Tweet
I'm reposting a good description of homologous recombination that Dr Niman provided in response to JJackson's post. It is from this thread:
Homologous vs Non-Homologous Recombination <hr size="1" style="color: rgb(204, 204, 204);"><!-- / icon and title --><!-- message --> Quote:
<table cellpadding="6" cellspacing="0" border="0" width="100%"> <tbody><tr> <td class="alt2" style="border: 1px inset ;"> Originally Posted by JJackson
I would like to thank Mellie for her link to http://www.cdc.gov/ncidod/EID/vol10no12/04-0743.htm
posted earlier in this thread. While this is not why she posted it turns out to be perfect for this tread.
As a layman trying to get to grips with the mechanism of mutation (Recombination vs Reassortment/point mutation) and its consequences for base sequences leading to Amino acid sequence leading to 3D protein shape and so to clinical effects ? infectivity & virulence this little story has it all.
Why do I like it so much?
1] It is free. The full text is available without subscription.
2] It is easy to follow even without any virology degrees
3] It is a slam dunk for Recombination.
In brief a farm had two chicken sheds, in shed 1 the flock seemed a little under the weather and then they had a massive die off in shed 2 so they culled all the birds, two of the cullers then became ill (not too seriously). The researchers had sequences for both workers and each shed. Shed 1 had an LPAI infection, shed 2 had the same HA strand but with a new insertion making it longer that it?s ?parent?, the two cullers both had the extended HA strand and a few point mutations and were slightly different from each other and shed 2. The inserted strand exactly matched a sequence of bases from one of the virus? other strands (21 bases from M).
While I am convinced by the sheer weight of examples of recombinant insertions listed on Dr Niman?s site we are not talking about a clear linkage that can be followed in parent/child/grandchild sense. It is more likely that we observe an atypical 30 base sequence in a Turkish duck and when we scan for that sequence in the literature we find an exact, or close, match in some Swan which died 3 years ago on a different continent. The linkage is not obvious but statistically sound due to the lower probability of that 30 base sequence recurring as 30 consecutive point mutations.
Where this example wins out is we have four sequences from the same location over the course of a few days. We can almost see the mutations occurring in real-time and we do not even need a second virus for the donor sequence it is provided from within the Matrix sequence. The short time scale makes the argument for accumulated point mutations fatuous. Once the mechanism is demonstrated - as I believe it is here ? then it is difficult to argue it does not occur elsewhere. The authors then go on to show how the insertion effects the structure of the HA protein and why that may account for the change from LPAI to HPAI.
</td> </tr> </tbody></table>
Reply from Dr Niman:
The above description is for non-homolgous recombination, which happens on RARE occasion. Most examples in flu are like the one above in which a new HA cleave site is created. The new cleavage site has many basic amino acids, so the virus with the new cleavage site has a major advantage because it can then infect many cell types.
The major driver of influenza evolution is HOMOLOGOUS recombination, which involves using stretches of sequence identity to copy part of one gene and then switch to another.
Homologous recombination happens when the polymerase is using one gene as a template. The newly created RNA the hops off the template 1 and lands on template 2 (the gene from the other virus in a dual infection).
Because the region (i.e the last 10 nucleotides) that was just copied is the same sequence on the second template, the newly created RNA finds the right place to bind to template 2 and then the polykmerase starts copying the gene again. However, now the new sequence is directed by template 2 so the new RNA has some genetic information from temoplate 1 and some from template 2. The new RNA is a recombinant.
This type of dual infectoion and jumping happens most often between the same serotype, because the two copies have a lot of identity so it is easy for the new RNA to find the correct spot on template 2. If template 2 was grossly different than 1, the new RNA would just find copies of template 1 and the new gene would be a copy of 1 and not have 2.
If the template 2 differs from 1 by just 1%, then most of the "new" information from 2 will just look like a point mutation.
That is why the newly released Indonesian H5N1 has so many changes that look like point mutations, but match other H5N1 sequences from nearby locations. The dual infections involve two closley related H5N1's, and the recombinant only has a few chnages from the new parent. However, dual infections are common, so there are many changes tracing back to several parents.
These changes are NOT random mutations. They are polymorphisms that can be easily found on other H5N1's.
In some cases, like the Canadian swine, the number of dual infections is limited, so there may be only one cross-over over a long time period. Therefore the two parents are exact matches over a large portions of the gene. This identity can be found in isolates more than 25 years apart, showing that copy errors are RARE, and the vast majority of changes are due to HOMOLOGOUS recombination, not random mutations.
