6
Biomacromolecules 2008, 9, 6–8
Macroscopic DNA Helices Juan A. Subirana,* Marc Creixell, Roberto Baldini, and J. Lourdes Campos* Department of Chemical Engineering, E.T.S.E.I.B. Universitat Politècnica de Catalunya, Av. Diagonal 647, 08028 Barcelona, Spain Received September 6, 2007; Revised Manuscript Received October 12, 2007
Presently there is great interest in the construction of nanostructures from DNA fragments. Here we report the preparation of much larger helical textures with different shapes from pure DNA fragments. We have observed them while preparing crystals of the dodecamer d(AAAAAATTTTTT), which only contains adenine and thymine. Noncoding regions of the genome are always rich in these two bases. We have found a strong influence of ions either monovalent or divalent, with which different crystalline structures are found. All of them contain long helical stacks of duplexes in the unit cell. The most remarkable structures are macroscopic helices with diameters and pitch in the range of 20–40 µm. Thus, pure DNA oligonucleotides may form a hierarchy of helical structures, going from the B-form double helix (pitch, p ) 33 Å) to helical stacks of duplexes (p ≈ 900 Å), and to macroscopic helices (p ≈ 300,000 Å). These different levels of organization are reminiscent of the different levels of organization of DNA in eukaryotic chromosomes. All-AT oligonucleotide duplexes tend to form crystals in which individual duplexes are stacked in columns. Such columns show a helical arrangement of stacked oligonucleotides (HASO structure), as described in detail elsewhere.1,2 In the crystal, the columns are parallel, with a large number of duplexes in the asymmetric unit. Stacking of duplexes is determined by the specific features of the virtual TA base step, which is found between neighbor duplexes in the stack, which has a negative twist of about -22°. In the case of all-AT dodecamers, the overall twist Ω between neighbor stacked oligonucleotide duplexes should be close to 36° × 11 – 22° ) 374°, equivalent to Ω ) 14°. For example, if Ω ) 13.8°, there should be 360/ 13.8 ) 26 dodecamers in the asymmetric unit, which generates a helical repeat. Small changes in Ω result in large changes of pitch.1 Because of this low value of Ω, dodecamers turn out to be a special case in the structures found in all-AT duplexes. They have very large unit cells in the direction of the helical stack, but comparatively small values in the perpendicular direction: only a few parallel helical stacks, usually one to three, are found in the unit cell. The structure of the crystals depends on the presence of divalent cations. When only monovalent cations are used, local packing in the crystals is similar, but no macroscopic helices have been found. Their structure will be described elsewhere. Presently, there is great interest in the construction of nanostructures from DNA fragments, as reviewed by different authors.3–6 Here we report the preparation of much larger helical textures with different shapes from pure DNA fragments. A few examples are given in Figure 1 and in the Supporting Information. We have observed them while preparing crystals of dodecamers that only contain adenine and thymine. Noncoding regions of the genome are always rich in these two bases. We should mention that helical crystals and related structures have also been described in other systems,7–11 including biological objects such as sperm heads.12 Their diameter is usually much smaller (0.1–1.0 µm) than that in the helices shown in Figure * Corresponding author. Phone: (+34) 934 016 688. Fax: (+34) 934 010 978. E-mail:
[email protected] (J.A.S.);
[email protected] (J.L.C.).
Figure 1. A gallery of crystals. (a) View of a crystallization drop demonstrating the variable appearance of the crystals. The approximately hexagonal crystal in the central region (arrow) gave rise to a twinned diffraction pattern, which corresponds to the monoclinic modified HASO structure described in the text. (b and c) Typical right handed corkscrew crystals obtained with NiCl2. Other divalent cations gave similar shapes. The rod-like crystals (arrows) with blunt ends apparent in c diffract as typical HASO structures.1 In c, a second helix growing on top of the original corkscrew can be seen in the left half of the structure. (d) A left handed cable-like structure, which is some times observed in early stages of crystallization, in this case with Sr2+. Additional images are given in Supporting Information.
