Invariant and variable base stacking geometries in B-DNA and A-DNA

Experimental detection of conformational transitions between forms of DNA: problems and prospects

Under different conditions, the DNA double helix can take different geometric forms. Of the large number of its conformations, in addition to the "canonical" B form, the A, C, and Z forms are widely known, and the D, Hoogsteen, and X forms are less known. DNA locally takes the A, C, and Z forms in the cell, in complexes with proteins. We compare different methods for detecting non-canonical DNA conformations: X-ray, IR, and Raman spectroscopy, linear and circular dichroism in both the infrared and ultraviolet regions, as well as NMR (measurement of chemical shifts and their anisotropy, scalar and residual dipolar couplings and inter-proton distances from NOESY (nuclear Overhauser effect spectroscopy) data). We discuss the difficulties in applying these methods, the problems of theoretical interpretation of the experimental results, and the prospects for reliable identification of non-canonical DNA conformations.

Thermodynamics and site stoichiometry of DNA binding by a large antiviral hairpin polyamide

PA1 (dIm-PyPyβPyPyPy-γ-PyPyβPyPyPyPyβ-Ta) is a large (14-ring) hairpin polyamide that was designed to recognize the DNA sequence 5'-W₂GW₇-3', where W is either A or T. As is common among the smaller 6-8-ring hairpin polyamides (PAs), it binds its target recognition sequence with low nM affinity. However, in addition to its large size, it is distinct from these more extensively characterized PAs in its high tolerance for mismatches and antiviral properties. In ongoing attempts to understand the basis for these distinctions, we conducted thermodynamics studies of PA1-DNA interactions. The temperature dependence of binding affinity was measured using TAMRA-labeled hairpin DNAs containing a single target sequence. PA1 binding to either an ATAT/TATA or an AAAA/TTTT pattern is consistently entropically driven. This is in contrast to the A/T pattern-dependent driving forces for DNA binding by netropsin, distamycin, and smaller hairpin polyamides. Analysis of the salt dependence of PA1-DNA binding reveals that within experimental error, there is no dependence on ionic strength, indicating that the polyelectrolyte effect does not contribute to PA1-DNA binding energetics. This is similar to that observed for smaller PAs. PA1-DNA recognition sequence binding stoichiometries were determined at both nM (fluorescence) and μM (circular dichroism) concentrations. With all sequences and under both conditions, multiple PA1 molecules bind the small DNA hairpin that contains only a single recognition sequence. Implications for these observations are discussed.

A targeted DNA substrate mechanism for the inhibition of HIV-1 integrase by inhibitors with antiretroviral activity

We recently reported that viral DNA could be the primary target of raltegravir (RAL), an efficient anti‐HIV‐1 drug, which acts by inhibiting integrase. To elucidate this mechanism, we conducted a comparative analysis of RAL and TB11, a diketoacid abandoned as an anti‐HIV‐1 drug for its weak efficiency and marked toxicity, and tested the effects of the catalytic cofactor Mg²⁺ (5 mm) on drug‐binding properties. We used circular dichroism and fluorescence to determine drug affinities for viral DNA long terminal repeats (LTRs) and peptides derived from the integrase active site and DNA retardation assays to assess drug intercalation into DNA base pairs. We found that RAL bound more tightly to LTR ends than did TB11 (a diketo acid bearing an azido group) and that Mg²⁺ significantly increased the affinity of both RAL and TB11. We also observed a good relationship between drug binding with processed LTR and strand transfer inhibition. This unusual type of inhibition was caused by Mg²⁺‐assisted binding of drugs to DNA substrate, rather than to enzyme. Notably, while RAL bound exclusively to the cleavable/cleaved site, TB11 further intercalated into DNA base pairs and interacted with the integrase‐derived peptides. These unwanted binding sites explain the weaker bioavailability and higher toxicity of TB11 compared with the more effective RAL.

