Chem. Res. Toxicol. 1993,6,519-523
519
The trans-syn-IThymine Dimer Bends DNA by =22" and Unwinds DNA by 4 5 " Cheng-I Wang and John-Stephen Taylor* Department of Chemistry, Washington University, St. Louis, Missouri 63130 Received April 9,1993 Irradiation of DNA with ultraviolet light leads to the formation of two classes of cyclobutane dimers a t adjacent thymines sites, of which the cis-syn is the major class and the trans-syn is the minor class. While the structure and properties of DNA containing cis-syn thymine dimers have been extensively studied, virtually nothing is known about DNA containing trans-syn thymine dimers. To investigate the bending and unwinding of DNA induced by the trans-syn-I thymine dimer, the electrophoretic properties of oligomers of trans-syn-I dimer-containing DNA duplexes were studied. Oligonucleotides 10, 11,and 12 bp in length containing a centrally located trans-syn-I thymine dimer were synthesized, polymerized,and analyzed by polyacrylamide gel electrophoresis. In contrast to the small bending angle ( ~ 7induced ~ ) by the cis-syn thymine dimer, we found that trans-syn-I thymine dimer bends DNA significantly (=22O). Both dimers, however, are found to unwind DNA by the same amount (=15O). On the basis of previous NMR studies, it appears that the bend of the trans-syn-I dimer is localized a t the Y-side of the dimer. Gel electrophoretic analysis of multimers of two 11-mers containing a cis-syn thymine dimer a t the 5'-end and a t the center of a d T ~ d A 6tract confirmed our previous estimates for the bending angles of thymine dimer-containing T6 tracts. The substrates reported may be useful in determining how general repair enzymes recognize DNA damage.
Introduction The predisposition to skin cancer of victims of the inherited disorder xeroderma pigmentosum has been attributed to defects in the biological systems required to repair the DNA damage caused by sunlight (for a general review of DNA damage and disease, see ref 1). The principal DNA photoproductsinduced by the UV radiation in sunlight are cyclobutane dimers and (6-4) products (2, 3), and a number of enzymes that recognize and repair DNA containing these photoproducts have been isolated and been the subject of intensive study (for a recent review see ref 4). Escherichia coli DNA photolyase is a flavoprotein that is capable of directly repairing cis-syn pyrimidine dimers via a photoinitiated electron-transfer pathway. T4 denV endonuclease V and Micrococcus luteus pyrimidinedimer glycosylases initiate repair of cissyn dimers by incising the DNA at the 5'-side of the dimer in a process involving initial hydrolysis of the glycosidic linkage of the 5'-pyrimidine of the dimer. The (A)BC excinuclease of E. coli is a general repair enzyme that recognizes bulky adducts and initiates repair of cis-syn and (6-4) products by cleaving phosphodiester bonds to the 5'- and 3'-side of the damage site. It has been proposed that repair enzymes locate the damage site by recognizing the bending or unwinding of the DNA caused by the damaging agent (5-8). What, in fact, is being recognized by these enzymes still remains unknown, however, due to a general lack of knowledge about the structure and properties of photodamaged DNA. The tram-syn thymine dimer is a minor photoproduct of DNA, being produced at about 12% the frequency of the cis-syn thymine dimer (Figure 1, structure 2a) in denatured DNA and less than 2% in native DNA (2). Model building suggests that it has the same stereochem-
* To whom correspondence should be addressed. 1
Kao, Nadji, and Taylor, Chem. Res. Toxicol., in press.