1. In most cases, no clear explanation of their helical origin has been provided. In this communication, we present the results obtained with the dodecamer d(AAAAAATTTTTT). The oligonucleotide was prepared by the phosphoramidite method. Crystals were obtained by vapor diffusion in hanging drops of 6–8 µL. Typically, the drops contained 0.4 mM oligonucleotide duplex, 0.3–0.5 mM spermine chloride, 7.5 mM M+2Cl2, and 25–50 mM Na cacodylate buffer at pH 6.5. They were equilibrated against a well with 10% 2-methyl-2,4-pentanodiol (MPD), at 13-20 °C. The concentration of MPD was increased with time up to 20–40%. Different divalent cations were used. At first, thin helical threads were observed, which grew in thickness with time. Often, it was found that the crystals acted as nucleation points for new crystals to grow (Figure 1c). Additional images
10.1021/bm700994v CCC: $40.75 2008 American Chemical Society Published on Web 12/15/2007
Communications
Biomacromolecules, Vol. 9, No. 1, 2008 7
Figure 2. Diffraction patterns. (a) A 10° rotation diffraction image from the monoclinic structure. Bragg spots are visible up to 6 Å resolution. A reflection due to base-pair stacking is also present in the diagrams, with a spacing of 3.3 Å (not shown). The pattern presented corresponds approximately to the b*c* plane, with the b* axis horizontal and oriented along the spindle axis. In this case, twinning is very limited. Although some inclined rows of spots can be detected in the lower left part of the image, they correspond to a twin region of the crystal. (b) A 15° rotation diffraction image from a corkscrew crystal is presented. The outer part is shown with a much higher intensity so that the diffraction arcs due to base-pair stacking can be clearly seen. They indicate that the oligonucleotide duplexes are inclined with respect to the axis of the corkscrew. In the center, fiber-like diffraction arcs, which correspond to twinned modified HASO structures, are also visible.
are given as Supporting Information. Rotation diffraction images from crystals frozen in liquid nitrogen were collected with wavelength λ ) 0.98 Å at the BM16 Spanish beam line at the European Synchrotron Radiation Facility (Grenoble, France). Individual images were processed with either DENZO13 or MOSFILM.14 Low resolution, high mosaicity, and twinning prevented automatic processing. Molecular models and their simulated diffraction were obtained with the CERIUS2 program (Acelrys, San Diego, CA). In the presence of divalent cations, three different related structures have been found for the dodecamer d(AAAAAATTTTTT). We assume that in all cases the dodecamer is in the standard B-form. It is unlikely that it could be in the Hoogsteen form, which has only been found in alternating AT sequences.15 Crystals of the three different types usually coexist in a single crystallization drop. They show the following features: Type 1, a conventional HASO structure, as found in general in all-AT oligonucleotides.1 It has a pitch of about 890 Å, with 23 duplexes in a helical turn. It has been described in detail in a previous paper (ref 1, Figure 9). It is found in crystals of oligonucleotides with blunt ends, some of which are shown in Figure 1c. Type 2, they show a modified HASO structure, in which neighbor columns of duplexes are no longer equivalent. They pack in a monoclinic unit cell, with space group C2 and parameters: 30.0 × 45.3 × 930 Å3, β ) 116°, and 12 stacked oligonucleotide duplexes in the asymmetric unit. Two parallel columns are found in the unit cell. They form helices with 24 duplexes in a helical turn. From these values, the volume occupied by a base-pair in the crystal can be calculated as 1972 Å3, which is much higher than the usual 1300–1400 Å3 found in conventional crystals of oligonucleotides. It indicates that they contain a large amount of solvent. A diffraction pattern is shown in Figure 2. It was obtained from an approximately hexagonal crystal such as the one shown in the center of Figure 1a. This structure has been found only in the presence of divalent
Figure 3. Model of the modified HASO structure. (a) View of the asymmetric unit. It contains 12 dodecamers with a rotation of Ω ) 15° between neighbor units. Each duplex has an inclination of 2° with respect to the c axis of the unit cell. In the standard HASO structure, the dodecamer axes are parallel to c, and a much simpler diffraction pattern is observed. (b) Projection of the structure on to the bc plane. The c axis is vertical. The slight distortion of the duplex columns is clearly apparent. One and a half unit cells in the vertical direction are shown. (c) Projection of the crystal structure on to a plane perpendicular to the c axis of the crystal. The pseudohexagonal packing of the duplex columns is evident. (d) Projection with a small inclination of two unit cells. The distortion of the duplex columns is clearly apparent. The drawings were prepared with the CERIUS2 program.
