Nucleic Acid Architecture
Tomado de http://www.chembio.uoguelph.ca/educmat/chm730/h730.htm


Nucleic acid structure


The double helical structure of DNA, as first proposed by Watson and Crick (Nature, 1953), was derived on the basis of X-ray diffraction from DNA fibers and i
ntelligent model building based on the known covalent structure of DNA.

From solution properties such as viscosity, it was known that DNA molecules were highly elongated. When a thread of concentrated DNA solution is pulled into a fiber, the elongated DNA molecules become aligned along the fiber axis in a sufficiently orderly way to produce a diffraction pattern.

The majority of the reflections were distributed in an X-pattern. It was known on theoretical grounds that a helix produces this X-distribution. The symmetry axes of the X are labelled the equator (transverse to the helix axis) and the meridian (parallel to the helix axis), and a pure helix produces no reflections on the meridian. The spacing of the first layer line from the equator allows the pitch of the helix to be calulated. If the helix is not a pure smooth object, but is divided into regularly repeating components, with multiple components per turn of the helix, this will produce a characteristic meridional reflection. If we simply treat the repeating components as a series of stacked undifferentiated planes, the Bragg law can be used to predict the spacing of reflections. The first layer line for the plane diffraction will be widely spaced, because the planes are close together, and the two are reciprocally related. This produces a strong reflection on the meridian. When the helix X-pattern and the planar pattern are combined, the ratio of plane spacing to helical pitch determines which layer line the meridional reflection falls on. For B-DNA, the meridional reflection falls on the 10th layer line, so there are 10 base pairs per turn of the helix.

helical diffraction

  The pattern is slightly more complex when there's a non-integral repeat per turn of helix, as for the peptide -helix with 3.6 amino acids per turn.

There are additional details in the B-DNA diffraction pattern due to the substructure of the deoxyribose-phosphate backbone, which are somewhat difficult to interpret given the low resolution for fiber diffraction. Those who were attempting to solve the pattern from first principles were making little progress, and the structure of DNA was determined more from the model building than from rigorous calculation. The X distribution of the diffraction pattern indicates 2-fold symmetry, hence an antiparallel pair of strands was used in the model.

twofold symmetry = antiparallel


The layer-line spacing indicates 34 per turn of the helix. The meridional reflection on the 10th layer line indicates 10 repeating units per turn of the helix, a separation of 3.4 along the helix axis.Pauling developed a plausible model fitting these constraints, based on a ribose phosphate core, and individual bases facing outwards, much as amino acid side chains face out from the core of the α-helix.

Watson and Crick made a crucial decision to place the ribose-phosphate backbone on the outside, based on the mutual repulsion of the negative charges of the phosphates. This meant that the bases had to face inwards, and pack together. The spacing of 3.4 represents the thickness of the van der Waals surface of a base, so bases could be arranged by stacking along the helix axis. However, there were problems with the contact between bases on opposite strands.

A casual conversation with a nucleotide chemist revealed that the contemporary textbook structures of guanine and thymine were incorrectly represented in the enol form; the equilibrium unexpectedly favours the keto forms by a factor of about 105.


keto enol tautomerism


With the corrected structures of the bases, Watson and Crick determined that bases could associate by complimentary hydrogen bonding such that adenine pairs with thymine in the opposite strand, and guanine pairs with cytosine. These base pairs allowed the DNA model structure to account for Chargaff's rule, determined from base analysis of DNA, that the content of A was equal to T, and content of G was equal to C.

watson crick basepairs

From model building, it was established that only right handed helices were compatible with the geometry of D-ribofuranose ring structures. More importantly, the Watson-Crick DNA model demonstrated how one strand could act as template to determine the sequence of the other strand, and hence how sequences could be replicated and inherited.
It was the potential for explaining biological function of DNA that led to the widespread acceptance of the Watson-Crick model rather than any compelling structural evidence. DNA structure was not rigorously determined by X-ray crystallography until the late 1970's.


double stranded DNA


Conformational considerations

The backbone of RNA and DNA consists of the alternating phosphate-ribose (or 2'-deoxyribose) chain. Conformational variation arises from restricted bond rotations within the sugar ring, giving rise to different ribose ring pucker, and torsional angles at bonds connecting phosphate to ribose.