2 cis-syn
3 trans-syn-l a Rl-DNA, Rp(-), R p D N A
4 trans-syn-ll b Rl-H, Rp(-), R3-H c Rl-DMT, R24CH3, R~-P(OCH~)(N(I-P~)Z)
Figure 1. The cis-syn (2a)and trans-syn-I (3a)photoproducts of a TpT site (la)and the correspondingbuilding blocks (2c and 3c)used to incorporate them into oligonucleotides by solid-phase synthesis. The trans-syn-I1 dimer (4b) has recently been prepared from a minor product formed during the synthesis of the cis-syn and trans-syn-I building blocks.'
istry as the major trans-syn isomer produced in the photolysis of TpT, which has been determined by NMR spectroscopy studies (&II) to have the tram-syn-I structure (Figure 1, structure 3b). Unlike DNA containing cis-syn thymine dimers, for which the structure, properties, and repair have been extensively studied, little is known about DNA containing trans-syn dimers. For example, though it has been reported that tram-syn thymine dimers are not repaired by DNA photolyase (121, it is currently unknown whether or not they can be repaired by the (A)BC excinuclease. One general problem with studying minor DNA damage products is the difficulty in preparing 0 1993 American Chemical Society
520 Chem. Res. Toxicol., Vol. 6, No. 4, 1993
DNA containing the product at a specific site in the purity and quantity needed for precise structure-activity experiments. As part of an overall program aimed at unraveling the structure-activity relationships of DNA photoproducts, we developed a building block for the sitespecific introduction of trans-syn-I thymine dimers into oligonucleotidesby automated solid-phase DNA synthesis (Figure 1, structure 3c) (13). By utilizing this building block, a duplex decamer fragment containing a site-specific trans-syn-I thymine dimer was prepared in sufficient quantity to be studied by 2D NMR2 spectroscopy (14). The NMR experimenb were interpreted to suggest that the trans-syn-I dimer is more distorting than the cis-syn dimer and introduces a kink or dislocation into B-form DNA. To more precisely assess the amount and type of distortion induced by the trans-syn-I dimer, we decided to utilize recently developed gel electrophoretic methods for determiningthe amount of bending (15)and unwinding (6,16) present in a DNA fragment. Herein, we report the synthesis and electrophoretic properties of multimers of duplexes containing site-specific trans-syn-I thymine dimers. By analyzing the electrophoretic mobilities of these multimers, we were able to determine that the transsyn-I dimer unwinds DNA by =15O, about the same amount as that previouslydetermined for the cis-syn dimer (17). In contrast to the small amount of bending caused by the cis-syn dimer ( ~ 7 " (18), ) we determined that the trans-syn-I dimer bends DNA by ~ 2 2 We ~ . also report the synthesis and electrophoretic properties of multimers of two 11-mer8containing a cis-syn thymine dimer at the center and at the 5'-end of a dTVd& tract. The study of these oligonucleotides was undertaken to confirm our previous estimates for the effect of cis-syn thymine dimer formation on A-tract bending that were based on the analysis of the electrophoretic mobility of multimers of 10-mers(18). Furthermore, the results of the latter study allowed us to confirm the previous estimate of =14O for the unwinding angle of the cis-syn thymine dimer (17).
Experimental Procedures Syntheses. cis-syn and trans-syn-I thymine dimer building blocks (Figure 1,structures 2c and 3c) were synthesized using modifications (19) of the original method (20). All oligonucleotides (Figure 2) were synthesized by standard automated solidphase DNA synthesis and were purified by reverse-phase HPLC. The oligonucleotides were then 5'-phosphorylated, annealed, polymerized, and electrophoresed in the same way as previously reported (18) with the exception that ligations of trans-syn-I dimer-containing oligonucleotides were conducted at 4 OC. To obtain a sufficient yield of multimers of TS-10, ligations were conducted at 0.5 mM ATP in the presence of 8% PEG 8000 [average M W 8000 poly(ethy1ene glycol), Sigma, St. Louis, Mol or 1 mM Co(NHs)&& (G. Frederick Smith Chemical Co., Columbus, OH) according to standard procedures for blunt end ligation (21). Calculations. Bending angles for a given DNA sequence were determined by first calculating the curvature of three in-phase multimers in the 120-145 bp range relative to that of T6 tract multimers from their RLdata (15). Second,the calculated relative curvatures were corrected by multiplying by the factor required to normalize the calculated relative curvature of the T6 m d t h e r s used as an internal electrophoresis standard to 1, and then by
* Abbreviations: 2D NMR, two-dimensionalnuclear magnetic resonance; HPLC,high-pressure liquid chromatography; ATP,adenwine triphosphate;PEG,poly(ethy1eneglycol);NOE, nuclear Overhausereffect.
