Table of Contents: Abstracts

As of February 12, 1999

Stabilization of Double Stranded Homologous poly(dA)·poly(dT) Strands by Taxol®

Gerlinde Bischoff*1, Ulrich Gromann1, Sabine Lindau1, Robert Bischoff2, Christian Bohley1, Walter-Vesély Meister1, Siegfried Hoffmann*1

1Martin Luther University Halle-Wittenberg,
Institute of Biochemistry, Kurt-Mothes-Str. 3,
D-06120 Halle (Saale), Germany
2Sensobi Sensoren GmbH,
Weinbergweg 22,
D-06120 Halle (Saale), Germany
*Author to whom correspondence should be addressed. Phone: ++49-345-55 248 55; E-mail: bischoff@fu-berlin.de

The binding of paclitaxel to synthetic (poly(dA)poly(dT), poly[d(A-T)]poly[d(A-T)], poly(dG)poly(dC), poly[d(G-C)]poly[d(G-C)]), and natural (calf thymus) double helical DNA has been characterized by UV-absorption and circular dichroism (CD) spectroscopy and thermal denaturation. Taxol", an oncological preparation containing the agent paclitaxel [1-3], ethanol and castor oil, and paclitaxel itself show molecular recognition to AT base pairs with a high affinity to homologous (dA)(dT) sequences, while no interaction with GC base pairs takes place. An astonishing stabilization of the DNA duplex up to deltaTm = 25 °C could be observed by Tm-measurements with poly(dA)poly(dT)-paclitaxel (and Taxol®) complexes. Contrary to this, circular dichroism (CD) measurements of the complexes result in reduced intensities of the A-T nucleic acid signals (also observed at poly[d(A-T)]poly[d(A-T)]-paclitaxel complexes), explained by a disorder of the A-T base pairings under influence of paclitaxel molecules (paclitaxel references considered).

However, the CD-signal of the optical active molecule paclitaxel (11 chiral atoms, CD-signal around 297 nm), does not change in presence of the DNA-drug complexes. So groups of the paclitaxel molecule, which have no optical activity, may perform the interaction with the DNA. In the paclitaxel molecule (Figure 1) hydrophilic groups (e.g., acetyl, C4 and C10 position), together with hydrophobic side groups (e.g., benzoate, C2 position) are laterally arranged by the taxane central part. This formation seems to have the possibility to interact with the hydrophilic phosphate groups at both sides of the DNA backbone (explanation for the strong Tm-shift) and the hydrophobic bases (explanation for the CD-signals). Our proposal consists of the speculation that a strong stabilization occurs by clipping together the backbones under the influence of paclitaxel. This will be supported by the narrower groove of d(AT) strands [4-6].

Figure 1: Chemical structure of paclitaxel and derivatives.

Paclitaxel, first isolated by Wani et al. [1] from the yew tree Taxus brevifolia, acts oncologically by increasing the stability of microtubules and preventing mitosis [2,3,7]. On the other hand supplementary to the microtubules focused studies, M.G. Solis Recendez et al. found 1996 a two fold higher concentration of paclitaxel in the nucleus than in the cytosol of a human lung tumor cell line [8]. This prompted our present study by investigating DNA-paclitaxel complexes. Recently paclitaxel derivatives, synthesized by other groups were described with their cytotoxicity and microtubules activity in the literature [9-13]. The developed derivatives are sometimes more effective in the cytotoxic activity, but not in all cases correlated with an increased microtubuli stabilization. Under consideration of this, we investigated the structures of the DNA-paclitaxel and DNA-paclitaxel derivative -complexes by molecular modeling calculations.

References and Footnotes

1. M. C. Wani, H. L. Taylor, M. E. Wall, P. Coggon, A. T. McPhail, J. Am. Chem. Soc. 93, 2325-2327, 1971.
2. P. Schiff, J. Fant, S. Horwitz, Nature 277, 665-667, 1979.
3. K. C. Nicolaou, W.-M. Dai, R. K. Guy, Angew. Chem. Int. Ed. Engl. 33, 15-66, 1994.
4. H. C. M. Nelson, J. T. Finch, B. F. Luisi, A. Klug, Nature 330, 221-226, 1987.
5. C. Yoon, G. G. Prive, D. S. Goodsell,R. E. Dickerson, Proc. Natl. Acad. Sci. USA 85, 6332-6336, 1988.
6. A. D. DiGabriele, T. A. Steitz, J. Mol. Biol. 231,1024-1039, 1993.
7. R. E. Smith, D. E. Thornton, J. Allen, Seminars in Oncology 22, 41-46, 1995.
8. M. G. Solis Recendez, F. Bichat, F. Grossin, H. Barbault, D. Khayat, G. Bastian, Anti-Cancer Treatment 15, 105, 1996.
9. A. C. Chaudhary, M. M. Gharpure, J. M. Rimoldi, M. D. Chordia, A. A. Leslie Gunatilaka, D. G. I. Kingston, J. Am. Chem. Soc. 116, 4097-4098, 1994.
10. X. Liang, D. G. I. Kingston, B. H. Long, C. A. Fairchild, K. A. Johnston, Tetrahedron Letters 43, 7795-7798, 1995.
11. X. Liang, D. G. I. Kingston, C. M. Lin, E. Hamel, Tetrahedron Letters 36, 2901-2904, 1995.
12. M. D. Chordia, D. G. I. Kingston, E. Hamel, C. M. Lin, B. H. Long, C. A. Fairchild, K. A. Johnston, W. C. Rose, Bioorganic & Medical Chemistry 5, 941-947, 1997.
13. M. D. Chordia, D. G. I. Kingston, Tetrahedron 53, 5699-5710, 1997.

DNA AT-Rich Regions: Some Structure-Function Relations

Anna Gabrielian
Institute of Molecular Biology,
NAS,Armenia,Yerevan

The detection and localization of DNA early-melting AT-rich regions, the study of their structural peculiarities and conformational fluctuations under different agents action are important because of these regions significant biological role. The purpose of this study is to receive the structure-function correlations, concerning the AT-rich regions of different DNA-s and these regions participation in DNA functional activity.

The improved UV-high resolution melting technique in combination with several other physical and physical-chemical approaches, including the method of selective fragmentation of AT-rich regions and following EM and electrophoretic analysis of remaining double-helical fragments, have been used.

The comparative characteristics of the primary structure peculiatities of the various DNA-s (phages, bacteria, plasmids) have been obtained. The light-melted AT-rich regions have been detected and localized. Their structural features in esch DNA apparently have relation to the activity shades. There is coincidence of AT-rich regions positions with integration-excision sites in P22 and dp Salmonella phages' DNA-s and with transposed genetic elements in highmolecular plasmid DNA-s (Tn3 in RP4 and its derivatives).

The complexes between double-helical DNA And DNA-binding 31 kDa protein from rat liver cell plasmatic membranes have been studied. The analysis of plasmid pAO3 DNA differential melting curve within the scope of the theoretical concepts of LAndo-Freedman,using also results of model investigations and binding constant determination allowed to elucidate the interaction process in details. This protein stabilizator was found as an AT-specific, extended ligand, cooperatively binding to double-stranded DNA. The interaction between protein molecules during their adsorption on DNA macromolecules was shown. A possible role of this internal membrane protein in transmembrane transfer process of exogenic DNA is discussed.

Effect of the Concentration of the Multimodal Ligands on the Stabilization of the Double Helix of DNA

Armen T. Karapetian1*, Grigor A. Manukin1 and Pogos O. Vardevanian2
1Yerevan Institute of Architecture and Construction,
105 Terian St., 2 bld.,
Yerevan 375009, Armenia
2Yerevan State University,
Yerevan 375049, Armenia
*Author to whom correspondence should be addressed. E-mail: ares@arm.r.am

Theoretical treatment of melting of complexes of DNA with the ligands having several types of binding sites on helical and single-stranded DNA (multimodal ligands) showed that the width of transition of the complexes becomes equal to that of "pure" DNA when the total concentration of the ligand on both forms of DNA are equal to each other at the transition point. Further increase in ligands concentration leads to a sharp increase of the width of transition of the complexes. In this case transition point of the complexes decreases which indicates that former stabilizer becomes a destabilizer [1].

We are going to present the experimental data proving the prediction of the theory at the case of Ethidium bromide and Actinomycine D.

We experimentally established the value of concentration at which these well-known stabilizers of the double helix of DNA begin to destabilize it.

Dependence on the ionic strength of the concentrations are also investigated.

