Unusual ring-shaped oligonucleotide crystals

 P. Satheesh Kumar and N. Gautham*

Department of Crystallography and Biophysics, University of
Madras, Guindy Campus, Chennai 600 025, India

We report here the preliminary characterization of hexagonal ring-shaped crystals of a hexadeoxyribonucleotide. This is the first report of crystals which grow naturally with this interesting ring morphology. We discuss the mechanism which could possibly lead to the growth of these rings.

SINCE the first oligonucleotide crystals were grown by Viswamitra et al.1 in 1978, there have been reports of close to 1000 other sequences of lengths ranging from 4 to 16 which have been crystallized2. Structures of longer oligomer sequences have been reported as protein–DNA complexes, as for example the 33-mer in the metal-ion-activated Diphtheria Tox Repressor/Tox DNA operator complex3. In the structures of oligonucleotides alone, owing to the approximately cylindrical shape of short DNA duplexes, there are limited number of packing modes for the molecules4,5 and this limitation is reflected, first in the crystal system and space group6 and then in the crystal morphology and habit7,8. Furthermore, the crystal packing modes have been shown to be dependent on the helical type4, i.e. on whether the DNA helix is of the right-handed A or B types, or of the left-handed Z type. This last helical type has been observed mostly in crystals of hexanucleotides with the sequence d(CGCGCG)2 or variations of it9,10. In a majority of the crystals of such hexamers, the packing mode is as follows: The Z-DNA duplexes stack on top of one another to form infinite columns. These columns are then bundled together in a hexagonal close packed arrangement to form the crystal. The external morphology of the crystal is principally hexagonal plates or prisms7,8. However, the space group to which the crystals belong is chiefly orthorhombic P212121, though others9, such as C2221, P21 and C2 have also been observed.

As part of our studies on the effect of AT base-pairs on the structure of Z-DNA we have been using X-ray crystallography to obtain the crystal structures of hexameric sequences in which a single AT base-pair replaces a CG base-pair at each position of the ‘canonical’ Z-DNA sequence d(CGCGCG)12. In this connection we have solved the structures of d(CGCACG).d(CGTGCG) and d(CACGCG).d(CGCGTG)11. The former duplex crystallizes in the monoclinic space group P21, while the latter crystallizes in orthorhombic P212121. In both cases, initial crystallization trials yielded hexagonal plates, but the crystals finally chosen for data collection were rectangular prisms8. Again in both cases, the duplexes pack in the crystal as hexagonal close packed columns. However, as discussed elsewhere5, a shift in the relative disposition of adjacent columns leads to the two different space groups. The third duplex in this series is d(CGCGCA).d(TGCGCG). Much to our surprise (and delight!) this sequence has yielded hexagonal ring-shaped crystals under specific and reproducible conditions.

The two hexamers, viz. d(TGCGCG) and d(CGCGCA) were purchased from M/s Microsynth, Switzerland and were used without further purification. A PAGE gel photograph supplied by the manufacturer, established that the purity was greater than 95%. The two sequences were carefully annealed to form the duplex which was then stored at 4C. Crystals of the duplex were grown in two different conditions. In the first set of experiments hexagonal plate-like crystals of size 0.2  0.2  0.1 mm were obtained at room temperature (300 K) by the hanging-drop vapour diffusion method. The drop contained 2 mM hexamer duplex, 100 mM sodium cacodylate buffer at pH 7.0 and 50 mM MgCl2 and was equilibrated for four days against the reservoir solution which contained 50% MPD. These crystals diffracted to a resolution of 2.0 . Structure analysis using data obtained from this set of crystals is in progress and will be reported elsewhere.

1076.jpg (19390 bytes)

1077.jpg (24257 bytes)

In order to improve the crystal quality, we performed a second set of crystallization experiments, replacing Mg2+ by cobalt hexamine. This ion is a strong inducer and stabilizer of the transition from right-handed to left-handed DNA13,14 and the experiments were based on the hypothesis that crystals grown from a drop containing [Co(NH3)6]3+ cations would diffract to a higher resolution, since the crystal packing would probably be less disordered. Also we hoped that the lack of disorder would lead to a single duplex in the asymmetric unit. This set of experiments consistently yielded the novel ring-shaped crystals as shown in Figure 1 a. Crystals of this morphology appeared in a small range of concentration of Co(NH3)6Cl3. The biggest crystals (of size 0.05  0.05  0.005 mm) were grown by the hanging-drop vapour diffusion method at 300 K. The drop contained 2 mM hexamer, 100 mM sodium cacodylate buffer at pH 7.0, 1 mM spermine and 20 mM Co(NH3)6Cl3 and was equilibrated against the reservoir solution of 50% MPD. The crystals were too small to be mounted for X-ray diffraction experiments with laboratory X-ray sources. The crystallinity was therefore checked by examining the crystals under cross-polarizers . The extinction observed indicated that they are crystals, and not glassy or gel-like aggregates. To establish that the ring was not an artefact of a particular view of the crystal, a thin glass fibre was inserted in the hole (Figure 2). It was possible to easily ‘string’ together a set of rings on the tip of the fibre. To establish that these are indeed crystals of the oligonuclotide and not of some salt or other impurity, a gel electrophoresis experiment was performed. The crystals were harvested from the drop and were dissolved in buffer solution after washing them thoroughly with MPD. A 20% poly acrylamide gel was cast according to established procedure15. Three lanes were run – lane A contained the dissolved ring-shaped crystals, lanes B and C contained 0.5 mM and 1.0 mM solution, respectively, of the hexamer as used in the crystallization experiment. The gel was stained with ethidium bromide and viewed under UV (Figure 3). It was clear from the gel that the crystals were those of the DNA hexamers.

