Skip to main content
NCBI home page
Nucleic Acids Research logoLink to Nucleic Acids Research
. 2000 Mar 1;28(5):1259–1265. doi: 10.1093/nar/28.5.1259

Intermolecular interactions and water structure in a condensed phase B-DNA crystal

George R Clark a, Christopher J Squire, L J Baker, Roger F Martin 1, Jonathan White 2
PMCID: PMC102594  PMID: 10666471

Abstract

By controlled dehydration, the unit cells of dodecamer DNA–drug crystals have been shrunk from 68 000 (normal state) to 60 000 (partially dehydrated intermediate state) to 51 000 Å3 (fully dehydrated state), beyond which no further solvent loss occurs. The total solvent content in the normal crystals is ~40% by volume, reducing to ~20% in the fully dehydrated phase. The 25% reduction in cell volume induced a dramatic enhancement in the resolution of the X-ray diffraction data (from 2.6 to beyond 1.5 Å). We have determined the structures of the normal, partially dehydrated and fully dehydrated crystals. Details of the ligand binding have been presented in the preceding article. The present paper describes the unique features of the structure of the fully dehydrated phase. This structure was refined with 9015 unique observed reflections to R = 14.9%, making it one of the most reliable models of B-form DNA available. The crystals exist as infinite polymeric networks, in which neighbouring dodecamer duplexes are crosslinked through phosphate oxygens via direct bonding to magnesium cations. The DNA is packed so tightly that there is essentially only a single layer of solvent between adjacent molecules. The details of the crystal packing, magnesium bridging, DNA hydration and DNA conformation are described and compared with other experimental evidence related to DNA condensation.

INTRODUCTION

The oligonucleotide d(CGCGAATTCGCG)2, alone and complexed with DNA-binding ligands, has been extensively studied by X-ray crystallography during the last 18 years (1–6). The crystals diffract poorly, usually only to 2–3 Å resolution, and the data sets consist of ~2500 unique reflections. Until recently, attempts to improve the X-ray resolution have not been very successful.

In the early 1980s Dickerson and colleagues prepared crystals of (CGCGAATTBrCGCG)2 by vapour diffusion at 7°C from 20–35% (v/v) 2-methyl-2,4-pentanediol (MPD) solutions (7). The crystals proved to be isomorphous with the native DNA. On transferring these crystals to a 60% solution of MPD in distilled water and soaking for several days at 7°C, the crystals underwent a partial dehydration transition that resulted in a reduction of the unit cell volume by 7%. Although this change was relatively small, the detailed structural analysis of this ‘MPD-7’ structure was significant, as it allowed the authors to derive a number of important properties relating to conformation and hydration of the B-DNA duplex.

In 1998, Shiu et al. reported an improved process for soaking d(CGCGAATTCGCG)2 dodecamer crystals in concentrated MPD prior to data collection to enhance the diffraction resolution (8). There was little change in unit cell volume, but they were able to refine the low temperature structure to a resolution of 1.4 Å employing over 11 000 reflections (8). This structure suggested that many sequence-dependent features of DNA conformation are mediated by site-specific binding of cations. Near atomic resolution (1.1–1.3 Å) dodecamer structures were also reported by Egli et al. in 1999 (9–11). These studies further probed the roles of cations in DNA conformation and packing interactions. The Shiu and Egli structures were obtained from cryo-cooled crystals and used rotating anode or synchrotron radiation.