Ring pucker affects the torsional relationships of adjacent C-H bonds in the sugar ring, and can be determined from spin coupling in NMR. Ring pucker also affects non-bonded H-H separations, which can be measured by NOESY.

Ribose ring pucker arises because a flat pentagon puts all ring C atoms into eclipsed conformations, despite the fact that the pentagon angle of 108o is very close to the C tetrahedral angle of 109.5o. Puckering relieves the steric conflict due to eclipsed groups, at the expense of a slight deviation from tetrahedral bond angle. The pucker is described in terms of the pentose ring atom that is displaced furthest out of plane. When an adjacent atom is slightly displaced to the opposite side, this gives the 'S' or 'N' pattern twists, according to the orientation of the zigzag.

When only a single atom is displaced, the pucker is decribed as an envelope pucker. In the first four views shown, C-4', O-4' and C-1' lie in the plane and C-2' and C-3' are displaced. Endo- pucker has the major displacement on the -face (same side as C-5' and the base N); exo- pucker goes to the opposite side of the ring. The fifth example has C-1' to C-4' in plane, and O-4' diplaced to the endo side

ribose ring pucker

The lowest energy conformations are 2'-endo and 3'-endo. B-DNA, the form described by Watson and Crick, adopts 2'-endo (example at right), whereas RNA and the A form of DNA adopt the 3'-endo twist. 2'-endo places the 1'- bond to the base at a shallow angle and the bond to 5'-C at a steeper angle relative to the plane of the pentose. 3'-endo shows the opposite angular relationships. The inset shows that if O-4' is displaced upwards slightly, C-2', C-3', C-4' and O-4' become coplanar, and the resulting pucker is 1'-exo

B-DNA 2'-endo pucker

Puckered forms interconvert by moving one atom at a time as demonstrated above. 2'-endo and 3'-endo interconvert via the sequence: 2'-endo - 1'-exo - O-4'-endo - C-4'-exo - 3'-endo with an energy barrier of only 2 kJ/mol, which is less than the barrier to rotation about single C-C bonds. 2'-exo and 3'-exo are only a few kilojoules less stable, and are also accessible states occurring in kinked or bent DNA.

Phosphodiester torsion angles

Torsion at the phosphodiester bonds are designated by dihedral angles α (bond to 5'-O) and ζ (bond to previous 3'-O).

DNA backbone torsion

Phosphate is tetrahedral like carbon, so favoured conformations occur at g (gauche, 60°) or t (extended, 180°). However unlike neutral C-C bonds, the bonded polar O atoms are biased towards the gauche states by dipolar interactions. The inset shows how a lone pair in the gauche O atom aligns with one of the other P-O bonds (the gauche effect). Polynucleotide geometry is most consistent with g (gauche, close to 60°).

gauche effect at phosphate

To be complete, there are actually four additional torsional angles for the C-O and C-C bonds of the pentose that link the phosphodiester. This provides so much flexibility that there is no inherent tendency to particular conformations as is the case for peptide chains.

In RNA (lower figure) or A-DNA, the 3'-endo pucker rotates the phosphate groups up and closer to the bases in the helix core than in B-DNA (upper figure). This movement repositions the phosphate away from steric interference with 2'-O in RNA. The inset shows the conflict between 2'-O and phosphate when the ribose is in the 2'-endo pucker state typical of B-DNA.

rna ribose phosphodiester organization

Base Orientation

Rotation about the C-1' bond allows two orientations: syn and anti base orientations


Free purine nucleosides, in particular guanosine, favour the syn- orientation, but adopt the common anti- orientation within most DNA and RNA helices.

Pyrimidines adopt anti- orientation almost exclusively, because of steric interference between O-2 and C-5' in the syn- orientation, shown here for 2'-endo pucker. Interference is even more severe in 3'-endo-.

