Did someone say “pylons”?
8 Likes
6 Likes
What’s more interesting is that Axe has done no followup to this allegedly important paper in 20 years, despite generous funding from the DI.
What do you make of the stark contrast with Axe (same one) et al. PNAS 93:5590-5594,1996?
Which paper is more thorough?
2 Likes
It’s even worse when you look at what Axe published (with supervision) but doesn’t cite:
https://www.pnas.org/doi/10.1073/pnas.93.11.5590
The title says it all, but the data make the contradiction even more glaring. Why doesn’t anyone at the DI ever cite that paper?
1 Like
You’re ceding far too much semantic ground there by adopting their absurd conflation of function and structure. AFAIK, there is no such thing as “the beta-lactamase fold,” because folds are structural, not functional classifications.
Here’s a good primer:
This part is correct. The problem is that you also are buying into conflating structure with function by using the term “the beta-lactamase fold.”
That’s just gobbledygook.
Correct, but more accurately, folds don’t simply have functions. They are structural classifications only.
IMO it is trash for a much more basic reason: he didn’t bother to measure enzymatic activity. It’s very important medically and assays are cheap. If I had reviewed it, I would have rejected it for that reason regardless of interpretation.
2 Likes
For those readers like me who need a refresher on TEM-1 β-lactamase and related terms, I posted this A.I. chat explanation as its own thread topic:
1 Like
I don’t agree that I’m doing that because I specifically made sure to point out it was the combination of structure and function that Axe purports to measure, at least according to Ann Gauger.
But more importantly, on nomenclature: According to SCOP that is literally the name for the fold. The structure is named by the functions known to be carried out by it.
Fold e.3: beta-lactamase/transpeptidase-like
And yes, they are not actually measuring this relationship. Truth is Axe has no idea what structure his mutant enzymes have adopted after mutagenesis. He never does anything like crystal structures or what have you. He just looks to see if colonies form on agar plates. Kinda strange for a piece of work advertised to probe the relationship between structure, function, and sequence space.
Yes and that’s the real thing that matters in evolution, function or degree of function.
He argues (completely unsuccessfully) in his 2004 paper that this is what he is trying to do, find a sort of base of the hill of activity that provides a fitness benefit because in his mind an enzyme like this would have to sort of evolve de novo (the actual main conceptual flaw of their argument), and this would be much more likely to find some weakly active variant than the global peak (true, but irrelevant).
You could argue that if he really wants to establish some estimate of how weakly functional an enzyme can get and still provide a fitness-benefit to the organism, and with this information try to estimate how many sequences could meet or exceed that threshold for functionality, his protocol would at some point have to include actual tests to see where the activity of his mutant enzymes provide enough of a fitness-benefit to incur a growth advantage under very low concentrations of antibiotic.
But that would also have to be well below the minimum inhibitory concentration (MIC). That is after all where bacteria are thought to develop resistance in the first place, where the concentration of antibiotic is low enough to not entirely arrest growth and kill them, but merely slow it down enough that mutations can result in beneficial effects large enough to be visible to selection.
It is interesting to consider in how many ways his paper fails to be what it claims to be, and what one should have done instead to at least try to estimate what they claim he estimated.
1 Like
No, it literally isn’t. That’s precisely why the name you cited literally includes transpeptidases and literally adds “-like.”
No, structures tend to be named for the protein or group(s) of proteins in which they were first described. Structural biologists have more open minds about this.
Not everyone agrees on the name, either. Some call it by its structural name, which includes the only widespread function (still likely not universal) of the fold, metal ion binding:
The group is in no way limited to lactamases and transpeptidases. It includes phosphodiesterases, ribonucleases, glyoxalases, phospholipases, and more. This is how protein evolution works.
This was known long before 2004, so Axe had no excuse. That’s why I’m pointing out that any adoption of the DI’s misuse of the term in 2024 is a rhetorical mistake.
There is a very partial list of enzymes in the Fig.1 legend of the paper linked above.
It would be better, but I’d have still rejected the paper if Axe had stopped with that, because beta-lactamase assays are available from Sigma-Aldrich and cheap.
Indeed. It was a complete failure of peer review.
It’s even more interesting if you contrast it with the careful work presented in Axe’s PNAS barnase paper, which included more senior coauthors.
2 Likes
12 posts were merged into an existing topic: The Argument Clinic