Sequence-specific physical properties of African green monkey alpha-satellite DNA contribute to centromeric heterochromatin formation

Satellite DNA, a major component of eukaryotic centromeric heterochromatin, is potentially associated with the processes ensuring the faithful segregation of the genetic material during cell division. Structural properties of alpha-satellite DNA (AS) from African green monkey (AGM) were studied. Atomic force microscopy imaging showed smaller end-to-end distances of AS fragments than would be expected for the persistence length of random sequence DNA. The apparent persistence length of the AS was determined as 35nm. Gel-electrophoresis indicated only a weak contribution of intrinsic curvature to the DNA conformations suggesting an additional contribution of an elevated bending flexibility to the reduced end-to-end distances. Next, the force-extension behavior of the naked AS and in complex with nucleosomes was studied using optical tweezers. The naked AS showed a reduced overstretching transition force (-18% the value determined for random DNA) and higher forces required to straighten the DNA. Finally, reconstituted AS nucleosomes disrupted at significantly higher forces as compared with random DNA nucleosomes which is probably due to structural properties of the AS which stabilize the nucleosomes. The data support that the AS plays a role in the formation of centromeric heterochromatin due to specific structural properties and suggest that a relatively higher mechanical stability of nucleosomes is important in AGM-AS chromatin.

The structure of SgrAI bound to DNA; recognition of an 8 base pair target

The three-dimensional X-ray crystal structure of the 'rare cutting' type II restriction endonuclease SgrAI bound to cognate DNA is presented. SgrAI forms a dimer bound to one duplex of DNA. Two Ca(2+) bind in the enzyme active site, with one ion at the interface between the protein and DNA, and the second bound distal from the DNA. These sites are differentially occupied by Mn(2+), with strong binding at the protein-DNA interface, but only partial occupancy of the distal site. The DNA remains uncleaved in the structures from crystals grown in the presence of either divalent cation. The structure of the dimer of SgrAI is similar to those of Cfr10I, Bse634I and NgoMIV, however no tetrameric structure of SgrAI is observed. DNA contacts to the central CCGG base pairs of the SgrAI canonical target sequence (CR|CCGGYG, | marks the site of cleavage) are found to be very similar to those in the NgoMIV/DNA structure (target sequence G|CCGGC). Specificity at the degenerate YR base pairs of the SgrAI sequence may occur via indirect readout using DNA distortion. Recognition of the outer GC base pairs occurs through a single contact to the G from an arginine side chain located in a region unique to SgrAI.

Efficient and specific internal cleavage of a retroviral palindromic DNA sequence by tetrameric HIV-1 integrase

Background

HIV-1 integrase (IN) catalyses the retroviral integration process, removing two nucleotides from each long terminal repeat and inserting the processed viral DNA into the target DNA. It is widely assumed that the strand transfer step has no sequence specificity. However, recently, it has been reported by several groups that integration sites display a preference for palindromic sequences, suggesting that a symmetry in the target DNA may stabilise the tetrameric organisation of IN in the synaptic complex.

Methodology/Principal Findings

We assessed the ability of several palindrome-containing sequences to organise tetrameric IN and investigated the ability of IN to catalyse DNA cleavage at internal positions. Only one palindromic sequence was successfully cleaved by IN. Interestingly, this symmetrical sequence corresponded to the 2-LTR junction of retroviral DNA circles—a palindrome similar but not identical to the consensus sequence found at integration sites. This reaction depended strictly on the cognate retroviral sequence of IN and required a full-length wild-type IN. Furthermore, the oligomeric state of IN responsible for this cleavage differed from that involved in the 3′-processing reaction. Palindromic cleavage strictly required the tetrameric form, whereas 3′-processing was efficiently catalysed by a dimer.

Conclusions/Significance

Our findings suggest that the restriction-like cleavage of palindromic sequences may be a general physiological activity of retroviral INs and that IN tetramerisation is strongly favoured by DNA symmetry, either at the target site for the concerted integration or when the DNA contains the 2-LTR junction in the case of the palindromic internal cleavage.