We have carried out a series of crystallographic investigations of complexes formed between d(CGCGAATTCGCG)2 and the iodinated Hoechst ligands 2′-(3-iodophenyl)-5-(4-methyl-1-piperazinyl)-2,5′-bi-benzimidazole (m-iodo Hoechst, IA) and 2′-(3-iodo-4-methoxyphenyl)-5-(4-methyl-1-piperazinyl)-2,5′-bi-benzimidazole (m-iodo, p-methoxy Hoechst, IB), using conventional data collection and structure solution procedures. We then subjected the capillary-mounted crystals of both complexes to dehydrating conditions as described in Materials and Methods of the preceding paper. The crystals shrank slowly as dehydration occurred, until their volume was ~75% of the initial size, after which no further dehydration or shrinkage occurred. The X-ray diffraction patterns were monitored at intervals during the dehydration process, revealing that the diffraction resolution improved markedly as the crystals shrank, from ~2.6 to beyond 1.5 Å. X-ray structures have been determined for the normal crystals (normal dodecamer crystals in this space group have a solvent content of ~40%), intermediate ‘partially dehydrated’ crystals (solvent content ~30%) and the ‘fully dehydrated’ crystals (solvent content ~20%). The high number of reflections per refined parameter for the fully dehydrated crystals makes these structures extremely reliable models of B-form DNA. Remarkably, the 1.5 Å data was collected without the use of synchrotron or even rotating anode radiation sources, but instead using a conventional sealed tube X-ray source of the type used for small molecule crystallography. The preceding companion paper presents all the experimental details and describes iodo Hoechst ligand binding in the normal, partially dehydrated and fully dehydrated crystals. The present paper details the processes that lead to the condensed phase DNA and describes the dodecamer geometry, cation interactions and water structure in the fully dehydrated crystals of the IB complex. The DNA geometry in the analogous IA structure is virtually identical.

MATERIALS AND METHODS

Details of the crystallography, crystal mounting and dehydration, data collection and DNA refinement are described in the preceding article.

RESULTS AND DISCUSSION

Crystal packing and the dehydration mechanism

The fully dehydrated crystals exist as infinite polymeric networks, in which neighbouring dodecamer duplexes are crosslinked by coordinate bonds through magnesium cations. The DNA molecules are packed much closer together than has been found in any previous oligonucleotide structure and there is essentially only a single layer of water molecules between adjacent DNA duplexes. The DNA concentration is comparable to that of condensed DNA in the bacteriophage T4 head (12). The closeness of the crystal packing and the spatial and molecular rearrangements of the DNA that take place during the dehydration process means that this particular structure can serve as a model for a condensed state of B-DNA.

Normal dodecamer DNA crystals contain ~40% by volume of solvent (water, hexa-hydrated magnesium cations and other droplet components). Water molecules either interact electrostatically with partially charged sites on the DNA or else occupy the interduplex interstices as ‘liquid’ water. We have found that dehydration by gradual loss of water molecules from the crystal lattice is accompanied by substantial rearrangements of the DNA molecules as they pack closer together in order to better fill the resulting interduplex voids. The DNA conformation and groove widths alter markedly in response to the changing intermolecular interactions.

During the transition from the normal to the partially dehydrated lattice, neighbouring duplexes are drawn closer together and we assume that repulsive charges between phosphate backbones are shielded by water molecules and aqueous magnesium cations. As the partially dehydrated crystals slowly lose further water and the duplexes approach even more closely, the magnesium ion activity increases within the crystal. Formation of the fully dehydrated structure cannot occur unless there are more specific intermolecular interactions or improved charge shielding at the closest contact points between the approaching duplexes. Additional attractive forces are provided by two magnesium cations (unambiguously located in the crystal structure) which form intermolecular bridges between DNA phosphate oxygens. These Mg interactions lock the fully dehydrated lattice together and generate the three-dimensional polymeric DNA crystal.

The normal crystals have unit cells with approximate dimensions of 25 × 41 × 66 Å, V = 68 000 Å3, whereas those in the fully shrunken phase are approximately 25 × 34 × 61 Å, V = 51 000 Å3. During the full shrinkage process the crystals have lost ~25% of their volume and half of their normal water content. The extent of dehydration and unit cell shrinkage that has occurred is shown in Figure 1. The overall crystal packing motif in our normal, partially and fully dehydrated crystals is the same as that found in all previous dodecamer and dodecamer–drug complexes which crystallise in space group P212121. All contain minor groove–backbone interactions (including intermolecular base pairing) between molecules related by 21 symmetry along the c-axial direction. One significant difference in the fully dehydrated crystal structure is the angle between the symmetry-related duplexes. A decrease in this angle from 150° for our normally hydrated crystals to 122° for our fully dehydrated crystals results in extreme shortening of the c-axial length. The more angular packing also allows compaction along the b-axial direction. It has been noted previously that an increase in hydration of normal dodecamer crystals causes an increase in the angle between adjacent molecules (13). Our results follow the same pattern, as the decrease in hydration causes a decrease in this angle.

Figure 1.

