Hacker News Reader

There are actually differences between using different codons for the same amino acid. Not all tRNA is present in the same amounts in the cell, using a rare codon will slow down translation because the tRNA for it is present in lower concentrations than the one for the more often used codon. So it will affect translation rates if you change the codons.

There are also cases where using a "slower" codon help the protein fold well by giving it a bit of time before continuing translation. Changing the kinetics of translation can affect the outcome.

Thanks for providing this insight. I was wondering why the resulting organisms were different. Your comment offers some solid suggestions as to why.

The article did mention the cells grow more slowly, maybe this explains why. They framed the impetus for this experiment as being "why do we have so many redundancies in codons", maybe the answer is that evolution favored generating new tRNAs for the more used amino acids, speeding up translation? This kind of thing could also be accomplished by changing the rate of generating tRNAs, and for all I know that plays a role too. I'd never really put much thought into why the code was set up that way, its interesting to think about.

> It should not be surprising that given a correct assembly the bacteria reproduced - by design the experiment only replaced codons with a different equivalent codon.

I am surprised. I was under the impression that while two codons may code for an equivalent protein, bacteria do play shenanigans with DNA, like re-reading the same genome multiple times with shifted read frame. I don't think codon equivalence is preserved when you shift the read frame by 1 or 2. Moreover, a DNA molecule with changed codons is slightly different chemically at that place - which means a protein that might have previously attached there for some reason now might not.

> Moreover, a DNA molecule with changed codons is slightly different chemically at that place - which means a protein that might have previously attached there for some reason now might not.

That shouldn't be a big concern. For one I think they only altered the base pairs inside the codons of the open reading frames. Most of the important protein binding during transcription and translation happens on regions outside the open reading frames which they left unaltered. Additionally the codons that code for the same amino acid share 2 out of the 3 bases, so they only had to alter a single base, which is usually not enough to change functionality a lot espacially when their main functional property is being translated.

Disclaimer: I only have access to abstract right now

Certainly that may be the case but at least with this strain that didn't affect the viability of the organism.

You are right. If you read the paper, they de-duplicated all out-of-phase (shift 1 or 2) overlapped genes. Simple search and replace wouldn't have been viable due to these overlapped genes.

Thanks. Should've read the paper before commenting.

Actually, your question and followup answer is valuable addition to the discussion.

Has anyone tried a similar technique to reduce code on a size limited computer demo?

They did though. They completely rewrote this organism's genome. That's a major technical achievement!

Imagine you're writing a text editor and you get it working well enough to replace spaces with tabs throughout its own source code. Sure, you're not changing the semantic meaning of the code --- but still you're demonstrating that your tool actually works and that it's capable of controlled global edits that don't completely wreck what you're trying to modify. That'd be a pretty major milestone in your text editor project.

This global DNA search-and-replace is the same kind of milestone.

Sounds like they refactor the DNA.

Basically correct, I think.

Nah. More like a clang-format pass. Refactoring would be more like rearranging genes on the chromosome.

I think the meaning of "synthetic genome" here is "synthesized in vitro", not "100% artificially engineered sequence".

I'm having difficulty understanding how a 100% artificially engineered sequence would come about? What is the raw material that would be used to start something like this?

I am not an expert (my biology degree is 30 years old) but generally the idea is that, if you create an artificial genome, you pop it it into a similar cell, which has previously had its native DNA removed) giving it the necessary machinery to express itself etc. Obviously, you're creating a hybrid, but over time the effect of the original host cell's makeup will diminish.

Very much depends what one would consider a "100% artificially engineered sequence". You could possibly swap out GATC for a different set of bases. You could also replace DNA with XNA. There are a whole lot exotic avenues for "artificially engineered", with each of them coming with its own mountain of challenges.

I think plonking is generous. They really chunked or slotted it in (I think, I haven't read the paper).

Running with it: They didn't do a sed in place, they split the file and did the sed in place in chunks.

"It should not be surprising that given a correct assembly the bacteria reproduced - by design the experiment only replaced codons with a different equivalent codon"

Codons are never really equivalent because of differences in tRNA abundances. So replacing a codon with another that codes for same amino acid can disrupt the rate of translation, and downstream processes like protein folding.

