Explanations of Mutation , Recombination , Reassortment , Recombination

[pjie2]

OK, some definitions for the audience. I'll lay out my credentials up front: I am a molecular biologist at Cambridge University working on analysis of male fertility and the sex chromosomes. Posters are welcome to send me a forum PM if they want my name and email address. I am not a flu expert, however these are standard biological definitions and do not depend on any particular specialism.

************************************************** *******

1) Mutation

At the most general level, this refers to a change in the nucleotide sequence of a gene, or the amino acid sequence of a protein. A mutation is called a de novo mutation at the moment it occurs.

There are a number of different types of mutation. I shall give examples of some different types of mutation in a DNA sequence. Please note that for clarity I am only showing one of the two DNA strands:

ATCGATC[b]G[/b]ATCGATCG  ->  ATCGATC[b]C[/b]ATCGATCG
[I]Substitution:[/I]  a G has been replaced by a C.  Bases can be purines 
(A and G) or pyrimidines (T and C).  Transition mutations, converting a purine
to a purine or a pyrimidine to a pyrimidine are much more common than
transversion mutations, which convert a pyrimidine to a purine or vice versa.

ATCGATC[b]G[/b]ATCGATCG  ->  ATCGATCATCGATCG
[I]Deletion:[/i]  a G has been removed from the sequence.  

ATCGATC[b]G[/b]ATCGATCG  ->  ATCGATC[b]GA[/b]ATCGATCG
[i]Insertion:[/i]  an A has been added in to the sequence.

When a nucleotide sequence is translated to make a protein, each triplet of nucleotides (called a codon) encodes a single amino acid in the resulting protein. This means that an insertion or deletion may change the "frame" of the readout and lead to a drastic change in protein sequence. Insertion/deletion of 3, 6, 9... nucleotides (or any multiple of three) will however preserve the reading frame.

In contrast, a single nucleotide substitution will only affect the single amino acid residue encoded by that triplet. This may have a significant effect on the protein fuction, or may have no effect at all, since there are several different codons which encode each different amino acid.

************************************************** *******

2) Recombination (standard definition)

This occurs when two DNA strands come together and exchange information. Here, I am showing both strands of the DNA molecule (note that A pairs with T, and C pairs with G).

AAAAAAAAAAAA                                AAAAAACCCCCC         
||||||||||||                                ||||||||||||
TTTTTTTTTTTT                                TTTTTTGGGGGG

    and            --recombination-->           and

CCCCCCCCCCCC                                CCCCCCAAAAAA
||||||||||||                                ||||||||||||
GGGGGGGGGGGG                                GGGGGGTTTTTT

Note that in the conventional description, recombination involves the exchange of information between two sequences, and involves large stretches of sequence information - typically many thousands or millions of base pairs in length.

THIS IS NOT THE DEFINITION OF RECOMBINATION THAT HENRY NIMAN IS USING. I am giving the standard definition here so people do not get confused when reading genetics textbooks, scientific papers etc. I will explain what Niman means by recombination later on in this post (or at least what I understand him to mean).

************************************************** *******

3) Reassortment

The flu virus genome is comprised of eight different strands. I will use normal alphabet letters to represent these rather than using DNA diagrams, for purposes of clarity

AAAAAAAAAAAA            aaaaaaaaaaaa
BBBBBBBBBBBB            bbbbbbbbbbbb
CCCCCCCCCCCC            cccccccccccc
DDDDDDDDDDDD            dddddddddddd
EEEEEEEEEEEE            eeeeeeeeeeee
FFFFFFFFFFFF            ffffffffffff
GGGGGGGGGGGG            gggggggggggg
HHHHHHHHHHHH            hhhhhhhhhhhh

The above represents the genomes of two different strains of flu, one in upper case and one in lower case. If a given person/animal is infected with BOTH strains of flu, and more particularly if the SAME CELL in the person/animal is infected with both strains then they can get confused and swap segments between each other, leading to the generation of reassorted virus particles such as the ones shown below.

AAAAAAAAAAAA            aaaaaaaaaaaa
bbbbbbbbbbbb            bbbbbbbbbbbb
CCCCCCCCCCCC            CCCCCCCCCCCC
DDDDDDDDDDDD            dddddddddddd
EEEEEEEEEEEE            eeeeeeeeeeee
fffffffffff             ffffffffffff
GGGGGGGGGGGG            GGGGGGGGGGGG
HHHHHHHHHHHH            HHHHHHHHHHHH

In this example, the virus particle on the left is mostly an upper-case strain, but has picked up the "b" and "f" strands from the lower-case strain. The particle on the right is mostly a lower-case strain, but has picked up the "C", "G" and "H" strands from the upper-case strain.