Wang and Taylor
Tb
CCGTTTTTTG CAAAAAACGG
M4
T_G C C G T-T T T T GGCAAAAAAC
Mdll
G G C C G T-T T T T T GGCAAAAAACC
T2M2
C C G T T T-T T T G CAAAAAACGG
T2DT2-11
C C G T T TIT T T G G CAAAAAACCGG
TSlO
CCAAGTSGGA TTCAACCTGG
Tsll
CCAAGEGGAC TTCAACCTGGG
TS12
CCAAGTZGGACG TTCAACCTGCGG
L
CGGGATCCCG GCCCTAGGGC T-T
cba~mymlnedmer
T"T
trantisyn thym~ne*er
Figure 2. The oligonucleotides used in this study. Data for DT, and TsDTz reported in Figure 4 were taken from our previous study (18). the factor required to normalize the number of monomer units/ bp to be the same as that of the To multimers (a factor of 1.1 when using multimers of 11-mers). The relative curvatures were then converted to absolute curvatures by multiplying by 1 9 O for the bend of aT6 tract (22) and averaged. Unwinding angles were calculated as described in the text. The bending and unwinding angles calculated in this manner from two independent seta of electrophoresis data were found not to differ by more than 1.5%.
Results and Discussion The oligonucleotide sequences in this study are shown in Figure 2. These oligonucleotides were purified, phosphorylated, annealed, and then polymerized with T4 DNA ligase and ATP. The apparent lengths of the resulting multimers on a nondenaturing electrophoresis gel (Figure 3) were determined by comparison to size markers (L), and then converted to RLvalues (the ratio of the apparent length to the actual length). The bending and unwinding angles were determined by analyzing plots of RL versus actual length (Figure 4) and plots of RL versus monomer length (Figure 5). Tramesyn-IDimer. The unwinding angle of the transsyn-I thymine dimer was determined by comparing the gel electrophoretic mobilities of multimers of TS-10, TS11,and TS-12 (Figure 3). The multimers with the highest electrophoretic anomalies (highest RLvalues) correspond to those in which the dimers are closest to being in phase with the helical turn, thereby allowing one to estimate the unwinding angle. The plots of RLversus actual length for the multimers of the oligonucleotides listed in Figure 2 are shown in Figure 4. To our surprise, the RL values of (TS-ll), are even greater than those of the well-characterized (T& bending sequence, indicating that the transsyn-I dimer bends DNA to a greater extent than a T6 tract. TS-10 did not polymerize under our standard experimental conditions, indicating that the trans-syn-I dimer induces a significant amount of structural distortion,
Chem. Res. ToxicoZ., VoZ. 6, No. 4, 1993 521
Thymine Dimer Bending and Unwinding Angles 2.4
c w
2.2
1
140 bp
\
I t 0.8
4
9.5
1
f
Figure 3. Autoradiograms of the acrylamide electrophoresis gels of the multimers of the oligonucleotideslisted in Figure 2. The multimers shown in the gel to the left were prepared by standard method, while the multimers of TS-10 and TS-12 shown to the right were prepared by the modified method that utilized PEG 8000. 2.8~ 2.61
/f
2.4+
20
40
60
80
100
120
140
11
11.5
12
4 12.5
Figure 5. Fits of RL versus monomer length data for 120,130, and 140 bp multimers of TS-10, TS-11, and TS-12 to quadratic equations.
