Friday, February 01, 2013

What Is a Mutation?

I've said it before and I'll say it again, biology is messy. It's really hard to rigorously define simple terms because there are always exceptions. Just think of the problems we've had trying to define a gene [What Is a Gene?].

"Mutation"¹ is almost as difficult. First, we want to distinguish between a mutation and DNA damage. DNA damage occurs when various enzymes make a mistake and damage the nucleotides in a DNA molecule. Damage also occurs when outside forces such as X-rays or chemical mutagens attack DNA. Examples are thymidine dimers or cleavage of a base from a nucleotide. DNA can also be broken into two or more pieces.

This damage is never copied and passed on to the next generation. Either it is fixed in some way or it is lethal. When the damage is fixed it may end up being identical to the original DNA molecular or it may be altered in some way that is passed on. Thus, mutation is (semi-)permanent change that is heritable.

In the example shown here, the damage is deamination of cytosine, a very common spontaneous reaction. It is usually repaired fairly quickly but if the DNA is replicated before repair it will result in a switch from a G/C base pair to an A/T base pair at the same site. This change is inherited in all subsequent generations ... it is a mutation.

The genetic material is DNA in most cases but RNA genomes (viruses) can also be mutated. There are many different kinds of "genomes" that have to be covered in our definition. This include virus genomes, mitochondrial genomes, chloroplast genomes, plasmids, and mobile genetic elements (mostly transposons).

Any alteration in the sequence of a genome counts as a mutation, not just those that occur in a gene (whatever that is!). This is important because some of the traditional definitions of mutation are restricted to genes.

Here's a good definition from the Wikipedia site ...
In genetics, a mutation is a change of the nucleotide sequence of the genome of an organism, virus, or extrachromosomal genetic element. Mutations result from unrepaired damage to DNA or to RNA genomes (typically caused by radiation or chemical mutagens), from errors in the process of replication, or from the insertion or deletion of segments of DNA by mobile genetic elements.[1][2][3] Mutations may or may not produce discernable changes in the observable characteristics (phenotype) of an organism.
The Understanding Evolution at UC Berkeley defines mutation as ...
A mutation is a change in DNA, the hereditary material of life. An organism's DNA affects how it looks, how it behaves, and its physiology. So a change in an organism's DNA can cause changes in all aspects of its life.
This isn't good because it doesn't cover RNA genomes and it doesn't distinguish between DNA damage and fixed, heritable, change.



-mutation types
-mutation rates
The Genetic Science Learning Center at the University of Utah offer this definition.
A mutation is a permanent change in the DNA sequence of a gene. Mutations in a gene's DNA sequence can alter the amino acid sequence of the protein encoded by the gene.
That's no good because it restricts mutations to protein encoding genes.

A quick Google search will reveal many other definitions but none as as good as the Wikipedia entry.

As usual, the standard dictionary definitions are not helpful. They are usually quite poor at defining biological terms. Merriam-Webster defines mutation as ...
a relatively permanent change in hereditary material involving either a physical change in chromosome relations or a biochemical change in the codons that make up genes
This is actually better than some of the definitions from scientists. It's main deficiency is that it restricts mutations to genes.

1. The word comes from the Latin mutare, to change.


  1. Why not say a mutation is any change in the nucleic acid genome of a biological entity. The various mechanisms by which one can arise, whether or not it is ultimately repaired or becomes fixed or is ultimately heritable would be part of an extended discussion about the nature of mutations. I understand why there would be an emphasis on heritability but from a mechanistic perspective I don't see why heritability would be a formal component of definition.

    Though I suppose by this definition a quite normal and often transient methylation of a base would also qualify as a mutation.

    Also, what about incorporation of an incorrect amino acid into a growing polypeptide by the ribosome that does not stem from an error in the mRNA codon? Though not heritable and generally of no consequence to the cell, it is surely a mutation in that particular protein and yields a mutant protein.

    1. I understand why there would be an emphasis on heritability but from a mechanistic perspective I don't see why heritability would be a formal component of definition.

      Because DNA damage is transitory as you mention below in the case of methylation. Most DNA damage is repaired and it gets confusing if you start talking about "genetic mutations" that are repaired.

      ... it is surely a mutation in that particular protein and yields a mutant protein.

      That depends on the definition of "mutation." :-)

      What we're talking about here is genetic mutation.

      I want a definition that excludes such cases. I'm happy to say that the protein is damaged by error-prone protein synthesis but I'm not happy to call it a genetic mutation.

  2. How about: A mutation is a change in nucleic acid sequence that is indistinguishable to enzymes from unchanged sequence? I don't think inheritance is essential - polymerase is just one of several enzymes that will cheerfully read any stretch of 'normal' bases, without cross-reference to a 'canonical' sequence.

    1. ...of course this would require an elaboration about what the meaning of "indistinguishable" is. To use RNA polymerase as an example, it will read through any most any mutation as well as it would the original sequence. But let that mutation be in the gene promoter and effect RNA polymerase binding to the promoter in the first place, then one might say there is a distinguishing element to that sort of mutation. Of course the polymerase will in no sense be said to sense a problem: it will just become the new normal at that particular promoter. Perhaps the same can be said of DNA polymerase and some sort of mutation at the chromosome origin of replication. Also, for both RNA and DNA polymerase, it is possible that a point mutation could generate a pause site that temporarily stalls the polymerase at the mutant sequence.

    2. I agree with SRM. It seems rather pointless to restrict the definition of mutation to only those mutations that essentially have no effect

    3. Of course they have an effect. Sequence is changed - that's the effect. It may or may not make a difference to the organism. The sequence at time t2 differs from that at time t1, which the organism cannot check. It may well be that the copy at time t2 is not the same copy as that at time t1 (the change has been replicated - inherited) but still, a change occurred some time between t1 and t2 that was not detected or successfully repaired by the cellular machinery, because it involved or resulted in a 'normal' base or base pair.

      The various scenarios that can be dreamt up based upon where exactly the change occurs wrt origins of replication etc are just the same general class of downstream consequences that affect (say) amino acid substitution in a gene - phenotypic effects of a genetic change.

      So for double-stranded DNA for example, normal pairs are indistinguishable from 'abnormal', unless there is a reference chromosome - even then, which is 'right'? Equally mispairs, and exotic bases, are detectable and potentially repairable - all of which endeavours may fail, and result in a 'true' mutation - the 'new normal', as SRM put it: an undetectable (to the cell) change in sequence.

    4. Of course they have an effect. Sequence is changed - that's the effect.

      Yes. Your original comment about enzyme reading cheerfully through a mutation made me think your definition was more specific and particular than you actually meant.

    5. SRM - OK. I'll drop the anthropomorphic language! Perhaps 'cellular machinery' might have been a better phrasing than 'enzymes'. Some issues are not detectable/repairable, even in principle, given the information the cell has to go on.

      An interesting additional layer is apoptosis - tissue mutation rates appear to decrease along the spermatogenesis pathway. Something, somewhere, appears able to do a 'cross-reference' and eliminate or influence elimination of, deviants.