To our knowledge, this is the first report ever of crystals which grow as rings. An extensive search of the literature, including over the internet, did not turn up any reference to such crystals. Quartz crystals have been cut into rings and used to establish frequency standards16, but there are no reports of quartz naturally grown as rings. One unusual, but natural morphology which may be relevant to the present discussion is that of the so-called ‘hollow crystals’17. These are crystals which grow as hollow tubes and have been reported in a wide variety of inorganic and organic substances18–21. Several mechanisms have been proposed for their unusual morphology22,23. The present flat, platy, ring-shaped crystals however, are unlikely to be a result of any similar mechanism, since the morphology is quite different. The schematic sketch of a ring crystal (Figure 1 b) indicates that one may consider the rings to be formed out of six identical rectangular plates which have been arranged in a hexagonal shape. This might suggest a twinning mechanism which leads to the formation of the rings. However, when the crystals are viewed under crossed polarizers, the extinction pattern does not support this explanation. Another possibility is that each ring grows around a hexagonal template which later disappears, leaving a hole in the center. Such an explanation finds some support from the fact that we noticed exactly similar ring-shaped crystals in a two-year old crystallization set-up of the hexamer d(CGCACG).d(CGTGCG), which contained Ru(NH3)6Cl3. This experiment (10 mM DNA, 100 mM Na cacodylate buffer at pH 7.0, 5 mM Ru(NH3)6Cl3, 0.5 mM spermine in the hanging-drop equilibrated against 45% MPD) first yielded regular hexagonal plates of size 0.1  0.15  0.20 mm within 5 days. These plates did not change for atleast six months. When the experiment was viewed again after two years, the drop now contained only ring-shaped crystals. Further support for a template-directed growth mechanism comes from the observation in many of the present cases of a thin spike running from one side of the hexagonal ring to another as sketched in Figure 1 b. This could be a remnant of the template.

To summarize, each ring-shaped crystal could be the residue of a flat hexagonal plate, the centre of which has somehow evenly dissolved back into the solution. This however, does not explain several features such as the notches seen at the six vertices of the hexagon. The role of the packing mode of DNA hexamers is also unclear. Further experiments are required to establish the molecular structure and packing of the DNA hexamers in these ring-shaped crystals and to arrive at satisfactory explanation for their formation.

   1.   Viswamitra, M. A., Kennard, O., Jones, P. G., Sheldrick, G. M. and Salisbury, S., Nature, 1978, 273, 687–688.

   2.   Berman, H. M., Olson, W. K., Beveridge, D. L., Westbrook, J., Gelbin, A., Demeny, T., Hsieh, S.-H., Srinivasan, A. R. and Schneider, B., Biophys. J., 1992, 63, 751–759.

   3.   White, A., Ding, X., Vanderspek, J. C., Murphy, J. R. and Ringe, D., Nature, 1998, 394, 502–506.

   4.   Wang, A. H. J. and Teng, M. K., J. Cryst. Growth, 1988, 90, 295–310.

   5.   Sadasivan, C., Karthe, P. and Gautham, N., Acta Cryst., 1994, 50, 192–196.

   6.   Kennard, O. and Hunter, W. N., Q. Rev. Biophys., 1989, 22, 327–379.

   7.   Karthe, P., Ph D thesis, University of Madras, 1997, pp. 10–18.

   8.   Sadasivan, C., Ph D thesis, University of Madras, 1994, pp. 20–34.

   9.   Shing Ho, P. and Blaine Mooers, H. M., Biopolymers (Nucleic Acid Sci.), 1997, 44, 65–90.

10.   Eichman, B. F., Schroth, G. P., Basham, B. E. and Shing Ho, P., Nucleic Acids Res., 1999, 27, 543–550.

11.   Sadasivan, C. and Gautham, N.,, J. Mol. Biol., 1995, 248, 918–930.

12.   Wang, A. H.-J., Quigley, G.-J., Kolpak, F. J., Crawford, J. L., van Boom, J. H., van der Marel, G. and Rich, A., Nature, 1979, 282, 680–686.

13.   Winkle, S. A., Aloyo, M. C., Lee-Chee, T., Morales, N., Zambrano, T. Y. and Sheardy, R. D., J. Biomol. Struct. Dyn., 1992, 10, 389–402.

14.   Lu, M., Kallenbach, N. R. and Sheardy, R. D., Biochemistry, 1992, 31, 4712–4719.

15.   Sambrook, J., Fritsch, E. F. and Maniatis, T., Molecular Cloning – A Laboratory Manual, Cold Spring Harbor Laboratory Press, 1989, pp. 6.39–6.44.

16.   Frerking, M. E., Proceedings of the 1996 IEEE International Frequency Control Symposium, 1996, pp. 33–46.

17.   Arivuoli, D., Ph D thesis, Anna University, 1986, pp. 106–129.

18.   Buckley, H. E., Z. Kristallogr., 1939, A82, 47–55.

19.   Mason, B. J. and Hallett, M. J., Proc. R. Soc. London, 1958, 247, 440–453.

20.   Scott, R. A. M. and Jarvis, M. G., J. Cryst. Growth, 1972, 16, 83–85.

21.   Rajan, S. S., per. commun., 1998.

22.   Cruceanu, E., Nicolesku, D. and Vancu, A., Kristallografiya, 1964, 9, 537–547.

23.   Sieker, L. C., J. Cryst. Growth, 1988, 90, 31–38.

 ACKNOWLEDGEMENTS.  We thank Department of Biotechnology, Government of India for financial support under project BT/R&D/15/31/94. P.S.K. thanks CSIR for the award of SRF.

 Received 19 May 1999; revised accepted 25 August 1999