helices for which every set of three adjacent helices frames a dihedral angle of 120 degrees (Figure Ia and Id). Adjacent double helices are held together by Holliday-junction crossovers that occur every 42 base pairs (Figure If). For each monomer, a 7308-base, M13-derived single-stranded circle of DNA (New England Biolabs) is employed as a "scaffold" and 168 single strands of DNA of

length 42 bases, programmed with complementarity to three separate 14-base regions of the scaffold, are employed as staple oligonucleotides ("staples") (Figure If). The distance between scaffold crossovers and oligonucleotide crossovers on adjacent lines is 42 base pairs, which results in good scaffold folding kinetics and thermodynamics. The staples self-assemble with the scaffold into the shape of six parallel double helices curled into a tube.

Each pair of adjacent helices should have at least two crossovers in order to enforce parallelism between the helices. The distance between oligonucleotide crossovers along a given line must be an even number of half-turns, usually 32, 42 or 52 base pairs, leading to a pitch spacing of 10.7 base pairs, 10.5 base pairs, or 10.4 base pairs, respectively. The distance between scaffold and oligonucleotide crossovers must be an odd number of half-turns. For the 32 base pair spacing, this corresponds to distances of 16 + 16 and 5 +27 base pairs. For the 42 base pair spacing, this corresponds to distances of 16 + 26 and 5 + 37 base pairs. For the 52 base pair spacing, this corresponds to distances of 26 + 26, 16 + 36, and 5 + 47 base pairs. In one embodiment, a five-helix bundle structure requires a 108 degree angle between any three adjacent helices. At 10.8 base pairs per turn, 14 base pairs yields 1.30 turns, yielding a rotation along the helix of 360 + 108 degrees. In another embodiment, a seven-helix bundle structure requires a 128.5 degree angle between any three adjacent helices. At 10.33 base pairs per turn, 14 base pairs yields 1.1355 turns, which is a rotation along the helix of 360 + 128 degrees.

In a preferred embodiment, a six-helix bundle requires a 120 degree angle between any three adjacent helices. With 42 base pairs between crossovers, the average twist of the helix is 10.5 base pairs per turn. At 10.5 base pairs per turn, 14 base pairs yields 1.33 turns, which is a rotation along the helix of 360 + 120 degrees. There are 42 base pairs between co linear crossovers (crossovers to the same adjacent helix), and either 14 or 28 crossovers along any helix to either adjacent helix. This implementation of the six-helix bundle uses oligonucleotides that are all 42 bases long, and whose ends line up with the positions of oligonucleotide crossovers on adjacent lines. This positioning is favorable in that chemical moieties added to the ends of the oligonucleotides will extend out from the helix orthogonal to the convex surface of the six-helix bundle.

In a six-bundle DNA nanotube, there are six DNA helices, and the scaffold is divided into six virtual strands. The top and bottom virtual strands depicted in Figure 2 A are continuous fragments of the scaffold. The middle four virtual strands each are composed of two pieces of the scaffold strand, separated by the strand seam. A 7308 base scaffold DNA strand results in virtual strands that are 7308/6 or 1218 bases long. Using a 42 base pair structure for the oligonucleotides, there are 29 pseudo-repeats of the basic staple oligonucleotide structure. The scaffold structure in the preferred embodiment is as follows, with the numbers representing the relative values of the base pair positions:

0-1217: virtual strand 1

1218-1829: upstream component of virtual strand 2 1830-2455: downstream component of virtual strand 3 2456-3109: upstream component of virtual strand 4 3110-3725 : downstream component of virtual strand 5

3726-4943: virtual strand 6

4944-5545: upstream component of virtual strand 5 5546-6109: downstream component of virtual strand 4 6110-6701: upstream component of virtual strand 3 6702-7303: downstream component of virtual strand 2

For helix 1, the first oligonucleotide strand attachment starts at position 16 from the 5' end of the virtual strand 1 at the proximal end of the nanotube, and binds a 14-base section of virtual strand 1 with a 14-base section of virtual strand 2 and a 14-base section of virtual strand 6. For helix 2, the first oligonucleotide strand attachment starts at position 26 from the 5' end of virtual strand 2 at the distal end of the nanotube, and binds a 14-base section of virtual strand 1 to a 14-base section of virtual strand 2 and a 14 base section of virtual strand 3. For helix 3, the first oligonucleotide strand attachment starts at position 2 from the 5 1 end of virtual strand 3 at the proximal end of the nanotube, and binds a 14-base section of virtual strand 3 to a 14-base section of virtual strand 4 and a 14-base section of virtual strand 5. For helix 4, the first oligonucleotide strand attachment starts at position 40 from the 5' end of virtual strand 4 at the distal end of the nanotube, and binds a 14-base section of

virtual strand 4 to a 14-base section of virtual strand 5 and a 14-base section of virtual strand 6. Most of the staple oligonucleotides in this embodiment are 42 base pairs long and attach to three non-contiguous sections of the scaffold DNA to produce the appropriate folding to generate the six-helix bundle nanotube.

