© 1990 Oxford University Press
Nucleic Acids Research, Vol 18, No. 6
1559
DNA with adenine tracts contains poly(dA).poly(dT)
conformational features in solution
Sabine Brahms and J.Georges Brahms
Institut Jacques Monod, CNRS, Universite Paris VII Tour 43, 2 Place Jussieu, 75251 Paris Cedex 05,
France
Received November 2, 1989, Revised and Accepted January 4, 1990
ABSTRACT
The conformation of DNA's with adenine-thymine tracts
exhibiting retardation in electrophoretic migration and
considered as curved were investigated in solution by
CD and RAMAN spectroscopy. The following curved
multimers with adenine tracts but of different flanking
sequences d(CA5TGCC)n, d(TCTCTA 6 TATATA 5 ) n ,
d(GA4T4C)n yield CD spectroscopic features indicating
a non-B structure of the dA.dT tract with similarities
to polyd(A).polyd(T). We suggest that adenine-thymine
bases in these multimers contain some of the
distinctive conformational features of polyd(A).polyd(T)
probably with large propeller twist found by NMR
(Behling and Kearns, 1987) and by X-ray diffraction on
oligonucleotides containing a tract of adenines (Nelson
et al. 1987, Coll et al; 1987; DiGabrlele ef al. 1989).
Some elements of distinctive CD features of the
contiguous adenines run are also observed in the
straight multi-9-mer d(CA5GCC)n which lacks in-phase
relation to the helical repeat. Despite the presence of
the TpA step in the straight multimer d(GT4A4)n, the
altered dA.dT conformation is not completely
destroyed. Interruption of adenine tract by a guanine
in d(CAAGAATGCC)n leads to a B-like conformation
and to a normal electrophoretic mobility. The Raman
spectra reveal a rearrangement of the sugar-phosphate
backbone of dA.dT tract in the multimer d(CA5TGCC)n
with respect to that of polydA.polydT. This is reflected
in the presence of an unique Raman band associated
to C2'-endo sugar with a predominant contribution of
C1'-exo puckering which is exhibited by the multimer
whereas two distinct Raman bands characterize
poly(dA).poly(dT) backbone conformation.
INTRODUCTION
DNA curving has been observed experimentally in a variety of
DNA containing homopolymeric (dA).(dT) tracts essentially by
anomalously slow electrophoretic migration on polyacrylamide
gels (1-5). In addition sequence directed curving has been
observed by electron microscopy (6, 7) and by demonstration
of higher circularization probabilities of ohgomers (8). Up to now
curved DNA structures were found composed essentially of short
runs of oligo(dA) ohgo(dT) spaced in phase with helical repeat.
Previously a weak periodicity of some dinucleotides was detected
in eukaryotic DNA which suggested that periodic positioning of
small bends could lead to unidirectional curvature (9). There is
a general agreement that the curvature is coming from the
differences in the roll angles of a sequence e.g. of An tract and
the roll angles in remaining DNA at a period equal to helical
repeat (14b). Different authors provide various models to
interpret the curvature (10, 11, 3, 12-14).
Poly(dA).poly(dT) has distinct structural and functional
properties; it adopts a modified B* form (15) and cannot undergo
a B - A transition upon lowering of relative humidity, as shown
by infrared dichroism and X-ray diffraction on DNA fibers (15,
16). Poly (dA). poly (dT) cannot be reassociated into nucleosomes
(17, 18). Results of X-ray diffraction on single crystal of
ohgomers with adenine tracts indicate that run of (dA).(dT) forms
a high propeller twist at each base pair facilitating the formation
of an additional bifurcated hydrogen bond (19-21). A large
propeller twist was also observed in poly(dA).poly(dT) in solution
(34). The results of crystallographic analysis of Nelson et al. (19)
and DiGabnelle (21) indicate that the A-tract exhibits very small
(close to zero) roll angles and the A-tract is straight. According
to the crystallographic results curving must occur outside of the
A tract (19).