cations (Mg2+, Ca2+, Ba2+, Sr2+, and Ni2+ but not Mn2+). In order to originate the complex diffraction pattern observed, the duplex columns must be slightly distorted. We have tentatively modeled the distortion with the individual duplexes inclined about 2° with respect to the c axis of the unit cell. A model is presented in Figure 3. The projection of the unit cell is quasihexagonal. This feature explains why the crystals are usually twinned, with three possible orientations of the ab sides of the unit cell. The well defined diffraction spots (Figure 2a) indicate that there is a very good molecular order. Whatever originates the distortion of the duplex columns gives rise to a rather stable structure. This observation is surprising since there are 48 duplexes in the unit cell. The only source of disorder appears to be twinning among different crystalline regions. Type 3, we have also detected macroscopic helical crystalline textures. A few examples are given in Figure 1 and in Supporting Information. Thin helical crystals of variable length appear first, later they increase in thickness, and with time, they tend to take the shape of corkscrews, with a pitch of 20–40 µm. A second helix may grow on top of the first one, as shown in Figure 1c. The process may continue so that the diameter of the helical texture may increase up to 100 µm. An example is shown in Supporting Information. The corkscrew crystals give rise to fiber-like diffraction patterns (Figure 2b). The equatorial arcs are consistent with the presence of distorted columns of
8
Biomacromolecules, Vol. 9, No. 1, 2008
duplexes, as found in the monoclinic structure described above. It appears that the corkscrew structures are formed by bundles of fibers with an organization similar to that shown in Figure 3. The crystallites now follow the helical path of the corkscrew texture, as indicated by the 3.3 Å reflection, due to base-pair stacking. It appears as a wide arc at both sides of the meridian (Figure 2). The shape of the helical textures varies. In particular, the helices may be either right-handed or left-handed. The most regular structures are those with a corkscrew shape, as shown in Figure 1b. We are presently trying to control the nucleation events so that unique shapes are obtained, but in general, different crystalline morphologies are found in the same crystallization drop. The results we have presented here cannot be interpreted in any simple straightforward manner. The different structures we have found are all based on long helical columns of stacked duplexes. It is clear that divalent ions play a critical role in the formation of the latter structure. In a crystal of an octanucleotide duplex, we found clusters of electron density, which we interpreted as clusters of ions.2 It is likely that the distortion of the duplex columns in the C2 structure is due to a nonuniform distribution of divalent cations. The formation of corkscrew crystals could be due to the presence of traces of intrinsic impurities which might generate an angle between neighbor crystallites. In any case, our results definitely show that the dodecanucleotide duplexes can generate macroscopic helices in the presence of the adequate counterions and that no proteins or other substances are required. Most work on DNA architecture has mainly dealt with nanostructures based on self-assembly through Watson–Crick base pairs.3–6 Large lattices have also been built.16 The macroscopic DNA helices we have described in this communication are much larger in three dimensions. They are assembled in a different way. The individual duplexes form helical stacks with a large pitch. They interact through specific base stacking between terminal base pairs in neighbor molecules and through ion-mediated interactions among columns of duplexes. No previous report is available showing such structures. They are unique for their size and shape as well as for the intermolecular forces that originate them. Because of the
Communications
large amount of solvent in the crystals, they might easily accommodate guest molecules. We think they may open the way to new approaches in nanotechnology/supramolecular chemistry. Acknowledgment. We thank Dr. I. Fita for advice and the staff of the BM16 Spanish beam line at the European Synchrotron Radiation Facility (Grenoble, France) for assistance in data collection. This work was supported by Ministerio de Educación y Ciencia, Grant BFU2006-04035. Supporting Information Available. A gallery of images of helices obtained with different cations. This material is available free of charge via the Internet at http://pubs.acs.org.
References and Notes (1) Campos, L.; Valls, N.; Urpí, L.; Gouyette, C.; Sanmartín, T.; Richter, M.; Alechaga, E.; Santaolalla, A.; Baldini, R.; Creixell, M.; Ciurans, R.; Skokan, P.; Pous, J.; Subirana, J. A. Biophys. J. 2006, 91, 892– 903. (2) Valls, N.; Richter, M.; Subirana, J. A. Acta Crystallogr., Sect. D 2005, 61, 1587–1593. (3) Seeman, N. C. Methods Mol. Biol. 2005, 303, 143–166. (4) Gothelf, K. V.; LaBean, T. H. Org. Biomol. Chem. 2005, 3, 4023– 4037. (5) Feldkamp, U.; Niemeyer, C. M. Angew. Chem., Int. Ed. 2006, 45, 1856–1876. (6) Jaeger, L.; Chworos, A. Curr. Opin. Struct. Biol. 2006, 16, 531–543. (7) Rybnikar, F.; Geil, P. H. Biopolymers 1972, 11, 271–278. (8) Li, C. Y.; Ge, J. J.; Bai, F.; Calhoun, B.H:; Harris, F. W.; Cheng, S. Z.; D, L.-C. C.; Lotz, B.; Keith, H. D. Macromolecules 2001, 34, 3634–3641. (9) Lotz, B.; Cheng, S. Z. D. Polymer 2005, 46, 577–610. (10) Che, S.; Liu, Z.; Ohsuma, T.; Sakamoto, K.; Terasaki, O.; Tatsumi, T. Nature 2004, 429, 281–284. (11) Moore, D.; Ding, Y.; Wang, Z. L. Angew. Chem., Int. Ed. 2006, 45, 5150–5154. (12) Maxwell, W. L. Proc. R. Soc. London 1974, 186, 181–190. (13) Otwinowski, Z.; Minor, W. Methods Enzymol. 1997, 276, 307–326. (14) Leslie, A. G. W. Newsletter on Protein Crystallography 1992, 26. (15) De Luchi, D; Tereshko, C; Gouyette, C; Subirana, J. A. ChemBioChem 2006, 7, 585–587. (16) He, Y.; Tian, Y.; Chen, Y.; Deng, Z.; Ribbe, A. E.; Mao, C. Angew. Chem., Int. Ed. 2005, 44, 6694–6696.
BM700994V