A monomer can be conceptualized as a series of 28 pseudo-repeat segments, each consisting of six parallel double helices that are 42 base pairs long, flanked by jagged overhangs on either end of the object (Figure Ib). Each segment can be conceptualized as a series of three subsegments, for which every double helix is 14 base pairs long (Figure If). Six of the twelve strands of a subsegment are provided by the scaffold strand, three are provided by one staple strand, and three by another staple strand. Adjacent subsegments are related by 120-degree screw pseudosymmetry. The scaffold generally does not cross over between helices, except for four times in the middle of each monomer to produce a "seam", and three times on each monomer end (Figure Ic).

DNA nanotube monomers can be multimerized using the appropriate design parameters. The inclusion of a seam in the design allows for the linkage of monomers in a head-to-tail fashion instead of in ahead-to-head fashion, as is evident from consideration of the polarity of the scaffold strand within each double helix (Figure 3e). Three extra staple strands block the head of the front monomer, and four extra staple strands block the tail of the rear monomer (Figure 3a and 3d). To facilitate heterodimerization, three extra staple strands with unpaired bases decorate the tail of the front monomer, and three extra staple strands with unpaired bases decorate the head of the rear monomer (Figure 3b and 3c).

The model is then converted into DNA sequences; this can be accomplished, for example, by coding performed by a Python program. An example is provided in Figure 4A-4D. The program performs the following tasks: (1) input the scaffold strand sequence; (2) break the scaffold strand sequence into virtual strands corresponding to each parallel double helix; (3) break each virtual strand into

complementary sequence tokens; and (4) generate the oligonucleotide sequences as catenated tokens.

For nanotube heterodimerization, a computer program can be written to generate staple strand sequences given the sequence of the scaffold (See Figures S3a - c). A first Python script can be used to generate front monomer core oligonucleotides and head caps (Figure 5A-5B). A second Python script can be used to generate rear monomer core oligonucleotides and tail caps (Figure 6A-6B). A third Python script can be used to generate front monomer tail connector oligonucleotides and rear head connector oligonucleotides (Figure 7A-7B). Using two cyclic permutations of the scaffold sequence as input to the program can generate independent sets of staple-strand sequences for folding two different monomer nanotubes. Therefore copies of the same scaffold molecule can be used to generate two chemically-distinct species.

It is understood that many possible sequence combinations exist which can give rise to a given nanotube structure. The initial choice of scaffold strand sequence will determine the sequences of the staple oligonucleotides. However, once a particular scaffold strand is chosen, any given point along the sequence of the scaffold strand can be chosen as a starting point to build the nanotube structure. The choice of scaffold sequence and starting point, together with the nanotube geometry and the number and position of crossovers and seams, will determine the sequences of the staple oligonucleotides. Furthermore, the sequences of staple oligonucleotides can be optimized in order to avoid unintended binding that can give rise to defective structures or poor assembly kinetics.

Length of the DNA nanotubes

The length of the DNA nanotubes is 50 πm or more. In one embodiment of the invention, the length of the DNA nanotubes is 200 run or more. In a preferred embodiment, the length of the DNA nanotubes is about 400 nm. A length of 400nm can be achieved, for example, with a scaffold DNA strand 7308 bases long folded into six strands to which complementary staple oligonucleotides are bound, forming six- helix bundles. Preferably, the length of the nanotubes varies by no more than 20% of the average length of the nanotubes, and more preferably by no more than 10 % of the

average length of the nanotubes. If a nanotube comprises helices of different lengths- then the length of the longest helix is considered the length of the nanotube.

The lengths of the nanotubes formed using this technique can also be modified through end-to-end multimerization of the bundle structures. In a preferred embodiment of the invention, head-to-head and tail-to-tail multimerization of the bundles can be generated from a scaffold DNA configuration in which the bends of the scaffold occur only at the ends of the bundles. In a more preferred embodiment of the invention, head-to-tail multimerization of the DNA bundles can be generated from a scaffold configuration in which some of the scaffold bends occur within the length of the DNA bundles, forming a seam across which the scaffold DNA does not cross. In a preferred embodiment, the virtual strands are connected to each other by the staple oligonucleotides in a staggered manner. Virtual strands 1 and 2 (see Figure 2B) each have 16 base pairs available for multimerization on the proximal end of the nanotube, and 26 base pairs available for multimerization on the distal end. Virtual strands 3 and 4 each have 2 base pairs available for multimerization on the proximal end of the nanotube, and 40 base pairs available for multimerization on the distal end. Virtual strands 5 and 6 each have 40 base pairs available for multimerization on the proximal end of the nanotube, and 2 base pairs available for multimerization on the distal end. Thus the connecting region for each multimerized strand is 42 bases long, maintaining a constant staple oligonucleotide length and allowing for head-to-tail multimerization of the DNA nanotubes.

