I've got a hypothesis that many microbes have proteins and pathways that are intended to mutate in response to stressors (or for mutants to already exist in any decently-sized population so that they can be available for selection). These are mostly repressors that offer a big "mutational target" and turn on a strong stress response when knocked out. This has a few advantages over normal transcriptional regulation, especially that even in a fairly small population a few cells will have picked up one of these mutants and already be expressing the stress response system as a kind of bet-hedging strategy.
Interesting find. Thanks for sharing. What are the mechanism(s) by which these proteins could mutate? How might the stressor signal trigger heritable changes in the genome?
Natural, random mutation would be enough to cause it. Using the assumptions of an 1000 bp repressor system and 1 mutation per doubling, then somewhere around one in a thousand or one in ten thousand cells would have it. That's enough that you don't need a specific mechanism to cause the mutation.
Then the stressor doesn't technically cause the heritable mutation, but it does strongly select for it, so the outcome is the same.
Technically there's no direct environmental transfer of information to the genome, but from the outside that's the apparent result. And biology doesn't care how it happens, just that it works.
What they found was that in some bacterial cells, to produce certain antiphage proteins, DNA is first transcribed into a noncoding RNA (ncRNA), which is converted back to DNA by a rolling circle reverse transcription reaction. The resulting DNA is subsequently transcribed into an mRNA that encodes the protein.
Interested by this: "the question of why biology has not evolved to alter DNA to swap between two forms of a protein."
I don't have a super strong foundation on protein structures and transcription, but wouldn't post-transcriptional modification provide a work around for this? Suppose I have two proteins with mostly the same structure—XYZ, XWZ. Could I not encode my DNA as XYWZ and via splicing construct XYZ and XWZ separately?
This would allow a perfect marriage between responsiveness within the organism and heredity of desirable proteins. Although I don't have much information on how post-transcriptional modification works, so perhaps there's some mechanism that limits its responsiveness.
That's a great question. In principle, yes, one gene can yield multiple mRNA transcripts (and thus make multiple protein isoforms.) However, alternative splicing is not always so simple. Sometimes gaining or losing an exon means new regulatory signals are required in the gene sequence, or the two protein variants are so different structurally that maintaining them in a single splice scheme makes mRNAs unstable. Maintaining these other splicing options in the cell can also be resource-intensive, especially if the other isoform is only made in rare situations.
Great piece. Right up until... "DNA is metabolically cheap to make, even if it does nothing (which explains why our DNA is full of meaningless repetitive sequences)".
Are you kidding me?!
Respectfully, I suggest he look deeper at the other "pseudo-dogma"- 98% junk DNA.
Peter Gariev's from Russia, and many others work has since debunked this poor scientific thinking. Crick was dealing with 2% of DNA with his hypothesis, and even he understood that the other 98% logically did "something" despite not having clarity on what that something was at the time.🤨😐
Hey there. We agree with you! The once prevalent idea that a majority of the genome is "junk" has not held up to scrutiny. But the article doesn't argue that; indeed, there do seem to be many repetitive sequences (probably viral-derived) that likely do nothing.
Also, at least for a single gene, the energetic costs of making proteins is far greater than making either DNA (through replication) or RNA (via transcription). https://pmc.ncbi.nlm.nih.gov/articles/PMC4697398/
It's an idea but their assumption is based on typical dogma around the Krebs cycle and ETC. We are learning that ATP is not the currency of the cell that we have been told. So I think they have a nice thought piece for discussions BTJMO😁
Yeah I wanted to come here to comment a similar sentiment. "Full of meaningless repetitive sequences" seems like it needs a bit more context as it may lead readers down the classic "dogma" of junk DNA. Maybe wording it to say that transposon remnants would presumably lower fitness but don't appear to, as they are metabolically cheap.
From my vantage point world seems increasingly awash in protein-driven DNA switches. I don't know whether I'd go so far as to argue they defy Crick's bar napkin, but they sure are cool.
I am very intrigued by “the question of why biology has not evolved to alter DNA to swap between two forms of a protein.”
Your hypothesis, based on the relative metabolic costs, suggests selection after sampling. But I wonder if nature has ever had access to mechanisms that would even allow that. Perhaps there’s an evolutionary bottleneck that hasn’t been breached to unlock that additional layer of complexity?
If that was possible, there could be many additional capabilities that it could enable of course, besides selection two forms of an enzyme. One can imagine that proteins that modify DNA in response to their environment could accelerate evolution itself. Path dependence would be less powerful. An organism could swiftly switch between phenotypes, and extinctions may be much rarer!
I've got a hypothesis that many microbes have proteins and pathways that are intended to mutate in response to stressors (or for mutants to already exist in any decently-sized population so that they can be available for selection). These are mostly repressors that offer a big "mutational target" and turn on a strong stress response when knocked out. This has a few advantages over normal transcriptional regulation, especially that even in a fairly small population a few cells will have picked up one of these mutants and already be expressing the stress response system as a kind of bet-hedging strategy.
