The potentially Z-DNA-forming sequence d(GTGTACAC) crystallizes as A-DNA

J. Mol. Sol. (1987) 197, 141-145

LETTERS TO THEEDITOR

The Potentially

Z-DNA-forming Sequence d(GTGTACAC) Crystallizes as A-DNA

(GT),/(CA), sequences have stimulated much interest because of their frequent occurrence in eukaryotic DNA and their potential for forming the left-handed Z-DNA structure. We here report the X-ray crystal structure of a self-complementary octadeoxynucleotide, d(GTGTACAC), at 2.5 a resolution. The molecule adopts a right-handed double-helical conformation belonging to the A-DNA family. In this alternating purine-pyrimidine DNA minihelix the roll and twist angles show alternations qualitatively consistent with Calladine’s rules. The average tilt angle of 9.3” is between the values found in A-DNA (19”) and B-DNA (-6”) fibers. It is envisaged that such intermediate conformations may render diversity to genomic DNA. The base-pair tilt angles and the base-pair displacements from the helix axis are found to be correlated for the known A-DNA double-helical fragments.

From various physicochemical studies we now recognize that DNA is significantly polymorphic, and it can adopt several right-handed as well as left-handed double-helical conformations (Dickerson et al., 1982; Kennard, 1983; Rich et al., 1984). This observed variability in DNA structure offers potential sequence-dependent diversity in atomic configuration of DNA that can be utilized by proteins for site-specific binding. Therefore, in recent years a considerable effort has been devoted towards an understanding of the determinants of DNA conformation. From single-crystal X-ray diffraction methods, structures of several DNA sequences have now been determined, and some generalizations have emerged. Firstly, in the crystalline state, Z-DNA is formed when the DNA sequence has an alternation of pyrimidines (py)t and purines (pu) (Rich et al., 1984). One base-pair out of alternation can sometimes be tolerated under conditions very favorable for formation of Z-DNA such as high salt concentration, presence of cobaltic hexamine trichloride, and methylation or bromination of cytosines (Wang et al., 1985). On the other hand non-alternating sequences have yielded right-handed A or B-DNA helices, and it has been suggested that the presence of G-G sequences induce formation of A-DNA (McCall et al., 1986). Contrary to these generalizations, our crystalloanalysis graphic on the self-complementary alternating pu-py sequence d(GTGTACAC) shows that it crystallizes as a right-handed double helix with an overall geometry similar to A-DNA. The oligodeoxynucleotide was synthesized using solid-support phosphoramidite chemistry (Dorman et al., 1984). The 5’-dimethoxytrityl derivative of the oligonucleotide was prepared on a 10 pmol scale by use of an Applied Riosystems synthesizer, purified by HPLC (Zon & Thompson, 1986), detritylated with acetic acid, and the resultant 5’-hydroxyl product was isolated in the sodium t Abbreviations (M,22-2S:j6/S7jl70145-0.5

used: py, pyrimidine(s); $03.00/O

pu, purine(s).

form by precipitation with ethanol/aqueous NaCl (20 to 30% overall yield). A crystal of the octamer GTGTACAC grew after several months in a crystallization droplet that contained DNA, magnesium chloride and spermine in a molar ratio of 2 : 3 : 2, while the vapor from the droplet was in equilibrium with that of a reservoir containing 30 y0 (v/v) 2-methyl-2,4-pentanediol. The tetragonal crystal bipyramidal of dimensions 0.5 mmx0.5mm x 1.2 mm diffracted out to 2 A resolution. Precession photography established that the crystal belongs to the tetragonal space group P4,2,2 (P4,2,2) and the cell dimensions are, a = 42.43& c = 24.76A. The space group was chosen to be P4,2,2 because of the similarity of our cell constants to other known octamer crystal structures (Kennard, 1983), and later confirmed by our structure analysis. A three-dimensional intensity data set was collected on an Enraf-Nonius oscillation camera in a coldroom at 4°C with an Elliott GX6 rotating anode X-ray source. The crystal was rotated around the [l lo] axis with each exposure covering 4” oscillation angle. After a total of 48” rotation the data collection was terminated because of crystal deterioration. Precession photographs of the hm and hhl planes taken earlier were used to augment the data. The films were scanned at the Univerisity of Chicago on an Optronix film scanner, and the data were indexed and reduced using the program DENZO. A total of 3708 reflections measured were above 30 level, of which 738 were unique. The effective resolution of the data was 2.5 A. The structure of the molecule was solved using molecular replacement methods. The volume of the unit cell (44,575 A3) is commensurate with the presence of only four molecules of the octamer duplex in the cell. This constrains the molecule to lie on the crystallographic dyad, and limits the variables to the rotation and translation of the molecule about the dyad. For trial models of A, B (Arnott & Hukins, 1972), D (Arnott et al., 1974) and 141

0

1987 Academic

Press Limited

142

8. Jain et al.

(c)

Fig. 1.