Base Pairing

The Watson-Crick DNA structure model was founded on the idea that adenine pairs to thymine by hydrogen bonding, and guanine to cytosine. The cyclic H-bonding pattern may account for some of the stability. However, these are not the only base pairs that form, nor is the AT pair even among the most stable.

cyclic H-bonding

Just about any base can hydrogen bond to any other base, including self pairs. Although the G-C is undoubtedly the most stable, several mispairs are stronger than the A-T pair.

A-A hydrogen bonding pair











The table shows the interaction energies for base pairs, including H-bonding and van der Waals forces, calculated in vacuo. The van der Waals contribution is about -3 kJ/mol in each case.

Base Pair
Mode
Energy
kJ/mol
G-C (wc) -69.9
A-T (wc) -29.3
A-T (rwc) -28.9
A-U (wc) -30.1
A-U (h) -27.6
A-U (rh) -28.4
A-A (wch) -23.4
U-U (rwc) -22.6
G-G (rwc) -67.0
C-C (rwc) -44.8
A-G (wc) -39.3
G-U (swc) -32.2

Base pairing is complicated by the fact that the purines possess two H-bonding faces, the Watson-Crick face, involving ring positions 1 and 6 for A and 1,2 and 6 for G, and the Hoogsteen face involving ring positions 6 and 7.


two faces of adenine

In the table of interaction energies, the pairing mode is indicated as wc where pairing occurs on the Watson-Crick face in the normal orientation, even for the mispair A-G. Reversed Watson-Crick pairing (rwc) forms on the Watson-Crick face, with one member of the pair inverted.

Hoogsteen AU base pairreversed Hoogsteen AU base pairreversed Watson Crick AU base pair

Base pairs can form on the Hoogsteen face either with the pyrimidine in normal orientation (mode h) or reversed (mode rh). Finally, the G-U pair forms hydrogen bonds across the normal Watson face in the normal orientation. However the U has slipped or sheared (swc) away from the normal location of C; this is also known as a wobble pair.

Wobble pair

Even more base pairings, including pairs with 3 H-bonds, can be devised if ionized or enol variants of the bases are considered. The unique character of the Watson-Crick A-T and C-G base pairs lies not in the hydrogen bonding, but in their shared geometry. Common base pair geometry

This means that sequences can mix A-T, T-A, G-C and C-G pairs in any combination with minimum perturbation of structural order of the helix backbone.

Sequences containing only these pairs result in the lowest energy state for the DNA helix, primarily by avoiding destabilizing backbone distortion. This means that when double stranded DNA is incubated at high temperatures, only complementary DNA can maintain its structure, and any destabilizing mispair will promote unravelling of the duplex.

Other combinations are two wide (purine-purine) or too narrow (pyrimidine-pyrimidine) or have wrong orientation to the helix backbone (A-C, G-U). It's primarily the function of DNA polymerase to maintain the precise complementarity by its replication fidelity. When single stranded DNA or RNA are incubated well below the critical temparature, duplex that includes or bypasses mispairs can form quite readily.



When difluorotoluene deoxyribosyl triphosphate is used as an analog of dTTP, DNA polymerase incorporates this efficiently opposite dA in the template, and incorporates dA specifically when the analog is in the template, despite the total lack of H-bonding (Moran and Kool, 1998). This demonstrates the importance of stereochemical fit rather than hydrogen bonding per se in matching up base pairs in duplex DNA.

van der Waals base pairing

Base pairs on the Hogsteen face or with reversed orientations are such radical misfits that they need not be considered in normal duplex. However, Hoogsteen pairing is important for triplex and quadruplex structures.

