pABA-based nonribosomal peptides exist—the cystobactamids—and wild-type E. coli ribosomes are happy to make aramids if you supply the right tRNAs. Making something structural through either of these pathways would definitely be a huge leap and unlikely to evolve.
But what gives Kevlar its strength is a whole lotta hydrogen bonds all lined up. That’s hardly outside of proteins’ central area of competency, which is how spider silk exists. So the hypothesis should not be flesh is weak because protein biochemistry is in a local optimum but evolution is smarter than you mere mortal and you have missed crucial dimensions of its optimization.
Now that de novo protein design is coming of age, there’s a big push towards new protein-based fibers, so proteins may take their crown back from the aramids yet. But this is outside my area of expertise.
As for graphene, if you wanted to ball out and do it with synbio, you could maybe, super speculatively, do not trust me hack it enzymatically with Diels-Alder cycloaddition. Diels-Alderases are borderline nonexistent in nature, but some of the early successes in de novo enzymes were Diels-Alderases.
The strongest spider silk is...weaker than Kevlar by density, but it’s comparable, and much more elastic. Not bad at all. And it can be spun from aqueous solutions, while Kevlar is spun in pure sulfuric acid. You also have to consider UV resistance here.
hack it enzymatically with Diels-Alder cycloaddition
Yeah, no.
If you’re interested in progress in aramid fibers, you could look up “5-(6)-amino-2-(4-aminobenzene)benzimidazole”.
TBC, nobody should think that making graphene enzymatically is in any way near-term plausible, or should be attempted as a project (god forbid), just that that’s the only way I’ve heard of making it biologically that isn’t an obvious type error. If someone wanted to write a sci-fi with biologically synthesized graphene, I would nod along if they said that.
Curious if you have any more thoughts on that—“maybe this could be done if they figure out the inverse folding problem idk” was my best impression when I looked into it 5 years and a bio degree ago, but I don’t know much about either the DA reaction or how carbon nanotubes are made etc.
(I’m bullish on the prospect of “Boltz/AF3/BioEmu successor models of the nearish future generalize enough that you can make an enzyme for darn near anything where you can make a theozyme/the thermodynamics work out,” and completely ignorant of DA thermodynamics)
pABA-based nonribosomal peptides exist—the cystobactamids—and wild-type E. coli ribosomes are happy to make aramids if you supply the right tRNAs. Making something structural through either of these pathways would definitely be a huge leap and unlikely to evolve.
But what gives Kevlar its strength is a whole lotta hydrogen bonds all lined up. That’s hardly outside of proteins’ central area of competency, which is how spider silk exists. So the hypothesis should not be flesh is weak because protein biochemistry is in a local optimum but evolution is smarter than you mere mortal and you have missed crucial dimensions of its optimization.
Now that de novo protein design is coming of age, there’s a big push towards new protein-based fibers, so proteins may take their crown back from the aramids yet. But this is outside my area of expertise.
As for graphene, if you wanted to ball out and do it with synbio, you could maybe, super speculatively, do not trust me hack it enzymatically with Diels-Alder cycloaddition. Diels-Alderases are borderline nonexistent in nature, but some of the early successes in de novo enzymes were Diels-Alderases.
The strongest spider silk is...weaker than Kevlar by density, but it’s comparable, and much more elastic. Not bad at all. And it can be spun from aqueous solutions, while Kevlar is spun in pure sulfuric acid. You also have to consider UV resistance here.
Yeah, no.
If you’re interested in progress in aramid fibers, you could look up “5-(6)-amino-2-(4-aminobenzene)benzimidazole”.
TBC, nobody should think that making graphene enzymatically is in any way near-term plausible, or should be attempted as a project (god forbid), just that that’s the only way I’ve heard of making it biologically that isn’t an obvious type error. If someone wanted to write a sci-fi with biologically synthesized graphene, I would nod along if they said that.
Curious if you have any more thoughts on that—“maybe this could be done if they figure out the inverse folding problem idk” was my best impression when I looked into it 5 years and a bio degree ago, but I don’t know much about either the DA reaction or how carbon nanotubes are made etc.
(I’m bullish on the prospect of “Boltz/AF3/BioEmu successor models of the nearish future generalize enough that you can make an enzyme for darn near anything where you can make a theozyme/the thermodynamics work out,” and completely ignorant of DA thermodynamics)
For some limits of enzymes, see these posts of mine:
https://www.lesswrong.com/posts/FijbeqdovkgAusGgz/grey-goo-is-unlikely
https://www.bhauth.com/blog/chemistry/metabolic engineering limits.html