Conformational Studies by1H NMR of the HIV Enhancer: the Transcription Factors NF-κB and Sp1 Binding Domains

MAGNETIC RESONANCE IN CHEMISTRY, VOL. 34, S67-S80 (1996)

Conformational Studies by 'H NMR of the HIV Enhancer : the Transcription Factors NF-KB and Spl Binding Domains

Muriel Delepierre* and Patrick Sodanot Laboratoire de Resonance Magnetique Nucleaire, UA 1129 CNRS, Institut Pasteur, 28 rue du Dr Row, 75724 Paris Cedex IS, France

Catherine Gouyette, Abdelkader Namane and Jean Igolen Unit6 de Chimie Organique, Institut Pasteur, 28 rue du Dr Roux, 75724 Paris Cedex 15, France

Jean-Louis Virelizier Unite d'Immunologie Virale, Institut Pasteur, 28 rue du Dr Roux, 75724 Paris Cedex 15, France

The solution structures of two DNA non-palindromic duplexes of 24 base pairs and corresponding to the 3' NF-KB- binding site and the adjacent 5' Spl-binding site of the HIV-1 long terminal repeat (SLTR) were analyzed by proton two-dimensional nuclear magnetic resonance spectroscopy. The first sequence d(GGGGACrrrCCAGGGAGGCGTGGC):d(GCCACGCCTCCCTGGAAAGTCCCC) corresponds to the wild-type LTR duplex (24mer-N), and the second sequence d(GCTCACTTTCCAGGGAGGCGTGGC): d(GCCACGCCTCCCTGGAAAGTGAGC) (24mer-M) contains a specific mutation (lightface letters) abolishing NF-KB binding. Assignment of protons essential for structure determination were obtained and structural features of both duplexes were analyzed. The overall structure of the duplex is of the B form but several significant local structural deviations were found. First, from analysis of NOE cross-peak intensities between adenine H2 base protons and sugar H1' protons, it was found that the CTC mutation results in widening of the minor groove with presumably narrowing of the major groove, thus impairing the binding of NF-KB to its responsive element in the HIV LTR. Second from chemical shift analysis of H1' sugar protons, some unusual structural features were found at the junction between the homopurine:homopyrimidine stretches, that is, at the junction of the NF-KB and Spl sites, consistent with the known synergism between NF-KB and Spl functions in the HIV LTR. The scopes and limitations of DNA fragment studies of such a sue are discussed.

KEY WORDS NMR; 'H NMR; homopyrimidine; homopurine DNA; local conformation

INTRODUCTION

The process of transcriptional activation in eukaryotes by site-specific DNA-binding proteins is a key step in gene regulation.'-6 Many eukaryotic genes which are regulated in an inducible or constitutive fashion have been analyzed. The major outcome is that eukaryotic transcription factors contain at least two distinct essential regions: a DNA-binding domain7 and a transcriptional activation r e g i ~ n . ~ Although a DNA-binding domain is necessary for transcriptional activation, it is not sufficient, and an important factor may be the possible interaction with other proteins that bind at adjacent sites. Thus, the formation and dissociation of protein-protein complexes are an integral part of the regulation of many cellular processes. The transcription factor NF-KB is such an example as it is regulated by highly specific association with the inhibitor IKBa, a

* Author to whom correspondence should be addressed. Present address: Centre de Biophysique Moleculaire, rue Charles

Sadron, 45071 Orleans Cedex 02, France

protein located in the cytoplasm.' Upon cellular activation the NF-KB : I K B ~ complex is disrupted allowing transport of the DNA-binding protein to the nucleus and subsequent transcriptional activation (8).

The transcription factor NF-KB controls, through specific interactions with appropriate DNA sequences, the expression of many cellular genes and viral genomes.' Its essential role in initiation transcription of the HIV genome is particularly well d o c ~ m e n t e d . ~ ~ ' ~ Transcription of the HIV genome entirely depends on the activation of its 5' regulatory region, the long terminal repeat (LTR)."*" The HIV LTR contains numerous sequences able to bind transcription factorsi3 (Fig. 1). Deletion or mutation of either NF-KB or Spl sequences abolishes LTR induction upon cell stimu- lation by phorbol esters or tumor necrosis factor (TNF).14 Moreover, mutations of both the 3' NF- KB - binding site and the adjacent Spl-binding box deeply suppress the replication competence of an HIV pro~irus . '~ We recently showed that a specific mutation of the HIV enhancer, abolishing NF-KB binding,16 blocks HIV provirus transcription in normal CD4 T lymphocyte^.'^ However, the activity of the HIV LTR promoter domain is inducible on its own, and we have

CCC 0749-1581/96/SIO067-14 0 1996 by John Wiley & Sons, Ltd.

Received 15 April 1996 Accepted (revised) 27 August 1996

S68 M. DELEPIERRE ET AL.

TAFs l B P

.1

Figure 1. Schematic representation of the HIV long terminal repeat involved in viral gene expression. Indicated are the trans- acting responsive element (TAR), the TATA box for the TATA box binding protein (TBP) and their associated factors (TAFs), the three consensus binding boxes for transcription factor Spl , the enhancer region containing the two directly repeated rcB sequences and the negative regulatory element (NRE).

NF-KB site and adjacent 5' Spl site, in comparison with that of the fragment in which mutations (GGG + CTC) were introduced in the 5'-terminal end, since this point mutation abolishes the NF-KB binding to DNA.,' Moreover, the presence of homopurine and homopyrimidine stretches is of interest for sequence-specific regulation of gene expression via triple helix forma- t i ~ n . ~ " ~ '

EXPERIMENTAL

shown that this inducibility is dependent on the Spl boxes placed upstream the TATA box.'8 Thus, HIV LTR activity is synergistically induced by signal targeting transcription factors on both the NF-KB- and Spl- binding elements. This prompted us to analyze the solution structure of a sequence containing, in their natural context of the HIV LTR, both NF-KB- and Spl- binding elements, rather than each separately.

The simplest DNA-binding form of NF-KB is com- posed of p50 homodimer. However, the functional DNA-binding form of NF-KB in nuclei of activated cells is the heterodimer p50-~65.~ The p65 subunit contains the IKB-binding domain and also a DNA-binding domain. In the NF-KB heterodimers, p65 modulates the DNA-binding specificity of p50. While, in the absence of p65, the p50 homodimer can bind with high affinity to completely palindromic sequences, the heterodimer binds with a much lower affinity to these sites but shows an increased affinity for the KB motif "GGGACTTTCC3'.'' Both p50 and p65 proteins possess a DNA-binding domain but the DNA binding activity of NF-KB is associated with the N-terminal region of p50. This region has no homology to other human DNA-binding motifs such as leucine-zipper, zinc-finger and helix-turn-helix, but contains a different type of conserved DNA-binding Only p50 binds with a very high affinity to the 5' consensus site (GGG) whereas p65 binds with a much weaker affinity to the 3' site.24 As opposed to what is usually observed, the DNA sequence recognized by these homo- or heterodimeric NF-KB proteins is not palindromic in the HIV LTR.25