It appears that more detailed studies of curved DNA structure
in solution are necessary to understand the structure of curved
sequences. We report here the investigation of the conformation
of curved repetitive DNA's with A-tracts in solution by
spectroscopic methods (far UV circular dichroism (CD) and
Raman scattering measurements). It is shown that the considered
curved DNA's d(CA5TGCC)n, d(TCTCTA6TATATA5)n and
d(GA4T4C)n display in solution some of the CD spectral
characteristics of poly(dA).poly(dT) homopolynucleotide
reflecting conformational similarity of the oligo(dA).oligo(dT)
runs of the different curved multimers. We have also investigated
A tract containing multimers which do not exhibit retardation
in gel electrophoresis either because of incorrect phase relation
or because of the presence of T4pA sequence or because of
interrupted A tract. Raman spectra show the rearrangement of
the sugar puckering of dA.dT tract to one main C2'-endo
conformation with predominance of Cl '-exo conformation in the
multimer.
1560 Nucleic Acids Research
MATERIALS AND METHODS
Preparation of synthetic oligomers and of corresponding
muitimers
Different oligomers and their complementary strands were
prepared on Applied Biosystems and Cruachem synthesizers.
Each oligonucleotide was synthesized at least twice using two
different solid state synthesizers. The correctness of the sequence
of one of the synthesized oligonucleotide CA5TGCC was
checked by direct sequencing. These oligomers were purified by
gel electrophoresis and by FPLC on a Mono Q anion exchange
chromatography (Pharmacia) following the instructions of the
manufacturer. The majority of oligomers were first submitted
to reverse phase C 18 HPLC chromatography, prior to
detntylation, with the generous help of Dr Ronald Frank (G.B.F.
Braunschweig).
Each of the oligomers was phosphorylated using T4
polynucleotide kinase during 1 hour at 37°C. The oligomers were
hybridized with their complementary strands heated to 60°C and
slowly cooled to room temperature to form double stranded self
ligated muitimers with sticky ends and left for 2 - 4 hours at 4°C.
The enzymatic ligation reaction was carried out with T4 ligase
(Pharmacia) at 12°C during 16 hours. It appeared of importance
in our approach to achieve a high degree of ligation of different
oligomers. The concentration of oligonucleotides in the ligation
reaction was at least 2 mg/ml in the presence of 10% of ethylene
glycol. After ligation the samples were deproteinized by phenol
followed by twofold chloroform-isoamylalcohol extraction.
Finally the muitimers were precipitated by ethanol. Under these
conditions the average size of ligated oligomers was estimated
-6-
1 6 0 M O
2 2 0 2 4 0
260
250
300
Figure 1. CD spectra of d(A 5 TGCQ n (dotted line) which exhibits retardation
in electrophoretic mobility and of d(CAAGAATGCC), (full line) of normal
mobility. The CD spectrum of B-form pBR322 (dashed line) is shown for
comparison. The spectra were measured at 20°C For preparation of muitimers
and other details see materials and methods
at about 500 to 700bp from gel electrophoresis The products
of ligation were run on 6% polyacrylamide gels in 75 mM Trisborate 1.2 mM EDTA pH 8 at room temperature. A center of
a band was excised from the gel and an additional electrophoresis
was performed in denaturating medium (formamide). The
mobilities of these muitimers were compared with fragments of
pBR322 digested separately with HaelTJ, Bgll and TaqI used as
size markers and with those of ligated products of 10 bp BamHI
linker.
Raman scattering spectra
Raman spectra were measured with a computer controlled Raman
spectrophotometer assembled from a Jarrel-Ash double
monochromator with an argon ion laser at 514 5 nm as a source
(22). The Raman spectrum represents an average over 40 scans
at a dwell time of 1 sec cm" 1 and is represented after a simple
base-line adjustment procedure. The multimer samples consist
of 100 ng of DNA in 5 ^1 of 40 mM KF pH 7 in either H2O
or 2H2O. The sample was held in a quartz capillary held in a
thermostated brass block at about 5°-10°C.
Circular dichroism spectra
Circular dichroism spectra were measured in the spectral range
from 170nm to 320 nm using the vacuum UV dichrograph
constructed in the laboratory (23). The sample absorbances were
about 0 25 to 0.50 at 260nm in 40 mM KF, 5 mM potassium
phosphate buffer pH 7. A CD spectrum of each mutimer
represents an average over three different DNA samples
measured at 20°C.