Letters to the Editor Z-DNA (Wang et al., 1979), these two parameters were systematically varied while the correlation crystallographic R-factor coefficient and (~(Fobs-F,,,I/~lF,,,I) were used as indicators of the agreement between the model and the diffraction data. In the initial searches with 88 low resolution data to 5 A, A-DNA gave slightly better agreement than the other models tried, but the correlation coefficient did not rise above 35%. A trial model was then constructed based on the crystal structures of GGGGCCCC (McCall et al., 1985) and ‘CCGG (Conner et al., 1982). The helical rise and twist of the central C-G step from the tetramer (‘CCGG) were applied to the central G-C dinucleotide unit of the octamer (GGGGCCCC) to obtain the tetramer model (GC),, which was then extended to the octamer (GC),. The latter model was then mutated to provide the sequence of our octamer, GTGTACAC. Searches using this model gave a correlation coefficient of 61 o/o and an R-factor of 39% in the best orientation. Refinement of this structure was then carried out using the program NUCLSQ (Westhof et at., 1985), which utilizes the constrained atomic refinement algorithms of Hendrickson & Konnert (1979). Initially, sugar puckers and torsion angles were kept free of constraints, but later some constraints were introduced to improve the geometry. The progress of the refinement was regularly examined on an Evans and Sutherland PS300 molecular graphics system using 2F,- F, maps, and some of the more pronounced deviations in the model were corrected manually. After several cycles of refinement the correlation coefficient was 91 o/o and the R-factor dropped to its present value of 23% without the inclusion of water molecules. The structural details reported here are likely to change only slightly on inclusion of water molecules and further refinement. Our present DNA structure shows several interesting features. Firstly, the twist angle between internal base-pair stacks alternates between a high value for the pu-py step, and a low value for the py-pu step, and the inverse alternation is seen for the change in roll angles. Thus, these results are in qualitative agreement with Calladine’s (1982) rules. Secondly, the average tilt angle of base-pairs is approximately 9*3”, which is significantly smaller than the value of 19” for the ideal A-DNA, and that of the other A-DNA fragments solved. Additionally, the average displacement of the base-pairs from the helix axis of approximately 3.1 A is also smaller than 4.5 A seen in fiber-diffraction A-DNA. Consequently the major groove of GTGTACAC is not as deep and the minor groove not as shallow as that of an A-DNA fiber (Fig. 1). Tilt and displacement are two parameters that seemingly distinguish A-DNA (tilt = 19”, displacement = 4.5 A), and B-DNA (tilt = -6”, displacement = 0 A) helices. From this criterion it

143

‘6 I1

8

IO 3

12 I

14 I

16 *

18 40 20

Bose-pair

22I1

24

26I

lilt (“1

Figure 2. Correlation plot between the base-pair tilt angle and the base-pair displacement from the helix axis. The correlation coefficient is calculated to be 0431. The 5 octanucleotides included are: ( n ) d(‘CCGG) (Conner et al., 1982); (0) d(GGCCGGCC) (Wang et al., ,l982a); (0) d(GGB’UAB’UACC) (Shakked et al., 1983); (0) r(GCG)d-

(TATACGC) (Wang et al., 1982b); (0) d(GTGTACAC) (present work); (8) ideal A-DNA. In the plot the data representing terminal base-pairs are not included because of possible end-effects. For some oligonucleotides fewer than expected data points are seen because of overlap. The line shown is the least-squares fit. of the data points.

appears that our octamer is an intermediate between A and B-DNA helices, with the tilt angle close to the halfway point between A and B-DNA, while the displacement is closer to the ideal A-DNA. The high tilt angle in A-DNA resulting from C-3’endo sugar pucker and the low glycosidic angle (Sundaralingam , 1969) leads to steric clashes between adjacent residues that can be relieved by displacement of the base-pairs away from the helix axis. Therefore, it would appear that the tilt angle and displacement are correlated. This is indeed observed in the A-DNA oligonucleotide structures that have been solved (Fig. 2). The correlation coefficient between tilt angles and displacement for all the inner base-pairs of five A-DNA oligonucleotides is 81%. It is seen (Fig. 2) that these parameters for the A-DNA fragments from singlecrystal studies are typically lower than those found in A-DNA fibers. The sequence of the octamer was chosen because (GT),/(CA), sequences are widespread in natural DNA (Trifonov et al., 1985). Approximately 50 base-pair long stretches of these sequences occur with a frequency of one per 30 x lo3 to 100 x lo3 basepairs in all eukaryotic genomes examined so far, including that of humans (Hamada et al., 1982). They have also been found to occur in DNA segments flanking the coding region (Nordheim & Rich, 1983). Additionally, the triplet GTGjCAC is found frequently in the genetic regulatory regions, and has been proposed