It has been shown by circular permutation mobility shift assaysz6 that DNA is strongly bent upon binding of NF-KB, whereas in the absence of protein it does not seem to be bent. However, the crystal structures of NF-KB p50 homodimer bound to either a palindromic decameric KB DNA site 5'd(GGGAATTCCC)3' '' or to an undecamer DNA sequence corresponding to the MHC H2TF1 KB motif that forms a duplex with a central A :A mistmatch," has been solved. In the first structure, a slight bend and a slight unwinding of the DNA are proposed" whereas in the second structure a bend of 15" is proposed at the junction between p50 and p65 binding sites20*21 that is, between the central AAT sequence and the GGGG flanking half sites. Recent circular dichroism spectroscopic studies have shown that the H2TF1 KB motif adopts an inherently unusual conformation whereas the DNA structure of the palindromic decamer is that of a B-DNA.28

The purpose of this work was to study by NMR the conformation in solution of the synthetic DNA sequence corresponding to the HIV LTR wild-type 3'

Sample preparation

The 24-bp, HIV LTR wild-type DNA duplex 5'd(GGGGACTTTCCAGG@4GGCGX33c)3' : 5'd(G- CCACGCCTCCCTGGAAAGTCCCC)3', where the NF-KB- and Spl-binding sites are underlined with a solid line and a dashed line respectively, and the same sequence, except for a GGG -, CTC mutation at the 5' end (bold letters) 5'd(GCTCACTTTCCAGGGAGGC- GTGGC)3' : "d(GCCACGCCTCCCTGGAAAGTGA GC)3' were synthesized 5'-DMTr-on with a 1 mmol column on an Applied Biosystems 380B system with phosphoramidites (average coupling yield 99.5% by trityl titration at 498 nm). Deprotection was carried out using concentrated ammonia at 55 "C overnight. After evaporation, the crude product was desalted on a Sep- hadex G-10 column eluted with triethylammonium bicarbonate (TEAB) 0.05 M. The eluted compound was lyoplilized and then purified using preparative HPLC. The purified 5'-DMTr-on oligonucleotide was detrity- lated using 80% acetic acid for 20 min at room temperature, co-evaporated with toluene, washed with diethyl ether and lyophilized. After a second purification using preparative HPLC, the oligonucleotide counter ions were exchanged to the NH, form using a Dowex 50W-X8 NH, column and lyophilized. Purity was checked by analytical HPLC and polyacrylamide gel electrophoresis of a 32P 5' end-labeled sample. During the purification steps the only difference between analytical and preparative runs was the amount of material injected. In both cases, the eluents and the chromatographic conditions were identical. The buffer used was 10 mM triethylammonium acetate (TEAA) (pH 7). The chromatographic conditions were a linear gradient of acetonitrile from 5% to 25% over 20 min. Purification was achieved on HPLC system (Perkin-Elmer) consisting of a Series 410 pump, an LC290 detector and a Model 7125 loop injection valve (Rheodyne). Detection was performed at 254 nm. HPLC analysis were carried out on a Hewlett-Packard 1090M system equipped with an auto sampler, an auto injector and a diode-array detector. For analytical runs a Nucleosil 5 mm CI8 column (150 x 4.6 mm i.d.) (Macherey-Nagel) was used with a flow rate of 1 ml min-'. For the preparative scale, a Nucleosil 5 mm CIS column (250 x 10 mm i.d.) with a flow rate of 5.5 ml min-' was used.

For NMR experiments, each separate strand was chromatographed over Chelex 100 resin equilibrated with sodium ions to remove dications and then dis- solved in 0.35 ml of aqueous solution containing 16 mM NaCl, 15 mM MgCl,, 100 mM KC1 and 0.125 mM EDTA

‘H NMR CONFORMATIONAL STUDIES OF HIV ENHANCER S69

at pH 7 to final concentrations of 0.25 and 0.8 mM for the wild-type (24mer-N) and the mutated (24mer-M) duplexes, respectively. As the DNA fragments to be studied were not self-complementary, the two strands were mixed in the proper ratio following the concentration by measuring the absorbance at the appropriate wavelength (260 nm). The complementary strands were annealed by incubation in a water-bath at 80°C for 10 min and then allowed to cool slowly. The NMR samples were prepared by lyophilizing the resultant solution and redissolving them in 750 pl of D 2 0 or 90% H20-10% D20. The final sample concentrations were 0.25 mM and 0.8 mM for the native and the mutated sequences, respectively.

NMR spectroscopy

NMR experiments were carried out on a Varian Unity 500 spectrometer and processed on a Sun workstation using VNMR software (Varian). All spectra were acquired in the phase-sensitive mode with the hyper- complex ~cherne.’~ The transmitter was placed on the water resonance. The sample temperature was adjusted to 25, 30 or 50°C in all experiments. Two TOCSY experiments using MLEV1733*34 to produce isotropic mixing were carried out with mixing times of 50 and 60 ms with 64 scans per t , increment.

For all NOESY experiments 512 complex pairs were collected in the t , dimension and 2048 data points in the t , dimension with 64 scans for each t , value. Several mixing times were used, 30, 50, 150, 300 and 500 ms at 30°C and 30, 60, 90 and 500 ms at 50°C, but the data obtained with 300 and 500 ms were chosen for assignment purposes at both temperatures. These spectra show a greater number of peaks due to spin diffusion, helping in the assignment procedure. A particularity of DNA is that spin-diffusion effects are always in the direction of the assignment pa th~ay . ’~ ’ ’~ NOESY spectra in H,O were recorded at 25°C using a com- bination of short-hard and spin-lock pulses in a sequence giving a binomial-like frequency response’ ’ with 3008 points in t , and 600 in C, for a spectral width of 10000 Hz. For each t , value 96 scans were collected with a relaxation delay of 1.3 s between transients. The mixing times were 120 and 300 ms. The z delay was 165 ms and a 2 ms spin-lock pulse was applied. The z delay value was chosen to have a minima at the water frequency and the first null at 10.8 ppm, i.e. between the amino and the imino proton region, while maxima are at 13.83 ppm in the imino proton region and at 7.8 ppm in the amino and adenine H2 proton region. Both dimensions were processed with a sine-squared 60” phase-shifted function. Finally, the shifted time-domain convolution procedure” was applied to the spectra with M = 21.

RESULTS AND DISCUSSION

NMR studies of DNA fragments of such a size are as difficult as, if not more than the structure determination of a protein of 140 amino acid residues. The duplexes having a molecular weight of 16000, we faced all prob-

lems inherent to the NMR of large molecules. First, it was important to establish the conformation type by mean of ”P NMR. The ’lP spectra run at two temperatures for the native fragment 24mer-N (Fig. 2) showed that all resonances were clustered in a very narrow bandwith, less than 1 ppm, indicating that the duplex contains a very large proportion of right-handed double-stranded helix. Similar results were obtained for the mutated sequence (data not shown). Then, optimum conditions for NMR experiments were obtained from the study of chemical shift variations with temperature. With the exception of the terminal residue protons, the chemical shift of non-exchangeable protons did not vary significantly between 25 and 70”C, indicative of a higher thermal stability of the molecule. Indeed, theo- retical cal~ulations’~ at the concentration of the NMR experiments showed that the mid-point of denatur- ation is well above 70°C for the two 24-mers and that at 50°C only a small fraction, less than 1%, of the nucleotide is single stranded (F. Schaeffer, personal communication).