References and Footnotes

1. A.T. Karapetian et al., J. Biomolec. Struct. Dyn. 14, 275-283, 1996.

Experimental Study of Triple Helices with Antiparallel and Parallel Third Strands Containing G and T Bases

E. Khomyakova1,3, H. Gousset2, J. Liquier2, V. Florentiev1, A. Mirzabekov2 and E. Taillandier2
1Engelhardt Institute of Molecular Biology,
Russian Academy of Sciences,
32 Vavilov St., 117984 Moscow, Russia
2Laboratoire de Spectrocopie Biomoléculaire,
URA CNRS 1430,
Université Paris Nord,
74 rue Marcel Cachin,
F93017 Bobigny Cedex, France
3. Argonne National Laboratory
Argonne, IL 60439

Many recent studies suggest that triple helices exist in eukaryotic cells and can play a role in transcriptional regulation and in replication. Besides the usual investigation techniques common for characterizing triple helical formation (UV melting, CD spectroscopy, chemical probing and S1 nuclease footprinting) we have used a new technique in which targeting of polypurine sequences in duplexes was demonstrated on oligonucleotide microchips by thermal denaturation. This technique is very successful to quickly test the formation of a large number of potential triple helices.

Classically recognition of a purine stretch in duplex DNA by an oligonucleotide residing in the duplex major groove can be achieved either by a pyrimidine sequence oriented in a parallel direction with respect to the purine duplex strand or by a purine sequence oriented antiparallelly. Among the possible sequences which could constitute a third strand, sequences containing G and T bases are very difficult to use as in particular they have a high affinity for self association. We have investigated in solution the formation of such (GT) triple helices for both orientations of the third strand. Different CD spectra have been proposed in the litterature for antiparallel GT triple helices. We shall present here the results obtained for antiparallel as well as parallel GT triplexes and discuss the characteristics of the CD spectra reflecting both structures. Modifications of the CD spectra with increasing temperature allow us to follow a biphasic melting of the triplexes. The melting temperatures measured are compared with the Tm classically determined by UV absorbance spectroscopy. Results obtained by S1 nuclease footprinting, KMnO4 and DMS chemical probing are consistent with the spectroscopic data and allow us to suggest base pairing models for the base triplets.

Crystal Structures of Highly Bent RNA Duplexes Forming Tertiary Interactions and Superhelical Coils

Shome Nath Mitra, Ke Shi, Roopa Biswas and Muttaiya Sundaralingam
The Ohio State University,
Departments of Chemistry and Biochemistry,
012 Rightmire Hall,
1060 Carmack Road,
Columbus, OH 43210

Crystal structures of the RNA octamers gugcacaC (I) and guauacaC (II), (II) having two independent molecules (A and B) in the asymmetric unit, have been determined. The structures consist of highly bent RNA duplexes of six Watson-Crick base pairs forming right-handed superhelices. The bent duplexes in the superhelices stack in a head-to-head fashion with g1 against g1 and g9 against g9. The right handed superhelical coils have been observed for the first time, contrary to the well documented rule of stacking of RNA duplexes into infinite pseudocontinuous columns. The deoxy-cytidines in the 3'-overhang terminal rAdC swing out and form antiparallel hemiprotonated trans (C C)+ pairs. The penultimate adenines at one end are syn while at the other are anti and form a*g c base triples in I. In the two independent molecules of II, one molecule is like I while in the other the anti adenine swings out forming a different conformation and stacks over the (C C)+ pair. The 3'- overhang rAdCs of one strand interpenetrate the minor groove of a stacked duplex related by a dyad. This results in ribose zipper interaction between the 3'-halves [r(a5c6a7)dC8] of the octamers. A total of four ribose zipper interactions are formed, the inner two are O2'-O2' interactions and the outer two are O2'-O3' interactions. In both cases the acceptor donates its proton to a base of the donor. The base triple interactions and the ribose zipper interactions will be compared with thos observed in the crystal structures of ribozymes (1) and the P4-P6 domain of group I intron (2).

Work supported by the NIH grant GM-17378 and an Ohio Eminent Scholar Chair and Endowment to MS.

References and Footnotes

1. Pley, H.W., Flaherty, K.M. and McKay, D.B., Nature 372, 68-74, 1994.
2. Cate, J.H., Gooding, A.R., Podell, E., Zhou, K., Golden, B., Kundrot, C., Cech, T. and Doudna, J. Science 273, 1678-1685 ,1996.

Solution Structure of an Antibody-bound HIV-1IIIB V3 Peptide: a Cis Proline Turn Linking Two Beta-Hairpin Strands

Vitali Tugarinov, Anat Zvi, Rina Levy and Jacob Anglister*
*Author to whom correspondence should be addressed. Phone: ++972 8 9343394; Fax ++972 8 9344136; E-mail: bpanglis@weizmann.weizmann.ac.il

Segments of the V3 loop or its immediate vicinity form the binding site for the gp120 co-receptors on T-cells and macrophages, and its sequence determines the phenotype of the virus. Most recently, the crystal structure of a CD4/gp120/antibody Fab complex was solved with the V3 loop as well as other variable loops deleted from the gp120 construct to enable crystallization. Using isotope filtered and isotope edited NMR experiments we determined the three dimensional structure of an 18-residue V3 peptide bound to the Fv fragment of a potent anti-gp120 HIV-neutralizing antibody. The refined solution structure of the HIV-1IIIB V3 peptide reveals an unexpected type VI beta-turn comprising residues RGPG at the center of a beta-hairpin. The central glycine and proline of this turn are linked by a cis peptide bond. The residues of the turn interact extensively with the antibody Fv. The solution structure is significantly different from the X-ray structures of V3MN peptides bound to anti-peptide antibodies where a type II beta-turn was observed in the GPGR sequence and the beta-hairpin conformation was not observed. These differences could be due to a two-residue (QR) insertion preceding the GPGR sequence in the HIV-1IIIB strain, and a longer peptide epitope immobilized by the anti-gp120 antibody. In view of its conservation and surface exposure it is very likely that the segment GPG plays an important functional role. Recent studies suggest that a cis-trans isomerization of the proline may be required for those isolates with type II beta turn.

A Method for Determining the Energy Hierarchy in Macromolecules of Globular Proteins. The Problem of Domain Structure Definition.

Igor N. Berezovsky*1,3, Vladimir A. Namiot2, Vladimir G. Tumanyan1 and Natalia G. Esipova1
1Engelhardt Institute of Molecular Biology,
Russian Academy of Sciences,
Moscow 117984, Russian Federation,
2Research Institute of Nuclear Physics of Moscow University,
Moscow 117234, Russian Federation,
3Department of Structural Biology,
The Weizmann Institute of Science,
P.O.B. 26, Rehovot 76100, Israel.

*Author to whom correspondence should be addressed. Phone:+972-8-9343367; Fax:+972-8-9342653; E-Mail: HYPERLINK mailto:csigor@wis.weizmann.ac.il; csigor@wis.weizmann.ac.il.

The conclusion that proteins consist of distinct, semi-independent, and stable structural fragments had been drawn from the results of limited proteolysis before the first protein X-ray structure was determined. Practically at the very beginning of X-ray studies on protein spatial structure, the concepts of domains and domain arrangement have been introduced. Despite the common use of this concept over three decades, there is yet no unequivocal and commonly accepted definition of the domain. Moreover, depending on the field of research (enzymology, biochemistry, sequence analysis, structural biology, microcalorimetry, etc.) and the experimental methods or theoretical approaches applied, domain may mean: functional sites of enzymes; cooperative units revealed by microcalorimetry; regions of amino acid sequences corresponding to the same structural motifs in different proteins; parts of spatial structures found as regions of high electron density in X-ray analysis; parts of structures calculated on the basis of inter-residue contacts, interface area, or compactness of structural units.

The reasons for advancing the new method are as follows:

1. Electron density alone is not satisfactory for unambiguous determination of the domain structure or a priori protein spatial structure determination.

2. Microcalorimetric experiments yield cooperative units which may correlate or not correlate with the structural domains. Moreover, the globule can melt as different sets of cooperative units depending on pH, ion concentration, etc.

3. Various energy domains and complex distribution of electron density are revealed in most of the multidomain globular proteins.

We propose a new method for determining the domain structure of globular proteins based on the entirely physical model of van der Waals interactions. Various levels of hierarchy in the protein spatial structure are delineated by analysis of the energy interaction between structural units of different scales. The algorithm readily and accurately locates domains formed by continuous segments of the protein chain as well as those comprising nonsequential segments. Our results demonstrate that division into domains which correlates with the X-ray definition of domains is attained at different levels of hierarchical organization. Besides, our method of domain identification includes geometrical methods of domain isolation as particular cases. We use the term "domain" to denote energy-independent parts of the protein spatial structure. Simultaneously, isolated fragments can be ordered into hierarchical series with certain relationships between the energies of domain boundaries. In this case, the notion "domain" is elaborated to the notion "domain of the particular level of the hierarchy". Thus, the notions "domain by electron density distribution" and "domain by geometrical criteria" are added as well. As a result, the notion "domain of the particular level of energy hierarchy" derived from the van der Waals energy capacity gives us a possibility to perform comprehensive analysis of not only the structural but also the energy aspects of the protein structure.