Figure 1

Open in a new tab

View down the a-axis of a 2 × 2 unit cell array for the normal (left), partially dehydrated (middle) and fully dehydrated (right) structures. The b-axis is aligned across the page, the c-axis up the page. Symmetry-related DNA molecules are shown as different coloured CPK models and were illustrated using Cerius2 (41). All non-DNA atoms have been omitted for clarity, including those of the IB ligand. As the crystals progress from normal to fully dehydrated, the crystal interstices occupied by water molecules become smaller, the DNA packs closer together in the b- and c-axial directions and the DNA backbones distort to better fill the spaces previously occupied by water molecules. The angles between DNA molecules related by 21 symmetry (i.e. red/green and yellow/magenta colour combinations) become smaller, allowing c- and b-axis compression.

We have observed that the crystals that shrink most readily to the fully dehydrated state originate from crystallisation droplets containing high initial concentrations of magnesium chloride. Methanol and spermine hydrochloride are also present in the crystallising solutions. Alcohol lowers the dielectric constant of aqueous DNA solutions and promotes divalent cation-induced condensation (14,15). Spermine too can cause DNA condensation in aqueous solution (16) and crystal structure determinations have demonstrated that it can induce tight packing of A-form DNA (17). As far as we can tell from our results, the minor groove-binding ligand plays no significant role in the dehydration process.

DNA parameters of the fully dehydrated structures

The combination of crystal packing effects, magnesium cation interactions and water structure induces unusual and substantial conformational changes in the DNA (Fig. 2). This is reflected in the root mean square (r.m.s.) deviations in atomic coordinates based on superposition of the central four nucleotides of the fully dehydrated DNA with those of the normal structure. r.m.s. values obtained are 1.38 Å for C1–G4/C21–G24, 0.75 Å for A5–T8/A17–T20 and 2.70 Å for C9–G12/C13–G16. The overall r.m.s. is 1.76 Å.

Figure 2.

Figure 2

Open in a new tab

Superposition of normal/partially dehydrated structures (left) and partially dehydrated/fully dehydrated structures (right) in stereo views drawn using Raswin (Berkeley Enhanced Rasmol for windows, available at http://mc2.cchem.berkeley.edu/Rasmol/v2.6/ ) and PDB output from CURVES. The CURVES output emphasises the calculated helical axes and the base/base pair geometries. Two views are shown (top and bottom), the lower after a 90° rotation of the DNA about the long helical axes. The normal, partially dehydrated and fully dehydrated structures are coloured black, magenta and yellow, respectively. All non-DNA atoms have been omitted for clarity.

Crystal structures of normally hydrated d(CGCGAATTCGCG)2 show a bend (~8°) in the helical axis around the junction of the G4–A5 sequence. In our partially dehydrated crystals and in the MPD-7 structure, this bend is absent (Fig. 2). The straightening of the helical bend is accompanied by concomitant changes in roll and propeller twist along the helix and through the central AATT region. The conformation adopted by the DNA in the fully dehydrated crystal appears to be an extension of the partial and MPD-7 helical straightening together with additional distortions of the helical backbone.

The DNA analysis programs SCHNAaP (18) and CURVES (19) were used to calculate the various DNA parameters. For the most part, the DNA in the fully dehydrated structures has standard base pair and helical parameters (20). Traditional base pairing and stacking interactions are retained throughout the structure and backbone torsion angles are within the normal range for B-DNA (21), except those around C3. Plots of χ versus δ, ɛ versus ζ and δ versus ζ (as described by Fratini et al.; 7) also show clear B-DNA characteristics, as do the intrastrand P–P distances as described by the SCHNAaP parameter Zp.

Although the individual backbone torsion angles are within the normal range, the combined effects of small deviations along the helix result in an unusual structure, especially near the ends of the helix where crystal packing interactions are most pronounced. Some of the phosphates swivel, resulting in conformational changes from BI to BII. In normal B-DNA, the backbone ɛ and ζ values are on average paired into two states, BI (trans, -gauche), and BII (-gauche, trans). In the BII state, the sugar pucker is strongly restricted to C2′-endo (δ = 144°) or C3′-exo (δ = 158°) values. A change from BI to BII is often accompanied by a narrowing of the minor groove as the phosphate group swivels into the mouth of the groove. The BI state is the most common in B-DNA, but there is little energy difference between the two states. The present fully dehydrated structure contains an unusually high number of BII backbone conformations, noticeable at C9, G10, G16 and C23 and to a lesser extent at G2.

