I know that parallel DNA helices exist and are governed by Hoogsten base pairing, but why can’t they be possible with Watson-Crick pairing? In the diagram below, if we were to flip one of the strands while keeping the other the same, it appears as though hydrogen bonding is still possible.

The only specific suggestions that I could find was because of the DNA replication process and the negative polarity of hydoxyl group on the phosphates. Moreover, after flipping one strand, the DNA nucleotides form enantiomers. Are these possible reasons, or are there others?

Anti-parallel and parallel DNA base-pairing

  • 1
    $\begingroup$ this question should cover it : biology.stackexchange.com/questions/27839/… $\endgroup$
    – gilleain
    Commented Feb 16, 2018 at 8:49
  • 1
    $\begingroup$ I have not previously reviewed that answer and many others but the questions stated above still remains. I would appreciate if someone could objectively answer them $\endgroup$ Commented Feb 16, 2018 at 10:03
  • $\begingroup$ Related: biology.stackexchange.com/questions/58156/… $\endgroup$
    – canadianer
    Commented Feb 16, 2018 at 23:17
  • $\begingroup$ @canadianer — I have changed the title again. Your revision ("Why isn't Watson-Crick base pairing used in parallel DNA helices?") had the falsse implication (I am sure unintended) that parallel DNA helices existed in the sense that they were reasonably widespread, or that those that had been engineered had complementary base-pairing, but of the non-WC type. My revision retains the emphasis on the lack of WC base pairing in parallel helices, but removes these implications, for the most part. Trust this is ok. $\endgroup$
    – David
    Commented Feb 17, 2018 at 10:21
  • $\begingroup$ @David Yes, sounds good to me. $\endgroup$
    – canadianer
    Commented Feb 17, 2018 at 17:57

1 Answer 1


“The only specific suggestions that I could find was because of the DNA replication process and…”

No. The explanation can have nothing to do with DNA replication. If the structure does not exist, you can’t replicate it, if it does, Nature will evolve a mechanism. (The related SE question, mentioned by @Gilleain, asked whether it could still replicate if it were parallel, i.e. using the enzymes that have evolved for parallel DNA.)

I know that parallel DNA helices exist…

Let us clarify this first. Perhaps the most extensive parallel duplex DNA, the structure of which has been determined, is that described by Parvathy et al.. The two parallel strands of this are shown below:

Parallel DNA sequence

The following points should be noted:

  1. This parallel DNA (and the shorter examples that preceded it) is not a pure stretch of complementary base pairs.
  2. It is stabilized by what the authors refer to as “CC+ clamps” at either end. One is left to conclude that without these the duplex would not form.
  3. All the complementary base-pairs are of the type AT (actually reverse Watson–Crick base pairs). Presumably GC base-pairs would have destablized the structure.

(You can inspect this structure in three-dimensions here. Choose ‘Licorice’ style and colour ‘by chain’ and notice the non-planarity of the three base ‘pairs’ at each end.)

So although the question refers specifically to parallel DNA helices with Watson–Crick base pairs, it should be recognized that extended parallel DNA helices composed of any kind of complementary AT and GC base pairs are not found, and the question applies equally well to them.

“In the diagram if we were to flip one of the strands while keeping the other same, hydrogen bonding is still possible”

The diagram in the question is two-dimensional; DNA is three-dimensional. It is only by considering the three-dimensional structure of DNA can you approach this question.

So how would one do that? One must consider the free energy of alternative structures in the relevant millieu to determine which will occur (i.e. be more thermodynamically stable).

  1. This will tell you whether single DNA strands with parallel sequences will form a double-stranded (ds) structure or not.

  2. This will tell you whether ds-parallel DNA is more or less energetically stable than a ds-antiparallel DNA. Hence, even if both can form (which I doubt, without some special circumstances*) the lower thermodynamic free energy of the anti-parallel dsDNA would give organisms adopting it an evolutionary advantage.

And the answer to the question?

It seems unlikely that one single factor is responsible or it would have been pointed out in elementary text books such as Berg et al..

To answer would require a complete theoretical analysis of the structure or structures. First one would have to build a model of a proposed parallel structure that could accommodate Watson–Crick base pairs. This in itself is a problem because there are likely to be many alternative structures. Perhaps there are computer programs that can find the structure with the lowest energy. This would be calculated in the classic manner, calculating the positive contribution of hydrogen bonding (which depends on distance and angle), ionic interaction etc.* against the negative contribution of charge and steric repulsions.

*Etc? The two-dimensional diagram fails to consider the contribution of base stacking (how could it?), which contributes to a considerable extent to the stability of nucleic acid helices, as the original cover design of Stryer’s Biochemistry is a constant reminder:

Cover of Stryer

  • $\begingroup$ Could you please explain what you mean by 'CC+ clamps on either end”? And what i meant was that the hydroxyl group, Now represented by a negative Oxygen molecule in the diagram attached to the phosphorus on either end are negatively charged, which might cause them to repel and therefore to some degree constitute the formation of Antiparallel DNA $\endgroup$ Commented Feb 16, 2018 at 16:42
  • $\begingroup$ do you think the parallel dna model would continue to form hydrogen bonds or would this be disrupted, moreover how would one calculate the free energy associated with the two dna models? $\endgroup$ Commented Feb 16, 2018 at 16:48
  • $\begingroup$ @AnamikaGhosh — 1. I have expanded my footnote to explain the CC+ clamp, which is in the paper I cited. 2. I have removed my reference to your remark about the phosphates as at is peripheral to the main argument and not worth getting bogged down in. 3. I will explain about calculating energies in a further footnote sometime this weekend. If you are impatient think about this section of Berg et al. $\endgroup$
    – David
    Commented Feb 16, 2018 at 22:01
  • $\begingroup$ Thank you, it makes a lot more sense to me now. $\endgroup$ Commented Feb 19, 2018 at 16:37
  • $\begingroup$ Glad it was of use. I learned something too. I have now added a link to a 3D view of the structure at the Protein Data Bank. $\endgroup$
    – David
    Commented Feb 19, 2018 at 18:35

Not the answer you're looking for? Browse other questions tagged .