The sequence-specific assignment of ‘H resonances of oligonucleotide proceeds in two stages. The first is identification via through-bond interactions of the two independent scalar coupling networks, that is, one corresponding to the pyrimidine bases and one corresponding to the sugar ring protons. The 17 pairs of H6-H5 cytosine connectivities were obtained from a TOCSY experiment (not shown). Thymine methyl group coupling with H6 base proton is quite small and much smaller than the linewidth, hence these could not

25‘C

4.4 4 .O 3.6 3.2 2.8 ppm Figure 2. 3‘P NMR spectra of the 24mer-N at 25°C (bottom) and 50°C (top) with proton decoupling and referenced to external H m ,


be detected at 25°C and only few were observed at higher temperature. Second, (i) the corresponding base and sugar protons for each nucleotide are matched via the intra-residue dipolar interaction between the H1’ or H2’ sugar proton and the H8 or H6 base protons and then (ii) the sequential assignment can be achieved via the inter-residue dipolar interactions using the four con- tinuous independent pathways, one in H 2 0 giving the interaction between adjacent imino protons, the three others in D,O via interactions between sugar protons H1‘ or H2’, H 2 with base protons of the 3‘ adjacent residue and via base-to-base proton (Fig. 3). The first step is conformation independent but the second step is not, emphasizing the necessity to define the conformation type initially. The connection between each of the assignment pathway can be used to cross-check the proposed assignments. Intra-residue and sequential (5’ to 3‘ direction) NOES are indicated using the notation defined by W u t h r i ~ h ~ ~ that is di(X;Y) and ds(X;Y), respectively.

Although sequence-specific assignment of resonances in small oligonucleotides is well documented41 and seems straightforward, for large duplexes containing more than 16 base pairs the assignment procedure remains a complicated task. Multi-dimensional NMR should be of help in doing so, in particular 3D NMR. However, here the quantity of nucleotides available, i.e. 0.25 mM (24mer-N) and 0.8 mM (24mer-M), precluded the use of 3D methods. Thus, the large size of the molecule, the fact that the nucleotide is not palindromic, the presence of long stretches of homopurine sequences always difficult to identify owing to the presence of only one base proton, and the small quantity available prompted us to use a strategy for assignments that differs from the standard way usually employed, and instead is similar to the strategy adopted for proton assignment of the Immobile Holliday J u r ~ c t i o n s ? ~ - ~ ~

H6/H8 \ , H6/H8 H1’

H2’ 7 HS/CH3 H2”

- H6/H8

\ H6/H8

H1‘ + H2’ - HS/CH3 -H5/CH3

H2” 9 HVCH3

Ei: I - HS/CH3

H2” H6/H8

H2 - HI’ H2 CL) H1’

Figure 3. Schematic representation of intra-strand sequential connectivities observed in a B-DNA (adapted from Ref. 41 ).

Non exchangeable proton assignments

When working with large DNA molecules, the problem of starting locations is particularly complex. Usually the chain-terminal nucleotides serve as the reference location for sequential assignment: since the terminal deoxyriboses are not phosphorylated, the corresponding H3’ and H5’ H5” resonate at higher field position compared to the corresponding protons of the other units. The H6 base protons of the two 3‘ end cytosines at 7.69 and 7.42 ppm give strong intra-residue interactions with the H3’ sugar protons at 4.54 and 4.48 ppm, respectively, while the H8 base protons of the two 5‘ end guanines at 7.97 and 7.83 ppm give intra-residue interactions with H5‘ and H5” protons at 3.75 and 3.66 ppm, respectively. Then, for these four residues, com- plete assignment of the sugar ring protons can be obtained from the TOCSY spectrum.

The location of the thymine methyl groups in a characteristic region of the spectrum serves at excellent starting point for the assignment of all XT dinucleotide units. Strong intra-residue dipolar interactions are expected between methyl groups and their corresponding base protons di(6;Me) since they are closely spaced (Fig. 4, left). A weak inter-residue interaction is also expected between the methyl group and the base proton of the 5‘ adjacent base ds(6/8 ; Me). The three sequential thymine T7, T8 and T9 residues in the coding strand are easily found through their mutual interactions di(Me ; 6) or ds(6 ; Me) (Fig. 4). There are three CT dinucleotides, C6-T7, C32-T33 and C36-T37, which can be found by matching the observed sequential effects ds(6;Me) in the methyl group region with the corresponding cytosine base proton region di(6; 5) (Fig. 4, right). The three CT dinucleotide assignments can be confirmed via the sequential interaction ds(5 ; Me) observed between the thymine methyl group and the cytosine H5 proton. Moreover, the C6 aromatic protons can be readily assigned to the 7.23 : 5.16 ppm pair as it precedes the three sequential thymines already assigned.

To proceed further in the assignment we used the effect of spin diffusion from cytosine H5 base proton to the 5’ adjacent residue base proton: a particularity of DNA molecules is that spin diffusion effects are always in the direction of the assignment ~ a t h w a y . ~ ~ . ~ ~ Effects can be observed between the base protons H6 or H8 of residue X and the cytosine H5 protons of the 3‘ adjacent residue allowing the assignment of all XC dinucleotides. Eleven H5 cytosine base protons but six gave an interaction with the preceeding H8/H6 base proton (Fig. 5). The XC sequences not identified here were mainly involved in TC or CC sequences and could be assigned via ds(5;Me) for T9-Cl0, T33-C34 and T44-C45 or via ds(5 ; 5) for C26-C27, C3 1-C32 and C45-C48.

The identification of the seven XT, 11 XC and a few CC as well as the chain-terminal residues gave a reason- able number of starting points to achieve the sequential assignment using the standard through-space connectivities between base protons and H1‘ sugar proton, di(8; 1’) and ds(1’; 8), in a NOESY spectrum, as shown in Fig. 6 (top) for the coding strand and Fig. 6 (bottom) for the other strand. The sequential resonance assignment of both strands started with the single cross peak

'H NMR CONFORMATIONAL STUDIES OF HIV ENHANCER

I I " . ' I . . . ' I . . . ' , . . . ' I ' ' . ' , . , , . I , . ' , l , , . , . . . . . . . . . . . . . . . . . . . .

8.3 8-2 8.1 8.0 7.9 7.8 7.7 7.6 7.5 7 . 4 7 . 3 7.2 7 . 1 7.0

S7 1

5.3-

5.4-

n 8 5.5-

5.6-

P * 5.7-

PPm

Figure 4. Through-space connectivities obtained at 30 "C with a mixing time of 150 ms. Left, interactions between thymine methyl groups and H 6 / H 8 base protons; right, interactions between cytosine H 5 base proton and cytosine H 6 base proton.

b

involving the H8 base proton of the 5'-terminal G1 and 8;6/8) (not shown) with the base to H1' sugar proton G25 di(8 ; 1') at 7.83 and 7.97 ppm, respectively. region (Fig. 6). Despite the methodology described here, However, the severe spectral overlap precluded a com- sequential assignment of the fragments T8-Cl0, G13- plete analysis of this characteristic region, and assign- G15, C26-C27, C32-C36 and C45-C46 was difficult as ment could only be obtained through parallel analysis these protons have either similar base proton chemical of different regions, mainly the base-to-base region ds(6/ shifts (G13 and G15, C26 and C27, C32 and C36, C45

*- 0. 2% -IpI ....................... .....................