Base Stacking

In aqueous solution H-bond formation must be balanced against desolvation, and H-bonding alone does not provide sufficient energy to make really stable helical duplex. Another interaction is needed to account for the full stability of the helix, and this is provided by stacking. As discussed earlier, the stacking interaction is primarily a dipole effect, aligning partial positively charged atoms at the edges of the bases with the pi electron system at the core. Bases show stronger stacking than ordinary aromatics because the N-H and C=O groups are more polarized and create better positive centers (at the H or C atom respectively) than plain hydrocarbon C-H.

stacked A-T pairs in DNA

The stacking interaction depend on which base pairs overly each other, and is classified according to the sequence step of two consecutive base pairs. The overlap occurring when G-C overlays C-G is not the same as C-G overlaid on G-C, and different values are determined. The table shown is for the base pair overlaps in B-DNA. Different helical arrangements will change the degree of overlap, so different values will be seen for RNA or A-DNA etc.

Step
Energy
kJ/mol
C-G
G-C
-61.6
C-G
A-T
or
T-A
G-C
-44.0
C-G
T-A
or
A-T
G-C
-41.1
G-C
C-G
-40.6
G-C
G-C
or
C-G
C-G
-34.6
T-A
A-T
-27.5
G-C
T-A
or
A-T
C-G
-27.5
G-C
A-T
or
T-A
C-G
-28.4
A-T
A-T
or
T-A
T-A
-22.5
A-T
T-A
-16.0

These values are calculated in vacuo, and will be partly counteracted by desolvation. Net free energy change for addition of a G-C base pair to a helix in aqueous solution averages to -15 kJ/mol, and about -6 kJ/mol for A-T pairs.

stacking of phosphate onto base

 

 




The charge distribution in the base ring also allows stacking interaction between base and phosphate.

This is not a factor in regular helix, but plays a role at helix boundaries, mispaired bubbles and bulges, and at terminal loops of hairpins. The example is taken from the TΨC loop of tRNA, and shows the phosphate group of G57 stacking onto U55, to terminate the helical stack at the end of the loop.

yeast initiator tRNA fMetU-phosphate stacking in tRNA


The organization of tRNA as two stacks joined at right angles is made apparent by stripping away the ribose phosphate backbone to reveal the bases alone.

The upper part of this image shows how the acceptor stem stacks smoothly into the TΨC arm with little discontinuity in the stack at the junction of sequences. U55 is at the end of this stack and is capped off by the phosphate group of G57. The inset shows a view along the axis of the acceptor stem.

stacking in tRNA

B-DNA

B-DNA is the most stable helical form adopted by random sequence DNA under physiological condtions (PDB 1BNA).

The helix averages 10 bp per turn of the helix, giving 36° twist and 3.4 rise per base pair (bp), although local sequence dependent variations range from 9 to 11 bp per turn. Bases are almost (within 1o) of being normal to the helix axis. Helix diameter is 24 .

B-DNA

The major groove is wide and shallow, and exposes the H-bonding faces, including the Hoogsteen face of the purines. The phosphate groups are turned outwards by the pucker (nominally 2'-endo, but with significant population of 1'-exo and O-4'-endo. This leaves the groove wide enough to accomodate a protein α-helix or -hairpin. The minor groove is narrow, and occupied by a spine of bound H2O and cations.

The view down the helix axis shows the base pairs (shown with thick bonds) are clustered along the central axis, and the backbone is at the periphery (thin bonds).

B-DNA viewed down helix axis

A-DNA and RNA

DNA adopts the A-conformation under dehydrating conditions (PDB 1ANA). This is also the stable conformation for RNA (PDB 1SDR)because it avoids the steric interference of 2'-O and phosphate. DNA-RNA hybrids also adopt a structure close to A-DNA.

A-form RNA

The helix averages 11 bp per turn of the helix or 32.5° per bp. In A-DNA, bases are tilted about 20° from the normal, reducing the rise to 2.6 per base pair. RNA bases are tilted about 15° from the normal, giving a rise of 2.8 per base pair.

The major groove is deep and narrow, and the phosphate groups are turned inwards by the 3'-endo pucker. This virtually excludes polypeptide access to the H-bonding face of the base pairs in the standard A-form. The A-form major groove is likely to be occupied and obstructed by metal cations which are tightly bound between the negative phosphates that line the inside edge of the groove like teeth. The minor groove is wide and close to the surface. The view down the helix axis shows the base pairs (shown with thick bonds) are distributed around the periphery of the helix with the minor groove face near the edge.