Dimerization of the DNA nanotubes can be achieved, for example, as shown in 3. Figure 3 shows schematic views of the pre-dimerization monomers that can combine to form heterodimer DNA nanotubes. Specifically, Figures 3a-3d show the scaffold-plus-staples schematic views of the front and rear overhangs of the monomers. One strand of each double helix can be contributed by the scaffold (darker lines in Figures 3a-3e), and the other strand can be contributed by a staple oligonucleotide. Base pairs in the Figure are depicted as short vertical lines between the paired strands. Helices 1-6 are labeled in the center of Figures 3a-b and 3c-d. Figure 3a shows the front monomer head segment. Three staple strands can serve to cap the front monomer head (see DNA sequences, Figure 8). Figure 3b shows the front monomer tail segment, which has three staple strands (see DNA sequences, Figure 9) with a total of 26 unpaired bases decorating the tail (2 bases in helix 2, 12

bases each in helices 5 and 6). The scaffold strand in this region is unpaired for 36 bases (12 bases each in helix 1, 3, and 4). Figure 3c shows the rear monomer head segment. Three staple strands (see DNA sequences, Figure 10) on this portion of the monomer have a total of 36 unpaired bases decorating the head. These unpaired regions are complementary to the corresponding 36 unpaired bases of the front monomer tail scaffold strand. The 26 unpaired bases in the rear monomer head scaffold strand can be complementary to the 26 unpaired bases of the three staple strands that decorate the front monomer tail. In the DNA nanotube heterodimer, these unpaired regions can match up to form the complete intermonomer junction, as shown in Figure 3e. Figure 3d shows the rear monomer tail segment. Four staple strands can serve to cap the rear monomer tail (see DNA sequences, Figure 11). Figure 3e shows the junction between the head and tail monomers forming the assembled heterodimer. The scaffold crossovers (darker vertical lines) that form an internal seam for each monomer occur at segments 14 and 15, as shown in Figure Ib and c. In the nucleic acid nanotubes of the inventions, the scaffold strand is arranged with base pair sequences optimized to avoid unintended binding events between staple strands and the scaffold strand, or between different sections of the scaffold strand. The scaffold strand can be derived from a natural source whose base pair sequences have been completely characterized. In one embodiment, the scaffold strand is derived from the M13mpl8 viral genome, which is well-characterized and relatively inexpensive to generate in large quantities. It is also amenable to recombinant approaches to insert or delete sequences. The scaffold strand can also be an entirely artificial sequence, a modified natural sequence, or any combination of natural and artificial sequences. In another embodiment, plasmids based on the pBluescript vector can be used where a shorter, exact number of bases is desired. This avoids having extra unscaffolded material that may interfere with folding of the scaffold. With pBluescript, there is more flexibility with inserting DNA's that are many kilobases in size, without concern about plasmid instability. To facilitate the excision of a single- strand DNA target insert from the generic vector, inverted repeat restriction sites can be introduced into the vector. Inverted repeat EcoR I sites separated by 20 base pairs can be added upstream of the target sequence. Inverted repeat Hind III sites separated by 20 base pairs can also be added downstream of the target sequence. In the single-

stranded DNA, the repeated sites fold up to form double-stranded sites that are recognizable by the appropriate restriction enzyme.

Assembly of the nanotubes The nucleic acid nanotubes of the present invention are self assembling. The scaffold strand and a molar excess of staple oligonucleotides are added to a desired buffer, which preferably contains MgCl 2 . The solution is heated to a temperature sufficient to denature all the nucleic acids contained therein {e.g., 9O 0 C), and then slowly allowed to cool. The step of cooling should be such that the solution returns to a temperature, e.g., room temperature, which permits assembly of the nanotubes over approximately 1 hour to 24 hours, e.g., over 2 hours or 20 hours.

Example 1

Preparation of DNA nanotubes M13 single stranded scaffold DNA (sequence shown in Figure 12 A-C) was obtained from phage produced from infected F+ bacteria grown in 2xYT media.

Purified single-stranded DNA was extracted from the phage using a Qiagen Gigaprep ion-exchange column. Six-helix bundle DNA nanotubes were folded directly from the eluate of a Qiagen Gigaprep ion-exchange column, eluted at 50 mM Tris pH 8.5 (Fisher Scientific), 1.6 M NaCl (Fisher Scientific), 15% isopropanol. In the folding reaction, the buffer was diluted to 1 M NaCl, 9% isopropanol, along with 50 nM of the organic chemical buffer HEPES (4-2-hydroxyethyl-l-piperazineethanesulfonic acid) pH 7.5 (Sigma), 10 mM MgCl 2 (Fisher Scientific). The scaffold concentration was at 6 nanomolar and the staple oligonucleotide (sequences shown in Figure 13A- D) concentrations were at 36 nM each. The isopropanol did not interfere with the folding. Folding was performed by heating the suspension in 100 mL Pyrex bottles in 2 L boiling water baths to 90° C, then covering the lid and allowing to cool to room temperature over the course of 20 hours.