.......... 8-31c a 23c 6 0

6 . 1 1 6 . 2 1

8

0

a

86 L

8 "

a

s12

- F l . Ippm): 5.2:

5.3-

5.4-

$ 5.5:

H g 5.71

CL 11.6-

a

i a 5 . 0 -

5.9-

M. DELEPIERRE ET AL.

'CmG'Tz2Gz'G 'C3'

6 . 0

6 1-

6 .2-

-0 e-

0 U

I

I

43

Y - l 3 . n t edb ' a n

i , , , , . , . . , . ~ I * ' * ' I ' ' ' X

8 3 8 2 8 1 8 0 7 9 7 8 7 7 7 6 7 5 7 4 7 3 7 2 7 1 7 C

F2 (DDm)

Bue protons

Figure 6. Sequential assignment from di(8; 1') and ds(1';8) interactions observed in the NOESY spectrum obtained at 30°C with a mixing time of 300 ms for the 24mer-N. The top spectrum represents the coding strand and the bottom spectrum the other strand. Starting locations, namely G1 and G25, are indicated with an arrow.

and C47) or H1' chemical shifts (G14 and G13 or C34 between G25 and C26 base protons and between A28 and C35). Fragment T8-ClO was easily assigned in a and C27 base protons. For the fragment G13-Gl5, G13 NOESY spectrum obtained at higher temperature and G15 were assigned from ds(6/8; 6/8) interactions (50°C) (data not shown). The dinucleotide C26-C27 observed between their base proton and the 3' and 5' was assigned from ds(6/8 ; 6/8) interaction observed adenine base proton, respectively, while the G14 loca-

'H NMR CONFORMATIONAL STUDIES OF HIV ENHANCER s73

13.0:

13.2:

13.4:

13.6:

13.8:

1 4 . 0 1

tion was deduced, at the end of the assignment procedure, by matching the number of assigned signals in the aromatic region with the integration curve of this region. The fragment C32-C36 was assigned from ds(5;Me) between C32 and T33, and between C36 and T37 and from ds(Me; 5) interactions observed between T33 and C34. The C35 assignment was deduced as it remained the last cytosine not assigned. The proposed assignment pathways have been cross-checked at several temperatures and on a sample in which all purine H8 base protons have been exchanged with deu- terium after heating at 70°C44 (data not shown). Base protons and sugar protons Hl', H2' and H2" and a few H3' protons are reported in Table 1.

:a

Adenine H2 base proton assignments

The seven adenine H2 protons are located in the spectrum via measurement of their characteristic long longitudinal relaxation time TI .45 The assignment of these protons proceeded via two pathways: (i) the three sequential adenines A40, A41 and A42 were assigned via direct through-space connectivities due to sequential intra-strand interactions between H2 base protons ds(2 ; 2), and (ii) the assignments of the remaining adenines were obtained via through-space connectivities between the base protons and the H1' sugar protons of intra-residue di(2 ; l), sequential intra-strand ds(2 ; 1') and inter-strand interactions dci(2 ; 1') distinguished by the additional subscript c according to Chen et al.42 The

long mixing time (zm = 300 ms) used allowed the observation of these three types of interactions in most cases, with the exception of A12. Its H2 proton reson- ated at 7.55 ppm in a very crowded region and was thus assigned by default. Assignments are reported in Table 1.

Exchangeable proton assignments

Finally, maybe the most important region in view of protein-DNA interaction studies is the imino proton region. Indeed, these are always located in a region free of protein signals and are very sensitive to the base-pair ring current effects. They can thus be influenced by changes in the local geometry.

The most direct sequential pathway to assign exchangeable protons is based on strong through-space interactions between imino protons of adjacent base pairs. These protons are separated by about 3.4 A and located in the uniquely low-field-shifted resonances region 12-15 ppm or via the intra-base-pair dipolar interaction between thymine imino proton and adenine H2 base proton dci(N3 ; 2) (about 2.5 A). Furthermore, as expected, the thymine N3H resonances (13.5-14.5 ppm) were well separated from the guanine NlH resonances (12.5-13.4 ppm) (Fig. 7, right). In the present study, these were used as starting points for assignments as all the adenine H2 protons were assigned independently. When peaks corresponded to a single proton, the assignment of the thymine N3H signal proton was as straightforward as for the A5.T44 at 13.67 ppm, the

P

*+-*-----.- I Po 8.4 8.2 8.0 7.8 7 .6 7.4 7 . 2 7 . 0 14.0 13.8 13.6 13.4 13.2 13.0 12.8 12.E

Fl (ppm) ~1 (ppm)

Figure 7. Through-space connectivities obtained at 25°C with a mixing time of 150 ms. On the right are shown the interactions between imino protons and on the left the interactions between adenine H2 base protons with the intra-imino base pair but also with adjacent imino protons

Tabl

e 1.

Che

mic

al sh

ift a

ssig

nmen

ts (p

pm) o

f the

24m

er-N

and

24m

er-M

oon

-exc

hang

eabl

e and

exc

hang

eabl

e pro

tons

Dup

lex

DN

A 2

4mer

-N

1G

2G

3G

4G

5A

6C

7T

8T

9T

1 oc

11c

12A

13

G

14G

15

G

16A

17

G

18G

19

c 20

G

21 T

22

G

23G

24

C

H8

H6

H5

7.83

7.

78

7.68

7.

67

8.1 2

7.

23

5.16

7.

37

1.48

7.

47

1.60

7.

41

1.63

7.

53

5.64

7.

43

5.60

8.

1 5

7.56

7.

58

7.58

7.

99

7.55

7.

56

7.86

7.85

7.

73

7.24

5.

19

7.12

1.

44

7.42

5.

40

H2

7.86

7.56

7.66

H1'

5.64

5.

61

5.68

5.

57

6.20

5.

77

6.03

6.

1 5

6.05

5.

88

5.27

5.

90

5.49

6.

01

5.49

6.

01

5.63

5.

84

5.65

5.

94

5.73

5.

65

5.97

6.

1 7

H2'

/H2"

2.46

2.61

2.52

-2.7

4 2.

51-2

.67

2.62

-2.7

4

2.67

-2.8

8 1.

97-2

.49

2.1 9

-2.6

0

2.1 8

-2.5

5 2.

1 1-2

.35

1.99

-2.2

6 2.

69-2

.83

2.83

-2.6

5 2.

462.

62

2.55

-2.8

2 2.

462.

64

2.46

2.67

2.

00-2

.39

2.59

-2.7

6 1.

95-2

.33

2.1 7

-2.6

3

2.47

-2.6

3

2.67

-2.7

5 2.

52-2

.70

2.1 9

-2.2

0

HJ

4.79

4.

96

5.01

4.

95

5.01

4.

65

4.86

4.

89

4.90

4.

82

4.79

5.

01

4.93

4.

93

4.93

5.

01

4.91

4.

93

4.83

4.

96

4.84

4.

98

4.95

4.

48

imin

os

(13.

12)

(13.

1)

12.8

1

14.0

7 13

.94

13.7

3

12.8

0

12.7

0

12.9

1 12

.86

12.7

6 13

.73

12.9

6 13

.07

i 2.

88

H8

25G

7.

97

26C

27

C

28A

8.

27

29C

30

G

7.82

31

C

32C

33

T 34

c 35

c 36

C

37

f

39G

7.

665

40A

8.

03

41A

8.

015

42A

7.

94

43G

7.

36

44T

45c

46C

47

c 48

C

386

7.77

H6

H5

7.52

5.

41

7.52

5.

64

7.19

5.

26

7.35

5.

27

7.53

5.

50

7.46

1.

59

7.58

5.