The region around the C3 residue shows the most deviation from standard B-DNA parameters. Here, the α, β and γ backbone torsions (125°, 130° and 174°, respectively) are highly unusual compared to the other residues within the structure and to most other B-DNAs (20,21). The parameters λ I and II (the C1′-C1′-N9 and C1′-C1′-N1 virtual valence angles) are good indicators of how well base pairs fit standard Watson–Crick geometry. Ideal values are in the range 54–56° and are symmetrical for λ I and II. In the fully dehydrated structure, λ I and II are noticeably asymmetric at the C3·G22 base pair with values of 62° and 56°, respectively. The partially and fully dehydrated structures are compared in the vicinity of the C3 residues in Figure 3. The backbone distortions of the fully dehydrated DNA result from a combination of interduplex DNA–DNA contacts, pushing the C1–C3 region downwards, and from DNA–magnesium interactions. The torsion angles of the strand backbone are transmitted through the sugar (which has tilted to the right in Fig. 3) to the C3 base, resulting in the non-standard C3·G22 base pair geometry. Base pair stacking along strand 1 between C3 and C1 is maintained and the G2·C23 and C1·G24 base pairs retain standard Watson–Crick geometry. Similar backbone movements are seen at the other end of the helix with large shifts in the positions of the terminal bases. These shifts can be traced back to the C9·G16 region where BI to BII transitions occur.

Figure 3.

Figure 3

Open in a new tab

A stereo view of the DNA around C3. The fully dehydrated DNA (shaded bonds) is drawn superimposed on the same region of the partially dehydrated structure (open bonds). Backbone torsions force the C3 sugar of the fully dehydrated DNA to lie to the right, affecting the C3 base and base pair geometries markedly.

Another unusual feature of the DNA duplex in the fully dehydrated crystal is the large value of opening for the two base pairs A6·T19 and T7·A18 (20° and 12°, respectively). It is difficult to isolate a single cause for these large values. It is likely that they result from a combination of the narrow minor groove and deep binding of the ligand. Similarly, propeller twists for A6·T19, T7·A18 and T8·A17 are very high (21°, 26° and 24°, respectively). The ligand hydrogen bonding to A6 N3, T7 O2, T19 O2 and T20 O2 in the floor of the minor groove (see preceding article) almost certainly causes some of this effect. The high propeller twist at the T8·A17 base pair may be a requirement for maintaining base stacking with the previous two base pairs. Conformational changes resulting from the coordinative magnesium binding to phosphates of T19 and T20 complicate the analysis.

The groove widths of B-DNA are an important and revealing parameter. Normally, the binding of a minor groove ligand causes an increase in minor groove width over the narrow central binding site, when compared to unliganded DNA. However, in the present crystals, the minor groove width is independent of the complexed ligand and is instead a consequence of the backbone distortions that result from the unusual crystal packing. Magnesium binding plays a key role in this, as is discussed in the following section. A plot of minor groove width based on P–P distances (see figure 3 of the preceding article) shows significant widening at the 3′-end of the AATT tract and narrowing at the 5′-end, when compared to the normally hydrated DNA. The major groove (not shown) has only small differences in width between the three hydration states, even though there is significant intrusion of symmetry-related DNA backbones into this groove in the fully dehydrated structure. These intrusions are accommodated easily within the wide major groove with minimal backbone deformation and without widening of the groove.

Effects of magnesium binding on DNA conformation in the fully dehydrated structures

The asymmetric unit of the structure contains two crystallographically distinct magnesium cations (Fig. 4). The first, Mg 26, forms an inner sphere coordination complex with two phosphate oxygens of symmetry-related DNA molecules. The bonds lengths are 2.1 and 2.0 Å for bonds to 12 O1P and #20 O1P [x + , 1 – y, –z] respectively. The magnesium adopts an octahedral geometry in which phosphate oxygens are situated trans and the remaining four coordination sites are filled with water molecules. The distance between 12 O1P and #20 O1P is only 4.2 Å and so the magnesium plays an essential role in shielding the two negative charges and allowing the two duplexes to come close together at that point. The second magnesium, Mg 27, also forms an inner sphere complex, but to three different symmetry-related DNA molecules, 10 O2P (2.1 Å), #19 O1P [x + , 1 – y, –z] (2.0 Å) and #23 O1P [ – x, 1 – y, z – ] (2.3 Å). This octahedral coordination geometry is more distorted than that of Mg 26. The phosphate oxygens are in a facial arrangement, with three water molecules occupying the remaining coordination sites. This magnesium also shields the negative phosphate charges, as distances between the facial phosphate oxygens are only 3.0–3.5 Å.