The folded six-helix bundle DNA nanotubes thus formed were separated from the excess oligonucleotides by precipitation with 40% ethanol. The DNA nanotubes survived desalting with a 75% ethanol wash followed by dehydration in a speedvac.

After drying, the DNA nanotubes were resuspended in a desired volume of buffer, without evidence of aggregation or other misfolding.

Example 2

Recombinant Ml 3 bacteriophage plasmid (p7308) construction

Recombinant Ml 3 filamentous bacteriophage was prepared by replacement of the BamHI-Xbal segment of M 13mp 18 by a polymerase chain reaction-generated 59 base pair (bp) fragment encoding a randomly-selected sequence (Figure 14), flanked by positions -25 to +25 of the middle of the Xbal cut site(T A CTAGA, or base 6258). A list of oligodeoxyribonucleotides that were used to construct the insert with flanking regions (109 bp total) is shown in Figure 15. Double-stranded (replicative form) bacteriophage M13 DNA bearing the 59 base insert was prepared as described in Sambrook, J. & Russell, D. Molecular cloning: A Laboratory Manual (Cold Spring Harbor Laboratory Press, Cold Spring Harbor, 2001). The 59 bp insert was verified by a double restriction digest with BamHI and Xbal, followed by sequencing. The result was a modified bacteriophage Ml 3 genome, 7308 bases in length. The full sequence is shown in Figure 16.

Example 3

Nanomole-scale production of M13 bacteriophage single-stranded DNA

Recombinant Ml 3 bacteriophage RF dsDNA was transformed into JMlOl cells and grown overnight at 37°C on an LB-agar plate (BD Diagnostics). A single, well-isolated plaque was used to inoculate 2 ml of 2xYT medium in a 14 mL sterile culture tube and agitated for 8 hours at 37°C. Bacterial cells were pelleted by centrifugation and phage was recovered from the supernatant by polyethylene glycol fractionation (incubation on ice for 30 minutes using a final concentration of 4% PEG8000, 0.5 M NaCl) followed by centrifugation. The phage was resuspended in lOOμL of 10 mM Tris-Cl pH 8.5 (Fisher Scientific) and labelled "pre-inoculation phage." E CoIi JM109 cells were grown overnight in 3 mL of 2xYT medium at 37°C. The 3 mL of JMl 09 culture was added to a 2 L flask containing 300 mL 2xYT medium supplemented with MgCl 2 to 5 mM final concentration and incubated at 37°C on a shaker at 300 rpm. When the bacterial culture reached Aβoo == 0.5, 50 μL of the "pre-inoculation phage" stock was added. The infected culture was grown at 37 0 C, shaking at 300 rpm for an additional 4 hours. Phage was recovered as described above, and resuspended in 3 mL 10 mM Tris-Cl pH 8.5 and labelled "inoculation

phage." Titer of "inoculation phage" was measured by plating out serial dilutions using saturated JMl 09 culture and LB-top agar plates. Titer of JMl 09 cells at A600 = 0.5 was measured by plating out serial dilutions on LB-agar plates. For nanomole- scale production of phage, twelve 2 L flasks each containing 30OmL 2xYT medium supplemented with 5 mM MgCl 2 , were inoculated with 3 mL overnight JMl 09 culture and incubated at 37°C shaking at 300 rpm. When density reached Agoo = 0.5, each flask was infected with "inoculation phage" at an MOI = 1. Phage was harvested as described, and resuspended in 0.5% of the original culture volume in 10 mM Tris-Cl pH 8.5. Single-stranded DNA was isolated from phage by alkaline/detergent denaturation as follows: Two volumes of lysis buffer (0.2 M NaOH, 1% SDS) were added to the resuspended phage, followed by 1.5 volumes neutralization buffer (3 M KOAc pH 5.5). Lysed phage was centrifuged for 10 minutes at 16000 rcf. The supernatant was combined with one volume of 200 proof ethanol and centrifuged for 10 minutes at 16000 rcf. Pelleted ssDNA was washed twice with 75% ethanol, centrifuged, and resuspended in 5% of the original culture volume in 10 mM Tris-Cl pH 8.5. The concentration of the recovered ssDNA was estimated on a UWvisible spectrophotometer (Beckman coulter) using an extinction coefficient = 37.5 μg/mL