to form a possible

for the proteins that interact

“detent”

with DNA (Lu et al.,

Figure 1. (a) A stereo view of the d(GTGTACAC) octamer duplex showing the major (deep) and the minor (shallow) grooves. (b) Stereo view of the octamer down the major groove showing base-stacking of the inner base-pairs. (c) Stereo view of the molecule down the helix axis showing the base-pair displacement from the helix axis.

144

S. Jain et al.

1983). Since these sequences are alternating pu-py, and they can be induced to form Z-DNA under high torsional stress of negative supercoiling, it has been suggested that they may play a biological role by undergoing a B to 2 transition (Nordheim & Rich, 1983). Fiber diffraction studies also revealed that poly(GT/AC) seq uences can form Z-DNA structure (Arnott et al., 1980). Nevertheless, experiments in viva failed to find any evidence for the existence of Z-DNA in these sequences (Rodriguez-Campos et al., 1986; Gross et al., 1985). It is interesting to note that our present octamer represents the first unmodified alternating pu-py DNA sequence that forms a right-handed double helix, and it is also the first one that begins with a purine on its 5’ end. From circular dichroism studies sequences in of (GC), and (CG), oligonucleotide solution, Quadrifoglio et al. (1984) have concluded that a 5’ purine start in an alternating pu-py sequence shifts the equilibrium towards a righthanded conformation from a left-handed one. This 5’ purine effect may explain our results as well. A rationale for the observed left to right helical shift may stem from the differential energetics of pu-py and py-pu dinucleotide blocks in the two different conformations. In Z-DNA. the two base-pairs of the py-pu unit show a small twist angle but are highly sheared, causing the interstrand pyrimidines to stack with each other, and the purines to form intrastrand sugar-base interactions (Wang et al., 1979). The net result is the formation of a compact py-pu block with good stacking. In contrast, in the pu-py dinucleotide unit of Z-DNA, the shear is absent and there is a high twist angle between adjacent base-pairs. Therefore, it appears that in Z-DNA, the py-pu unit shows somewhat better stacking than the pu-py unit. Conversely, in righthanded DNA the degree of base-pair stacking is substantially better the py-pu sequence pu-py unit more

(Manzini

in the pu-py sequence than in (Bugg et al, 1971), making the stable than the py-pu unit

et al., 1987). This could translate

into a

greater stability of d(pu-py), in the right-handed conformation as it has n pu-py blocks and only n-l

py-pu blocks, whereas the opposite will be the case for the d(py-pu), sequence. The above phenomenon may also have some biological implications, since in a contiguous genomic DNA an alternating py-pu sequence, whenever flanked by a purine on its 5’ side, may required higher energy to form the lefthanded Z-DNA. Previous X-ray crystallographic studies have shown that many oligonucleotides of differing sequences adopt A -DNA conformation. Our present study demonstrates that an alternating pu-py sequence can also form an A-DNA structure. These of diverse oligonucleotide crystal structures sequences show large variability in many of the double-helical parameters, which in some instances take intermediate values between A and B-DNA. B-DNA may also be expected to show similar It then variations in helical parameters. appears that A-B junctions may not be high-

energy conformations. It can therefore be envisaged that in local regions of genomic DNA, a large number of possible permutations and combinations of the A-DNA (C-3/-e&o) and B-DNA (C-2’-endo) nucleotide conformations (Sundaralingam, 1969) can give rise to a rich diversity of configurations, quite distinct from B-DNA. In conclusion, local different regions of DNA can have inherently conformations as determined by the base sequence that can be exploited for recognition by DNA binding proteins. We thank Mr Z. Otwinowski and Professor P. Sigler of the Univerisity of Chicago for the use of their film

processing programs and film scanner, Dr S. T. Rao for helpful assistance, the National Institutes of Health for research grants GM-17378 and GM-18455, and the College of Agricultural and Life Sciences and the University Graduate School for their continued support.