-150
-50
10.5
Monomer Length
I:rqf -100
10
160
Ts-l'
180
for (TS-ll), is almost parallel to that of (T&, both of which do not plateau within the observation range, allow us to conclude that the length of TS-11 monomer is approximately equal to that of one helical turn and that TS-10 and TS-12 are out of phase. Consistent with this conclusion is the presence of a number of high molecular weight products in the polymerization of TS-11 (Figure 3) that do not fit the RLversus actual length curve (Figure 4). These are likely to be circularized multimers which arise because of the almost perfectly in-phase relationship between the bends in TS-11 (22,24). By plotting the RL value for each of the three oligonucleotides for a given multimer length, the length in bp corresponding to an exact repeat (the pitch, P) can be calculated by fitting to a quadratic equation and solving for the monomer length in bp corresponding to the maximum of the curve (6)(Figure 5). From the calculated value of 10.94 f 0.02 bp/turn3 for the pitch of a duplex containing a trans-syn-I dimer and assuming a value of 34.3'/bp for the average base pair twist (t) of the corresponding undamaged sequence [derived from a pitch of 10.5 for heterogeneous B-form DNA (25,26)], we can calculate the unwinding angle (At) of 15.2 f 0.7' for the trans-syn-I thymine dimer according to eq 1. The un-
Pt - At = 360°/turn
200
220
ACTUAL LENGTH
Figure 4. Plots of RL (apparent length/actual length) versus actual length in base pairs for the multimer bands shown in Figure 3. The curves for DT4 and TzDTz multimers are included for comparison and are from our previously reported study (18).
in accord with a previous NMR and melting temperature study (14). Polymerization of TS-10 could, however, be effectedby conductingthe ligation in the presence of either poly(ethy1ene glycol) (PEG) or cobalt(II1) hexamine, conditions that have been effectively used for blunt end polymerizations (21). Both the curves for (TS-lo), and (TS-12), appear to plateau when the multimer length is longer than ~ 1 4 bp 0 (12-14 repeats), an indication that the dimers are not perfectly in phase (23). These observations,together with the observationthat the curve
(1)
winding angle calculated in this manner is rather sensitive to value of the pitch, such that a 1%error in the pitch leads to a 20% error in the unwinding angle. In support of the calculated unwinding angle of 15', however, one need only note that the RLvs L curve of TS-11multimers is very similar to that of cis-syn dimer-containing multimers of T2DT2-11 and that the cis-syn dimer has been determined by a different method to unwind DNA by 14' (17). On the basis of the RL values for multimers of TS-11 and the value of 19O for an &-tract bend (22),we calculate (15) that the trans-syn-I thymine dimer bends DNA by 21.6 f 1.0'. The structural basis for this bend is unknown at the moment, but certain suggestions as to its nature come from an NMR study that we had conducted on d(CGTAT[t,s]TATGC).d( GCATAATACG) (14). In that study we had discovered an interruption in the NOE sequential assignment pathway on both strands at the 5'-side of the trans-syn-I dimer. On the basis of the lH NMR chemical shift and NOE data, we had proposed that 3 The standard deviationgiven is that for the maxima of the quadratic curves used to fit the data of Figure 5 and does not take into account any other sources of error, which would make the true error higher.