A-form RNA viewed down helix axis

The different accessibility of the major groove in B-DNA and A-form RNA is more easily seen by tilting the helix to view the groove profile directly. This has considerable bearing on the mechanisms used by DNA and RNA binding proteins to locate specific sequences of nucleotides.

tilted views of BDNA and A-DNA helix

So-called C-DNA is a state found by crystallization at high ionic strength and low relative humidity. It has only 9.3 bp per turn of helix, 3'-exo pucker (close to 2'-endo) and bases are tilted 11o, i.e. in the opposite sense to A-DNA. It is otherwise very similar in appearance to B-DNA.

Sequence dependent DNA structures.

D-DNA

A- and B- DNA structures accomodate arbitrary sequences. However there are a number of additional structures that are sequence specific. D-DNA is formed in sequences that contain alternating A and T, typical of promoter sites. AT rich regions are less stable than GC containing sequences. These sequences form normal B-DNA, but in high ionic strength, low humidity or under the influence of specific binding proteins, can be triggered into the D-DNA form which is overwound with only 8 bp per turn, and with wildly tilted base pairs.



Z-DNA

Z-DNA is formed at high salt concentration from sequences which are GC rich and have strictly alternating purine-pyrimidine pairs, e.g. GCGCATGCGC. The diffraction pattern shows a strong meridional reflection on the twelth layer line, indicating 12 bp per turn of the helix, however a weak but undeniable meridional spot on the sixth layer line at first defied explanation. This implies a repeating unit occurred 6 times per turn of the helix.

The puzzle was solved by Alexander Rich (Rich et al. 1984) who determined that the repeating unit was two consecutive base pairs in different conformations, and that this pattern circumvented the inability of a single uniform deoxyribose phosphate conformation to form left handed helix.

The structure,(e.g. PDB 2ZNA) is a left handed helix, with the purines in syn orientation (strongly favoured by G) and 1'-exo pucker while the pyrimidines are in the normal anti orientation and with 2'-endo pucker.

This results in a backbone which traces a zig-zag path, hence Z-DNA. Since the complementary strand has C opposite G, one strand zigs where the other strand zags. Even though a purine could be placed in the anti step of one strand, this is disallowed, because the complementary pyrimidine in the opposite strand would be forced into the disfavoured syn orientation.

Z-DNA

The helix is very thin and extended in Z-form, with bases being highly exposed to solvent in both major and minor grooves.

The biological relevance of Z-DNA remains controversial. One suggested mechanism is that the Z-state is determined by the action of Z-DNA binding proteins. Introduction or removal of a segment of left handed Z-DNA in a longer B-DNA region would affect the state of DNA winding and supercoiling, and this in turn might affect access for transcription or DNA condensation within the chromosome structure.

DNA Triplex

The wide major groove of B-DNA allows access to the Hoogsteen face of the base pair purines. This allows a third nucleic strand to hydrogen bond by forming base triplets (e.g. Tarkoy et al. 1998).

The original models (e.g. PDB 1D3X) had a run of purines in one strand, since only purines have the extra Hogsteen H-bond acceptor on N-7. The third strand (yellow in the figure) runs parallel to the purine strand (red). This strand echoes the conformation of the purine strand (compare the deoxyribose of A with the deoxyribose of the Hoogsteen paired T below.

triplex DNA

The triplet T-AxT (xT denotes the Hoogsteen paired base) is the most stable triplet; C-GxC+ requires a non-ground-state protonated form of C, although G can also Hoogsteen pair with the modified pyrimidine pseudouridine. However, reversed Hoogsteen pairs T-AxI (I = inosine), T-AxT, T-AxA and C-GxG also allow triplet formation with normal tautomers of the purines in the third strand, which runs antiparallel to the purine strand of the duplex.