65

7.53

5.

55

7.52

5.

56

7.27

1.

62

7.20

1.

10

7.56

5.

58

7.55

5.

61

7.59

5.

71

7.69

5.

83

H2

7.74

7.20

7.

1 1

7.35

H1'

5.99

6.

06

5.43

6.

1 9

5.56

5.

83

5.91

5.

83

6.04

5.

95

5.93

5.

85

5.56

5.

50

5.33

5.

78

5.82

5.

94

5.77

6.

01

5.96

5.

92

6.03

6.

25

H2'

/H2"

2.64

-2.7

7 2.

1 8-2

.46

2.07

-2.3

7 2.

73-2

.86

1.88

-2.3

2 2.

29-2

.68

2.20

-2.4

8 2.

1 1-

2.49

2.

22-2

.53

2.22

-2.4

2 2.

21-2

.55

2.04

-2.4

1

2.58

-2.6

6 2.

43-2

.61

2.60

-2.8

0 2.

57-2

.80

2.56

2.81

2.

352.

66

2.1 3

-2.5

2 2.

23-2

.49

2.22

-2.4

7 2.

21-2

.46

2.28

1.91

-2.2

3

H3'

4.84

4.

85

4.84

5.

02

4.83

4.

94

4.76

4.

72

4.84

4.

95

4.70

4.

76

4.81

4.

94

4.94

5.

01

5.01

4.

98

4.77

4.

84

4.83

4.

82

4.82

4.

54

imin

os

12.8

6 F 5

13.7

8 rm 21 E 6 m

13

.94

12.8

1 5

12.5

8 1

P

12.8

1 13

.67

Tabl

e 1.

(C

ontin

ued)

Dup

lex

DN

A 2

4mer

-M

1 G

2c

3T

4c

5A

6C

7T

8T

9T

1 oc

11c

12A

13

G

14G

15

G

16A

17

G

18G

19

c 20

G

21 T

22

G

23G

24

C

H8

7.96

8.27

8.1 5

7.

58

7.57

7.

58

7.98

7.

53

7.56

7.86

7.85

7.

74

H6

H5

H2

7.54

5.

37

7.44

1.

60

7.51

5.

71

7.29

5.

25

7.37

1.

49

7.44

1.

59

7.41

1.

62

7.51

5.

64

7.43

5.

60

7.65

7.56

7.66

7.24

5.

19

7.12

1.

43

7.40

5.35

H1'

5.45

6.

09

6.08

5.

52

6.20

5.

76

6.01

6.

1 2

6.05

5.

87

5.26

5.

90

5.49

6.

01

5.49

6.

01

5.58

5.

84

5.65

5.

94

5.75

5.

65

5.97

6.

1 6

H2'/H

2"

H3

2.46

-2.6

1 5.

01

2.1 7

-2.5

7 4.

81

2.1 6

-2.5

5 4.

86

2.75

-2.8

9 5.

04

1.97

-2.4

9 4.

65

2.1 5

-2.5

8 4.

85

2.1 7

-2.6

1 4.

89

2.18

-2.5

2 4.

91

1.95

-2.2

5 4.

80

2.70

-2.8

3 5.

01

2.50

-2.6

3 4.

96

2.34

-2.4

2 4.

93

2.46

-2.5

6 4.

94

2.56

-2.8

5 5.

01

2.45

-2.6

5 4.

94

2.46

-2.6

7 4.

92

2.00

-2.3

9 4.

83

2.59

-2.7

6 4.

95

2.67

-2.7

5 4.

97

2.53

-2.7

2 4.

96

2.1 6

-2.2

0 4.

48

2.01

-2.3

7 4.

84

2.1 2

-2.4

0 4.

84

1.92

-2.3

4 4.

85

imin

os

13.8

0

14.0

6 13

.92

13.6

8

12.8

8 12

.86

12.6

7

12.8

9 12

.88

12.7

3 13

.70

12.9

3 13

.1

25G

26

C

27 C

28

A

29C

30

G

31 C

32

C

33T

34c

35c

36C

37

T 38

G

39G

40

A

41 A

42

A

43G

44

T 45

G

46A

47

G

48C

H8

H6

7.96

7.

51

7.50

7.20

7.36

7.

52

7.46

?

7.55

7.

53

7.51

7.

27

8.27

7.82

7.76

7.

66

8.03

8.

00

7.92

7.

37

7.85

8.

1 0

7.62

7.02

7.35

H5

H2

5.39

5.

64

5.25

5.27

5.

48

1.58

5.

65

5.55

5.

56

1.62

7.74

7.1 9

7.

08

7.34

1.13

7.69

5.28

H1'

5.99

6.

08

5.43

6.

20

5.55

5.

84

5.91

5.

83

6.04

5.

95

5.93

5.

86

5.56

5.

50

5.34

5.

80

5.82

5.

96

5.74

5.

73

5.41

6.

06

5.80

6.

1 3

H2'

fH2"

HJ

2.64

-2.7

8 4.

84

2.20

-2.4

6 4.

86

2.73

-2.8

6 5.

03

1.87

-2.3

0 4.

79

2.60

-2.6

8 4.

94

2.20

-2.4

8 4.

76

2.23

-2.5

4 4.

87

2.22

-2.4

2 4.

95

2.05

-2.4

3 4.

76

2.07

-2.3

7 4.

84

2.1 1

-2.4

9 4.

72

2.21

-2.4

7 4.

70

1.91

-2.2

3 4.

79

2.58

-2.6

6 4.

94

2.46

-2.6

2 4.

95

2.55

-2.8

0 5.

01

2.50

-2.8

1 5.

02

2.33

-2.6

4 4.

83

2.65

-2.7

5 4.

98

2.65

-2.8

6 5.

05

2.45

-2.6

5 4.

94

2.1 2

-2.2

0 4.

45

2.57

-2.8

1 4.

98

1.87

-2.3

2 4.

84

imin

os

12.8

5

13.7

5

13.9

0 12

.78

12.5

5

12.7

0 13

.48

12.5

5

12.9

5


A16.T33 at 13.78 ppm and the T7.A42 at 14.07 ppm. There were two overlapping peaks at 13.94 and 13.73 ppm, of which one, at each frequency, belonged to the unique fragment containing three sequential T.A pairs. This allowed us to assign the two peaks at 13.94 ppm to the T8.A41 and the A12.T37 pairs and the two peaks at 13.73 ppm to the T9.A40 and to the T21.A28 pairs. However, whereas the T.A base pairs could be easily identified, the strong overlap of peaks in the G.C region (12.5-13.3 ppm) permitted only limited information. An alternative to direct imino-imino assignment involves a two-step pathway, first to establish intra-base-pair correlations from the imino proton to the cytosine 4NH2 resonances within G.C base pairs dci( 1 ; 4NH2) or to the H2 resonance within A.T base pairs dci(N3 ; 2), and then to follow sequential through-space interactions from H2 resonances, for example, to adjacent imino protons dcs(Nl,N3 ; 2) and dcs(2 ; N1,N3).36,42,46 These sequential NOES are directional and can be used to distinguish a G.C from a C.G or an A.T from a T.A.42 In addition, the dci(N3 ; 2) distance is much shorter than the dcs(Nl,N3 ; 2) or the dcs(2 ; Nl,N3) distance. The first starting point is the unique (T.A)3 sequence (Fig. 7). To overcome the strong overlapping problem in the GC region assignments were completed either from a spectrum obtained at a lower temperature (10 "C) or through relayed information under conditions of extensive spin diffusion from the guanine imino protons to the cytosine H5 base protons via amino protons. The remaining imino protons that could not be assigned using this procedure involve G2 and G3 NH1, which must be at 13.10 and 13.12 ppm as deduced from 1D spectrum integration curve, while the Gl.C48 and C24.G25 imino must be exchanged due to fraying.