Wang and Taylor
522 Chem. Res. Toxicol., Vol. 6, No. 4, 1993
this interruption might be due to either a kink or a dislocation in the DNA duplex at this position induced by the methyl group of the 5'-thymine unit of the dimer. This thymine unit is constrained to adopt a syn glycosyl orientation relative to the normal anti conformation found in B DNA, which causes the hydrophobic methyl group of this thymine to be thrust into what is normally the base pairing region of a B DNA duplex. The finding that the trans-syn-I dimer induces a sizable bend suggests that the trans-syn-I dimer induces a kink rather than a dislocation in the DNA at the 5'-end of the dimer. The direction and precise location of this bend remain to be determined, however. cis-syn Dimer-Containing T6 Tracts. Recently, we reported the bending angles of dimer-containing T6 tracts based on the gel mobility data of multimers of decamers containing cis-syn thymine dimers at all five possible sites within a dAs.dT6 tract (18).Since the helical repeat of the T6 multimers is estimated to be 10.3 bp (27)and the cissyn thymine dimer has been determined to unwind DNA the phase angle between the bends of all five by 14' of the dimer-containing decamers is -25'. As a result (23),the curves in the plots of RL (apparent length/actual length) versus actual length for the multimers of the decamers containing cis-syn thymine dimers a t different sites with the T6 tract appear to plateau when the lengths of the multimers are longer than ~ 1 4 bp 0 (14 repeats). Consequently, the bending angles caused by the dimercontaining T6 tracts could not be directly determined by comparing their multimer RL values to those of a more nearly in-phase T6 tract. We did observe, however, that the curve of DT4 multimers in the plot of RL versus actual length was very similar to that of multimers of &Ns (28), in which the phase angle between the TstraCts is calculated to be ==+25', equal in magnitude, but opposite in sign, to that between DT4 tracts. Because of this similarity, we concluded that a DT4 and an & tract bent DNA to approximately the same extent. The curvature of each dimer-containing T6 tract was then calculated by assuming that RL - 1 is directly proportional to the square of the relative curvature (15) and that an A6 tract bends DNA by =19' (22). In order to test if a DT4 tract really bends DNA to the same extent as an &tract, 11-mers corresponding to two of the five previously studied cis-syn dimer-containing T6 tracts were synthesized so as to more nearly put the dimercontaining T6 tracts into phase (Figure 2). Multimers of DT4-11were calculated to have a phase angle of =+loo between DT4 tracts, based on an unwinding angle of 14' for the cis-syn dimer ( I 7). This phase angle is identical in magnitude, but opposite in sign, to that calculated for the phase angle betweenT6 tracts in multimers of T6 (-10'). Instead of showing a plateau as does (DT4)n,the curve for (DT4-ll), is virtually superimposable with that of (T6)n (Figure 4), indicating that the RL values and, hence, the curvatures of the multimers of these two sequences are the same. Because 10 repeats of DT4-ll has the same curvature as that of 11 repeats of T6, an individual DT4 tract must have a bend 11/10 as large as that of a T6 tract, or 20.6 f l.Oo, assuming that a T6 tract bends DNA by 19'. This new value can be compared to our originally reported value of 19'. The fact that the curves for (T6)n and (DT4-llInare superimposable is consistent with the
(In,
4 Svoboda, Smith,
Taylor, and Sancar, J. Biol. Chem. (in press).
previously reported unwinding angle of 14' for the cissyn thymine dimers (I7). The phase angle between TzDT2 tracts in (T2DT2-11)nis likewiseclose to ' 0 as the RLvalues do not show a plateau in the range examined. The bending angle of T2DT2-11 calculated from the RL values of its multimers is 8.7 f 0.8', identical to our previous estimate of 9' derived from analysis from the RL data of T2DT2. This result confirmed our original conclusion that the bending of the TzDT2 sequence is due almost entirely to the cis-syn thymine dimer, which has a bending angle of 7' ( I @ , and that all intrinsic bending of the & tract is abolished by dimer formation at the center of the tract.
