From an information perspective, are both strands of DNA necessary?

Question

I am learning about the genetic code, replication, and transcription, and I have a question about whether or not both strands of DNA are really "necessary".

In replication, at a high level, we are taking 2 strands of DNA and creating 4 of them. Part of my question is: suppose hypothetically that the genome only consisted of a single strand (instead of the double helix), couldn't replication just as well occur by turning 1 strand into 2 (instead of 2 into 4)? I suppose this could be done by taking advantage of complementary base pairing as usual, and then splitting the two strands?

In transcription, I believe only one strand of a gene is read by RNA polymerase, so I don't think think the second strand is "necessary" here.

Can someone shed some light on this for me? I realize this is pretty speculative, and DNA is what it is, and does what it does. but I think this sort of question will allow me to see deeper into tits finer workings.

Answer 1

From a strictly information theoretical perspective, no, a second strand does not provide any additional information that cannot be inferred from the complementary strand. But as many comments have pointed out, there are many practical reasons for DNA's double strandedness.

• DNA is biochemically more stable in a double stranded state.
• The redundancy of two strands allows for maintenance and repair mechanisms to ensure the integrity of the genome.
• Proteins that bind to DNA can only recognize the sequence of that strand, so even though the complementary strand can always be reconstructed from the primary strand, DNA binding proteins cannot target that complementary strand unless it physically exists. In some cases, a DNA binding site encompasses both strands simultaneously.
• Proteins (such as polymerases) traverse DNA in the 5' to 3' direction, so if information from the complementary strand was required it would have to be read "backwards".

Answer 2

Although the title refers only to an “information perspective”, the question itself brings in the processes of replication and transcription. It is not entirely clear whether the poster envisages ‘piggybacking’ on existing DNA- and RNA polymerases which operate on double-stranded DNA (dsDNA) or whether he is envisaging completely different enzymes. I shall consider how the existence of single-stranded DNA and RNA viruses make it possible to argue that, both scenarios are possible in theory.

My answer to the question, therefore, like that of @DanielStandage, is no.

Single-stranded genomes relying on a double-stranded replicative form

Many viruses exist in which the genome packaged in the virion is single-stranded DNA (ssDNA). Replication occurs using either virally-encoded or host-coded DNA polymerases, often going through a double-stranded replicative intermediate (but see next section). Transcription normally occurs using the nuclear host DNA-dependent RNA polymerase, acting on the dsDNA.

In order to have all the information for transcription and translation on a single strand in a non-cryptic form, the individual genes would all have to have the same orientation so that the codons of all the transcribed single-stranded mRNA made sense. There seems no reason why this should not be so, although in practice most DNA genomes of which I am aware have genes in both orientations. However the banana buchy top virus has a genome composed of six individual ssDNAs, each apparently with a single gene.

Single-stranded genomes without a double-stranded replicative form

The rolling circle mechanism of DNA replication involves generation of a single stranded form without a double-stranded replicative intermediate, and, although the most well characterized examples are dsDNA viruses, the ssDNA banana buchy top virus replicates by the same mechanism. If we think of what might be possible in linear ssDNA genomes by extrapolating from what exists in linear ssRNA viruses, it is not hard to envisage replication mechanisms involving the production of many copies of a ssDNA of opposite sense to a single copy of a ssDNA template. Replication is therefore not a problem with ssDNA-dependent ssDNA polymerases analogous to extant viral RNA-dependent RNA polymerases. For transcription it would seem more logical to make a further step of a ssDNA-dependent ssRNA polymerase, rather than ‘mutating’ extant dsDNA-dependent RNA polymerases.

Why bother?

As things don’t operate this way it may seem to be a waste of time discussing them. However:

1. The poster is reassured that his reasoning is — in principle — correct.
2. The poster and reader is made aware of the varieties of viral genomes and replication strategies of which he might previously have been ignorant.
3. The answer provides food for thought in relation to the idea that a world in which cellular organisms had RNA genomes preceded the current one in which they have DNA genomes.

A final word about information

The way in which the question is posed relates to information as a linear sequences of nucleotides in DNA, specifically codons. This can be extended to other signals such as transcriptional start sites, splicing sites, polyadenylation/cleavage sites etc. However it is also possible to conceive of three-dimensional information. Although three-dimensional structures are possible with ssDNA, some. e.g. cruciform DNA do require dsDNA.