RESULTS AND DISCUSSION
Conformation of A.T bases of curved and straight muitimers
with adenine tracts
We have investigated the conformation of synthetic DNA's with
repetitive adenine-thymine sequences which may or may not
exhibit retardation of electrophoretic mobility. In figure 1 we
compare the CD spectra of d(CA5TGCC)n known to produce
retardation of mobility and considered as curved (2) and of
d(CAAGAATGCC)n in which bending is eliminated by
interruption of the adenine tract by a single guanine residue and
exhibits normal mobility in electrophoresis (3). Figure 1 shows
that the curved multimer d(CA5TGCC)n display in near UV a
particular CD spectrum with four characteristic bands at about
245nm, 258nm, 272nm and at 282nm (dotted line) which shape
is distinct from a normal B-conformation CD spectrum (dashed
line). The bands at 258nm at 272nm in the curved muitimers
d(CA5TGCC)n are distinct and particularly weak compared to
the CD of an average DNA like pBR322 of B-form (fig. 1). The
bands at 258nm and 272nm were assigned to adenine and thyrrune
respectively (24). hi contrast the spectrum of the straight multimer
d(CAAGAATGCC)n (fig. 1, full line) is of entirely different
shape and indicates a conformation approaching that of a normal
type B-DNA when compared to the CD spectrum of pBR322
of 'conservative' character composed of two bands at 275nm and
245nm of positive and negative sign and of nearly equal rotational
strength (fig. 1 dashed line). We may consider that the CD
spectrum of the muitimers is composed quite closely of two
contributions i.e. of dA.dT base pairs of the adenine tract and
of the base pairs of flanking sequences. The presence of the
295nm band of negative sign is related to the GC flanking
sequences outside of the dA tract and appears similar in the two
muitimers. The difference in the two muitimers CD spectra on
Nucleic Acids Research 1561
(TCTCTA TATATA )
9M
O
220
2<0
260
w«v«l*noth
300
< n
Figure 2. Comparison of CD spectra of poly(dA) poly(dT) (dashed line) with
that of two curved multimers (full line) (A) dfTCTCTAjTATATAj),, and (B)
d(CA 5 TGCQ n both exhibiting retardation of electrophorctK mobility. The
conditions are as in figure 1.
figure 1 mainly results from the presence of a guanine which
interrupts the A-tract and should be the main cause of the collapse
of the fine structure and of a significant modification in CD shape.
The intense far U V bands below 200 nm are indicative of good
base pairing and stacking in a right handed helix in all these
multimers.
The CD shape of the curved multimer suggests that the adenine
tract may produce a distinguishable fine structure in the CD of
curved multimers. To obtain more information on conformational
features of curved DNA two multimers of different sequences
dCTCTCTAfiTATATAj),, (ref. 8) and d(CA5TGCC)n were
compared (Fig. 2). Figure 2 shows that the CD shape of these
two multimers d(CA5TGCC)n and d(TCTCTA<;TATATA5)n
resemble that of poly(dA).poly(dT) but their CD fine structure
is attenuated as compared to the homopolymer partly as a result
of superposition of the two CD contributions: of adenine tract
and of flanking base pairs. Nevertheless the CD spectra indicate
that some of the essential features like the particularities of basebase arrangement of poly(dA).poly(dT) already exist to some
extent in adenine tracts in both curved DNA's
dCTCTCTA^TATATA,),, (Fig. 2A) and d(CA5TGCC)n (Fig.
2B). This presumably gives rise to a CD shape (dotted line) with
splitted 285nm-245nm region into four positive and negative
bands different from the B-type spectra. Haran and Crothers (31)
have shown a cooperativity in building of an A-tract structure
which is in agreement with our results. In the straight
ohgonucleotide d(CAAGAATGCC)n the required minimum of
three A:T base pairs is no longer present to satisfy the
cooperativity of the adenine tract which may explain the observed
difference in conformation (Fig. 1). In the straight multimer the
interruption of the adenine tract by a single guanine residue causes
a transition to a B-like conformation.
According to their respective CD (Fig. 2), the tracts of six
and five adenines in both multimers considered as curved have
conformational features different from a B-DNA. We suggest
that the A-tract of the multimers shown in Figure 2 adopts a
structure approaching that of poly (dA). poly (dT). Our results are
in agreement with recently reported data on CD of poly
d(An).poly dCTJ composed of ohgodA.oligodT of various
lengths (35). The authors observed a rapid change and increase
of the 258nm band intensity with an increase of the number of
contiguous adenine residues starting from the tnmer.