Figure 4.

Figure 4

Open in a new tab

Mg 26 (top) and Mg 27 (bottom) in the fully dehydrated IB structure are drawn overlaid with a 2Fo – Fc map calculated without phasing from the magnesiums. Mg 26 coordinates two symmetry-related phosphate oxygens in a trans arrangement and has a regular octahedral geometry. Mg 27 forms coordinate bonds to three symmetry-related phosphate oxygens in a facial arrangement, resulting in a distorted octahedral geometry. Electron density is drawn at the 1.5σ level using TURBO (42). The quality of the electron density map is indicative of the high resolution of the data collected from the fully dehydrated crystals. Symmetry-related residues are labeled with a # symbol.

Figure 5a shows how the magnesium interactions ‘pull’ the DNA backbone in various directions and cause the large deviations in DNA conformation from that of the normal structure. There is a strong correlation between magnesium binding and minor groove width. The groove is very narrow at the top end of the duplex where Mg 26a crosslinks the minor groove by hydrogen bonding to phosphates of A5 and G24 (Fig. 5b). At the opposite end of the duplex, Mg 27b bonds directly to the phosphate on T19 while Mg 27c bonds directly to the phosphate on G10 of the second strand. The effect of these two Mg 27 interactions is to pull open the minor groove in the vicinity of the IB ligand piperazine group (Fig. 5c).

Figure 5.

Figure 5

Open in a new tab

(a) View of the fully dehydrated DNA structure including all symmetry-related magnesium contacts. Orange coloured arrows show the directions in which regions of the DNA have been ‘pulled’ during dehydration. Magnesiums are drawn as magenta coloured spheres and the IB ligand as a magenta stick model. Two close-up views are shown in (b) and (c) where the magnesium cations are shown as octahedral ball and stick figures with the coordinated waters coloured red. The magnesium-coordinated phosphate oxygens of symmetry-related DNA duplexes are shown as green spheres. Hydrogen bonding interactions (broken lines) across the minor groove between A5 and G24 cause groove narrowing (b), while magnesium bridging to neighbouring duplexes causes the groove to open between G10 and T19 (c).

Water structure in the fully dehydrated DNA

The first real evidence that hydration impacts strongly on DNA structure came from the initial X-ray diffraction studies of DNA fibres, when Franklin and Gosling found different diffraction patterns depending on the relative humidity (22). The term ‘economy of hydration’ was coined by Saenger et al. to describe the different hydration patterns and the relationship between hydration and DNA conformation (23). In high humidities, the phosphate groups are fully and individually hydrated and B-DNA predominates. If the water activity is reduced, water bridges form between phosphate groups and the hydration is deemed more economical. The phosphates move closer together, which greatly influences sugar puckers and the overall DNA conformation. The more economical hydration in A- and Z-DNA is the underlying cause of the B→A and B→Z transitions.

The differences in the P–P distances between adjacent phosphates is a reflection of the economy of hydration. For our fully dehydrated structure, the average distance between adjacent phosphorus atoms is 6.8 Å, which is typical of B-form DNA. However, the water content of the fully dehydrated crystals is exceptionally low and thus the economy of hydration is far higher than in the normally hydrated B-DNA state. This is achieved by a spectacular increase in the number of interduplex bridging interactions between water molecules and symmetry-related DNA molecules (Table 1). These waters, which shield the repulsive interactions between phosphate backbones, comprise ~19% of the total water molecules in the crystal (compared with ~3% in normal hydration structures). In this way, phosphates of adjacent nucleotides within the same DNA strand are not required to approach closely and the normal B-DNA conformation is retained. The restraints of crystal packing probably preclude major changes in conformation (such as a B→A-form transition) as the preformed crystals dehydrate. The hydration at the phosphate groups is conserved, but there is less hydration of the DNA bases.