The structural basis of DNA flexibility

Although the average physico-chemical properties of a long DNA molecule may approximate to those of a thin isotropic homogeneous rod, DNA behaves more locally as an anisotropic heterogeneous rod. This bending anisotropy is sequence dependent and to a first approximation reflects both the geometry and stability of individual base steps. The biological manipulation and packaging of the molecule often depend crucially on local variations in both bending and torsional flexibility. However, whereas the probability of DNA untwisting can be strongly correlated with a high bending flexibility, DNA bending, especially when the molecule is tightly wrapped on a protein surface, may be energetically favoured by a less flexible sequence whose preferred configuration conforms more closely to that of the complementary protein surface. In the latter situation the lower bending flexibility may be more than compensated for on binding by a reduced required deformation energy relative to a fully isotropic DNA molecule.

Expansion during PCR of short single-stranded DNA fragments carrying nonselfcomplementary dinucleotide or trinucleotide repeats

We performed PCR of many DNA fragments of 6-32 nucleotides in length. Some of the fragments expanded into kilobase lengths even in the absence of the complementary strand. The dramatic expansion was observed for (CA)8, (TG)8, (CA)4, (CA)6, (CA)12, (TG)4, (TG)6, (TG)12, (TC)10, (GA)10 and other single strands. Similar expansions were exhibited by related trinucleotide repeats (TTG)5, (CAA)5, (TGG)5, and (CCA)5 as well. However even small perturbations of the strict repetitive nature of the DNA primary structure substantially reduced the expansions. The expansion products had properties characteristic for normal Watson-Crick duplexes. Hence either the Taq polymerase and/or other components of the PCR buffer promote homoduplex formation of the nonselfcomplementary fragments, which is necessary to prime the synthesis of the complementary DNA strand, or the Taq polymerase is able to copy the single-stranded DNA template without any priming effect. The present observations have implications for the evolution of genomic DNA, microsatellite length polymorphism as well as the pathological expansions of trinucleotide repeats in the human genome.

Free energy and structural pathways of base flipping in a DNA GCGC containing sequence

Structural distortions of DNA are essential for its biological function due to the genetic information of DNA not being physically accessible in the duplex state. Base flipping is one of the simplest structural distortions of DNA and may represent an initial event in strand separation required to access the genetic code. Flipping is also utilized by DNA-modifying and repair enzymes to access specific bases. It is typically thought that base flipping (or base-pair opening) occurs via the major groove whereas minor groove flipping is only possible when mediated by DNA-binding proteins. Here, umbrella sampling with a novel center-of-mass pseudodihedral reaction coordinate was used to calculate the individual potentials of mean force (PMF) for flipping of the Watson-Crick (WC) paired C and G bases in the CCATGCGCTGAC DNA dodecamer. The novel reaction coordinate allowed explicit investigation of the complete flipping process via both the minor and major groove pathways. The minor and major groove barriers to flipping are similar for C base flipping while the major groove barrier is slightly lower for G base flipping. Minor groove flipping requires distortion of the WC partner while the flipping base pulls away from its partner during major groove flipping. The flipped states are represented by relatively flat free energy surfaces, with a small, local minimum observed for the flipped G base. Conserved patterns of phosphodiester backbone dihedral distortions during flipping indicate their essential role in the flipping process. During flipping, the target base tracks along the respective grooves, leading to hydrogen-bonding interactions with neighboring base-pairs. Such hydrogen-bonding interactions with the neighboring sequence suggest a novel mechanism of sequence dependence in DNA dynamics.