Fold-Back DNA Tetraplex Specieses in DNAse I-Resistant DNA Isolated From Hela Cells

En-hua Cao1, Xue-guang Sun1, Xiao-yan Zhang1 and Chun-li Bai2
1Institute of Biophysics,
Chinese Academy of Sciences,
Beijing 100101, China
2Institute of Chemistry,
Chinese Academy of Sciences,
Beijing 100080, China

Recent studies in vitro reported that (1) the telomere of the linear chromosomes of eukarotes are composed of tandem repeats of short DNA sequences (TTAGGG)n, one strand being rich in guanine. Telomeric G-strand DNAs from a variety of organisms adopt compact structures, the most stable of which is explained by the formation of G-quartets; (2) A major nuclease­hypersensitive element III (sequence: 5'-TGGGGAGGGTGGGGAGGGTGGGGAAGG-3') termed NHE, which corresponds to bases 2186-2212 in the sequence of the human c-myc locus of the human c-myc oncogene has an unusual strand asymmetry: one strand is an almost perfect homopurine tract. It adopts an intrastrand fold-back DNA tetraplex. We believe that formation of the tetraplex may play an important role in vivo. It is unclear if there is such fold-back DNA tetraplex in cells or DNA sample isolated from cells. We present experiments indicating that DNA sample from HeLa cellS has a DNase I -resistant species, which were about 40 -50 bp fragment sizes responding to double-strand DNA marker by gel electrophoresis. The image of atomic force microscopy showed that this species has was assumed to be quadruplex structure according to its apparent width and height, which was observeously different from the standard duplex DNA. Its UV, CD spectrum also exhibited characteristics different from the standard duplex DNA. Results from HPLC analysis showed it has higher G content. A syntheric 27-base-long oligonucleotide (5'-TGGGGAGGGTGGGGAGGGTGGGGAAGG-3') or the human telomere sequence (TTAGGG)4 could be hybridized to isolated DNase I -resistant species. All results strongly suggest that the DNase I -resistant DNA specie isolated from HeLa cell has at least two components which adopt an intrastrand fold-back DNA tetraplex. Their sequences were similar to human telomere or NHE of the human c-myc oncogene, respectively.

Novel Sheared A:C Pairing in DNA

Shan-Ho Chou
Institute of Biochemsitry,
National Chung-Hsing University,
Taichung, 40227, Taiwan.
*For author correspondence. Phone: 04-285-3486, Fax: 04-285-3487, E-mail: shchou@dragon.nchu.edu.tw

Recent progress in SELEX methodology has allowed one to select ssDNA molecules with catalytic activities or ssDNA with strong affinity toward any target (monoclonal aptamers). The structures or conformation these biologically active ssDNA molecules are able to adopt are thus of quite interests. It is not unreasonable to expect that they may adopt as many structural varieties as those in protein. In the past, several stable mismatches have been discovered in DNA duplex. The most notable are the so-called sheared G:A mispair, which when present in tandem pair, can be stably incorporated into a duplex without any destabilization. The extension of the sheared pairing family to A:A pair or G:G pair, or the (GGA)2 motif with the central two unpaired guanosines intercalating with each other and with the two guanosines participating in the bracketing sheared GA pairs, have also been found to be stable and exist in biologiclaly important DNA regions. We now extend these sheared purine:purine pairing to a purine:pyrimidine (A:C) pairing. Several cross base-pair NOEs derived from H2O- or D2O-NOESY spectra, i.e. the AH2 to CH5, CNH2 to AH1', AH4', and CNH2 to AH2 etc. clearly define these novel shearing nature. This sheared A:C pair can exist alone to close DNA hairpin loops with single residue, or adjoin with another sheared GA pair to form a tandem sheared base pairs. These motifs are possibly adopted in the loop of a general RNA-cleaving DNA enzyme, in the DNA topoisomerase cutting site of a cruciform promoter, and in the strong human HIV-1 reverse transcriptase inhibitors.

References and Footnotes

1. Chou, S.-H., Zhu, L. & Reid, B. R., J. Mol. Biol., 267, 1055-1067 (1997).
2. Chou, S.-H. & Tseng, Y.-Y., J. Mol. Biol., 285, 41-48 (1999).
3. Chou, S.H., Tseng, Y.-Y. & Wang, X-W., J. Mol. Biol., (in press, 1999).

Structure of Photoreactive Binary System of Oligonucleotide Conjugates Assembled on the Target Nucleotide Sequence

Michail I. Dobrikov1, Elena V. Bichenkova2, Kenneth T. Douglas2, Timur I. Gainutdinov1 and Valentin V. Vlassov1
1Institute of Bioorganic Chemistry,
Russian Academy of Sciences,
8, Lavrentiev Ave.,
Novosibirsk 630090, Russia.
2School of Pharmacy/Pharmaceutical Sciences,
University of Manchester,
Manchester, M13 9PL, UK

Recently we have developed an approach to superspecific photomodification of nucleic acids by binary systems of oligonucleotides conjugated to precursor groups capable of assembling into photoactivatable structure upon simultaneous binding of the conjugates to the target [1-3]. In the binary system shown on the figure one of the groups is a sensitizer (S) which absorbs long wavelength light and transfers the absorbed energy to juxtaposed photoreagent (R), triggering crosslinking of the latter to the target nucleic acid. The advantage of the binary systems are their enhanced reaction efficacy and greater specificity which is determined by independent recognition of a long target sequence by two oligonucleotides.

The initial rate of the sensitized photomodification is greatly affected by the structure of the photoactive complex, because the efficiency of the energy transfer is dependent on the inverse sixth power of the distance between sensitizer and photoreagent. Detailed knowledge of the spatial arrangement of all components of the complex formed by the target nucleic acid sequence and two oligonucleotide conjugates are needed for design of new binary systems with improved properties.

We have investigated the structure of a model binary system (1):(2):(3), where (1) is the target 12-mer 5'-pdGTATCAGTTTCT, (2) is the photoreactive conjugate 5'-dAGAAACp-NH(CH2)2NH-R and (3) is the sensitizing conjugate 5'-Pyr-pdTGATAC (R is p-azidotetrafluorobenzoyl group and Pyr is the pyrenyl-1-methylamino group) by means of high-resolution 2D NMR spectroscopy [4]. Data of 1H NOESY experiments evidence that the Pyr group is located in the minor groove of the complex close to T4, C5, A6 and G7 residues.

The photoreaction within this complex results in crosslinking of reagent 2 with N7-position of the G7 residue thus indicating that the photoreagent R is located in the major groove near G7 residue. The distance between the centroids of the Pyr and R moieties was estimated to be 8-10 A0.

These results and computer modeling studies allowed us to design a new more efficient binary system of oligonucleotide conjugates containing perfluoroarylazide and perylene residues. This system can be activated by visible light (450-580 nm) irradiation, reacts 300000 times faster than the azide conjugate in the absence of the perylene conjugate and provides highly efficient (up to 99%) photocrosslinking with target ssDNA.

This research has been supported by the INTAS-RFBR grant N95-0653.

References and Footnotes

1. Dobrikov, M. I., Gaidamakov, S. A., Koshkin, A. A., Gainutdinov, T. I., Luk'yanchuk, N. P., Shishkin, G. V., Vlassov, V. V., Russian J. Bioorgan. Chem. 22, 166-173 (1996).
2. Vlassov V. V., Dobrikov M. I., Gaidamakov S. A., Gaidamakova E. K., Gainutdinov I., Koshkin A. A., DNA and RNA Cleavers and Chemotherapy of Cancer and Viral Diseases, Ed. D. Meunier. Dordrecht: Kluver Acad. Publ. P. 195-210 (1996).
3. Dobrikov, M. I., Gaidamakov, S. A., Gainutdinov, T. I., Koshkin, A. A., Vlassov, V. V., Antisense Nucl. Acid Drug. Develop. 7, 309-317 (1997).
4. Bichenkova, E. V., Marks D. S., Lokhov S. G., Dobrikov M.I., Vlassov V.V., Douglas K.T., J. Biomol. Struct. Dyn. 15, 307-320 (1997).