************************************************** *******

4) Recombination (Niman definition)

Niman holds that when virus particles are copying themselves within the cell, if there is a dual infection, the synthesis machinery may make a mistake and copy part of the "other" strain's genome before switching back to the original template. Thus, rather than whole segments being reassorted, smaller sequence stretches - perhaps as small as a single nucleotide - can move between strains.

AAAAAAAAAAAA            aaaaaaaaaaaa                        AAAAAAAAaAAA
BBBBBBBBBBBB            bbbbbbbbbbbb                        BBBBBbBBBBBB
CCCCCCCCCCCC            cccccccccccc                        CcCCCCCCCCCC
DDDDDDDDDDDD            dddddddddddd         ---->          DDDDDDDDDdDD
EEEEEEEEEEEE            eeeeeeeeeeee                        eEEEEEEEEeEE
FFFFFFFFFFFF            ffffffffffff                        FFFFFFfFFFFf
GGGGGGGGGGGG            gggggggggggg                        GGGGgggGGGgG
HHHHHHHHHHHH            hhhhhhhhhhhh                        HHHHHHHHHHHH

In this example, an "upper-case" strain of the virus has picked up some patches of sequence from a "lower-case" strain in most of its segments. This is not actually recombination in the normal sense: there is no suggestion that an "AAAAAAAAAAAA" molecule and an "aaaaaaaaaaaa" molecule line up next to each other and exchange information. There is no reciprocal product generated. Rather, it is a case of gene conversion, where the information from one strain over-writes the information in the other strain. I'm happy to believe that this may be a case of terminology differing between fields though: as I say, my own field is fertility, in which the meaning of recombination is as given in (2).

[pjie2]

Re: Origins and evolutionary genomics of the 2009 swine origin H1N1 influenza A (pandemic)

Further to the above, some thoughts on Niman's recombination hypothesis. I have no dog in this fight: I am not a flu expert and have no reason to suspect that the template-switching process does or does not occur. An obvious point that occurs to me is that it requires the same preconditions as reassortment: a double infection of the same cell of the same animal with two viruses. In addition it requires the growing nucleic acid chain to swap templates at least once during synthesis. This allows two predictions:

* "recombination" (defined per Niman) should occur at the same or lower frequency than reassortment. This follows from the fact that this form of recombination requires a double infection AND template switching, while reassortment only requires a double infection

* The length of the conversion tracts should be biased, with shorter stretches of gene conversion (recombinations) being less common than longer stretches. This follows from a consideration of the proposed template-switching mechanism. It seems inherently unlikely that a growing strand will swap templates for a single nucleotide and then swap back again immediately.

These might be testable in culture: infect an egg with two viral strains and use PCR to look for viral strands that have swapped sequence. Not sure how easy that would be to do, but it's something I'd consider if I were working in the area. Similar techniques are used (for example) to estimate recombination rates during sperm production - take the starting genotype and make PCR primers specific for a recombined genotype, and see if you can find it.

Set against these predictions, however, there is the question of selection on the resulting virus particles in terms of ability to reproduce. One of the reasons reassortments are rare is that the segments have been evolving separately and may not "play nice" with each other in terms of making a viable virus. The same would presumably apply to "recombination" events, favouring smaller conversions that bring in smaller stretches of new sequence at a time. This can also be tested: if there is selection for short conversion tracts, then you would expect shorter conversion tracts in coding sequence than in noncoding sequence.

**********************

Niman raises above the question of the H274Y polymorphism that confers resistance to oseltamivir. This has (apparently) appeared in multiple different strains of flu, including those from countries not using oseltamivir, where you would not expect selection for resistance. His contention is that this represents recombination events which have spread this polymorphism between disparate flu strains. The alternative is that the spread of this change represents de novo mutation in each strain: i.e. they have independently acquired this change without sharing information between each other.

It strikes me that the proposed recombination events do not so much explain the mystery as replace it with another. Instead of the question: "Why have unrelated strains acquired H274Y by de novo mutation if there is no oseltamivir use to select for this change?", we have "why have unrelated strains acquired H274Y by recombination if there is no oseltamivir use to select for this change?" The only message I can take is that we do not know why H274Y is appearing in different strains. We might surmise that there is some other selective factor driving it to fixation, or it might be hitch-hiking on other mutations, or it may be pure coincidence.