There's a few papers that have shown things like this, but the most recent one I've read was https://www.nature.com/articles/s41467-020-19713-w
Interesting find. Thanks for sharing. What are the mechanism(s) by which these proteins could mutate? How might the stressor signal trigger heritable changes in the genome?
Natural, random mutation would be enough to cause it. Using the assumptions of an 1000 bp repressor system and 1 mutation per doubling, then somewhere around one in a thousand or one in ten thousand cells would have it. That's enough that you don't need a specific mechanism to cause the mutation.
Then the stressor doesn't technically cause the heritable mutation, but it does strongly select for it, so the outcome is the same.
Technically there's no direct environmental transfer of information to the genome, but from the outside that's the apparent result. And biology doesn't care how it happens, just that it works.
Great article as always. I'd like to point out a recent article from Feng Zhang's lab that could add to the discussion here. Would love to have your take on it. https://www.science.org/doi/abs/10.1126/science.adq3977
Wow - the rabbit hole always goes deeper! Excited to read this one. We'll post an update here this week.
What they found was that in some bacterial cells, to produce certain antiphage proteins, DNA is first transcribed into a noncoding RNA (ncRNA), which is converted back to DNA by a rolling circle reverse transcription reaction. The resulting DNA is subsequently transcribed into an mRNA that encodes the protein.
Interested by this: "the question of why biology has not evolved to alter DNA to swap between two forms of a protein."
I don't have a super strong foundation on protein structures and transcription, but wouldn't post-transcriptional modification provide a work around for this? Suppose I have two proteins with mostly the same structure—XYZ, XWZ. Could I not encode my DNA as XYWZ and via splicing construct XYZ and XWZ separately?
This would allow a perfect marriage between responsiveness within the organism and heredity of desirable proteins. Although I don't have much information on how post-transcriptional modification works, so perhaps there's some mechanism that limits its responsiveness.
That's a great question. In principle, yes, one gene can yield multiple mRNA transcripts (and thus make multiple protein isoforms.) However, alternative splicing is not always so simple. Sometimes gaining or losing an exon means new regulatory signals are required in the gene sequence, or the two protein variants are so different structurally that maintaining them in a single splice scheme makes mRNAs unstable. Maintaining these other splicing options in the cell can also be resource-intensive, especially if the other isoform is only made in rare situations.
😐🤨
Great piece. Right up until... "DNA is metabolically cheap to make, even if it does nothing (which explains why our DNA is full of meaningless repetitive sequences)".
Are you kidding me?!
Respectfully, I suggest he look deeper at the other "pseudo-dogma"- 98% junk DNA.
Peter Gariev's from Russia, and many others work has since debunked this poor scientific thinking. Crick was dealing with 2% of DNA with his hypothesis, and even he understood that the other 98% logically did "something" despite not having clarity on what that something was at the time.🤨😐
Hey there. We agree with you! The once prevalent idea that a majority of the genome is "junk" has not held up to scrutiny. But the article doesn't argue that; indeed, there do seem to be many repetitive sequences (probably viral-derived) that likely do nothing.
Also, at least for a single gene, the energetic costs of making proteins is far greater than making either DNA (through replication) or RNA (via transcription). https://pmc.ncbi.nlm.nih.gov/articles/PMC4697398/
It's an idea but their assumption is based on typical dogma around the Krebs cycle and ETC. We are learning that ATP is not the currency of the cell that we have been told. So I think they have a nice thought piece for discussions BTJMO😁
Yeah I wanted to come here to comment a similar sentiment. "Full of meaningless repetitive sequences" seems like it needs a bit more context as it may lead readers down the classic "dogma" of junk DNA. Maybe wording it to say that transposon remnants would presumably lower fitness but don't appear to, as they are metabolically cheap.
Great article and comments just what I wanted from substack.
Invertons are an exciting new confirmation of the stress-responsive-mutation hypothesis:
https://pmc.ncbi.nlm.nih.gov/articles/PMC6543533/
Playing with viral sequences has got me wondering whether ribozymes might be doing inverton-like jobs in animal germlines:
https://cbuck.substack.com/p/can-self-cleaving-dna-resurrect-lamarck?r=5cli6
And of course let's not forget the APOBECs!
https://pubmed.ncbi.nlm.nih.gov/29746834/
From my vantage point world seems increasingly awash in protein-driven DNA switches. I don't know whether I'd go so far as to argue they defy Crick's bar napkin, but they sure are cool.
I am very intrigued by “the question of why biology has not evolved to alter DNA to swap between two forms of a protein.”
Your hypothesis, based on the relative metabolic costs, suggests selection after sampling. But I wonder if nature has ever had access to mechanisms that would even allow that. Perhaps there’s an evolutionary bottleneck that hasn’t been breached to unlock that additional layer of complexity?
If that was possible, there could be many additional capabilities that it could enable of course, besides selection two forms of an enzyme. One can imagine that proteins that modify DNA in response to their environment could accelerate evolution itself. Path dependence would be less powerful. An organism could swiftly switch between phenotypes, and extinctions may be much rarer!
(I explore path dependence in the way the human-made world evolves: https://gairiksachdeva.substack.com/ )