Solution Structure and Base-pair Opening Rates of GGCC Containing Oligonucleotides

Utz Dornberger1, Mikael Leijon2 and Hartmut Fritzsche1*
1Institut für Molekularbiologie,
Friedrich-Schiller-Universität,
Winzerlaer Str. 10, D-07745 Jena, Germany
2Stockholm University,
Department of Biophysics,
Arrhenius Laboratory,
S-10691 Stockholm, Sweden
*Author to whom correspondence should be addressed. Phone: +49-3641-657560, Fax: +49-3641-657520; E-mail: hfri@hodler.molebio.uni-jena.de

There is now ample evidence that the sequence-dependent local DNA structure is correlated with biological functions. Bending induced by phased runs of four to six adenine base-pairs (A-tracts) or GGGCCC sequence elements (GC-tracts) was demonstrated by an anomalously slow electrophoretic mobility (Crothers et al., 1990; Brukner et al., 1993). Recently, bending induced by phased runs of TGGCCA sequence elements was suggested based on electrophoretic mobility data (3). These data suggest that a special helical conformation may be responsible for the observed bendability and flexibility of the TGGCCA sequence. Previously, a crystal structure of the DNA decamer d(CATGGCCATG)2 was presented as an example of intrinsic bending (4).

We used 2D 1H- and 31P-NMR spectroscopy combined with restrained molecular dynamics and distance geometry calculations to describe the solution structure of the same DNA decamer d(CATGGCCATG)2. We present a detailed analysis of the helical parameters of the NMR-derived solution structure compared with the crystal structure of this DNA decamer (5). The NMR structure and the previously published crystal structure (4) of the DNA decamer display a specific pattern of stacking interaction in the central GGC base triplet. Considerable differences among the two structures occur in the local conformation of the TG/CA base-pair step. The solution structure of the TG/CA base-pair step obtained by high resolution NMR studies is characterized by a positive roll angle whereas in the crystal this base-pair step tends to adopt unusually high twist angles. We suggest a flexible TpG and a peculiar stacking in GGC tracts as a source of bendability and rapid base-pair opening.

Sequence-dependent structural features of the DNA double-helix have a strong influence on the base-pair opening dynamics. Here we report a detailed study of the kinetics of base-pair breathing in tracts of GC base-pairs in DNA duplexes derived from 1H-NMR measurements of the imino proton exchange rates upon titration with the exchange catalyst ammonia (6). Unexpectedly, the base-pair lifetimes in G-tracts are about ten times shorter than for isolated GC base-pairs. The base-pair lifetimes in the GC-tracts are below 5 ms for almost all of the base-pairs, and the lifetimes decrease even further when the tracts are of the type GnCn (7). The unusually rapid base-pair opening dynamics of GnCn -tracts is in striking contrast to the behavior of AT tracts, where very long base-pair lifetimes are observed. The implication of these findings for the structural principles governing spontaneous helix opening as well as the DNA-binding specificity of the cytosine-5 methyltransferases, where flipping of the cytosine base has been observed, are discussed.

Work supported by grants of DFG (Fr876/5-1), Deutscher Akademischer Austauschdienst and Fonds der Chemischen Industrie (to H.F.).

References and Footnotes

1. Crothers, D. M., Haran, T. E. & Nadeau, J. G., J. Biol. Chem. 265, 7093-7096, 1990.
2. Brukner, I., Dlakic, M., Savic, A., Susic, S., Pongor, S. & Suck, D., Nucleic Acids Res. 21, 1025-1029, 1993.
3. Chastain, P. D. & Sinden, R. R., J. Mol .Biol. 275, 405-411, 1998.
4. Goodsell, D. S., Kopka, M. L., Cascio, D., Dickerson, R. E., Proc. Natl. Acad. Sci. U.S.A. 90, 2930-2934, 1993.
5. Dornberger, U., Flemming, J., Fritzsche, H., J. Mol. Biol. 284, 1453-1463, 1998.
6. Guéron, M., Leroy, J.L., Methods in Enzym. 261, 383-413, 1995.
7. Leijon, M., Dornberger, U., Fritzsche, H., J. Biol. Chem. (in press).

Specific Binding of Lexitropsins to Okazaki Fragments Studies by PFG NMR

William H. Gmeiner1*, Christopher Hudalla1, and J. William Lown2
1Eppley Institute,
University of Nebraska Medical Center,
Omaha, NE 68198-6805
2Department of Chemistry,
University of Alberta,
Edmonton, Alberta Canada T6G-2G2

Okazaki fragments occur as intermediates during lagging strand DNA replication. Alterations in Okazaki fragment structure may contribute to the anticancer activities of nucleoside analogs such as cytarabine, a potent anti-leukemic agent that inhibits lagging strand replication. We recently reported the NMR solution structures for two model Okazaki fragments, [OKA] and [ARAC] (Gmeiner et al., (1999) Biochemistry, In Press). These sequences are derived from a frequent initiation site for primase during replication of the SV-40 viral genome. The sequence of [ARAC] differs from [OKA] only by substitution of cytarabine for one deoxycytidine. The structure of each model Okazaki fragment was elucidated using NMR spectroscopy and restrained molecular dynamics simulations. The solution structures of [OKA] and [ARAC] each consist of two distinct domains: a DNA duplex region (DDR) and an RNA:DNA hybrid duplex region (HDR). The DDR of [OKA] adopts geometry similar to B-form except for variations in helical parameters, especially twist and roll, which occur in the purine-tract increasing base overlap among the five consecutive purines. The helical axes for the DDR and HDR of [OKA] are bent 22° relative to one another. We have applied a variety of solution NMR techniques including VAST and PFG-NMR to analyze the binding of sterically restricted lexitropsins to [OKA]. Lexitropsins are oligopeptidic molecules designed to bind in the minor groove of duplex DNA. The variability in minor groove width between the DDR and HDR, and the bending of the helical axis, makes the identification of structural motifs that bind these unusual structures a challenging NMR structural problem. PFG-NMR has provided insight into the spatial restriction of oligopeptidic moieties needed to preferentially bind Okazaki fragments.

Electrophoretic Mobility of DNA and Protein-DNA Complexes

U. Mohanty
Eugene F. Merkert Chemistry Center,
Boston College,
Chestnut Hill, MA 02167

Ogstron and "tube" reptation models are generalized to describe the mobility of ss oligomeric DNA in polyacrylamide gel. The model predicts, in agreement with experiments, that significant anomalous migration exists with sequences about six residues such that the electrophoretic mobility of a 3-residue fragment is coparable to a 14-residue fragment (1).

A model for the free solution electrophoretic mobility of oilgo ds DNA under buffer and pH conditions relevant to capillary electrophoresis is described. We find, in agreement with experiments that the free solution mobility of ds DNA increases as the molecular weight increases up to a few hundred bp (2,3).

We have developed a model for gel retardation of intrinsically curved DNA molecule, such as the A-tracts, that takes into account in an approximate way the low and the high frequency viscoelastic response of the polyacrylamide gel matrix and the transverse frictional drag forces. Dynamical motions of the "probe" chain are accompanied by viscoelastic response of the "vicinal" matrix chains by reptation. In agreement with Crothers and coworkers, analysis shops that the gel mobility of DNA with a bend at the center is slower than that of the same molecule with a bend at its end (4).

Finally, we will also describe a novel theoretical model of gel-shift retardation of protein-DNA complexes in polyacrylmide gel.

References and Footnotes

1. Mohanty, U., Serls, T., McLaughlin, L., J. Am. Chem. Soc., 120, 8275-8276, 1998.
2. Mohanty, U., Stellwagen, N.C., Biopolymers, in press, 1999.
3. Stellwagen, N.C., Gelfi, C., Righetti, P.G., Biopolymers 42, 687-703, 1997.
4. Mohanty, U., unpublished, 1998.

Optical Probes of DNA Sequence-Directed Bending and Dynamics

Catherine J. Murphy*
Department of Chemistry and Biochemistry,
University of South Carolina,
Columbia, SC 29208 USA
*For author correspondence. Phone: 803-777-3628; Fax: 803-777-9521; E-mail: murphy@psc.sc.edu

Abundant experimental evidence supports the notion that certain DNA sequences are inherently curved. Here we summarize studies from our laboratory that use photoluminescent probe molecules to interrogate these structures in solution. Noncovalent probes include organic intercalators, metallointercalators and protein-sized inorganic nanoparticles; covalent probes include the fluorescein-rhodamine energy transfer dye pair and an unnatural fluorescent base pair analog. Steady-state data suggest that the double-stranded 5'-GGCC-3' motif is quite bent in solution itself. Temperature-dependent optical studies of an organic intercalator in generic DNA suggest that the inside of the DNA base stack is highly mobile at room temperature. The extension of these general conclusions to the development of experimental methods to probe sequence-dependent dynamics on the picosecond time scale will be discussed.