It was of interest to compare the conformation of multimers
in which oligomeric components were propagated in phase with
the helical repeat to that propagated out of phase. Stable axial
curvature is expected to have inherent phase relation to the helical
repeat. In contrast the 9-mer multimer d(CA5GCC)n in which
such phasing is absent migrates with normal mobility. Figure
3 shows the CD spectrum of the 9-mer multimer which to some
extent resembles that of the decamer d(GA5TGCC)n displayed
on figure 2, both exhibiting a non-B-type structure of the A-tract.
Despite the existence of certain conformational features of the
adenine tract susceptible to generate a local alteration of the helix
direction, the 9-mer multimer is not producing a macroscopic
curvature as a consequence of the absence of phase coherence.
An example of sequence dependent differences concerns two
multimers, d(GA4T4C)n showing retardation in electrophoretic
migration and d(GT4A4C)n which have normal mobility (5b).
Differences between df^GA^C),, and d(GT4A4Qn were also
observed in cutting reactivity by the hydroxyl radical and were
interpreted on the basis of differences in the narrow minor groove
width of the former and of normal width of the latter as a
consequence of TpA step (25).
The comparison of the CD spectra of the curved multimer
d(GA4T4C)n with the straight multimer d ^ T ^ Q n shows that
both multimers exhibit a conformation different from B-form
DNA (Fig. 3). The two spectra show some differences, but one
can observe in both CD the presence of the specific conformation
of the adeninerthymine tract reflected in the maxima at about
258nm, 272nm and 282nm. In both multimers the dA.dT tract
seems to be in a non-B conformation as shown by the presence
of the 258nm band which is much strongly attenuated in
d(GT4A4C)n. The contribution of G-C sequences are reflected
by the presence of the negative band at about 295nm.
1562 Nucleic Acids Research
1000
800
>, 600
'm
c
4)
£ 400
200
65O
7OO
75O
BOO
S5O
cm-1
180
200
220
240
250
280
300
Figure 3. Circular dichroism spectra of d(GA4T4C)n (dotted line) exhibiting
anomalously slow electrophoretic migration and of dfGT^C),, (dashed line)
with normal migration Trie nanomer multimer d(CAjGCC)n (full line) is lacking
macroscopic curvature and does not show retardation in electrophoretic mobility
The experimental conditions are as in figure 1
Calorimetric measurements show that the 5'TpA step has lower
energy of stacking interaction (AH= 6.0 kcal/mol) than ApT
(AH= 8.6 kcal/mol) (27). This difference in stability is most
probably a consequence of a smaller stacking overlap for TpA
than for ApT (28, 29,30a). Possibly the TpA step destrabilizes
also the spine of hydration of poly(dA).poly(dT) structure (30b).
We found (fig. 3) that the TpA step does not destroy completely
the adenine-thymine tract structure in d(GT4A4C)n which
maintains some conformational features of a modified B-DNA
form. Our results are also in good agreement with NMR data
of base pair lifetime determination of adenine lmino-protons of
d(A4T4) and d(T4A4) oligomers (26). The exchange time of
adenine iminoprotons of d(A4T4) was found to be long whereas
in dC^A,,) even if appeared much shorter was found also
anomalous. The interpretation of Gu6ron et al. is very similar
to ours i.e. in 'dOVA*) also the non-B structure is formed either
incompletely or only part of the time' (26). Measurements of
anomalous migration of multimers containing TnAn tracts and
A,,Tn tracts reported by Haran and Crothers (31) also lead to an
existence of a structure different from B within the components
of TnAn blocks.
The specific CD band shape in the 250nm-280nm range
suggests that some of the particular features of adenine-thymine
base-pairing of a modified B-DNA form found in
poly(dA).poly(dT) tend to be adopted by the adenine-thymine
tract in all investigated curved DNA.
We suggest that oligodA.oligodT tract in curved multimers
adopt in solution a non-B conformation approaching
polyd(A).polyd(T) type. The CD spectrum of poly(dA).poly(dT)
(Fig. 2) of unusual shape with sphtted three bands in the region
255nm-3OOnm is not encountered in any other DNA. We may
Figure 4 Raman scattering spectra of curved multimer (CA5TGCC)n (dashed
line) and of poly(dA) poly(dT) (full line) in the frequency range
500cm~' — 900cm
of base and sugar-phosphate backbone vibrations The
spectra were normalized with respect to the 1094cm ~' phosphate vibrations band
(data not shown)
assume that this particular shape of the homopolydeoxynucleotide
results from stacking of bases with relatively large propeller twist.