Sanjeev Jain’ Gerald Zod Muttaiya Sundarakingam ’ t 1Department of Biochemistry College of Agricultural and Life Sciences University of Wisconsin Madison, WI 53706, U.S.A. ‘Applied Biosystems Foster City, CA 94494, U.S.A. Received 25 November 8 May 1987 t Author

to whom

1986, and in revised form

correspondence

should

be addressed.

References Arnott, 8. & Hukins, D. W. L. (1972). Biochem. Biophys. Res. Commun. 47, 1504-1509. Arnott, S.; Chandrasekaran, R., Hukins, D. W. L., Smith, P. ,J. C. & Watts, L. (1974). J. Mol. Biol. 88, 523533. Arnott, S., Chandrasekaran, R., Birdsall, D. L., Leslie, A. G. W. & Ratliff, R. L. (1986). Nature (London), 283, 743-745.

Bugg, C. E., Thomas, J. M., Sundaralingam, M. & Rao, S. T. (1971). Biopolymers, 10, 175-219. Calladine, C. R. (1982). J. Mol. BioZ. 161, 343-352. Conner, B. N., Takano, T., Tanaka, S., Itakura, J. & Dickerson, R. E. (1982). Nature (London), 295, 295299. Dickerson, R. E., Drew, H. R., Conner, B. N., Wing, R. M., Fratini. A. V. & Kopka, M. L. (1982). Science,

216, 475485. L. J. & Dorman, M. A., Noble, S. A., McBridge, Caruthers, M. H. (1984). Tetrahedron, 40, 95-102. Gross, D. S., Huang, S. Y. & Gerrard, W. T. ( 1985). J. Mol. Biol. 183, 251-265. Hamada, H., Petrino, M. G. & Kakunaga, T. (1982). Proc. Nat. Acad. Sci., U.S.A. 79, 646-469. Hendrickson, W. A.‘. & Konnert, J. (1979). In Biomolecular Structure, Conformation, Function and Evolution (Srinivasan, R., ed.), pp. 43-57. Pergamon Press, Oxford. Kennard, 0. (1983). In Nucleic Acids; The Vectors oj Life (Pullman, B. & Jortner, J., eds), pp. 3348, Reidel Publishing Company, Dordrecht.

Letters to the Editor

1,~~ P., Cheung, S. & Amdt, K. (1983). J. Biomol. Struct. Dynam. 1, 509521. Manzini, G.. Xodo, L. E., Quadrifoglio, F., van Boom, J. H. & van der Marel, G. A. (1987). J. Biomol. Struct. Dynam. 4, 651-662. McCall. M., Brown. T. & Kennard, 0. (1985). J. Mol. Biol. 183, 385396. McCall, M.. Brown, T., Hunter, W. N. & Kennard, 0. (1986). Nature (London), 322, 661-664. Sordheim, A. & Rich, A. (1983). Proc. Nat. Acad. Sci., U.S.A. 80, 1821-1825. Quadrifoglio, F., Manzini, G. & Yathindra, N. (1984). J. Mol. BioZ. 175, 41!+423. Rich, A., Nordheim, A. & Wang, A. H. J. (1984). Annu. Rev. Biochem. 53, 791-846. Rodriguez-Campos, A., Ellison, M. J., Perez-Grau, L. t Azorin, F. (1986). EMBO J. 5, 1727-1734. Shakked, Z., Rabinovich, D., Kennard, O., Cruse, W. B. T., Salisbury, S. A. & Viswamitra, M. A. (1983). J. Mol. BioZ. 166, 183-201.

145

Sundaralingam, M. (1969). Biopolymers, 7, 821-860. Trifonov, E. N., Konopka, A. K. & Jovin. T. M. (1985). FEBS Letters, 185, 197-202. Wang, A. H. J., Quigley, G. J., Kolpak, F. J., Crawford, J. L., van Boom, J. H., van der Marel, G. t Rich, A. (1979). Nature (London), 282, 68W586. Wang, A. H. J., Fujii, S., van Boom, J. H. & Rich, A. (1982a). Proc. Nat. Acad. Sci., U.S.A. 79, 3968-3972. Wang, A. H. J., Fujii, S., van Boom. ,J. H., van der Marel, G. A., van Boeekel, S. A. A. & Rich, A. (19826). Nature (London), 299, 601-604. Wang, A. H. J., Gessner, R. V., van der Marel, G. A., van Boom, J. H. & Rich, A. (1985). Proc. Nat. Acad. Sci., U.S.A. 82, 3611-3615. Westhof, E., Dumas, P. & Moras, D. (1985). J. Mol. HioZ. 184, 119145. Zon, G. & Thompson, J. A. (1986). BioChromatogr. 1. 2225.

Edited by A. Klug