APPENDIX B - General Discussion

Al and I would like to begin the general discussion by first focusing in on the material already presented, and then expanding this to include possible mathematical physics approaches to understand the nature of ‘discrete breathers’ in DNA.

With this in mind, we begin by reviewing the concepts proposed to understand how simple intercalators bind to DNA.

These involve the following steps:

  1. Simple intercalators (i.e., ethidium) begin by "pinning" premeltons that exist with highest probability at particular DNA regions. These premeltons correspond to the micrococcal nuclease hypersensitive sites in naked DNA molecules ("NUCLEATION").
  2. As the drug/DNA binding ratio increases, ethidium molecules continue to bind to DNA within these premeltons, sequentially “pinning” expanding regions of the high energy beta-DNA form. These regions remain surrounded by kink-antikink phase boundaries, which continue to move apart as complex formation proceeds ("PROPAGATION").
  3. At saturation, almost all the DNA is complexed with ethidium, DNA being held in organized domains (each, perhaps, being as large as a couple of hundred base pairs) that contain an elongated and unwound high-energy beta-DNA form complexed to ethidium.
  4. This sequence of events describes a structural phase transition, in which the ethidium: beta-DNA complex emerges as the dominant phase.

A similar mechanism can explain how simple intercalators bind to double-stranded RNA.

The model provides an explanation for the Langridge-Lippard fiber diffraction data of the terpyridine platinum (II) organometallo-intercalator: DNA complex, observed at high drug/DNA binding ratios (this platinum organometallo-intercalator is another example of a simple intercalator). The sharpness of the diffraction pattern in this study (the 10.2 Angstrom and the higher order 5.1 Angstrom periodicities reflect primarily the platinum-platinum scattering vectors present between “nearest-neighbor” platinum organometallo-intercalators in DNA) indicates the complex to have a microcrystalline domain size of between one to two hundred base pairs.

Leonard Lerman asks, “The binding of ethidium to DNA is known to give a hyperbolic, rather than a sigmoidal binding curve. This has been interpreted to mean a process that involves the random insertion of ethidium into DNA, rather than an organized process such as the one you describe. How do you explain these experimental observations?”

Henry Sobell replies, “A key issue is the number of nucleation sites that initialize ethidium-DNA binding. If ethidium begins by binding to a large number of different nucleation sites (i.e., the micrococcal nuclease sensitive sites), this would decrease the cooperativity observed for the binding reaction, and, in so doing, diminish the sigmoidal nature of the binding curve. A careful study of the binding reaction should therefore confirm the presence of a small but significant departure from true hyperbolic binding”.

Alex Rich asks, “A large number of crystal structures containing daunomycin (and its closely related derivative, nogalomycin) complexed to several different self-complementary oligonucleotides have been determined by X-ray crystallogaphy. In every case, these molecules intercalate into a distorted B-DNA like conformation – not to the beta-DNA conformation. How do you explain these observations?”

Henry Sobell replies, “Intercalation by daunomycin and related compounds is atypical in that the aglycone anthracycline chromophore lies at right angles to the long dimension of the DNA base pairs. This orientation is not expected to provide adequate stacking energies to "pin" the premelton.

For this reason, I propose the initial complex to be unstable, leading to a spontaneous rearrangement in which daunomycin remains attached to a lower energy B-like conformation.

The process is reversible. A premelton spontaneously reappears, allowing the complex to undergo rearrangement to its original form. After a breathing motion of sufficient magnitude, daunomycin (or nogalomycin) can then detach from DNA.”

Alex Rich asks, “We have recently observed the B- to Z- junction in an oligonucleotide duplex (stabilized in part by a Z-binding protein) by X-ray crystallography. The junction contains B-DNA juxtaposed to Z-DNA, but separated by one denatured base pair (this involves two extruded bases). There is little or no base pair unstacking at (or around) the junction. This differs from the junction you propose. How do you explain this?”

Henry Sobell replies, “My mechanism describes how long chain poly d (GC) undergoes the B- to Z- transition in the presence of high salt and/or negative superhelicity. This would require the presence of a moveable junction connecting B-DNA with Z-DNA for the phase transition to be able to occur.

I estimate the minimum size of this B- to Z- junction to be in the order of 50 base pairs, considerably longer than the oligonucleotide chain length used in your model study (15 base pairs). This observation, combined with the known ability of ethidium to bind to poly d (GC) under Z- form conditions with extremely high cooperativity to give rise to neighbor exclusion binding at saturation (i.e., clearly indicating the presence of a structural phase transition), continues to support my mechanism to understand the B- to Z- transition.”

Aaron Klug asks, “What is the difference between the alternating B-DNA conformation, and the beta-DNA conformation?

Henry Sobell replies, “The alternating B-DNA conformation contains a base-paired dinucleotide that contains both C3’ endo and C2’ endo sugar puckers in the asymmetric unit. This results in the helix having an alternating pattern of sugar puckering down its polynucleotide backbone. Base pairs remain perpendicular to the helix axis, adjacent dinucleotides being related by a twist of 72 degrees and a translation of 6.8 Angstroms along the helix axis. The alternating B-DNA conformation is proposed to be a stable variant of the B-form.

The beta-DNA conformation contains the beta-structural element as the asymmetric unit of the helix. This is a base-paired dinucleotide that contains both C3’ endo and C2’ endo sugar puckers within the asymmetric unit, neighboring sugar residues being linked: C3’ endo (3’-5’) C2’ endo.

In its lowest energy form, base pairs within the beta-structural element are inclined 40 degrees to one another (“roll”), and 33 degrees to a plane perpendicular to the helix axis (“tilt”). Neighboring beta-structural elements are twisted 68.5 degrees and translated 5.9 Angstroms along the helix axis.

In its highest energy form, base pairs within the beta-structural element are inclined 10 degrees to one another (“roll”), and 5 degrees to the helix axis (“tilt”). Neighboring beta-structural elements are related by a twist of 46 degrees and a translation of 9.8 Angstroms along the helix axis. This highest energy form is stretched approximately one and one-half times and unwound -26 degrees every other base pair, relative to B-DNA.

The beta-DNA conformation is both metastable and hyperflexible and is, therefore, distinctly different from A-, B- or Z- DNA. Each of the latter structures is stable at well defined thermodynamic conditions.”

Michel Peyrard asks,” I have looked over your website, and like the general features of your model very much. My only hesitation is that “discreteness” may interfere with the movement of kink and antikink. How do you propose to test this?”

Henry Sobell replies, “Indirectly, by doing a combination of both experiments 1 and 2, in which both ethidium and irehdiamine serve as competitive inhibitors of the micrococcal nuclease enzyme. This will answer the question about the size of the premelton (estimated to be about 50 base pairs), and whether the central beta-structural element alternates between its lowest and highest energy forms. Additional information will have to come from studies with the pancreatic DNase; however, for reasons of brevity, I did not mention this (see however, the Experimental Predictions section in my website for further information).”

Michel Peyrard comments, “I also enjoyed learning about your model to understand the (conceptual) physics of the attachment by the RNA polymerase to DNA, a model you appropriately call the “zip-lock attachment model”. I have also been thinking along these lines for many years (i.e., the RNA polymerase – promoter interaction being a structural phase transition), but have yet to pursue this in any detail.”

We would like to encourage additional questions or comments from the audience, particularly from those interested to discuss possible mathematical approaches to understand the nature of ‘discrete breathers” in DNA.

Thanks again for your attention.

Henry Sobell

Alwyn Scott