This is a long one. (TLDR at the bottom)

The prospect of alternative genetic polymers has attracted the interest of people for almost as long as we have known about DNA. Indeed, many alternative polymers have already been found, with comparable properties to DNA (some of which are mentioned here). The one I will focus on today is PNA.

Peptide Nucleic Acid (which really should be called Peptide Nucleic Polymer, since the peptide backbone is not really an acid) has a number of interesting properties when used in conjunction with DNA.

  • It is more chemically stable than the DNA backbone.
  • It forms a more stable duplex with DNA i.e the melting temperature (which is the temperature at which the two strands of the duplex fall apart) for a DNA-DNA duplex is 10°C, while for a PNA-DNA duplex it is 31°C.
  • PNA-DNA duplexes are more intolerant of base mismatches. A mismatched base pair will cause the duplex to become more unstable than a DNA-DNA duplex or even a PNA-PNA duplex (Refer to this article for more details, scroll down to the discussion section for a quick summary).

How is this relevant to us, you say?

Well, imagine a hypothetical cell containing a PNA-DNA duplex as the genetic polymer. This cell would have a number of advantages over more conventional cells.

  • It would have a genetic data storage polymer that is more resistant to oxidative and hydrolysis damage. The PNA backbone of the duplex is more chemically stable and more hydrophobic than the DNA backbone. So in the event of damage to the DNA, the repair enzymes can use the PNA side to reconstitute the DNA side.
  • The heteroduplex is more intolerant of base mismatch, to the extent of structure destabilization, which means that error correction enzymes will have an easier time to get to and correct such mistakes.
  • It gets rid of the weirdness of DNA replication.

The last one requires a bit more explanation. Refer to this webpage, this Wikipedia article and this video for more details, but the short version is that new nucleotides can only be added to the DNA in the 5' to 3' direction. This causes speedy replication of the leading strand, but the lagging strand has to contend with stuff like Okazaki fragments and multiple extra enzymes to be replicated successfully. Not to mention that all the extra time spent unwound from a duplex creates a higher chance of introducing mutations to the strand.

However, a PNA backbone will have no such requirements. The N-(2-aminoethyl)-glycine monomer is the basic building block of the PNA backbone, and it can be extended in any direction with the right enzyme, which can be a ribosome like enzyme, maybe? Since they are quite good at creating peptide polymers, aka proteins, and has the benefit of already existing. Let's call it a PNA polymerase.

So in this genetic polymer replication system,

  1. There will be two distinct replication enzymes. One is PNA polymerase, which will use the DNA strand as a template. The other is DNA polymerase, which will use the PNA strand as a template.
  2. The leading strand will be the PNA backbone, and the DNA polymerase will assemble the DNA backbone like it normally and quickly does (in the 5' to 3' direction).
  3. The lagging strand will be the DNA backbone, and here the PNA polymerase will build the PNA backbone like assembling a protein.

There will be no looping strands, no Okazaki fragments, no extra enzymes to deal with all the weirdness of the lagging strand. I'm guessing this will lead to a major simplification of the replication machinery complex, as well as improving replication speed, efficiency and accuracy.

Sounds wonderful, right? I think so too. But as it happens, I'm not a biochemist and I have the tendency to skip out on crucial details in pursuit of the perfect solution. That's where you guys come in, to give me a (pseudo?) reality check. Does my idea have merit or is it completely fanciful? Any input will be appreciated. Thanks!

TLDR: a hypothetical PNA-DNA duplex as a genetic polymer would be superior to our regular DNA duplex in terms of stability, damage resistance, error correction and replication efficiency.

  • $\begingroup$ It'd be more energetically favorable to make it likely that fossils of ancient organisms carrying this might be sequenced. But I'm not sure what the question is. You obviously seem to think it's viable and your argument stands.... do you have a specific concern? $\endgroup$ Commented Mar 29 at 8:50
  • $\begingroup$ How do you propose to keep the DNA polymerase from building on unwound DNA strands and the PNA polymerase from building on PNA strands? Or worse, one of each trying to work on the same strand at the same time? $\endgroup$
    – Cadence
    Commented Mar 29 at 9:49
  • $\begingroup$ @Escapeddentalpatient. My concern is that I might have missed something major that renders my idea unviable, or something that I can correct or remove. If the idea works, great! $\endgroup$ Commented Mar 29 at 10:59
  • $\begingroup$ @Cadence Answering the second part of your question first, once a polymerase has worked on a strand, another polymerase cannot touch the same strand simply because the strand is now part of a finished duplex. Polymerases need unwound single strands to work. $\endgroup$ Commented Mar 29 at 11:02
  • $\begingroup$ @Cadence To answer the first part of your question, I'd like to point you to how reverse transcriptases used by retrovirii work. The reverse transcriptase itself is a polymerase that transcribes RNA to DNA first, which the host DNA polymerase then uses to synthesize DNA which is spliced onto the genome. $\endgroup$ Commented Mar 29 at 11:06