T-A-T tripletC-G-C triplet

Triplets may exist naturally in various complex RNA structures such as pseudoknots (to dealt with later). Another application involves making oligonucleotide probes that bind to, but do not disrupt the normal duplex order of the target DNA. Binding of the triplex strand is expected to inhibit expression of target genes or binding of transcription factors to target sequences.

DNA quadruplex and G quartets

Guanine quadruplexes are four stranded structures(e.g. PDB 139D) found naturally as terminating sequences at the ends of eukaryotic chromosomes or telomeres (Wang and Patel, 1993). Telomeres contain repetitive sequences such as G4T4G4, and are generated by the enzyme telomerase which uses an internal RNA template to generate the sequence. Telomeres solve a problem with terminating the lagging strand by the normal replication mechanism, which may otherwise stop short without copying the last nucleotides on the parental 3-'end. Telomeres extend the genomic sequence so that the loss of nucleotides from the 3' end is harmless.

G quadruplex

Runs of 4 Gs in sequence may also be involved in recombination processes and centromere linkage. The four strands of the quadruplex associate through guanine quartets, in which each guanine uses its Watson-Crick face to H-bond to the Hoogsteen face of its neighbour.

G quartet

At the center of each quartet are four inward pointing O-6 atoms of the guanine. Two adjacent quartet layers can trap a metal cation, e.g. K+, in an octahedral cage of O ligands at the center of the quartet

Quadruplex strands may be arranged parallel or antiparallel in several patterns, depending on the connectivity. When strands are antiparallel, the neighouring guanines must be arranged syn- (favoured by guanine) on one strand, and anti- on the other.


quadruplex strand arrangements

Often, the end of a quadruplex strand carries a T. Four thymines meet in a plane, and may associate either by the hydrophobic effect of the converging methyl groups, or by H-bonding .

T quartet, perpendicularGCGC quartet

In some cases, the thymines may align perpendicular to the plane of the quartet, with the methyl groups clustered towards the interior of the quadruplex.

T quartetT quartet, H-bonding

Sequences containing the occasional C can form quartets between diagonally opposed G-C pairs, and normal Watson-Crick H-bonds form within the G-C pair.

Inter-pair H-bonding occurs between the other C -NH2 proton and N-7 of the G. Since these quartets are under-sized, they are preferred at the top or bottom layer of the quadruplex. Two GC quartet layers form a small tetrahedral oxygen cage, so binding favours smaller ions such as Na+.

Four cytosines in a layer can also form a hydrogen bonded quartet similar to the thymine quartet.

C quartet, H-bonding

Keep your terminology straight!
Quadruplex: a four stranded DNA structure.
G-quartet or tetrad: four guanines from four separate strands, hydrogen bonding in one plane of the quadruplex

Intercalated C motifs

Replication of established telomeric DNA sequences creates runs of C in the complementary strand. When the G-rich strand forms quadruplexes, the C-rich strand forms a four-stranded structure known as the intercalated C motif, ICM (Cai et al. 1998).

intercalated C motif
Intercalated C-stacking

Diagonally opposite strands run parallel and make a C-C base pair in reversed Watson-Crick mode (PDB 294D). The middle H-bond implies that the pair is protonated on one of the opposed N-3 atoms. Since the proton-N bond can be delocalized, this should be easier to form than a single CytH+.

reversed Watson Crick C-C pair

When the backbones are simplified, they can be seen to run almost parallel to the central stack. Bases on any one strand are about 6.5 apart, double the usual separation, and this leaves room for intercalation by another base pair from the other diagonal pair of strands.

isolated parallel pair from ICM


In the Tetrahymena ICM sequence, each segment teminates in a pair of adenines (complementary to the TT pair in the G-rich quadruplex). The adenines do not fit in the structure, but stack to each other at the ends of the strand.

backbone of intercalated C motif

The central C-stack contains alternating diagonally opposed pairs of cytosines, stacked so that the amino hydrogens overlay the negative electrons of the pyrimidine ring in the adjacent layer. Strands can be arranged in various ways to generate ICM from continuous sections of DNA. Free ends tend to contain A, and tight connecting loops tend to contain T.