Example 4

Preparation of 800 nm DNA nanotube heterodimers

A detergent-resistant liquid crystal of 0.8 μm-long DNA nanotubes has been designed and constructed, and has been shown to induce weak alignment of membrane proteins. The nanotubes are heterodimers of 0.4-μm-long six-helix bundles each self-assembled from a 7.3-kilobase scaffold strand and over 170 short oligonucleotide staple strands. Desalted DNA oligonucleotides, normalized by concentrations to 150 uM, were purchased from Invitrogen (see DNA sequences of Figures 8-11, 17, 18). Equal volumes of each staple oligonucleotide were pooled into groups: front monomer staple stock (DNA sequences of Figures 8, 9, 17) and rear monomer staple stock (DNA sequences of Figures 10, 11, 18). The front monomer staple stock includes front head cap staples (DNA sequences of Figure 8), front core staples (DNA sequences of Figure 17), and front tail connector staples (DNA sequences of Figure 9). The rear monomer staple stock includes rear head connector

staples (DNA sequences of Figure 10), rear core staples (DNA sequences of Figure 18), and rear tail cap staples (DNA sequences of Figure 11). Concentrations of the pooled staple stocks were estimated on a UV/visϊble spectrophotometer using an extinction coefficient = 33 μg/mL for A 26 o = 1- Front and rear monomers were prepared with their respective staple stocks, but otherwise using an identical protocol. The front monomer folding mixture was prepared by combining p7308 ssDNA (3OnM), front monomer staple stock (300 nM each staple strand), 50 mM HEPES pH 7.5, 50 mM NaCl, and 30 mM MgCl 2 in a final volume of 76.8 mL. The folding mixture was aliquoted into four 96-well plates (Molecular BioProducts) (200 μL per well), and folded on a thermal cycler (MJ Research Tetrad) with the following program:

1. 80 0 C for 5:00

2. 80 0 C for 2:00 (-1 0 C per cycle)

3. Go to 2, 60 times 4. End

Folded material was pooled into a 250 mL polypropylene centrifuge bottle. Folded nanotubes were separated from excess staple strands via PEG fractionation as follows: 19.2 mL of 20% PEG8000 (Fisher Scientific), 2.5 M NaCl was added to mixture, which was then centrifixged at 15000 rcf for 15 minutes. The supernatant was discarded, and the nanotube pellet was resuspeπded in 38.4 mL 50 mM HEPES pH 7.5, 50 mM NaCl, and 30 mM MgCl 2 . A second PEG fractionation was carried out as follows: 9.6 mL of 20% PEG8000, 2.5 M NaCl was added to mixture, which was then centrifuged at 15000 rcf for 15 minutes. The supernatant was discarded, and the nanotube pellet was resuspended in 38.4 ML 50 mM HEPES pH 7.5, 50 mM NaCl, and 30 mM MgCl 2 .

Nanotube heterodimers were formed by combining front and rear monomer mixtures together and incubating at 37°C for 2 hours. Two volumes of equilibration buffer (75OmM NaCl, 50 mM MOPS, pH 7.0, 15% isopropanol, 0.15% Triton X-100 (v/v)) were added to the mixture. Heterodimers were loaded on a Qiagen-Tip 10000 gravity-flow ion-exchange column equilibrated with 75 mL equilibration buffer. The column was washed with six 100 mL volumes of 1 M NaCl, 50 mM MOPS, pH 7.0,

15% isopropanol (v/v). Then the nanotubes were precipitated by addition of one volume 200 proof ethanol, centrifuged at 15000 rcf for 15 minutes, washed twice with 75% ethanol, and resuspended in 3 mL ImM NaH 2 PO 4 pH 7.0, 1 mM MgCl 2 - The nanotube concentration was estimated via UV absorbance at 260 nM assuming an extinction coefficient OfA 26 O = 1 for 50μg/ml. The nanotube heterodimer mixture was then concentrated by Speedvac vacuum centrifugation to a final volume of 300μL. Front and rear monomers were folded in separate chambers via heat denaturation followed by cooling for renaturation.

The front and rear monomers were mixed to self-assemble heterodimers (Figure la-c, and Ie). The joining of the tail of the front monomer to the head of the rear monomer should generate a 42-base-pair pseudo-repeat segment (Figure Ie). In this segment, all six staple strands bridge the two scaffolds, although by varying numbers of base pairs. In total, a net 62 base pairs must be broken to sever the linkage between successfully heterodimerized monomers.