Designed Two-Dimensional DNA Holliday Junction Arrays

Chengde Mao, Weiqiong Sun and Nadrian C. Seeman*
Department of Chemistry,
New York University,
New York, NY 10003, USA
*Author to whom correspondence should be addressed. Phone: 212-998-8395; Fax: 212-260-7905; E-mail: ned.seeman@nyu.edu.

A two-dimensional DNA crystal has been designed and constructed from Holliday junction analogs that contain two helical domains twisted relative to each other. The Holliday junction is not an inherently rigid system, but it can be made less flexible if it is combined into a larger construct. We have fused four junctions into a rhombus-like molecule consisting of four six-turn helices, two on an upper layer and two on a lower layer; the branch points, which define vertices, are separated by four double helical turns each. Ligation of the rhombus-like motifs produces no cyclic species, when assayed by ligation-closure experiments. Self-assembly of the rhombuses in one dimension leads to a linear pattern. The rhombuses can be directed to self-assemble by hydrogen bonding into a 2-dimensional periodic array, whose spacing is six turns in each direction. The expected spacing is seen when the array is observed by atomic force microscopy (AFM). Variation of the dimensions of the parallelogram from 4 turns x 4 turns to 4 turns x 6 turns results in the expected increase in unit cell dimensions. Hence, it is possible to assemble periodic arrays with tunable cavities using these components. This system also provides the opportunity to measure directly the angles or torsion angles between the arms of branched junctions; here we measure the torsion angle between the helical domains of the Holliday junction analog. We find by AFM that the torsion angle between helices is 63.5°, in good agreement with previous estimates.

This research supported by grants from the NIGMS, ONR, NSF/DARPA and USAF.

Modifying the Surface Features of Two-Dimensional DNA Crystals

Furong Liu, Ruojie Sha and Nadrian C. Seeman*
Department of Chemistry,
New York University,
New York, NY 10003, USA
*Author to whom correspondence should be addressed. Phone: 212-998-8395; Fax: 212-260-7905; E-mail: ned.seeman@nyu.edu.

DNA double crossover molecules (DX) are rigid DNA motifs that contain two double helices linked at two different points. It is possible to form hydrogen bonded two-dimensional crystals from DX molecules and to observe those arrays by atomic force microscopy (AFM). The sticky-ends that hold the arrays together can be varied, so as to include diverse periodic arrangements of molecules in the crystal. The inclusion of extra DNA hairpins designed to protrude from the plane of the crystal provides a topographic label that is detected readily in AFM images: By using these labels, it is possible to produce stripes at predicted spacings on the surface of the crystal. The experiments presented here demonstrate that it is possible to modify these patterns, by both enzymatic and non-enzymatic procedures. We show that a hairpin containing a restriction site can be removed quantitatively from the array. We also demonstrate that a sticky end protruding from the array can be ligated to a hairpin containing its complement. In addition, it is possible to anneal a hairpin to the crystalline array by hydrogen bonding, both in solution and after deposition on a mica surface. The ability to modify these arrays increases the diversity of patterns that can be produced from an initial set of DX components. Thus, a single array can be modified in a large number of ways that can alter its physical or chemical features.

This research supported by grants from NIGMS, ONR, NSF/DARPA and USAF.

Parallel Helical Domains in DNA Branched Junctions Containing 5', 5' and 3', 3' Linkages

Ruojie Sha1, Furong Liu1, Michael F. Bruist2 and Nadrian C. Seeman1*
1Department of Chemistry,
New York University,
New York, NY 10003, USA
2Department of Chemistry and Biochemistry,
University of the Sciences in Philadelphia,
600 South 43rd St.,
Philadelphia, PA 19104, USA
*Author to whom correspondence should be addressed. Phone: 212-998-8395; Fax: 212-260-7905; E-mail: ned.seeman@nyu.edu.

The Holliday junction is a central intermediate in genetic recombination. It contains four strands of DNA that are paired into four double helical arms that flank a branch point. In the presence of Mg2+, the four arms are known to stack in pairs forming two helical domains whose orientations are antiparallel, but twisted by about 60°. The basis for the antiparallel orientation of the domains could be either junction structure or the effect of electrostatic repulsion between domains. To discriminate between these two possibilities, we have constructed and characterized an analog, called a Bowtie junction, in which one strand contains a 3', 3' linkage at the branch point, the strand opposite it contains a 5', 5' linkage, and the other two strands contain conventional 3', 5' linkages. Electrostatic effects are expected to lead to an antiparallel structure in this system. We have characterized the molecule in comparison with a conventional immobile branched junction by Ferguson analysis and by observing its thermal transition profile; the two molecules behave virtually identically in these assays. Hydroxyl radical autofootprinting has been used to establish that the unusual linkages occur at the branch point, and that the arms stack to form the same domains as the conventional junction. Cooper-Hagerman gel mobility analyses have been used to determine the relative orientations of the helical domains. Remarkably, we find them to be closer to parallel than to antiparallel, suggesting that the preferred structure of the branch point dominates over electrostatic repulsion. This finding suggests that control of branch point structure alone can lead to parallel domains, which are generally consistent with recombination models derived from genetic data.

This research supported by grants from NIGMS, ONR, NSF/DARPA and USAF.

Two Dimensions and Two States in DNA Nanotechnology

Nadrian C. Seeman1*, Furong Liu1, Chengde Mao1, Xiaoping Yang1, Lisa A. Wenzler1, Roujie Sha1, Weiqiong Sun1 and Erik Winfree2
1Department of Chemistry,
New York University,
New York, NY 10003 USA
2Computation and Neural Science,
California Institute of Technology,
Pasadena CA 91125 USA
*Author to whom correspondence should be addressed. Phone: 212-998-8395; Fax: 212-260-7905; E-mailned.seeman@nyu.edu.

The minimal requirements for components of designed crystals are [1] programmable interactions, [2] predictable local intermolecular structures and [3] rigidity. The sticky-ended association of DNA molecules fulfills the first two criteria, because it is specific and diverse, and it results in the formation of B-DNA. Stable branched DNA molecules permit the formation of networks, but simple branches are too flexible. Antiparallel DNA double crossover (DX) molecules can provide the necessary rigidity, so we can use components with which we tile the plane. It is possible to include DNA hairpins that act as topographic labels for this 2-D crystalline array, because they protrude normal to it. By altering the sticky ends, it is possible to change the topographic features, and to detect these changes by means of AFM. We can modify the array by restricting hairpins or by adding them to the array. In the drawing, the dark circle in Bo represents a hairpin and the light circle in D* can represent a hairpin, a hairpin containing a restriction site or a sticky end to which a hairpin can be added. Parallelograms of single-crossover molecules are sufficiently rigid that they can be used to produce 2D arrays with tunable cavities.

The rigidity of the DX motif can also be exploited to produce a nanomechanical device predicted on the B-Z transition:

This effect is shown by fluorescence energy resonance transfer, because two dyes (filled and empty circles above) have different separations in the two states

This research supported by grants from NIGMS, ONR, NSF/DARPA and USAF.

Conformational Dynamics of a Nucleic Acid Molecule over Hydration-Dehydration Cycle: A Theoretical Study

M.Ye. Tolstorukov, S.V. Gatash and V.Ya. Maleev
Chair of Molecular and Applied Biophysics,
Kharkov State University, Svobody
Sq. 4, 310077, Kharkov, Ukraine

The nucleic acid hydration shell is one of the main factors determining its conformation. Thus, the interaction of water molecules and nucleic acids is of great biological importance. Convenient samples to study nucleic acid-water system are wetted nucleic acid fibers and films. The hydration shell stabilizes the ordered conformations of nucleic acids forming the regular structures due to H-bonds among sorbed water molecules, ''water backbones'', which are specific for every conformation. The hydration shell is formed with the increasing of the water content of the sample, and nucleic acid transits from unordered to A- and then to B- form, in the case of DNA and salt contents similar to the in vivo conditions. Dehydration of nucleic acids results in the reverse conformational transitions but they occur at the values of relative humidity less then ones for direct transitions. Thus, there is a conformational hysteresis over the hydration-dehydration loop. The sorption isotherms of the nucleic acids also demonstrate the hysteresis phenomenon. These hysteresis phenomena are reproducible and do not depend on the size of the experimental sample.

In the present paper a semi-phenomenological model taking into account the mutual dependence of hydration and nucleic acid conformation transition processes and its verification with experimental water sorption isotherms of DNA are offered.