[pjie2]

Re: Explanations of Mutation , Recombination , Reassortment , Recombination (Niman)

Adding to this thread after the sticky:

The mechanism proposed by Dr Niman is most usually called "copy choice recombination", although it is (very rarely) called homologous recombination. The latter terminology is misleading as it invites confusion with the normal DNA definition of homologous recombination, as given in example (2) above.

[Farmer]

Re: Explanations of Mutation , Recombination , Reassortment , Recombination (Niman)

Thanks pjie2. Very clear presentation!

[Lizw]

Re: Origins and evolutionary genomics of the 2009 swine origin H1N1 influenza A (pandemic)

When a nucleotide sequence is translated to make a protein, each triplet of nucleotides (called a codon) encodes a single amino acid in the resulting protein. This means that an insertion or deletion may change the "frame" of the readout and lead to a drastic change in protein sequence. Insertion/deletion of 3, 6, 9... nucleotides (or any multiple of three) will however preserve the reading frame.

I'm a computer person from a long way back (the days when sequential storage devices were standard, and programming was done with numbers, not words). I've been struck every time I've seen an explanation like this by how similar the processes are and how similar the kinds of errors are to the early computer days. A "frame" error would lead to nonsense in program execution, and I imagine it does most of the time in virus replication as well.

Thank you for a very clear explanation, and for prompting me to read further. We're just big biological computers after all.

[the doctor]

Re: Explanations of Mutation , Recombination , Reassortment , Recombination (Niman)

The rapid adoption of seasonal H1N1 of H274Y is an anomaly IMO. While a supporter of Dr. Niman, his notion that this polymorphism is the result of it "hitch hiking" on to a closely related conserved polymorphism does not wash.

While what he says, and I am going way out on the ignorance limb here since I am among the genetically challenged, is not supported by the facts. If he could point to the closely related sequence that was conserved and explain why it was, then this would provide support for the hitch hiker view. As far as I know, and remember I am ignorant, he has not identified this key segment.

I come at this from the macro view as someone with a background in biology, medicine and a love of history, politics and economics. The comments made by pjie2 that the usual source of a change in the virus like this would require that it be exposed to an environment where there was extensive use of oseltamivir. This is my view. I do not think seasonal H1N1 could become universally resistant to this antiviral in such a short time, from one flu season to the next, unless it experienced this specific environmental condition.

As pointed out by pjie2, there is no evidence that seasonal H1N1 was ever in this type of environment therefore this is probably not the reason for its acquisition of this polymorphism. In response, I would like to remind him or her that absence of proof is not proof of absence. Furthermore, there is Okam's razor to consider. After all other explanations for an observation have been eliminated, the one remaining is probably correct.

I would also add that as a practicing physician, with the exception of Japan where oseltamivir has been used considerably more clinically than in other developed nations, the use of this drug in patients has been very low. Most treating physicians don't prescribe it since by the time they see the patient, the 48 hour efficacy window has long passed. Even if the patient is seen when the window is open, most treating physicians don't prescribe it out of ignorance.

What I am proposing is that the most likely reason for seasonal N1N1 for obtaining H274Y from the macro biologically perspective remains environmental pressure from extensive use of oseltamivir over a large area for a prolonged period of time.

Given what both pjie2 and I have said about the limited use of this drug, how could this have occurred?

IMO, the only logical answer to this is that the Chinese government manufactured large quantities of oseltamivir off-licence and added it to animal feed destined for use by poltrey and swine.

Are there any precedents for this? Yes, the Chinese did this earlier in the century with amantidine when H5N1 was devastating these mainstays of their agricultural economy resulting in the seasonal flu affecting humans developing complete resistance to this antiviral.

The Chinese were roundly criticized for this action then so why would they make the same ill advised decision again with oseltamivir? If indeed this is what happened, only they know why but IMO it had to do with the Beijing Olympics. I think they blanketed the countryside with oseltamivir to do everything possible to damp down H5N1 outbreaks because they knew that if there were reports of this in the media in the months leading up to the games, it could reduce attendance. While many may think this a trivial excuse, for the Chinese leadership it was anything but trivial. A successful 2008 games for them and the nation has been on the top of their domestic and foreign policy agenda over the past decade.