Example 5

Demonstration of the liquid crvstallinity of DNA nanotubes

Six helix bundle DNA nanotubes from Example 1 were incubated in 25% ethanol, causing selective precipitation of the nanotubes, and leaving behind the excess unbound staple oligonucleotides. Nine milligrams of the DNA nanotubes were resuspended in a volume of 6 mL 2.5 mM HEPES pH 7.5, 2.5 mM NaCl, 0.5 mM MgCl 2 , and then dehydrated in a Savant speedvac concentration system to achieve a final concentration of 30 mg/ml (300 μL of a 3% suspension) and final buffer concentrations of 50 mM HEPES pH 7.5, 50 mM NaCl, and 10 mM MgCl 2 . The liquid crystallinity of the DNA nanotube suspension in an NMR tube was verified by observation of birefringence under crossed polarizers. A low-salt, aqueous suspension of DNA-nanotube heterodimers at a concentration of 28 mg mL "1 forms a stable liquid crystal, as indicated by strong birefringence observed through crossed polarizers, as shown in Figure 19c. (Figure 19c). The liquid crystals were diluted by 10% with deuterated water, and were aligned for three hours in a 600 MHz NMR spectrometer. Strong birefringence was observed when the sample tube was placed at 45 degrees to the crossed polarizers.

Further evidence for liquid crystallinity of the DNA nanotube monomers was obtained by NMR spectroscopy, measuring quadrupolar splitting of the deuterium, where a coupling constant of 6.7 Hz was observed (Figure 20A). Next, 1-myristoyl- 2-hydroxy-sn-glycero-3-[phospho-RAC-(l-glycerol)] (LMPG) detergent was added to 100 mM. After addition of the detergent, the coupling constant dropped to 5.1 Hz, which is consistent with the 16% dilution of D 2 O in the sample upon addition of the detergent suspension. The liquid crystals remain stable over at least 24 hours in the presence of the detergent (Figure 20C).

When the suspension of 800 nanometer heterodimers is aligned in an 11.4 Tesla magnetic field in the presence of 10% D 2 O, the weakly-oriented HDO yields 2 H quacirupolar splitting of 5.56 Hz (Figure 19d). The ID 2 H spectrum shown in Figure 19d was obtained from a 10 mM NaH 2 PO 4 , 10 mM MgCl 2 , 90% H 2 O/10%D 2 O sample containing 28 mg/mL DNA nanotube heterodimers. NMR spectra were processed and analyzed using NMRPipe. Fitting of the dipolar couplings to the known ζ-ζ homodimer structure was done by singular value decomposition (SVD), using the program PALES. The goodness of fit was assessed by both Pearson correlation coefficient (r) and the quality factor (Q).

Example 6 Characterization of DNA nanotubes

Folded DNA nanotubes were analyzed using agarose gel electrophoresis and negative-stain electron microscopy using uranyl formate (Pfaltz & Bauer) as the stain. Gel electrophoresis experiments indicated that the majority of scaffold molecules are folded as monomers, as they produced a single band upon agarose gel electrophoresis in the presence of 10 mM MgCl 2 (Figure 21a). Further analysis of folding and heterodimerization of DNA nanotubes was conducted via electrophoresis in a 2% agarose gel containing HmM MgCl 2 , 0.5 μG /mL ethidium bromide, 45 mM Tris base, 45 mM boric acid, and 1 mM EDTA (pH 8.0), and is shown in Figure 19a. The majority of DNA objects migrate as a single band in agarose-gel electrophoresis (Figure 19a). This population presumably represents well-formed nanotube monomers, while slower migrating species apparent on the gel presumably represent

misfolded or multimerized structures. Lane M is the marker lane with DNA size standards denoted by number of base pairs shown to the left of the lane. Lane 1 shows the M13-derived single-stranded DNA scaffold. Lanes 2 and 3 show the front and rear DNA monomers (including scaffold plus staples). Lanes 4 and 5 show the front and rear monomers after PEG fractionation. Agarose-gel electrophoresis of heterodimers assembled from the two monomers indicates that the majority of DNA objects migrate as a single band (Figure 19a, Lane 6), although some misfolded objects are evident, as are a small population of monomeric nanotubes. Lane 6 shows the heterodimers after a two-hour incubation of mixed monomers at 37°C. Electron micrograph analysis was carried out using Image SXM. The lengths of 20 well-isolated DNA nanotube monomers and 20 well-isolated DNA nanotube dimers in several separate electron micrographs were measured manually using the segmented-line tool. Following the dimerization step, DNA nanotube dimers were diluted to InM concentration and prepared for imaging by negative stain with 0.7% uranyl formate (Pfaltz & Bauer) as previously described. (Obi, M., Cheng, Y., WaIz, T. Biol. Proc. Online 6, 23-24 (2004)). Gilder Fine BarGrids, 400 mesh, 3.05 mm O.D. (Ted Pella) were used. Imaging was performed on a Tecnai G 2 Spirit BioTWIN.