We consider a finite space, which contains the nucleic acid sample and is in contact with a bath of water or water vapor. That allows one to support the relative humidity inside the experimental space at the constant level and change it when necessary. Such a scheme corresponds to the real experiments with wetted nucleic acid samples. A nucleic acid molecule is simulated by the sequence of monomer units (nucleotide pairs). Every unit can be found in the unordered state or in one of the ordered conformations. Units can reversibly transit from one state to another. We consider three types of water sorption on biopolymer units: sorption with and without saturation and multilayer sorption. The temporal evolution of the system is described by set of the partial differential equations of diffusion type. To take into account the mutual dependence of the sorption and conformation transitions we assume that equilibrium constants of sorption depend exponentially on the conformational variable and equilibrium constants of conformational transitions depend exponentially on the sorption variables.

Both cooperative and non-cooperative conformational transitions of the nucleic acid can be described in the framework of the model. The analysis of kinetic equations for the conformational transitions and water sorption process shows non-trivial bifurcation behavior of the system, which leads to multistability. This fact allows one to explain the hysteresis phenomena observed. The sorption isotherms have been obtained for fibers of calf thymus DNA by using the gravimetry technique over the whole (0-100-0% of relative humidity) and restricted (0-90-0%, 0-65-0%) hydrationvdehydration cycles. All isotherms demonstrate sorption hysteresis phenomenon, except ones for the 0-65-0% cycle. Theoretical and experimental data are in good agreement over every relative humidity interval considered. The results obtained allow us to conclude that the hysteresis phenomena are caused by the cooperative conformational transitions of DNA molecule. The spatially extended model allows one to describe possible autowave processes in the system, if diffusion of sorbed water molecules is taken into account. It has been established that propagation of a new conformation along the nucleic acid molecule during structural rebuilding is an autowave process of the travelling front type. The time evolution of the conformational perturbation of the nucleic acid molecule has been investigated under various relative humidity values. For the multistable region it has been shown that when a fluctuation appears in the system, growing, it can occupy the whole system or collapse for finite time. Which scenario will be realized depends on the relative stability of conformational states, which in its turn is determined by the water content of the samples. Also, the spatially extended model allows describing the conformational dynamics of the nucleic acids with inhomogeneous primary structures.

M-DNA: A Physiologically Stable New Form of DNA Which is an Effective Molecular Wire

Palok Aich1, Shaunivan L. Labiuk1, Krista Wooller1, Les W. Tari3, Kenneth J. Falk2, Louis J. T. Delbaere1, William J. Roesler1, Ronald P. Steer2 and Jeremy S. Lee1
1Department of Biochemistry,
2Department of Chemistry,
University of Saskatchewan,
Saskatoon, Saskatchewan, S7N 5E5,
3Department of Biological Sciences,
Calgary, Alberta, CANADA.

*Author to whom correspondence should be addressed. Phone:+1-306-966-4374; Fax:+1-306-9664390; E-mail: aich@duke.usask.ca

Electron transfer through the p-stack of DNA base-pairs has become an important area of research in recent years. It has been proposed that the stacked aromatic bases of DNA may act as a 'p -way' for the efficient transfer of electrons. There are examples in the literature that photoinduced electron transfer occurred between donor and acceptor molecules separated by a few Å on a DNA molecule. However, kinetic analysis of distance-dependent electron transfer in a DNA hairpin suggested that DNA is only somewhat more effective than proteins as a conductor of electrons.

We report here a new form of DNA, M-DNA, that not only contains the p-stack but also an interchelated metal ion (Zn2+, Co2+, or Ni2+), making it an excellent candidate for the conductance of electrons. Efficient electron transfer is observed between two fluorophores separated by at least 150Å in an M-DNA duplex. Moreover, addition of a sequence-specific DNA-binding protein prevents the flow of electrons. Also time-resolved fluorescence studies on a 20-nt deoxyribooligonucleotide labeled with fluorescein (FAM) and rhodamine (TAMRA) on the opposite ends show a very fast component of lifetime (~10psec) consistent with electron transfer upon formation of M-DNA. Therefore, M-DNA behaves as a molecular wire and could be manipulated to prepare self-assembling electronic circuits.

Triplex DNA and Terminal Differentiation

Claude E. Gagna*, Hon-Reen Kuo and W. Clark Lambert
Department of Pathology and Laboratory Medicine,
University of Medicine and Dentistry of New Jersey-Medical School,
Newark, N.J. 07103
*Author to whom correspondence should be addressed: Phone: 201-224-7364; Fax: 201-568-3846

Triplex DNA has become the topic of exhaustive research, and these nucleic acids can be developed from synthetic pyrimidine-purine (pyr-pur) DNA (1). Triplexes based on T-A-T and C-G-C base triads can be formed easily and are stabilized in the presence of physiological concentration of polyamine (2). Supercoiling provides the energy needed to drive the unwinding of the DNA molecule for triplex DNA development (1). Long tracts of pyr-pur occur in eukaryotic chromosomes and could constitute 1% of the genome (3). The human genome has more than 100,000 factors of pyr-pur tracts, which are 200-300 bp long (4). This raises the possibility that triplex DNA may have a biological function (1). Triplex DNA may be involved in the in vivo regulation of gene expression and in chromosome organization (1).

Figure 1: Diagrammatic cross-section of the adult ocular lens, defining the supramolecular order: epithelium [central zone (CZ), preequatorial zone (PZ), germinative zone (GZ), meridional rows (MR), and differentiating cells (DC)], bow and equator regions, capsule, phase transition zone #1 (PTZ), PTZ #2, superficial fiber cells (SF), middle fiber cells (MF), deep fiber cells (DF), secondary anucleated fiber cells (SAF), and primary anucleated fiber cells (PAF).

Adult dog ocular lens tissue (Figure 1) processed with 10% neutral buffered formalin (NBF) allowed for either a very weak immunohistochemical staining pattern, or none at all (data not shown). However, the use of an alcohol fixative, Carnoy's fluid, allowed for a strong staining signal and easy quantification (Leica DM-RB microscope and Leica Quantimet 500 + image analysis system) of anti-triplex DNA immunohistochemistry (Figure 2 and 3). We speculate that aldehyde fixatives (10% NBF) preserve triplex DNA binding proteins, thus reducing or completely preventing anti-triplex DNA immunoreactivity. However, alcohol fixatives remove proteins and may expose triplex DNA binding sites.

Figure 2: Triplex DNA immunohistochemistry of dog lenses deep fiber cells (DF), located after the PTZ (Carnoy's). Tissues counterstained with H&E for superior visibility. Epithelium = E, Fibers = F, small arrows are pointing to immunostained reactions and large arrow indicates direction of terminal differentiation. Bar = 150 µm.

ur group has discovered the presence of a phase transition zone (PTZ) in the adult dog ocular lens (noncataractous), at a depth of about 100 (m (5) (Figure 1). Prior to the PTZ, in the superficial fibers (SF), the DNA is double-stranded (ds-), but after the PTZ the DNA becomes progressively more denatured [single-stranded (ss-) DNA] (5) (Figure 1). Prior to the PTZ, left-handed ds-Z-DNA and alpha-crystallin protein are immunohistochemically detectable, but not after the PTZ (5) (Figure 1). After the PTZ we observe a precipitous decrease in ds-B-DNA, followed by a gradual decline in ds-B-DNA (5). We believe that the ss-DNA initiates terminal differentiation by destroying intact B-DNA and by preventing Z-DNA formation, which consequently inhibits protein synthesis, such as alpha-crystallin (5). The addition of cell layers on the lens surface and the continuous displacement of older secondary fiber lamellae inward also regulates terminal differentiation (Figure 1). Eye lens nucleated secondary fiber cells (postmitotic) undergo a specialized form of cell death, terminal differentiation, in which lens fibers gradually undergo denucleation, but with no damage to the fibers. These previous results (5) support the data presented here. We show that less triplex DNA immunohistochemistry was observed prior to the PTZ (Figure 3), but after the PTZ, we detected a precipitous increase, followed by a gradual addition in triplex DNA formation, and then a sudden decline, within the lens fibers (Figure 3).

Figure 3: Image analysis data. Quantification of Z-RNA immunohistochemical staining intensities within the lenses SF, MF, differentiating fibers #1 (DF-1), and differentiating fibers #2 (DF-2) (deeper into the lens) (Carnoys). Also shown are the two zones: PTZ #1 (~100 µm) and PTZ #2 (~1301 µm). The intensity of the immunohistochemical staining products is measured in mean optical density units.

The PTZ #1 (~100 (m) we discovered in this study (Figure 3) corresponds precisely to the PTZ we uncovered in our previous work (5), using anti-ss-DNA, B-DNA and Z-DNA antibodies. However, in this study we determined an additional zone, the PTZ #2 (~1,301 (m) (Figure 3). The presence of triplex DNA extends deeper (~1398 (m) into the lenses secondary anucleated fiber cells (Figure 1). Our results suggest that in vivo triplex DNA helix formation may be used for the regulation of gene expression in the lenses SF. It may also be used during terminal differentiation to repress in vivo transcription of lens fiber cells so as to aid the normal process of fiber denucleation (Figure1).