Genetics feels like an art of working with an extreme version of a legacy code. No documentation, and side effects literally everywhere :)

I agree with your statement. The core machinery for transcription and translation is still the same -- I will be impressed when a synthetically designed ribosome or DNA/RNA polymerase is created.

The redundancy is probably there to so that the DNA->RNA copying is more robust against chance mutations in DNA that occurred during cell division.

Well, one thing they've proved by doing that is that the redundant codings didn't do anything unexpectedly important.

I don't think that's entirely correct.

> The bacteria are alive, though unusually shaped and reproducing slowly.

If I understand correctly, they completely removed two serine encodings and one stop codon encoding. Given how important tertiary DNA and RNA structure is, I think it's remarkable that they're alive at all.

It's e coli. Theyre like weeds, plus we coddle them in the lab.

E. Coli is super robust even in adverse conditions, and a lot of the coding regions are essentially to support those adverse conditions.

Lab conditions are super friendly to E. coli - they're plated on agar that is tailored to them, kept at a temperature that is tailored to them, and the plates are sterilized so that they have no competitors.

So even if the initial cells were super unhealthy, as long as it wasn't directly fatal the bacteria would grow, and bacteria evolves quickly - fast enough that I'd be interested in seeing how much change was accumulated after 24 hours.

That said, they edits were deliberately just replacing like with like, so it would be expected that they survive.

Interesting follow on research: If you repeatedly ensured that they dropped the alternate serine codings, could you convince it to treat the alternate codings as being something other than serine: the careful engineering would have removed the selective pressure on those codings being serine.

> That said, they edits were deliberately just replacing like with like, so it would be expected that they survive.

To the extent that we only consider encoded proteins, sure. But DNA and RNA has secondary and tertiary structure that does things - it isn't just bits on a hard drive. Riboswitches and Catalytic RNA are good examples.

For example: (https://www.nature.com/scitable/topicpage/rna-functions-352)

> the elevated temperature inside the host's body melts the secondary structure of a segment in the 5' untranslated region of the mRNA produced by the bacterial prfA gene. As a result of this alteration in secondary structure, a ribosome-binding site is exposed, and translation of protein can take place

Riboswitch overview: (https://www.nature.com/scitable/topicpage/riboswitches-a-com...)

> as long as it wasn't directly fatal the bacteria would grow

The abstract states that 18,214 occurrences were edited. The fact that such drastic structural changes to DNA and RNA weren't immediately fatal is what surprises me. It's believable enough, but I would not have predicted it.

I wonder if they just got lucky with (or perhaps strategically chose) the specific codon substitutions to perform? Would a different set of synonym substitutions have been nonviable?

> I'd be interested in seeing how much change was accumulated after 24 hours.

Very much this. Also, a sibling comment mentions the E. coli long-term evolution experiment; I'd be very interested to find out what changed to compensate for this over hundreds or even thousands of generations.

> I wonder if they just got lucky with (or perhaps strategically chose) the specific codon substitutions to perform? Would a different set of synonym substitutions have been nonviable?

Recoding scheme was strategically chosen. They cite 2016 paper (PMC5035903) which showed substituting AGA with CGA (synonymous codon for arginine) on E. coli was nonviable.

That is a really neat paper.

> Successful replacement codons tended to conserve local ribosomal binding site-like motifs and local mRNA secondary structure, sometimes at the expense of amino acid identity. Based on these observations, we empirically defined metrics for a multidimensional “safe replacement zone” (SRZ) within which alternative codons are more likely to be viable.

Why is it surprising that it isn't fatal? They picked relatively rare codons, and they slotted the changes in sequentially. The process of inserting dna is selective (plating lawns of bacteria and picking colonies), probably the fastest growing colonies had the opportunity to build in compensatory mutations if there were subtle secondary or tertiary XNA features that were mildly deleterious.

> Given how important tertiary DNA and RNA structure is, I think it's remarkable that they're alive at all.

It'd be interesting to add these things to the famous E. coli long-term evolution experiment [1] and see how many of the removed codons reappear due to beneficial spontaneous mutations that revert Dr. Chin's changes.

[1] https://en.wikipedia.org/wiki/E._coli_long-term_evolution_ex...

To establish that the organism would have to be observed for some time to rule out exceptional responses/relationships to other microbial communities, and materials. Life is problematically complicated and confounding over time, it doesn't lend itself to neat proofs.