A large propeller twist of adenine thymine tracts (18° -24°) was
found by X-ray diffraction analysis of single oligomers crystals
(19-21). The presence of a large propeller twist is also confirmed
by energy calculation from fiber diffraction and NMR data in
solution (32, 33, 34). Interruption of the adenines tract e g. by
a guanine residue in a straight d(CAAGAATGCC)n restores a
B type conformation.
Our interpretation is also in agreement with the recent results
of CD spectra investigation of a series of oligomers
ohgod(A)n.oligod(T)n (35). The authors observed an abrupt
change between ohgod(A)2.oligod(T)2 and higher oligomers
which subsequently generate continuously changing spectra for
higher oligomers. This was interpreted as reflecting appearance
and change of propeller twist at AT base pairs with increasing
length of contiguous adenosines (35) starting from the trimer.
The presence to some degree of a type of conformation
poly(dA) poly(dT)-like in the adenine tract of the 9-mer multimer
d(CA5TGC)n supports the idea that the structural elements of the
distortion of the helical axis do exist but sine they are not
propagated in phase with the helical repeat the macroscopic
curvature can not occur.
Our data clearly show that despite the low conformational
stability at the TpA step probably resulting from its lower energy
of stacking interaction (27-30) in d(GT4A4C) the non-B
structure of the adenine tract is not destroyed but less complete.
This interpretation of our results on d(T4A4)n is in good
agreement with NMR data of base pair lifetime determination
(26). To some extent this specific structure is retained in the
adenine run of the 9-mer d(CA5GCC)n and in d(GT4A4C)n both
with normal electrophoretic mobility. The spectra of straight
d(CAAGAATGCC)n have lost the spectral features of
polyd(A).polyd(T). In d(CAAGAATGCC)n the presence of a
Nucleic Acids Research 1563
guanine may interrupt the A-tract structure presumably the
network of propeller twisted adenines (19-21).
Sugar-phosphate backbone and bases conformation — Raman
scattering investigation
Raman scattering spectrum of d-(CA 5 TGCC) n in the
conformationally sensitive region of base and of phosphate-sugar
backbone vibrations 620cm" 1 -900cm" 1 is shown in Fig. 4
(dotted line) where it is compared to that of the homopolymer
polyd(A) polyd(T) (Fig. 4 full line).
In the frequency range characteristic of sugar-phosphate
backbone vibrations the multimer yields one main broad band
at 828cm"1. This band approximately corresponds to the
frequency of vibrations of sugar in C2'-endo furanose ring pucker
family (36, 37). The displacement of this band from 835cm~'
typical of C2'-endo furanose pucker to a band centered at
828cm "' suggests a distribution of sugar comformations around
C2'-endo family presumably including a main contribution of
Cl'-exo deoxyribose pucker. These results in solution are in
agreement with crystallographic analysis data on a similar
sequence containing A-tract d(CGCA6GCG) by Nelson et al
(19).
In this frequency range polyd(A).polyd(T) exhibits clearly two
distinct bands contribution at 816 cm" 1 and at 841cm"1 (Fig.
4; ref. 38). This may suggest that in polyd(A).polyd(T) two
populations of deoxynbose conformations contribute i.e. one at
816"' closely related to 04'-endo and the second at 841 cm" 1
of the C2'-endo family. The band at 816cm"1 is tentatively
assigned to an intermediate conformation presumably to O4'-endo
deoxyribose pucker since it occur between 809cm" 1
characteristic of C3'-endo ring pucker and 835cm"1 assigned to
C2'-endo (36, 37). The O4'-endo pucker is intermediate between
C2'-endo and C3'-endo puckering in terms of pseudorotation
angle (39). This comparison suggests the occurrence of a
rearrangement of the backbone conformation of adenine tract
from two distinct furanose pucker populations in
poly(dA).poly(dT) homopolymer to one main population of the
C2'-endo family with Cl'-exo deoxyribose conformer
contribution in the curved multimer of d(CA5TGCC)n.
Further indications of the sugar-phosphate backbone
rearrangement of the A-T tract of the multimer are also suggested
by the changes of the Raman band at 750cm"1 of thymine
which is sensitive to sugar C2'-endo/T conformation (40). The
relative intensity of the 750cm"1 band sharply decreases in the
spectrum of the multimer d(CA5TGCG)n with respect to that of
poly(dA).poly(dT) (Fig. 4).