The mutated sequence 24-merM

The whole strategy developed for the 24mer-N was used to assign the protons of the mutated sequence, giving an additional cross-check of the validity of the proposed assignments. These assignments are reported in Table 1.

During the completion of this paper, two NMR studies of the conformational characteristics of the HIV-1 long terminal repeat first NF-KB site4' and of the Spl binding enhancer element4* were published. Taking into account that the size of the duplexes is different and that these structural studies were performed using different buffer conditions, in particular in the absence of magnesium, the structural features of the two shorter fragments are similar. Although this statment is particularly true for the NF-KB site some differences were observed between the decamer corresponding to the Spl site and the related sequences in the 24-mers. Indeed, in the decamer the G8 H1' resonance displays a large downfield shift (0.62 ppm) compared with the resonance of the corresponding residue in the 24-mers (G22). Similarly, the A14 H2 base proton resonance is shifted upfield by 0.6 ppm in the decamer compared with its position in the 24-mers (A28). Of course, the decamer was studied at 25°C and end effects might be important for a fragment of this size. However, the upfield position of the A14 H2 base proton resonance, a base surrounded by two pyrimidine bases appears quite

surprising. Similarly, the H 1' resonance position of G8 is well outside the expected range for the H1' sugar proton of a purine. Because our assignments were obtained independently on the two 24-mers, we are con- fident in their validity.

Structural features of the 24-mers

Once assignments had been obtained, the structural features of both sequences could be analyzed. This was achieved by monitoring the characteristic distances for each type of known helices family. Stronger effects observed at short mixing time (50 ms) between base protons and H2'/H2 sugar protons suggest a right- handed conformation. The lack of strong interactions between aromatic protons and the H3' sugar proton excludes a C3' endo conformation for the sugar and consequently A-type helices. This statement has to be treated with caution as the distinction between A-DNA and B-DNA has been established largely from the conformation of the sugar phosphate backbone and from the average rise between base pairs. As already sug- gested, if base stacking is taken into consideration, the demarcation between the two forms is not so clear.49 To illustrate this, the few weak interactions observed for some of the residues between H1' and H 4 sugar protons, a distance very sensitive to the sugar conformation, are an indication of sugar phase angle variations that differ in B- and A-DNA conformation^.^^ Unfortunately, the lack of a full assignment for the H4' sugar protons precludes any definitive conclusion to be drawn from this observation.

A detailed analysis of the structure at particular points with respect to sequence-specific variations would require a quantitative analysis of cross-peak intensities of the NOESY spectrum, dihedral bond angle information from homonuclear proton coupling constants and also from heteronuclear proton- phosphorus coupling constants together with the use of back calculation refinement methods. Although NOESY experiments were performed at several temperatures and with mixing times ranging from 50 to 500 ms, a quantitative analysis of cross-peak intensities was not possible for several reasons. First, we have to face severe overlaps, second, the size of the nucleotide is such that spin diffusion occurs since 50 ms, and finally, with the shape of the molecule that is a long cylinder, the hypothesis of isotropic motions for reorientation no longer hold for distance evaluation.

Neverthelesss, in this study our interest was the comparative analysis of the two 24-mers structures in solution. Thus, a qualitative analysis of the NMR parameters indicated that, although the overall structure of the duplex is of the B form, there are several significant local structural deviations. These were evi- denced from (i) analysis of NOE cross-peak intensities between adenine H2 base proton and sugar HI' protons and (ii) from chemical shift analysis of H1' sugar proton.

(i) It has been proposed that the sequence-dependent digestion of DNA by various nucleases is due to sequence-dependent variations in DNA structures such as variations in the helix groove widths, in radial asym- metry and in phosphate acce~siblity.~ ls2 Since

'H NMR CONFORMATIONAL STUDIES OF HIV ENHANCER s71

0 8

"A

7 . 6 7 . 5 7 . 4 7.3 7 . 2 7 . ;

5 . 4 @ 5 . 6

5 7 4

5 .e -

5.9-

6 . 0 -

6.1-

Q

8

8

Q

@

00 0

Q 6

8

0 Ill '

00

6

4 9

0

8 3 8 2 8 1 8 0 7 9 7 8 7 7 7 6 7 5 7 4 7 3 7 2 7 1 7 0

F2 (WIT)

Figure 8. Through-space connectivities obtained at 50°C with a mixing time of 300 ms. Interactions between adenine H2 base proton with H1' sugar proton are indicated: top spectrum for the 24mer-N in which the purine H8 base protons have been partly exchanged, bottom spectrum for the 24mer-M. Similar results were obtained at lower temperatures but then overlapping signals hampered a full analysis, which was easier at 50°C.


AH2-H1’ interactions are sensitive to the DNA conformation and minor groove width, measurements of these cross-relaxation rates may permit an evaluation of the local onf formation.'^-'' The adenine H2 interactions can give rise to three types of NOE with sugar HI’ protons, two involving intra-strand resonances, that is, with its own Hl’, and with the H1‘ of the 3‘ flanking residue and one involving inter-strand resonances. The adenine H2 base protons are isolated from other non- exchangeable protons, regardless of DNA conformation so the contribution of spin diffusion is minimal.53 Although the AH2-H1’ interaction is difficult to quan- tify because the NOE intensity reflects not only the cross relaxation rate constant but also the longitudinal relaxation time saturation effects that can seriously distort the apparent NOE. However, here we were interested in a comparative study and experiments were run under identical conditions on duplexes of the same size. A semiquantitative interpretation of NOE cross- peak intensities in terms of distances can thus be achieved. Interestingly, while the intensities of the observed effects were quite similar in the two sequences for residue A5, we found a large difference for the cross- strand effect (Fig. 8). The strong effect observed between the adenine 5 H2 base proton and the H1’ of residue 45 in the 24mer-N sequence, and indicative of a narrow minor groove, is absent in the 24mer-M sequence indicating a wider minor groove (Fig. 8). It has been shown previously that the cross-strand effects probe the adenine environment and that this distance is large for all pyrimidine-purine steps such as CA for the 24mer-M and small or large, depending on the flanking sequences, for purine-purine steps such as GA in the 24mer-N sequence. 54 The results presented here agree well with the sequence analysis, suggesting that pyrimidine-purine steps widen the minor grooves4 and are in favour of a narrower minor groove in the native sequence.

It seems likely that the narrower the minor groove, the wider the major groove will be, and vice versa, although this interpretation requires more data to be proved decisively, and may not be valid for all sequence types. Indeed, a recent analysis of the geometry of DNA grooves has shown that there is no simple correlation between major and minor groove ~ i d t h s . ’ ~ , ’ ~ The large width for the minor groove in the mutated sequence would induce modifications for the major groove, thus preventing association with the protein since it binds via the major groove. This cannot be excluded, and it is more likely that the lack of binding is simply related to the absence of the two adjacent guanines in the mutated sequence and not to structure alteration, since the x-ray structures published on the NF-rcB :DNA complexes20-22 report hydrogen bond interactions with the two adjacent guanines on the coding strand. However, none of these complexes has been solved with the HIV-rcB DNA sequence.