Homologous vs Non-Homologous Recombination <hr size="1" style="color: rgb(204, 204, 204);"><!-- / icon and title --><!-- message --> Quote:
<table cellpadding="6" cellspacing="0" border="0" width="100%"> <tbody><tr> <td class="alt2" style="border: 1px inset ;"> Originally Posted by JJackson
I would like to thank Mellie for her link to http://www.cdc.gov/ncidod/EID/vol10no12/04-0743.htm
posted earlier in this thread. While this is not why she posted it turns out to be perfect for this tread.
As a layman trying to get to grips with the mechanism of mutation (Recombination vs Reassortment/point mutation) and its consequences for base sequences leading to Amino acid sequence leading to 3D protein shape and so to clinical effects ? infectivity & virulence this little story has it all.
Why do I like it so much?
1] It is free. The full text is available without subscription.
2] It is easy to follow even without any virology degrees
3] It is a slam dunk for Recombination.
In brief a farm had two chicken sheds, in shed 1 the flock seemed a little under the weather and then they had a massive die off in shed 2 so they culled all the birds, two of the cullers then became ill (not too seriously). The researchers had sequences for both workers and each shed. Shed 1 had an LPAI infection, shed 2 had the same HA strand but with a new insertion making it longer that it?s ?parent?, the two cullers both had the extended HA strand and a few point mutations and were slightly different from each other and shed 2. The inserted strand exactly matched a sequence of bases from one of the virus? other strands (21 bases from M).
While I am convinced by the sheer weight of examples of recombinant insertions listed on Dr Niman?s site we are not talking about a clear linkage that can be followed in parent/child/grandchild sense. It is more likely that we observe an atypical 30 base sequence in a Turkish duck and when we scan for that sequence in the literature we find an exact, or close, match in some Swan which died 3 years ago on a different continent. The linkage is not obvious but statistically sound due to the lower probability of that 30 base sequence recurring as 30 consecutive point mutations.
Where this example wins out is we have four sequences from the same location over the course of a few days. We can almost see the mutations occurring in real-time and we do not even need a second virus for the donor sequence it is provided from within the Matrix sequence. The short time scale makes the argument for accumulated point mutations fatuous. Once the mechanism is demonstrated - as I believe it is here ? then it is difficult to argue it does not occur elsewhere. The authors then go on to show how the insertion effects the structure of the HA protein and why that may account for the change from LPAI to HPAI.
</td> </tr> </tbody></table>
Reply from Dr Niman:
The above description is for non-homolgous recombination, which happens on RARE occasion. Most examples in flu are like the one above in which a new HA cleave site is created. The new cleavage site has many basic amino acids, so the virus with the new cleavage site has a major advantage because it can then infect many cell types.
The major driver of influenza evolution is HOMOLOGOUS recombination, which involves using stretches of sequence identity to copy part of one gene and then switch to another.
Homologous recombination happens when the polymerase is using one gene as a template. The newly created RNA the hops off the template 1 and lands on template 2 (the gene from the other virus in a dual infection).
Because the region (i.e the last 10 nucleotides) that was just copied is the same sequence on the second template, the newly created RNA finds the right place to bind to template 2 and then the polykmerase starts copying the gene again. However, now the new sequence is directed by template 2 so the new RNA has some genetic information from temoplate 1 and some from template 2. The new RNA is a recombinant.
This type of dual infectoion and jumping happens most often between the same serotype, because the two copies have a lot of identity so it is easy for the new RNA to find the correct spot on template 2. If template 2 was grossly different than 1, the new RNA would just find copies of template 1 and the new gene would be a copy of 1 and not have 2.
If the template 2 differs from 1 by just 1%, then most of the "new" information from 2 will just look like a point mutation.
That is why the newly released Indonesian H5N1 has so many changes that look like point mutations, but match other H5N1 sequences from nearby locations. The dual infections involve two closley related H5N1's, and the recombinant only has a few chnages from the new parent. However, dual infections are common, so there are many changes tracing back to several parents.
These changes are NOT random mutations. They are polymorphisms that can be easily found on other H5N1's.
In some cases, like the Canadian swine, the number of dual infections is limited, so there may be only one cross-over over a long time period. Therefore the two parents are exact matches over a large portions of the gene. This identity can be found in isolates more than 25 years apart, showing that copy errors are RARE, and the vast majority of changes are due to HOMOLOGOUS recombination, not random mutations.