Table 1. Numbers of phosphate bridging water molecules in high resolution structures of A-DNA (43), B-DNA (8) and the present fully dehydrated IB structure.

  A-DNA B-DNA Fully dehydrated
Resolution of X-ray structure 1.4 1.4 1.6
Unique reflections for refinement 13095 11438 10031
No. of water molecules located 135 160 48
Average intrastrand P–P distance (Å) 6.0 6.7 6.8
1 water intrastrand P–W–P bridge 6 0 0
Intrastrand P–P gaps bridged by one water (%) 22 0 0
1 water intermolecular P–W–P bridge 6 5 9
Waters in intermolecular P–W–P (%) 4 3 19
Open in a new tab

The magnesium cations effectively block water molecules from approaching 20% of the phosphate groups. Because of their positive charges and coordinate interactions with the DNA, the magnesium cations exert a far greater stabilising effect on the DNA structure and crystal packing than would water molecules similarly located. However, these magnesium cations alone are not sufficient to balance the 22 negative charges on each duplex. It seems probable that some of the bridging water molecules, particularly those making short contacts and with high coordination numbers, are in fact cations at partial occupancy.

While it might appear that relatively few water molecules have been located in the fully dehydrated structure, this is not a measure of the data quality but an indication of the extreme dehydration of this crystal.

Biological implications of the fully dehydrated structure

DNA condensation is a complex but vitally important biological process. Recent interest has been directed at the role of multivalent cations (24,25), packaging of DNA in bacteriophage (12,26) and viruses (27), and chromatin condensation (28). There is anticipation that condensation in vitro could provide a novel and practical mechanism for packaging DNA into cationic liposomes for delivery as gene therapy agents in vivo (29).

Crystalline DNA, like condensed DNA, exists in an intermediate stage between the solid and solution states and thus crystal packing in oligonucleotide structures could show structural similarities to condensed DNA in vivo. Dodecanucleotide structures which crystallise in space group R3 have previously been compared to the structure of condensed DNA (30–32). The crystal structures showed an interlocking groove–backbone interaction where sections of the DNA backbone of one helix invaded the major groove of other symmetry-related helices. These crystal packing interactions demonstrated that the self-fitting properties of B-DNA can influence DNA condensation and recombination processes.

In our fully dehydrated crystals in space group P212121, the DNA backbones intrude into the major grooves of symmetry-related DNA duplexes. While these backbone intrusions are not deep and do not cause base pair melting, they do demonstrate one way in which B-DNA chains may coalesce into a more condensed state.

The groove–backbone interactions in the fully dehydrated structure are geometrically different from the R3 structures (where the DNA helices cross each other at an angle) because the DNA duplexes in P212121 structures are lined up approximately parallel to each other. Yet, magnesium cations play an integral role in the self-fitting process in both of these structures. The groove–backbone interactions in the R3 structure are mediated by bridging hexa-hydrated magnesium cations and there is a strict requirement for at least 16 magnesium cations per oligonucleotide in order to obtain well-diffracting crystals (30). The fully dehydrated structure has demonstrated that coordinative magnesium bridging can deform the DNA backbone, neutralise charge repulsion and stabilise a three-dimensional network structure, and in doing so, allow the DNA to coalesce into a highly condensed state.

Hydration plays an integral role in DNA condensation in vitro and in vivo. There is evidence that the various polyvalent ligands that induce DNA condensation act by reconfiguring the water between macromolecule surfaces to create long range attractive hydration forces (33). As DNA condensation progresses, water molecules surrounding the DNA must rearrange. Our results show that the B-form conformation can be retained in the condensed state by preferential hydration of the charged phosphate groups (over the base and sugar moieties) and by sharing of water networks between phosphates widely separated in sequence.

DNA condensation processes in vivo are mediated by multivalent metal cations, polyamines such as spermine and by proteins. Yet even in some protein-induced condensation experiments, the size and morphology of the condensates is dependent on the ratio of positive to negative charges and the ionic strength of the surrounding medium, and is independent of the DNA or protein (34). Magnesium cations have been shown to enhance the side-by-side association of DNA segments (35,36) and to cause aggregation of melted DNA by forming crosslinks (37). The aggregation studies did not define an explicit role for the divalent cations studied yet determined that crosslinking results in a strongly stabilising free energy. Our observation of interduplex magnesium bridging indicates that the role of divalent cations in DNA condensation may be more important and diverse than previously thought, and we recommend that this be taken into consideration in designing and interpreting future experimental studies.