strand arrangements in ICM

Peptide nucleic acid

Peptide nucleic acid, PNA, is an artificial construct in which the pentose-phosphate backbone is replaced by polyaminoethylglycine, a polymer of peptide-like bonds (Nielsen, 1999). The bases are coupled to the glycine N as acetyl derivatives. This was designed to mimic the geometry of the ribose phosphate. Since PNA are not degraded by any known enzyme, they offer unique advantages as probes or antisense expression blockers.

PNA backbone


When the backbone was first designed, carbon chain lengths were varied at the points labelled n1, n2, and n3, so aminopropyl was tested in place of aminoethyl, -alanine in place of glycine, and propionyl instead of acetyl derivatives of the bases. The oligonucleotide analog strands based on these derivatives were tested for ability to hybridize to complementary DNA strands, and the aminoethylglycine structure illustrated gave the highest melting points. Melting points were 8o lower for the aminopropyl, 10o lower for -alanine, and 21o lower for propionyl derivatives of bases.

PNA backbone

The hybrid adopts a loose A-form with the phosphates turned intowards the major groove (PDB 1PDT). However, the helix is very underwound, so the major groove remains wide and accessible.


pna-dna hybrid


The PNA strand is conformationally flexible, and adapts easily to the target. As a result PNA-DNA hybrid is more stable than RNA-DNA, and is readily able displace its target strand from duplex DNA.

pna targets purine rich dna as a triplex

PNA works most effectively as a probe for purine rich sequences. Two strands of complementary pyrimidine containing PNA readily form a triplex, displacing the native pyrimidine strand. The strength of the hybrid is sufficient to tolerate a few mispairs due to the odd pyrimidine in the target sequence. PNA was envisaged as an antisense and antigene reagent or drug. The limiting factor is resistance to uptake by organisms, although techniques are being devised to "smuggle" PNA in by attaching to peptide fragments that are taken up and degraded by the host.

References

L.Cai, L. Chen, S. Raghavan, R. Ratliff, R. Moyzis, A. Rich, Intercalated cytosine motif and novel adenine clusters in the crystal structure of the Tetrahymena telomere. Nucleic Acids Res 26: 4696 (1998)

M. Eriksson, P.E.Nielsen, Solution structure of a peptide nucleic acid-DNA duplex. Nature Struct Biol 3: 410 (1996).

P.E. Nielsen. Peptide Nucleic acids as therapeutic agents. Curr. Opin. Struct. Biol. 9: 353-357 (1999).

H.Robinson, Y.G. Gao, B. S. McCrary, S. P. Edmondson, J. W. Shriver, A. H. Wang,.: The hyperthermophile chromosomal protein Sac7d sharply kinks DNA. Nature 392: 202 (1998)

A.A. Rich, A. Nordheim, A.H. Wang. The chemistry and Biology of Z-DNA. Ann. Rev Biochem. 53: 791-846 (1984)

M. Tarkoy, A.K.Phipps, P. Schultze, J.Feigon, Solution structure of an intramolecular DNA triplex linked by hexakis(ethylene glycol) units: d(AGAGAGAA-(EG)6-TTCTCTCT-(EG)6-TCTCTCTT). Biochemistry 37: 5810 (1998)

Y. Wang, D.J. Patel, Solution structure of a parallel-stranded G-quadruplex DNA. J Mol Biol 234: 1171 (1993)

Reading list

C. Branden and J. Tooze. Introduction to Protein Structure. 2nd Ed.Chapter 7 pp.121-126. Publ. Garland, NY (1999)

D.E. Gilbert and J Feigon Multistranded DNA structures. Current Opin.Struct. Biol 9: 305-314 (1999)

S. Moran and E.T Kool, Efficient replication of a DNA base pair between non-hydrogen bonded nucleoside analogs. Nature Struct Biol. 5: 950-954 (1998).