Interestingly the mutation that abolishes the NF-KB binding (that is, the GGG replacement by a CTC) has introduced a CAC site. These sites are known to have unusual structures, probably owing to the unfavorable stacking of CAC base^.'^^'^ In addition, CAC triplets are encountered in a wide variety of protein:DNA recognition ~ i t e s . ~ ~ . ~ ~

(ii) It has been already reported that in DNA the resonance position of a particular proton is determined essentially by its local chemical environment, therefore by the nucleotide sequence and subsequently by the local conformation.62 Analysis of H1‘ chemical shifts obtained for the two 24mers shows unusual positions for the C11 H1’ sugar proton (5.27 ppm) and also mar- ginally for the G39 H1’ proton (5.33 ppm). These two residues are involved in a CpA or a GpA step. From a statistical analysis of over 600 nucleotides it has been shown that the H1’ resonance positions appear to be sensitive to their own nucleotide and to the one on their 3‘ side,62 and ring current effects induced by adenine are the largest ones.63 In the nucleotides studied here there are two other GpA sequences (4G5A or 15G16A) and one other CpA sequence (27C28A) for which these upfield shifts are not observed. A literature survey of the assignments published for DNA sequences of at least 10 base pairs is in agreement with the fact that the C11 H1’ position is well outside the chemical shift probability profile defined for cytosine and to a lesser extent the G39 Hl’. High-field shifted resonances have already been observed previously for H1’ sugar protons. This was the case for C3 H1’ sugar proton in the d(CGAAAAATCGG) : (CCGATTTTTCG) sequence which had been described as a putatively bent DNA64 and also for the G2 H1’ sugar proton in the d(CCAAAATTTTCG), sequence for the same reason in relation to the presence of an adenine t r a ~ t . ~ ’ Although, in the absence of protein, there is no evidence here of a bent structure for the 24mers, it is interesting to note that the two residues involved, i.e. C11 and G39, occupy positions n on the coding strand and n + 1 on the opposite strand, respectively, and are localized at the junction between homopyrimidine and homopurine

homopurine : homopyrimidine sequences are known to be frequent in eukaryotic genomes. They are assumed to adopt noncanonical B-DNA structures that can act as regulatory signals in gene expression66 via binding to double-stranded DNA in a sequence-specific manner to form triple helices.67 Our results did not allow us to link these high-field shifted resonances to a peculiar structure or to an equilibrium between different confor- mations. However, taken together, the observation of low-field signals in the 31P spectrum (Fig. 2) and the higher sugar phase values expected in the homopurine stretch compared with homopyrimidine stretch here is a favorable case for a BI to BII t r a n s i t i ~ n . ’ ~ ~ ’ ’ ’ ~ ~ - ~ ~

Finally, the C11 and the G39 nucleotides are very close to the junction between the NF-rcB and the Spl binding sites and belong to the p65 weak binding site. Knowing that the activity of the HIV LTR is synergistically induced by signals targeting transcription factors on both the NF-rcB and the Spl binding sites, this local structural modification might play an important role in protein-protein interactions.

stretches. These long stretches of

CONCLUSION

With the strategy described in this paper, it was possible to obtain assignments of protons essential for structure

'H N M R CONFORMATIONAL STUDIES O F HIV ENHANCER s79

determination and to analyze structural features of both duplexes. It was a complicated task in the case of these large duplexes for the following reasons: (i) the large size and the shape of the molecules render the dipolar interaction analysis difficult as spin diffusion occurs at short mixing times and rotational motion are not isotropic; (ii) the nucleotides are not palindromic; (iii) the presence of long stretches of homopurine sequences which are always difficult to identify owing to the presence of only one base proton; and (iv) the small quantity available. Nevertheless, information on the solution structure could be obtained and a qualitative analysis of the NMR parameters indicated that although the overall structures of the duplexes are of the B form, there are several significant local structural deviations. Of course, one can wonder why is it necessary to study such large fragments instead of shorter oligomers of 10 or 12 base pairs. Indeed, such studies have already been conducted on the two consensus sequences, NF-KB and Spl. However because the HIV LTR activity is synergistically induced by signal targeting transcription factors on both the NF-KB and the Spl binding ele-

ments, it is important to have structural informations on the junction between the two elements. The peculiar structure suspected at base pairs C10-G39 and C11-G38 reinforced the usefulness of this kind of study. Only limited conformational information can be derived here but it provides good support for further studies on shorter fragments. This study is now being pursued in our laboratory on 16mer fragments encompassing the ten base pairs of the KB site (bold letters)

CCCAG) with three extra base pairs at both extremities to avoid end effects but also to include the junction.

d5'(CTGGGGAC"l'CCAGG) : d(CCTGGAAAGTC-

Acknowledgements

We thank Dr Brigitte Hartmann and Dr Fernando Arenzana- Seisdedos for stimulating discussions. This work was supported by grants from the Institut Pasteur, the Agence Nationale pour la Recherche sur le SIDA and the Centre National de la Recherche Scientifique. P.S. acknowledges the Agence Nationale pour la Recherche sur le SIDA for a fellowship.

REFERENCES

1. S. C. Harrison, Nature (London) 353,715-71 9 (1 991 ). 2. C. 0. Pabo and R. T. Sauer, Annu. Rev. Biochem. 61, 1053-

8. 9.

10.

11 .

12.

13. 14. 15.

16.

17.

18.

19. 20.

21.

22.

23. 24.

25. 26.

27.

1 095 (1 992). 3. C. Wolberger, Curr. Opin. Struct. Biol. 3, 3-1 0 (1 993). 4. M. Ptashne, Nature (London) 335,683489 (1988). 5. P. J. Mitchell and R. Tjian, Science 245, 371-378 (1989). 6. P. Lamb and S. L. McKnight, Trends Biochem. Sci. 16,

7. K. Struhl, Trends Biochem. Sci. 14,137-140 (1 989). 41 7422 (1 991 ) . P. A. Baeuerle. Biochim. Bioohvs. Acta 1072,6340 (1 991 ). M. J. McElrath, J. E. Prueiand Z.A. Cohn,.Proc. N h . Acad. Sci. USA 86,675-679 (1 989). P. Simmonds, P. Balfe, J. F. Peutherer, C.A. Ludlan, J.O. Bishop and A. J. Leigh Brown, J. Virol. 64,864-872 (1 990). W. C. Greene, E. Bonlein and D.W. Ballard, Immunol. Today