Electron microscopy experiments (Figures 21b, 21c) showed that the DNA nanotubes are much more rigid than double helices. If double helices are assumed to be 2 nm wide and 0.34 ran per basepair, then the predicted width would be 6 nm, and the predicted length would be either 200 nm or 414 nm. The length and width of the imaged objects approximately matched the predicted dimensions. The DNA nanotube heterodimer mixture was also analyzed using negative-stain electron microscopy, and the results are consistent with a large fraction of intact nanotubes of length 402 ± 6 nanometers (Figure S2, a, c, and e). This measured length is in good agreement with the predicted length of 400 nanometres for 28 segments that are 42 base pairs long, assuming a rise of 0.34 nanometers per base pair. Negative-stain electron microscopy also revealed nanotubes of length 813 ± 9 nanometers, as shown in Figure 19b (scale bar = 500 nanometers) and Figure S2, b, d, and f). This measured length agrees well with the predicted length of 814 nanometers for 57 segments that are 42 base pairs long.

Example 7

Solution NMR methodology for membrane-protein structural determination

Membrane proteins play important roles in cell biology and medicine. For example, over half of hormones and neurotransmitters studied to date transduce signals through members of the G-Protein Coupled Receptor (GPCR) family of membrane proteins. Similarly, over half of all commercial drugs target GPCR's. Despite their importance, structures of only ~100 membranes proteins have been solved to date (http://blanco.biomol.uci.edu/Menibrane_Proteins_xtal.html holds a tally that is updated regularly). The slow rate of membrane-protein structure determination represents a significant bottleneck for both basic and applied bioscience discovery. This bottleneck largely derives from difficulties in forming well-ordered three-dimensional crystals of membrane proteins (Caffey M, Membrane protein crystallization, J. Struct. Biol. 142, 108-132, 2003). Solution NMR presents a potentially-attractive alternative for the study of many membrane proteins, as high- resolution structural information can be obtained for systems up to 80 kD in size without the need for crystallization.

Solution NMR has advanced to the point where structure determination of 30- kD water-soluble proteins has become routine. This has not been the case, however. for membrane proteins. For solution NMR, macromolecules must be solubilized in water to facilitate fast tumbling; the faster the tumbling, the better the spectra. To promote water solubility, membrane proteins must be complexed with detergent micelles. The micelle-protein complex is considerably larger than the protein alone, and tumbling is relatively slow as a result. This increase in effective size is especially problematic for a-helical membrane proteins greater than 15 kDa in size, where resonance peaks are closely spaced and become unresolvable with the fast coherence relaxation of the slowly-tumbling macromolecules. Some of the larger alpha-helical membrane proteins whose structures has been solved by solution NMR include the Mistic membrane-surface-associating protein (13 kDa) (Roosild TP, Greenwald J, Vega M, Castronovo S, Riek R 5 Choe S, NMR structure of Mistic, a membrane- integrating protein for membrane protein expression, Science 307, 1317—1321, 2005) and subunit c of the ATP synthase (7 kDa) (Girvin ME, Rastogi VK, Abildgaard F, Markley JL, Fillingame RH, Solution structure of the transmembrane H+-transporting

subunit c of the FIFO ATP synthase, Biochemistry 37, 8817-8824, 1998). Recently, our collaborators in Dr. Chou's laboratory have used solution NMR for the de novo structure determination of the phospholamban pentamer, a 30-kD channel-like protein that spans the sarcoplasmic reticulum membrane (Oxenoid K, Chou JJ, The structure of phospholamban pentamer reveals a channel-like architecture in membranes, Proc Nat'l Acad Sci U S A 102, 10870-10875, 2005). In that case, however, the NMR spectra were simplified because of the five-fold rotational symmetry in the complex. For conventional NMR spectroscopy, the Nuclear Overhauser Effect (NOE) provides the only experimentally-measurable distance restraint for tertiary structure determination. Successful structure determination requires the correct assignment of most of the proton resonances, a demand that can be almost impossible to meet for poorly-resolved spectra such as those recorded for a-helical membrane proteins. Furthermore, NOE's only are detectable for distances shorter than five angstroms, thus determination of the global shape of extended proteins is subject to compounded errors.

RDCs encode global orientatioπal constraints that enable structure determination with only limited NOE assignments required. If a large number of accurate RDCs can be measured, then a full analysis of the NOESY spectra — which may in practice be unobtainable — becomes unnecessary. In this case, it will be sufficient to measure NOE's after selective labeling of amino acids, which simplifies the spectrum, or to measure semi-quantitative distance constraints from paramagnetic- broadening techniques.