Dehydrated crystals are of particular interest in the field of DNA radiation chemistry, where it has been found that the yield of radical species in irradiated DNA (38) and the product profile (39) is dependent on the extent of hydration, with data available for Γ values (number of water molecules per DNA base pair) of a few to >20. Most of these studies have used DNA preparations extracted from biological material, but experiments have also been carried out using crystals of synthetic oligodeoxynucleotides (40). One aim of such experiments is to study systems at varying degrees of hydration (W.A.Bernhard, personal communication) and in this context the fully dehydrated crystal structures outlined in this paper could be of considerable interest.

Acknowledgments

ACKNOWLEDGEMENT

We thank the University of Auckland Research Committee for financial support (to C.J.S.).

NDB accession nos DD0003–DD0009

REFERENCES

  • 1.Berman H.M., Olson,W.K., Beveridge,D.L., Westbrook,J., Gelbin,A., Demeny,T., Hsieh,S.-H., Srinivasan,A.R. and Schneider,B. (1992) Biophys. J., 63, 751–759. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Wing R.M., Drew,H.R., Takano,T., Broka,C., Tanaka,S., Itakura,K. and Dickerson,R.E. (1980) Nature (Lond.), 287, 755–758. [DOI] [PubMed] [Google Scholar]
  • 3.Drew H.R., Wing,R.M., Takano,T., Broka,C., Tanaka,S., Itakura,K. and Dickerson,R.E. (1981) Proc. Natl Acad. Sci. USA, 78, 2179–2183. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Drew H.R. and Dickerson,R.E. (1981) J. Mol. Biol., 151, 535–556. [DOI] [PubMed] [Google Scholar]
  • 5.Drew H.R. and Dickerson,R.E. (1981) J. Mol. Biol., 149, 761–786. [DOI] [PubMed] [Google Scholar]
  • 6.Dickerson R.E. and Drew,H.R. (1981) Proc. Natl Acad. Sci. USA, 78, 7318–7322. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Fratini A.V., Kopka,M.L., Drew,H.R. and Dickerson,R.E. (1982) J. Biol. Chem., 257, 14686–14707. [PubMed] [Google Scholar]
  • 8.Shui X., McFail-Isom,L., Hu,G.G. and Williams,L.D. (1998) Biochemistry, 37, 8341–8355. [DOI] [PubMed] [Google Scholar]
  • 9.Tereshko V., Minasov,G. and Egli,M. (1999) J. Am. Chem. Soc., 121, 470–471. [Google Scholar]
  • 10.Tereshko V., Minasov,G. and Egli,M. (1999) J. Am. Chem. Soc., 121, 3590–3595. [Google Scholar]
  • 11.Minasov G., Tereshko,V. and Egli,M. (1999) J. Mol. Biol., 291, 83–99. [DOI] [PubMed] [Google Scholar]
  • 12.Mullaney J.M. and Black,L.W. (1998) J. Mol. Biol., 283, 913–929. [DOI] [PubMed] [Google Scholar]
  • 13.Chatake C., Sunami,T., Ono,A., Ueno,Y., Matsuda,A. and Takenaka,A. (1999) Acta Crystallogr., D55, 873–876. [DOI] [PubMed] [Google Scholar]
  • 14.Votavova H., Kucerova,D., Felsberg,J. and Sponar,J. (1986) J. Biomol. Struct. Dyn., 4, 477–489. [DOI] [PubMed] [Google Scholar]
  • 15.Arscott P.G., Ma,C., Wenner,J.R. and Bloomfield,V.A. (1995) Biopolymers, 36, 345–364. [DOI] [PubMed] [Google Scholar]
  • 16.Gosule L.C. and Schellman,J.A. (1978) J. Mol. Biol., 121, 311–326. [DOI] [PubMed] [Google Scholar]