Other differences concern the band at 728cm"1 of adenine in
plane vibrations of the multimer d(CA5TGCC)n which exhibits
Raman hyperchromism (Fig. 4). This suggest that short tracts
of five adenines in the multimer d(CA5TGCC)n are slightly less
stacked with respect to the continuous dA.dT pairs in the
homopolymer. The Raman band at 670cm"1 assigned to
thymine is well resolved and of intermediate intensity between
poly(dA) poly(dT) and DNA, as one could expect from the
dependence of intensity on A-T base content (40). The
contribution of G-C bases is reflected by the presence of a
682cm"1 band assigned to guanine and at 782cm~' assigned to
cytosine.
Raman scattering is particularly sensitive for detection of short
life states of the order of 10"12sec. The contributions of
841cm"1 and of 816cm"1 bands in polyd(A).polyd(T) may
represent two families of deoxynbose conformations and/or may
reflect rapid interconversion of furanose puckering which may
not be observed by other methods e.g. NMR, particularly nuclear
Overhauser effect which measures over 10~3sec. As noted
previously (38) it is not surprising why NMR studies on
polyd(A).polyd(T) report only one C2'-endo sugar conformation
(34) whereas Raman scattering may detect contribution of two
populations of furanose puckering.
Our results about the backbone flexibility of DNA containing
adenine tracts are in general agreement with the data of X-ray
diffraction analysis of Dickerson et al. (41, 42) and their recent
refinements by Westhof (43). The variability of furanose pucker
in DNA was also observed in solution of supercoiled and relaxed
pBR322 (44) and of DNA's (40) from the analysis of band shape.
It is to be noticed that our present method is direct since DNA's
with adenine tracts yielded Raman bands with well defined
maxima (Fig. 4).
The Raman spectra reveal rearrangement of sugar-phosphate
backbone of the dA.dT tract of the multimer d(CA5TGCG)n
from two Raman bands in poly(dA).poly(dT) to one band at
828cm"1 assigned to C2'-endo family with predominant
Cl'-exo sugar pucker. This is also manifested by the changes
of the Raman band at 750cm"1 of C2'-endo/thymine of this
multimer. In the multimer the constraint introduced by the
presence of mixed sequences of guanine and cytosine exerted
a pressure on adenine-thymine tract which was absorbed by the
relatively flexible backbone, leading to a modification of furanose
ring pucker; this is shown by one broad Raman band at
828cm~' reflecting the predominant Cl'-exo sugar pucker of
C2'-endo family conformation.
The present results on the DNA backbone conformations with
oligo(dA).oligo(dT) tract in solution are in good agreement with
the data of crystallographic analysis of Nelson et al. (19) who
found on a similar sequence d(CGCA6GCG) that sugar puckers
around Cl '-exo and C2'-endo for adenine and thymine strands.
REFERENCES
1 Manni J C , Levene.S D , Crothers.D M and Englund.P T (1986) Proc
Nail Acad Sa. USA 79, 7664-7668
2 Wu,H M and Crothers.D M (1984) Nature 308, 509-513
3 Koo.H S , Wu,H N and Crothers.D M (1986) Ataure 320, 501-506
4a Diekmann.S and Wang,J C (1985)7 Mol Biol 186, 1-11
4b Diekmann.S (1987) Nucl. Acids Res 15, 247-265
5a Hagerman.P J (1984) Proc Natl. Acad Sa USA 81, 4632-4636
5b Hagerman.P J. (1986) Nature 321, 449-450.