Cataract is defined as visual impairment as a result of a disturbance of ocular lens transparency. The means to inhibit gene transcription during cataractogenesis is a very desirable outcome in rational drug design. Antigene strategies, using synthetic oligonucleotides (6) as instruments for the control of gene expression could be one possible way to inhibit cataractogenesis.

References and Footnotes

1. Sinden, R.R., in DNA Structure and Function. Academic Press, NY., pp. 217-258 (1994).
2. Hampel, K.J. et al., Biochemistry 30, 4455-4459 (1991).
3. Birnboim, H.C. et al., Eur. J. Biochem. 98, 301-307 (1979).
4. Hoffman-Liebermann, B. et al., Mol. Cell. Biol. 6, 3632-3642 (1986).
5. Gagna, C.E. et al., J. Histochem. Cytochem. 45, 1511-1521 (1997).
6. Helene, C. et al., Ann. N.Y. Acad. Sci. 660, 27-36 (1992).

Protein-Modulated DNA Electron Transfer

Scott R. Rajski,1 Sanjay Kumar,2 Richard J. Roberts2 and Jacqueline K. Barton1
1Division of Chemistry and Chemical Engineering,
California Institute of Technology,
Pasadena, CA 91125
2New England Biolabs,
Beverly, MA 01915

Oxidative DNA damage promoted from a remote site occurs as a result of electron migration through the pi-stacked array of heterocyclic DNA base pairs (1,2). Studies with intervening DNA base pair mismatches and bulges have shown that charge migration relies heavily upon the integrity of the DNA pi-stack (3-7). Gaps in the pi-stack would also be expected to inhibit the electron transfer process. The dynamic structure of DNA dictates however, that such gaps in otherwise B-form DNA are often transient and lead to other structural forms (8). The HhaI methyltransferase presents an opportunity to explore the effect of gaps within the DNA pi-stack on DNA-mediated electron transfer. Upon binding, the protein induces and stabilizes a pi-gap using a novel DNA base-flipping mechanism (9,10).

Long-range oxidative damage to DNA has been examined in assemblies containing a tethered photooxidant (shown above in blue) spatially separated from 5'-GG-3' sites (depicted above in red) as a function of M.HhaI binding to an intervening sequence (M.HhaI binding sequence shown above in yellow). We find that, in binding to its target site, M.HhaI inhibits oxidative damage to DNA past the enzyme-induced gap. In contrast, we find that a M.HhaI mutant capable of preserving the pi-stack via insertion of tryptophan into the cavity created by base-flipping, does not inhibit DNA-mediated electron transfer. Furthermore, tryptophan insertion by this mutant (bottom right panel of figure shown above) yields a charge-transfer-dependent DNA damage event within the protein-DNA binding region. Damage to the guanosine 5' of the putative tryptophan insertion site occurs and is unique to this mutant of M.HhaI. Long range DNA-mediated electron transfer can therefore be modulated both by inhibition and activation through an intervening site upon M.HhaI-DNA binding.

References and Footnotes

1. Holmlin, R. E., Dandliker, P. J. & Barton, J. K,. Angew. Chem. Int. Ed. Engl. 36, 2714-2730 (1997).
2. Hall, D. B.; Holmlin, R. E.; Barton, J. K., Nature 382, 731-735 (1996).
3. Hall, D. B. & Barton, J. K., J. Am. Chem. Soc. 119, 5045-5046 (1997).
4. Arkin, M.R., Stemp, E. D. A., Pulver, S. C. & Barton, J. K., Chem. Biol. 4, 389-400 (1997).
5. Kelley, S. O., Holmlin, R. E., Stemp, E. D. A. & Barton, J. K., J. Am. Chem. Soc. 119, 9861-9870 (1997).
6. Gasper, S. M. & Schuster, G. B., J. Am. Chem. Soc. 119, 12762-12771 (1997).
7. Okahata, Y., Kobayashi, T., Tanaka, K.& Shimomura, M., J. Am. Chem. Soc. 120, 6165-6166 (1998).
8. Withka, J. M., Wilde, J. A., Bolton, P. H., Mazumber, A. & Gerlt, J. A., Biochemistry 30, 9931-9940 (1991).
9. Cheng, X., Kumar, S., Posfai, J., Pflugrath, J. W. & Roberts, R. J., Cell 74, 299-307 (1993).
10. OíGara, M., Klimasauskas, S., Roberts, R. J. & Cheng, X., J. Mol. Biol. 261, 634-645 (1996).

Molecular Modelling of the HIV-1 Tat Protein - TAR RNA Complex and the Design of an Inhibitor of the Tat/TAR RNA Interaction

Gita Subba Rao*, Mohd. Imran Siddiqui and Sneh Arora
Department of Biophysics,
All India Institute of Medical Sciences,
New Delhi-110029, India
*Author to whom correspondence should be addressed. Phone: 91-11-659 4816; Fax: 91-11-686 2663; E-mail: gitarao@medinst.ernet.in

HIV-1 Tat protein (trans-acting transcriptional activator) regulates transcription in the virus upon binding to the transactivation response (TAR) RNA sequence, which is located in the HIV-1 long terminal repeat. Tat is a small protein of 86 residues. Mutation data, binding studies and the use of chemical probes have shown that specific binding occurs between the region centered on a U-rich bulge of TAR RNA and the arginine rich basic region (residues 49-57) of Tat. The fact that Tat protein plays a critical role in HIV-1 replication suggests that small compounds that disrupt the Tat-TAR interaction will inhibit viral replication. In order to design such a compound, the first step is to model the Tat-TAR complex.

The NMR structure of a 29-mer of TAR RNA that contains the minimal RNA region recognised by the Tat protein is known (1). However, the structure of the basic region of Tat is not well established. We have, therefore carried out a conformational analysis of the 9-residue long basic region that has the sequence RKKRRQRRR. The results indicate that this region has an extended structure. This result is consistent with CD spectroscopic data (2).

We have also carried out a conformational analysis of five synthetic 9-residue peptides whose transactivation abilities have been measured (3). A superposition of the energy minimized structures of the synthetic peptides with the 9-residue Tat peptide shows that the r.m.s. deviations of the residues located in the positions 5-8 of the highly active peptides from the corresponding region in Tat are about 0.7Å, whereas the r.m.s. deviations are higher (about 1.5Å) in the case of the peptides with low activities. Thus the conformation of the tetrapeptides is found to correlate well with activity data.The above conclusion is supported by molecular modelling of the complex of the 9-residue Tat with the averaged NMR structure of the 37-mer TAR RNA which has been done by a manual docking procedure. It is observed that the maximum number of interactions of the Tat peptide with the recognition site on TAR are with residues 5-8 (RQRR) of the Tat peptide.

Based on these observations and using the folding rules for dehydro-alanine (Delta-Ala) derived from our earlier studies, we have designed a tetrapeptide with a sequence Arg-DeltaAla-DeltaAla-Arg. Energy minimization and molecular modelling of the interaction of the designed tetrapeptide with the TAR recogniton site indicate that the tetrapeptide not only has interactions with the recognition site on TAR but also has interactions with additional bases of TAR in the stem region as compared to those with Tat. We therefore expect the designed peptide to be an effective inhibitor of the Tat/TAR RNA interaction.

References and Footnotes

1. Aboul-ela ,F., Karn,J. & Varani, G., J. Mol. Biol. 253, 313-352 (1995).
2. Loret, E.P., Georgel,P., Johnson, W.C. Jr. & Ho, P. S., Proc. Natl. Acad. Sci. USA., 89, 9734-9738 (1992).
3. Calnan, B.J., Tidor, B., Biancalana, S., Hudson,D. & Frankel, A.D., Science 252, 1167-1171 (1991).

Assembly of Topologically Linked Structures Between Duplex DNA and a Circular Oligonucleotide

Heiko Kuhn, Vadim V. Demidov and Maxim D. Frank-Kamenetskii
Center for Advanced Biotechnology,
Department of Biomedical Engineering,
Boston University,
36 Cummington St.,
Boston, MA 02215

A novel supramolecular structure was assembled by topological linkage of duplex DNA with a circular oligodeoxyribonucleotide (cODN). The topological link was thereby immovably formed at a precise position on the double helix.