  • 17.Tippen D.B. and Sundaralingam (1997) J. Mol. Biol., 267, 1171–1185. [DOI] [PubMed] [Google Scholar]
  • 18.Xiang-Jun L., El Hassan,M.A. and Hunter,C.A. (1997) J. Mol. Biol., 273, 668–680. [DOI] [PubMed] [Google Scholar]
  • 19.Lavery R. and Sklenar,H. (1996) CURVES version 5.0. Institut de Biologie Physico-Chimique, Paris, France.
  • 20.Berman H.M. (1997) Biopolymers, 44, 23–44. [DOI] [PubMed] [Google Scholar]
  • 21.Schneider B., Neidle,S. and Berman,H.M. (1997) Biopolymers, 42, 113–124. [DOI] [PubMed] [Google Scholar]
  • 22.Franklin R.E. and Gosling,R.G. (1953) Acta Crystallogr., 6, 673–677. [Google Scholar]
  • 23.Saenger W., Hunter,W.N. and Kennard,O. (1986) Nature, 324, 385–388. [DOI] [PubMed] [Google Scholar]
  • 24.Bloomfield V.A. (1996) Curr. Opin. Struct. Biol., 6, 334–341. [DOI] [PubMed] [Google Scholar]
  • 25.Bloomfield V.A. (1997) Biopolymers, 44, 269–282. [DOI] [PubMed] [Google Scholar]
  • 26.Grimes S. and Anderson,D. (1997) J. Mol. Biol., 266, 901–914. [DOI] [PubMed] [Google Scholar]
  • 27.Tuma R., Bamford,J.H.K., Bamford,D.H. and Thomas,G.J. (1996) J. Mol. Biol., 257, 102–115. [DOI] [PubMed] [Google Scholar]
  • 28.Laitinen J., Stenius,K., Eloranta,T.O. and Holtta,E. (1998) J. Cell. Biochem., 68, 200–212. [DOI] [PubMed] [Google Scholar]
  • 29.Huang L., Klibanov,A.L., Mori,A. and Gao,X. (1993) J. Liposome Res., 3, 505–515. [Google Scholar]
  • 30.Timsit Y., Westhof,E., Fuchs,R. and Moras,D. (1989) Nature, 341, 459–462. [DOI] [PubMed] [Google Scholar]
  • 31.Timsit Y. and Moras,D. (1991) J. Mol. Biol., 221, 919–940. [DOI] [PubMed] [Google Scholar]
  • 32.Timsit Y. and Moras,D. (1995) J. Mol. Biol., 251, 629–647. [DOI] [PubMed] [Google Scholar]
  • 33.Rau D.C. and Parsegian,V.A. (1992) Biophys. J., 61, 246–259. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Garcia-Ramirez M. and Subirana,J.A. (1994) Biopolymers, 34, 285–292. [Google Scholar]
  • 35.Adrian M., Ten Heggeler-Bordier,B., Wahli,W., Stasiak,A.Z., Stasiak,A. and Dubochet,J. (1990) EMBO J., 9, 4551–4554. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Shaw S.Y. and Wang,J.C. (1993) Science, 260, 533–536. [DOI] [PubMed] [Google Scholar]
  • 37.Duguid J.G. and Bloomfield,V.A. (1995) Biophys. J., 69, 2642–2648. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Milano M.T. and Bernhard,W.A. (1999) Radiat. Res., 151, 39–49. [PMC free article] [PubMed] [Google Scholar]
  • 39.Swarts S.G., Becker,D., Sevilla,M. and Wheeler,K.T. (1996) Radiat. Res., 145, 304–314. [PubMed] [Google Scholar]
  • 40.Milano M.T., Hu,G.G., Williams,L.D. and Bernhard,W.A. (1998) Radiat. Res., 150, 101–114 [PubMed] [Google Scholar]
  • 41. Molecular Simulations (1997) Cerius2, version 3.5. Molecular Simulations Inc., San Diego, CA.
  • 42. BioGraphics (1997) TURBO-FRODO. AFMB-CNRS, Marseille, France.
  • 43.Gao Y.-G., Robinson,H.H. and Wang,A.H.-J. (1999) Eur. J. Biochem., 261, 413–420. [PubMed] [Google Scholar]

Articles from Nucleic Acids Research are provided here courtesy of Oxford University Press

RESOURCES