B. Berkhout, A.Gatignol, A.B. Rabson and K-T. Jeang, Cell

R. Gaynor,AlDS 6,347-363 (1992). J.-L. Virelizier, Curr. Opin. Immunol. 2, 409-41 3 (1 990). E. K. Ross, A. J. Buckler-White, A. B. Rabson, G. Englund and M. A. Martin, J. Virol. 65, 4350-4358 (1 991 ). G. Nabel and D. Baltimore, Nature (London) 326, 71 1-71 3 (1987). J. Alcami, T. Lain de Lera, L. Folgueira, M. A. Pedraza, J. M., J. M. Jacque, F. Bachelerie, A.R. Noreiga, R. T. Hay, D. Harrich, R. B. Gaynor, J.-L. Virelizier and F. Arenzana- Seisdedos, EMBO J. 14,1552-1 560 (1 995). J. Vlach, A. Garcia, J.-M. Jacqub, M. S. Rodriguez, S. Michel- son and J.-L. Virelizier, Virology 208, 753-761 (1 995). M. B. Urban and P.A. Baeuerle, Cell 61,255-265 (1 990). C. Muller, F. Rey, M. Sodeka, G. Verdine and S. C. Harrison, Nature (London) 373,311-31 7 (1 995). C. Muller and S. C. Harrison, FEBS Lett. 369, 1134-1 17 (1 995). J. Kuriyan and D. Thanos, Curr. Opin. Struct. Biol. 3, 1351 41 (1 995). A. K. Aggarwal, Struct. Biol. 2,1861 86 (1 995). G. P. Nolan, S. Ghosh, H-C. Liou, P. Tempst and D. Baltimore, Cell 64,961-969 (1991). S. McKnight, Pourla Science 164,4654 (1991). R. Schreck, H. Zorbas. E.-L. Winnacker and P.A. Baeuerle, Nucl. Acids Res. 18, 6497-6502 (1 990). G. Ghosh, G. Van Duyne, S. Ghosh and P. Sigler, Nature (London) 373,303-31 0 (1 995).

10,272-278 (1989).

62,757-767 (1 990).

28.

29. 30. 31.

32.

33.

34.

35.

36.

37.

38.

39. 40.

41.

42.

43.

44.

J. R. Matthews, J. Nicholson, E. Jaffray, S. M. Kelly, N. C. Price and R . T. Hay, Nucleic. Acids Res. 23, 3393-3402 (1 995). R. Sen and D. Baltimore, Cell 46,705-71 6 (1 986). H. E. Moser and P. B. Dervan, Science 238,645-650 (1987). T. L. Doan, L. Perrovault, D. Praseuth, N. Habhoub, J.-L. Decout, N. T. Thuong, J. Lhomme and C. HBIBne, Nucleic Acids Res.l5,7749-7760 (1 987). D. J. States, R. A. Haberkorn and D. J. Ruben, J. Magn. Reson. 48,286-292 (1 982). A. Bax and D. G. Davies, J. Magn. Reson. 65, 355-360 (1 985). C. Griesinger, G. Otting, K. Wuthrich and R. R. Ernst, J. Am. Chem. SOC. 110,7870-7872 (1988). D. R. Hare, D. E. Wemmer, S. H. Chou, G. Drobny and B. R. Reid, J. Mol. Biol. 171,319-336 (1 983). K. Wuthrich, NMR of Proteins and Nucleic Acids. Wiley, New York (1986). P. Sodano and M. Delepierre. J. Biomol. NMR 3, 471477 (1 993). P. Sodano and M. Delepierre, J. Magn. Reson. A 104, 88-92 (1 993). F. Schaeffer, A. Kolb and H. Buc, EMBO J. 1.99-105 (1982). D. Wemmer and B. R. Reid, Annu. Rev. Phys. Chem. 36, 105-137 (1985). A. M. Gronenborn and G. M. Clore, Prog. Nucl. Magn. Reson. Spectrosc. 17, 1-32 (1 985). S-M. Chen, F. Heffron and W. J. Chazin, Biochemistry 32, 31 9-326 (1 993). S- M. Chen and W. J. Chazin, Biochemistry 33,1145S11459 (1 994). J. M. Benevides, D. Lemeur and G. J. Thomas, Jr, Bio- oolvmers 23.1 01 1-1 024 I1 9841.

45. N. Assa-Munt and D. R.'Kearns, Biochemistry 23, 791-796 (1 984).

46. G. Otting, R . Grutter, W. Leupin, C. Minganti, K. N. Ganesh, B. S. Sproat, M. Gait and K. Wuthrich, Eur. J. Biochem. 166, 21 5-220 (1 987).

47. M. P. Singh, N. L. Fregeau, R. T. Pon and J. W. Lown, J. Biomol. Struct. and Dyn. 13,269-284 (1 995).

48. M. P. Singh, R. T. Pon and J. W. Lown, J. Biomol. Struct. and Dynam. 13,553-564 (1995).

49. A. A. Travers, Curr. Opin. Struct. Biol. 2, 71-77 (1 992). 50. F. J. M. Van de Ven and C. W. Hilbers Eur. J. Biochem. 178,

1-38 (1 988).

sso M. DELEPIERRE ET AL.

51. H. R. Drew and A. A. Travers, Nucleic Acids Res. 13, 4445- 4467 (1 983).

52. U. Heinemann and M. Hahn, J. Biol. Chem. 267, 7332-7341 (1 992).

53. R. Grutter, G. Otting, K. Wuthrich and W. Leupin, Eur. Bipophys. J. 16,279-286 (1 988).

54. V. P. Chuprina, A. A Lipanov, 0. Y. Fedoroff, S.-G. Kim, A. Kintanar and 8. R. Reid, Proc. Natl. Acad. Sci. USA 88, 9087- 9091 (1991).

55. V. P. Chuprina, E. Stetten and 0. Y. Fedoroff, J. Biomol. Struct. Dyn. 10,693-707 (1 993).

56. E. Stofer and R. Lavery, Biopolymers 34,337-346 (1 994). 57. N. Boutonnet, X. Hui and K. Zakrwewska, Biopolymers 33,

479-490 (1993). 58. D. J. Patel, L. Shapiro and D. Hare, in Unusual DNA sfruc-

fures, edited by R. D. Wells and S. C. Harvey, pp. 115-161. Springer, New York (1978).

59. A. Bolshoy, P. McNamara, R. E. Harrington and E. N. Trifinov, Proc. Natl.Acad.Sci. USA 88,2312-2316 (1991).

60. A. M. Barber and V. B. Zurkin, J. Biomol. Sfruct. Dyn. 8, 21 3-232 (1 990).

61. M. E. Donlan and P. Lu, Nucleic Acids Res. 20, 525-532

62. F. J. M. Van de Ven and C. W. Hilbers, Nucleic Acids Res. 16.

63. C. Giessner-Prettre and 8. Pullman, Biochem. Biophys. Res.

64. A. Kintanar, R. Klevit and 8. R. Reid, Nucleic Acids Res. 15,

65. M. P. Singh, 8. Plouvier, G. C. Hill, J. Gueck, R. T. Pon and J.

66. R. D. Wells, D. A. Collier, J. C. Hanvey, M. Shimizu and F.

67. E. M. Evertz, K. Rippe and T. M. Jovin. Nucleic Acids Res. 22,

68. M. Gochin, G. Zon and T. L. James, Biochemistry 29, 1161-

69. 8. Hartmann, D. Piauola and R. Lavery, Nucleic Acids Res.

70. P. Sodano, 8. Hartmann, T. Rose, S. Wain-Hobson and M. Delepierre, Biochemistry 34,6900-691 0 (1 995).

(1 992).

571 3-5726 (1 988), and references cited therein.

Commun. 70,578-581 (1976).

5845-5862 (1 987).

W. Lown, J. Am. Chem. Soc. 116.70067020 (1994).

Wohlrab, FASEB J. 20,2939-2949 (1988).

3293-3303 (1 994).

1171 (1990).

21,561-568 (1993).

Conformational Studies by1H NMR of the HIV Enhancer: the Transcription Factors NF-κB and Sp1 Binding Domains

Documents