To construct the assembly, dsDNA (I, in Figure 1) was locally opened using a pair of peptide nucleic acid (PNA) clamps (1). The PNA 'openers' were bound next to each other to one of the two DNA strands leaving an extended stretch of the opposite strand exposed and thus accessible for hybridization with an ODN (II, in Figure 1). An ODN was designed in a manner that its termini are complementary to the displaced DNA target and in juxtaposition upon hybridization. After hybridization, the ODN is circularized by enzymatic ligation. The result is a true topological link of the cODN with one of the two DNA strands (III, in Figure 1). Depending on the nature of the DNA targeted (linear or closed circular dsDNA), the assembled structure, resembling an earring, is either a locked pseudorotaxane or a locked catenane (Figure 2). Virtually complete formation of III using plasmid DNA target was obtained. A target containing a single mismatch between the two PNA binding sites was readily discriminated against the correct, fully complementary target.

Figure 1: Successive steps for formation on 'earring' tags on dsDNA.

The true topological link makes our assembly different from structures obtained by others. Landegren et al. circularized ODN probes on single-stranded DNA targets (2-4). In these so-called padlock constructions, the introduced cODN tags may travel considerable distances during post-assembly manipulations. A sliding clamp approach was described by Ryan and Kool, leading to a pseudorotaxane structure in which a preformed cODN is threaded on the dsDNA target (5). Obviously, only linear dsDNA fragments can be targeted by this approach. In this assemblage, like in the case of padlock structures, the cODN can also slip away from its target site.

Figure 2: Pseudorotaxane (A) or catenane (B) assembly on dsDNA.

On the contrary, in our construction the introduced cODN tag remains at a precise position as long as the participating dsDNA macromolecule retains its native duplex structure. This new structure may find applications in DNA diagnostics and in DNA nanotechnology.

References and Footnotes

1. N.O. Bukanov, V.V. Demidov, P.E. Nielsen, M.D. Frank-Kamenetskii, Proc. Natl. Acad. Sci. USA 95, 5516-5520 (1998).
2. M. Nilsson, H. Malmgren, M. Samiotaki, M. Kwiatkowski, B.P. Chowdhary, U. Landegren, Science 265, 2085-2088 (1994).
3. M. Nilsson, K. Krejci, J. Koch, M. Kwiatkowski, P. Gustavsson, U. Landegren, Nature Gen. 16, 252-254 (1997).
4. J. Banér, M. Nilsson, M. Mendel-Hartvig, U. Landegren, Nucleic Acids Res. 26, 5073-5078 (1998).
5. K. Ryan, E.T. Kool, Chem. & Biol. 5, 59-67 (1998).
6. This work was supported by the NIH.

Structure and Motion of an Intramolecular i-motif of Human Telomeric DNA

Anh Tuân Phan*, Jean-Louis Leroy and Maurice Guéron
Groupe de Biophysique de l'Ecole Polytechnique et de l'UMR C-7643 du CNRS,
91128 Palaiseau, France
*Author to whom correspondence should be addressed. Phone: 33-1-69 33 44 12; Fax: 33-1-69 33 30 04; E-mail: pat@pmc.polytechnique.fr

At slightly acidic or neutral pH, C-rich deoxy-oligonucleotides form four-stranded structures built by head-to-tail intercalation of two parallel-stranded duplexes with hemi-protonated C.C+ pairs. This so-called i-motif may be formed in a tetramer of an oligonucleotide containing a stretch of adjacent cytidines, or in a dimer of a oligonucleotide containing two stretches of cytidines, or by intramolecular folding of an oligonucleotide containing four stretches of cytidines.

Such oligonucleotides are found notably in telomeres, the ends of eukaryotic chromosomes, which are essential for stable chromosome maintenance. Telomeric DNA almost invariably consists of one strand of short cytosine-rich repeats and of the complementary G-rich strand. In humans the telomeric repeat is (5'-CCCTAA):(5'-TTAGGG).

We present the structure of a human telomeric sequence d[CCCTAACCCTAACCCTAACCCT] in solution. This structure is stable up to pH 6.5 at 0 °C. It includes an i-motif core of six C.C+ pairs, similar to that observed previously in i-motif tetramers. At one end (the "top"), the central TAA linker loops across one of the narrow grooves, and the core is extended by stacking of the T10.T22 pair and of A11. At the bottom of the core, the two other TAA linkers loop across the wide grooves. The two intercalated duplexes and the two linkers form a pseudo-symmetric structure with a vertical two-fold rotation axis in the center of the i-motif stack. Whereas most of the structure is quite rigid, the NMR spectra reveal motions in the two bottom loops on the microseconds to milliseconds scale, precluding detailed determination of their structures. The spectra remain broad even at low temperatures (-10 °C) and 800 MHz.

In order to determine the structures of the bottom loops, we synthetized sequences with substitutions within the loops (e. g. T16 -> U) or close to them (e. g. C7 -> 5mC), which destroy the pseudo-symmetry between the two loops. With the two substitutions indicated above, the NMR spectrum corresponds to two forms of the telomeric strand in slow exchange, with a switching time of 3 ms at 0°C. In each form, a high-definition structure of the loops is obtained. The loops are well stacked. The two forms differ principally in one loop, by the orientations of A18, which are respectively syn and anti. The chemical shifts evaluated for the two orientations of A18 agree with those observed in the two components of the NMR spectrum, providing an independent confirmation for the proposed structures.

Parallel-Stranded DNA with Mixed AT/GC Content: Role of Trans G·C Pairs in Specific Sequence Dependence of the Double Helix Stability

Anna K. Shchyolkina1, Olga F. Borisova1, Michael A. Livshits1, Galina E. Pozmogova2, Reinhardt Klement3 and Thomas M. Jovin3
1Engelhardt Institute of Molecular Biology,
Russian Academy of Sciences,
117984 Moscow, Russia
2 Centre "Bioengineering",
Russian Academy of Sciences,
Prospekt 60-let Oktyabrya, 7/1,
117334 Moscow, Russia
3|Department of Molecular Biology,
Max Planck Institute for Biophysical Chemistry,
D-37018 Göttingen, Germany

Parallel-stranded (ps) DNAs with mixed AT/GC content comprising GáC pairs in a various context have been investigated at neutral pH in the presence of NaCl and LiCl. Oligonucleotides having two 10-nt complementary strands in parallel orientation bound with triethyleneglycol linkers have been synthesized:

ps-N1: 3'-d(CTATAGGGAT)-L-d(ATCCCTATAG)-3'
ps-N2: -d(CTGAGTAGAT)-L-d(ATCTACTCAG)-3'
ps-N4:                3'-d(CTGATAGGAT)-L-d(ATCCTATCAG)-3'
ps-AT: 3'-d(AT)5-L1-(AT)5-3'
ps-N6:                    3'-d(TTATAGGGAT)-L-d(ATCCCTATAA)-3'
ps-N7:                    3'-d(TTGAGTAGAT)-L-d(ATCTACTCAA)-3'

Predominant formation of the intramolecular hairpins was facilitated by the chosen experimental conditions and controlled by measuring rotational relaxation times of the molecules and percentage of unpaired bases. Specific CD spectral signature of the trans GáC pair was obtained. A multistate mode of the ps hairpins melting was visualized with temperature gradient gel electrophoresis (TGGE). Stability of the ps hairpins with mixed AT/GC content was shown to depend strongly on the nucleotide sequence, presenting a peculiar feature of a parallel-stranded double helix. The entire set of melting data unambiguously and model-independently arrays the six investigated ps hairpins by their stability: ps-N7<ps-N2<ps-AT<ps-N4<ps-N6£ps-N1. This reveals the context-dependent stabilization and destabilization effects of incorporation of trans G·C pairs into a row of AT-pairs: a separate GáC pair destabilizes the duplex while any block of consecutive GáC pairs exerts a stabilization effect.

A multi-state heterogeneous zipper model for melting the hairpins was derived and thermodynamic parameters of the ps hairpins formation were determined from experimental thermal denaturation curves for six sequences by global minimization procedure. Formation of a trans GáC pair in homo GG/CC context was estimated to be for 3 kJ (in 0.1 M LiCl, at 3¡C) more energetically advantageous than formation of a trans AáT pair in heteroAT/TA context. Substantially, an unfavorable additional free energy of a GC/AT contact was found to be about 2 kJ. Consequently, the base pairs sequence, specifically quantities of GáC pairs and AT/GC contacts, determines an overall stability of ps-DNA with a mixed AT/GC sequence. Thus, stability of ps-DNA comprising successive three or two GáC pairs is even higher than that of ps-DNA with alternating AT sequence, whereas increasing number of AT/GC contacts in the nucleotide sequence with isolated GáC pairs facilitates destabilization of ps-DNA.

The study was partially supported by the Russian Foundation for Basic Research, Russian Foundation "State support of the Russian leading scientific schools", and the Max Planck Society.

Go to next page

Back to Table of Contents