6 GnffithJ , Bleyman.M , Rauch.C A Kjtchen.P A and Enghind.P T (1986)
Cell 46, 717-724
7 Tteveny.B , Coulaud.D , Le Bret.M and Revet.B (1988) in Structure and
Expression 3, 1 —16, ed Olson, Sarma, Sarma and Sundaralingham Adenine
Press
8 Ulanovsky.L , Bodner,M , Tnfonov.E N and Choder.M (1986) Proc Natl
Acad Sa USA 83, 862-866
9 Tnfonov.E N. and SussmanJ L (1980) Proc. Nail. Acad. Sa USA 77,
3816-3820
10 Prunell.A , Goulet.I , Jacob.Y and GoutorbeJ (1984) Eur J Biochem.
183, 253-257
11 Ulanovsky.L and Tnfonov.E N (1987) Nature 326, 720-722
12aKoo,HS and Crothers.D.M (1988) Proc Natl Acad. Sa USA 85,
1763-1767
12b Koo.H S and Crothers.D M. (1987) Biochemistry 26, 3745-3749
13. Kietzing,E von and Diekmann.S (1987) Eur Biophys 7 15, 13-26.
14a Calladine.C R (1982)7 Mol Bio! 161,343-352
14b.Calladine,C.R and Drew.H R (1986) 7. Mol Biol 192,907-918
14c.Calladine,C.R., Drew.H.R. and McCall.M.J (1988) 7 Mol. Biol 201,
127-137
15. PileU BlicharskiJ and BrahmsJ.G (1975) Biochemistry 14, 1869-1875
16 Arnott.S. and Selsing.E. (1974)7 Mol Biol. 88, 509-521.
17 Kunkel.G R and Martmson.H G. (1981) Nud Acids Res 9,6869-6888
1564 Nucleic Acids Research
18 PruneU.A. (1982) EMBO J 1, 173-179
19 Nelson.H C M , FinchJ T , Luisi.B F. and Klug.A (1987) Nature 330,
221-226
20 CoU,M., Fredenck.C A., Wang.A H J and R>ch,A. (1987) Proc NatL Acad
Sa. USA 84, 8385-8389
21 DiGabnellc.A D , Sanderson.M R. and Stertz.T A (1989) Proc Nail Acad.
Sa USA 86, 1816-1820
22 Brahms.S , Vergne,J , Brahms.J G., Di Capua.E. and Koller.T (1982) J
Mol Biol. 162, 473-493.
23 Brahms.S and Brahms J G (1980)/ Mol Biol. 137, 149-178
24 Edmondson.S.P and Johnson,W.C (1985) Biopolymers 24, 825-841.
25 Burkhoff.A M and Tulhus.T D (1988) Nature 331, 455-457
26 Leroy J L , Charretter.E , Kochoyan.M and Gueron.M (1988) Biochemistry
27, 8894-8898
27 Brcsslauer.K J , Frank.R., Blocker.H and Marky.L A (1986) Proc Noll.
Acad Sci. USA 83, 3746-3750.
28 Klug.A , Jaclc,A , Viswamitra.M A , Kennard.O , Shakked.Z and
Steitz.T A (1979) J Mol. Biol 131, 669-680
29 Viswamitra.M.A , Kennard.O , Jones.P G , Shcldnck.G.M , Salisbury.S ,
Falvello.L and Shakked.Z (1978) Nature 273, 687-688
30a Yoon.C , Pnvif.G G , GoodseU.D S. and Dickerson.R E (1988) Proc NatL
Acad Sa USA 85, 6332-6336
30b Chupnna.V P (1987) Nud Acids Res. 15,293-311
31 Haran.T E. and Crochers.D M. (1989) Biochemistry 28, 2763-2767
32 Alexeev.D.G , Lipanov.A A and SkuratovsIu.I Ya (1987) Nature 325,
821-823
33 Lipanov.A.A and Chupnna.V.P (1987) Nud. Acids Res 15,5833-5844
34. Behling.R and Keams.D R (1986) Biochemistry 25, 3335-3345
35. Gudibande.S R , Jayasena.S D. and Behe.M J (1988) Biopolymers 27,
1905-1915
36. Goodwin.D C and BrahmsJ G (1978) Nucl Acids Res 5, 835-850
37. Erfiirth.SC, Bond.P J and PeUcolas.W L (1975) Biopolymers 14,
1245-1254
38 Wartell.R M and HarreU T. (1986) Biochemistry 26, 2664-2671
39. Saenger.W. (1984) in Principles of Nucleic Acid Structure, Springer Verlag,
New York.
40 Benevides.J M , Wang.A H J , van der Marel.J H van Boom, and
Thomas.G J (1988) Biochemistry 27, 931-938.
41 Dickerson.R E and Drew.H R (1981) / . Mol Biol 149, 761-768.
42. Fratini.A V , Kopka.M L , Drew,H R and Dickerson.R.E (1985)7. Biol
Qiem. 257, 14686-14707
43 Westhof.E (1987)7. Biomol Str and Dynam. 5,581-600
44 Brahms.S., Nakasu.S., Kikuchi.A and Brahms J G (1989) Ear J. Biochan.
184, 297-303