Conclusion Gel electrophoretic analysis of in phase duplex DNA multimers is a very useful method to quantify DNA curvature. Since the relative mobilities of the multimers is very sensitive to the phasing of the bends, the unwinding angle can also be readily estimated by varying the length of the monomers. We found that trans-syn-I thymine dimer bends DNA by =22', possibly in the form of a kink at the 5'-end of the dimer, in contrast to =7' for the cissyn dimer. The directionality of the bends remains to be determined and is currently under study. Both dimers, however, were found to unwind DNA by the same amount (=15'). By using 11-mers as monomers, we were able to refine our previous estimates of the bending angles for thymine dimer containing T6 tracts and conclude that a T6 tract with a cis-syn dimer at the 5'-end bends DNA (=21°) by almost the same amount as does a trans-syn-I dimer. With these three substrates and a previously reported substrate containing a site-specific cis-syn dimer in a nonbending sequence (29) it might now be possible to determine whether or not general repair enzymes such as the (A)BC excinuclease locate damaged DNA by (1) local bending (the trans-syn-I dimer is repaired the fastest), (2) net bending (the trans-syn-I dimer and the cis-syn dimer in DT4 are repaired the fastest), (3) local unwinding (the trans-syn-I dimer and the cis-syn dimer in any sequence are repaired equallyfast), (4) net unwinding (the trans-syn dimer and the cis-syn dimer in the heterogeneous sequence are repaired the fastest), or (5) some other mechanism. In this regard, we have recently found that the trans-syn-I dimer is repaired by the uvr(A)BC enzyme at a 6-fold faster initial rate than the cis-syn dimer in the nonbending sequence context d(CGAAT=TAAGC)? This indicates that duplex unwinding is not the major determinant in the recognition of damaged DNA by the (A)BC excinuclease as has been proposed (6). What is the major determinant remains to be determined.
Acknowledgment. This investigation was supported by PHS Grant R01-CA40463, awarded by the National Cancer Institute, DHHS. References (1) Friedberg, E. (1985) DNA Repair, W. H. Freeman and Co., New
York.
(2) Patrick, M. H., and Fiahn, R. 0. (1976)Photochemistry of DNA and
Polynucleotides: Photoproducts. In Photochemistry and Photobiology of Nucleic Acids (Wang, s. Y., Ed.) Vol. 11, pp 35-95, Academic Press, New York. (3) Cadet, J., and Vigny, P. (1990)The photochemistry of nucleic acids. InBioorganic Photochemistry (Morrison, H., Ed.) Vol. 1,pp 1-272, John Wiley & Sons,New York. (4) Sancar, A., and Sancar, G. B. (1988) DNA repair enzymes. Annu. Rev. Biochem. 57,2947.
Thjvmine Dimer Bending and Unwinding Angles (5) Sancar, A., and Rupp, W. D. (1983) A novel repair enzyme: uvrABC
excision nuclease of Escherichia coli cuts a DNA strand on both sides of the damaged region. Cell 33, 249-260. Bellon, S. F., Coleman, J. H., and Lippard, 5. J. (1991) DNA unwinding produced by sibspecific intrastrand cross-links of the antitumor drug cis-diamminedichloroplatinum(I1). Biochemistry 30,8026-8035.
Van Houten, B., Gamper, H., Sancar, A., and Hearst, J. E. (1987) DNase I footprint of ABC excinuclease. J.Biol. Chem. 262,1318013187.
Hueain, I., Sancar, G. B., Holbrook, 5.R., and Sancar, A. (1987) Mechanism of damage recognition by Escherichia coli DNA photolyase. J. Biol. Chem. 262, 131M-13197. Liu, F.-T., and Yang, N. C. (1978) Photochemistry of cytosine derivatives. 1. Photochemistry of thymidylyl-(3’+5’)-deoxycytidine. Biochemistry 17,4865-4876. Kemmink,J., Boelens, R., and Kaptein, R. (1987) Two-dimensional 1H NMR study of two cyclobutanetype photodimersof thymidylyl(3’-5’)-thymidine. Eur. Biophys. J. 14, 293-299. Kan, L., Voituriez, L., and Cadet, J. (1988) Nuclear magnetic resonance studies of cis-syn, trans-syn, and 6-4 photodimers of thymidylyl(3’-S’)thymidine monophosphateand cis-syn photodiers of thymidylyl(3’-5’)thymidine cyanoethyl phosphotriester. Biochemistry 27,5796-5803. Ben-Hur, E., and Ben-Ishai, R. (1968) tram-syn-Thymine dimers in ultraviolet-irradiated denatured DNA identification and photoreactivability. Biochim. Biophys. Acta 166,945. Taylor, J.-S., and Brockie, I. R. (1988) Synthesis of a tram-syn thymine dimer building block. Solid phase synthesis of CGTAT[t,slTATGC. Nucleic Acids Res. 16, 5123-5136. Taylor,J.-S.,Garrett,D. S.,Brockie, I. R., Svoboda,D. L., andTelser, J. (1990) ‘H NMR assignment and melting temperature study of cis-syn and trans-syn thymine dimer containing duplexes of d(CGTATTATGC).d(GCATAATACG).Biochemistry 29, 88588866.