Importance of the conserved CA dinucleotide at Mu termini

The dinucleotide CA found at the termini of transposable phage Mu also occurs at the termini of a large class of transposable elements, including HIV, all retroviruses and many retrotransposons. In order to understand the importance of this sequence conservation, the activity of all 16 dinucleotide permutations of the termini was first examined using a sensitive plasmid-based in vivo transposition assay. The reactivity of these substrates varied over several orders of magnitude in vivo, with substitutions at the A position being more severely impaired than those at the C position. The same general hierarchy of reactivity was observed in vitro using mutant oligonucleotide substrates. These experiments revealed that CA was not important for the chemistry of strand transfer, and that the block in the activity of the mutant substrates was at the stage of assembly of a stable transpososome. Given that DNA at the Mu-host junctions is melted/distorted concomitantly with transpososome assembly, we consider the hypothesis that the CA dinucleotide has been selected at transposon termini primarily for its significant conformational mobility.

Solution structure of the dodecamer d-(CATGGGCC-CATG)2 is B-DNA. Experimental and molecular dynamics study

The DNA duplex d-(CATGGGCCCATG)2 has been studied in solution by FTIR, NMR and CD. The experimental approaches have been complemented by series of large-scale unrestrained molecular dynamics simulation with explicit inclusion of solvent and counterions. Typical proton-proton distances extracted from the NMR spectra and the CD spectra are completely in agreement with slightly modified B-DNA. By molecular dynamics simulation, starting from A-type sugar pucker, a spontaneous repuckering to B-type sugar pucker was observed. Both experimental and theoretical approaches suggest for the dodecamer d-(CATGGGCCCATG)2 under solution conditions puckering of all 2'-deoxyribose residues in the south conformation (mostly C2'-endo) and can exclude significant population of sugars in the north conformation (C3'-endo). NMR, FTIR and CD data are in agreement with a B-form of the dodecamer in solution. Furthermore, the duplex shows a cooperative B-A transition in solution induced by addition of trifluorethanol. This contrasts a recently published crystal structure of the same oligonucleotide found as an intermediate between B- and A-DNA where 23 out of 24 sugar residues were reported to adopt the north (N-type) conformation (C3'-endo) like in A-DNA (Ng, H. L., Kopka, M. L. and Dickerson, R. E., Proc. Natl. Acad. Sci. U S A 97, 2035-2039 (2000)). The simulated structures resemble standard B-DNA. They nevertheless show a moderate shift towards A-type stacking similar to that seen in the crystal, despite the striking difference in sugar puckers between the MD and X-ray structures. This is in agreement with preceding MD reports noticing special stacking features of G-tracts exhibiting a tendency towards the A-type stacking supported by the CD spectra also reflecting the G-tract stacking. MD simulations reveal several noticeable local conformational variations, such as redistribution of helical twist and base pair roll between the central GpC steps and the adjacent G-tract segments, as well as a substantial helical twist variability in the CpA(TpG) steps combined with a large positive base pair roll. These local variations are rather different from those seen in the crystal.

Crosslinking of the complementary strands of DNA by UV light: dependence on the oligonucleotide composition of the UV irradiated DNA

UV light crosslinks the complementary strands of DNA. The interstrand crosslinks may contribute to the biological and pathological effects that UV irradiation is known to bring about. Here alkaline agarose gel electrophoresis was used to assess the crosslinked fraction of 31 selected restriction fragments of six viral and plasmid DNA molecules exposed to UVC light irradiation. As many as 17 independent experiments were performed with the particular DNA fragments to get sufficiently precise data suitable for quantitative analyses. The data were used to determine how the crosslinked fraction depended on the dinucleotide, trinucleotide and tetranucleotide contents of the irradiated DNA fragments. This analysis demonstrated that DNA conformation and/or flexibility, rather than the local double helix thermostability, governed the phenomenon of crosslinking. For example, (GA).(TC) suppressed the crosslink formation in DNA more than any dinucleotide composed of only G and C. In addition, (CTAG).(CTAG) promoted crosslinking much more than any other tetranucleotide, including e.g. (TATA).(TATA), whereas the closely related (CATG).(CATG) belonged among the tetranucleotides that most suppressed the UV light induced crosslinks between the complementary strands of DNA. The present data reproduced crosslinking of the analyzed 31 restriction fragments with a correlation coefficient exceeding 0.90. This result will be useful to predict crosslinking along the whole human genome.