Residual dipolar coupling leads to informative resonance frequency splitting. In the presence of an external field B that points in the z-direction, the z-component of the magnetic field from nucleus S will change the magnetic field at I such that the resonance frequency of I will shift by a quantity that depends on the internuclear distance and on the internuclear angle with respect to the z-axis. If the protein is undergoing rapid isotropic rumbling, then the average perturbation averages to zero. In order to obtain information about the internuclear angles, then, each protein must be made to tumble in a weakly-ordered regime. Too much ordering and dipolar couplings become so strong that peaks are unresolvable, while too little ordering leads to undetectable levels of dipolar coupling. The appropriate weak ordering, about

0.1%, can be achieved by dissolving the protein in the right concentration of a suitable alignment material. For example, water-soluble proteins can be aligned weakly by the required amount with -1.5-2% PfI filamentous phage, which forms a liquid crystal at that concentration. Membrane proteins can be weakly aligned. The easiest method for weak- alignment of proteins is through mixing the protein with a liquid-crystalline medium, such as PfI filamentous phage, DMPC/DHPC bicelles, C12E5 polyethylene glycol, or cellulose crystallites. However, none of these media are compatible with detergent- solubilized membrane proteins. The only method currently available for weak alignment of membrane proteins involves the use of radially-compressed polyacrylamide gels (Oxenoid K, Chou JJ, The structure of phospholamban pentamer reveals a channel-like architecture in membranes, Proc Nat'l Acad Sci U S A 102, 10870-10875, 2005; Chou JJ, Gaemers S, Howder B, Louis JM, Bax A, A simple apparatus for generating stretched polyacrylamide gels, yielding uniform alignment of proteins and detergent micelles, J Biomol NMR 21, 377-382, 2001; Chou JJ,

Kaufman JD, Stahl SJ, Wingfield PT 5 Bax A, Micelle-induced curvature in a water- insoluble HIV-I Env peptide revealed by NMR dipolar coupling measurement in stretched polyacrylamide gel, J Am Chem. Soc 20, 2450-2451, 2002; Tycko R, Solid- state NMR as a probe of amyloid fibril structure, Curr Opin Chem. Biol 4, 500-506, 2000). A technical problem encountered during the weak alignment of phospholamban was that the maximum protein concentration obtainable in the gel was —0.2 mM, despite soaking in a solution with a protein concentration of 1—2 mM. Because of the low concentration, the signal-to-noise ratio of the NMR signals was low. Ixmg data acquisition times were required, and the resultant RDC measurements were of limited accuracy. The difficulty in soaking detergent-solubilized membrane proteins into polyacrylamide gels is a well-known problem in the NMR community. The six-helix bundle DNA nanotubes described herein represent a detergent- resistant shape mimetic of the PfI filamentous phage. These DNA nanotubes have similar liquid-crystalline behavior as PfI, but are completely resistant to strong detergents such as SDS.

Example S

Use of DNA nanotube liquid crystal to measure backbone RDCs for the transmembrane domain of the T-cell receptor

All NMR experiments were performed on Bruker spectrometers equipped with cryogenic TXI probes at 30 0 C. The RDCs were obtained from subtracting J or J+D couplings of the aligned sample from that of unaligned sample. The 1 H- 15 N RDCs were obtained from VNH / 2 and ( 1 JNH + 12, which were measured at 600 MHz ( 1 H frequency) by interleaving a regular gradient-enhanced HSQC and a gradient- selected TROSY, both acquired with 80 ms of 15 N evolution. The 1H α - 13 C α RDCs ('.DcαHα) were measured at 500 MHz ( 1 H frequency) using a 2D CACONH quantitative 1 J CaHa experiment with interleaved spectra recorded at 1 J Ox H a modulation times of 1.83, 3.63, and 7.12 ms. This experiment was modified from the 3D CBCACONH quantitative J CH experiment 30 used primarily for measuring protein side-chain 1Hβ- 13 Cp RDCs. The CACONH was optimized for measuring the backbone 1H α - 13 C α RDCs only. Since the ζ-ζ transmembrane (TM) domain is a homodimer obeying two -fold rotational symmetry, the same RDCs are assigned to both sύbunits. The frequency labeled dimensions in this experiment are 1H N (direct) and 15 N (indirect).

The DNA-nanotube liquid crystal enables the accurate measurement of backbone N H and C α H α RDCs for the detergent-reconstituted ζ-ζ transmembrane domain of the T-cell receptor. The measured RDCs validate the high-resolution structure of this transmembrane dimer. The DNA heterodimer nanotubes were tested for weak alignment of the transmembrane (TM) domain (residue 7-39) of the ζ-ζ chain of the T-cell receptor complex reconstituted in mixed dodecylphosphocholine (DPC) / sodium dodecyl sulfate (SDS) detergent micelles. The measured 1 H- 15 N and 1 H 01 - 13 Ca RDCs agree very well with the known NMR structure of the ζ-ζ TM domain, with a correlation coefficient of the Singular Value Decomposition (SVD) fit, RSVD > of 0.98, or a free quality factor, Qf ree , of 16% (Figure 23a). The magnitude of the alignment tensor, D a , is -9.9 Hz (normalized to ∑hm), which is ideal for RDC measurement and structure calculation. Ih addition, the axis OfC 2 rotational symmetry of ζ-ζ is parallel to the largest principal axis, A 22 , of the alignment tensor (Figure

23b). This result is expected from the rotational averaging of the dimeric complex around its C 2 axis in the aligned medium.