3 Answers 3


Are you sure you want more stability and better error repair?

At the end of the fair evolution is caused by random mutations which are caused by instability and/or non perfect repair.

Too many mutations are detrimental for the life of the individual, but not enough mutations are detrimental for the life of the species and its genetic code.

In short: maybe DNA is the sweet spot between not having errors and having too many errors.

If nature had wanted a better solution, it would have found it.

  • $\begingroup$ Re: the last point. It's a major error to believe that evolution finds the best solutions, especially for fundamental things like this. Natural selection can only ever move in directions that are locally favoured. Once a molecule is in use, changing it would have far too high a cost for evolution to favour a mutant that moved in the direction of using it. $\endgroup$ Commented Apr 17 at 14:52

If PNA synthesis happened in the cell in the same way peptide synthesis does it would have to be N terminus to C terminus, the reason is explained well here. While PNA can bind parallel or antiparallel, antiparallel is preferred and has the N terminus where the 5' end would be were it DNA. So the direction of replication is the same as in normal DNA in the preferred orientation.

It's already been mentioned that more stability is bad for evolution, but on top of that remember that to replicate or read a double helix you're going to need to spend energy to unwind it. A lot more energy actually, PNA sticks pretty tightly to DNA as you know, so your organism is burning through energy faster without much immediate benefit. It would probably lose out to an organism evolved to use a DNA duplex since its advantages are either longer term or actually disadvantages unless it's in an environment where it's constantly exposed to mutagens, where those tradeoffs would actually make sense.

In principle, those issues can be solved through more energy and time (or being designed, for the evolution issue). A bigger problem is that PNA is poorly soluble in water, and it gets worse as the length of the PNA increases (and as you add more purines). The easiest way to solve the solubility issue is to add charged amino acid side chains at the end, but that doesn't scale up well.

The free energy of DNA mostly comes from the phosphates in the backbone interacting with the water, followed by stacking interactions between bases. What that means is that DNA (and RNA) will try to keep the backbone exposed to the water, which keeps is from aggregating and keeps the shape fairly uniform (so it can be read and copied by the same mechanism regardless of what information it stores). With a backbone that lacks those regular ionic charges, the backbone will start to stick together due to hydrophobicity, so large stretches of PNA will wind up peeling away from DNA and crumpling up due to that.

Adding something hydrophilic to the gamma carbon helps solubility (γ-PNA). L-amino acids work better as precursors than D, and L are what's found in Earth life, which is convenient. Some examples of commercially available PNA derivatives with improved solubility are here, but for something that would be liable to appear in early life, an ethanol group on the gamma carbon can be made from fairly simple precursors and has pretty good solubility (Cγ-CH2-OH).

Some more info on PNA here that might help you with whatever you're writing. And for some fun trivia, there's a version of the RNA world hypothesis that suggests PNA came first and was replaced by RNA.

  • $\begingroup$ Thanks! Your response added a lot more clarity to the situation. Guess I'll have to rework/abandon the idea. $\endgroup$ Commented Apr 18 at 17:45

The constant separation of strands is vital to DNA's function

DNA is not a dead molecule, idly sitting by in the cell until it is time for replication, it is a molecule that is constantly being separated into strands so that it can be read for transcription into RNA. Increasing the difficulty of doing that would actively make it worse at its primary role - directing the behaviour of the cell.

5' to 3' isn't a property of DNA, it's a property of replication enzymes

PNA doesn't solve any problem here, you still need different enzymes and enzyme complexes to build the strands in different directions. That means evolving two parallel copying systems, each with dozens of components, which - given evolution has never favoured it - is probably not a advantage compared to the minor difficulties of Okazaki strands.


You must log in to answer this question.

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