Koo, HA., and Crothers,D. M. (1988) CalibrationofDNAcurvature and a unified descriptionof sequence-directedbending. hoc. Natl. Acad. Sci. U.S.A. 85,1763-1767. Lee, C.-S., Sun, D., Kizu, R., and Hurley,L. H. (1991) Determination that are responsible for of the structural features of (+)-CC-1065
Chem. Res. Toxicol., Vol. 6, No. 4, 1993 623 bending and winding of DNA. Chem. Res. Toxicol. 4,203-213. Ciarrocchi,G., andPedrini,A. M. (1982) Determinationofpyrimidine unwinding angle by measurementof DNA electrophoreticmobility. J. Mol. Biol. 155, 177-183. Wang, C.4, and Taylor, J.4. (1991) Sibspecific effect of thymine dimer formation on A-tract bending and its biological implications. Roc. Natl. Acad. Sci. U.S.A. 88,9072-9076. Taylor, J.-S.,and Nadji, S. (1991) Unravelingthe origin of the major mutation inducedby ultraviolet light, the C-T transition at dTpdC sites. A DNA synthesis building block for the cis-syn cyclobutane dimer of dTpdU. Tetrahedron 47, 2579-2590. Taylor, J.-S.,Brockie, I. R., and ODay, C. L. (1987) A building block for the sequence-specific introduction of cis-syn thymine dimers into oligonucleotides. Solid-phaseSynthesis of TpT[c,slpTpT. J. Am. Chem. Soc. 109,6735-6742. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecular cloning: A laboratorymanual,ColdSpringHarborLaboratoryPreea, Cold Spring Harbor, NY. Koo, H.-S., Drak, J., Rice, J. A., and Crothers, D. M. (1990) Determinationof the extent of DNA bendingby an adenine-thymine tract. Biochemistry 29,4227-4234. Calladine,C. R., Drew, H. R., and McCall, M. J. (1988) The intrinsic curvature of DNA in solution. J. Mol. Biol. 201, 127-137. Taylor,W. H., and Hagerman, P. J. (1990) Application ofthe method of phage T4 DNA ligase-catalyzed ring-closure to the etudy of DNA structure. 11. NaC1-dependence of DNA flexibility and helical repeat. J. Mol. Biol. 212, 363-376. Wang, J. C. (1979) Helical repeat of DNA in solution. hoc. Natl. Acad. Sci. U.S.A. 76, 200-203. Rhodes, D., and Klug, A. (1980) Helical periodicity of DNA determined by enzyme digestion. Nature 286,573-578. Drak, J., and Crothers, D. M. (1991) Helical reDeat and chirality effects on DNA gel electrophoreticmobility. hdc. Natl. Acad. Sci. U.S.A. 88, 3074-3078. Koo, HA., Wu, H.-M., and Crothers, D. M. (1986) DNA bending at adenine-thymine tracts. Nature 320, 501-506. Taylor, J.-S.,and ODay, C. L. (1990) Cis-syn thymine dimers are not absolute blocks to replication by DNA polymerase I of Escherichia coli in vitro. Biochemistry 29,1624-1632.
