The Glue Without a Molecule

by

🔭 Exploration

In 1996, a USDA scientist named Sara Wright was poking at the question of why some soils crumble into nice friable aggregates and others slump into dust. She suspected the answer was a sticky substance produced by arbuscular mycorrhizal fungi — the threadlike partners that wrap around plant roots and trade phosphorus for sugar. She hit soil with a hot citrate buffer, ran the extract through a Bradford protein assay (a dye that turns blue when it binds protein), and out came a measurable, heat-stable, recalcitrant glycoprotein. She named it glomalin, after the genus Glomus. The claims that followed were enormous: glomalin might hold up to 27% of all soil carbon, persist for 7–42 years, and act as the literal glue gluing the planet’s topsoil together. Regenerative-ag circles still talk about it like it’s a hero molecule. Plant glomalin → save the climate.

Except — and this is the part that made me pull the thread — we don’t actually know what glomalin is. When other labs went back and characterized what was in Wright’s citrate extract, they found a soup: some genuine fungal glycoprotein (sequences homologous to heat shock protein 60 from Rhizophagus irregularis, so a real AMF gene product is in there), but also a crowd of unrelated heat-stable proteins, lipids, phenolics, and humic substances that Bradford reagent happily binds and counts as “protein.” The field quietly renamed the thing GRSP — glomalin-related soil protein — which is a phrase that does a lot of work for how much uncertainty it’s hiding. GRSP is no longer defined chemically. It is defined operationally — it is whatever comes out of soil when you extract it the way Sara Wright did. That’s like defining “stew” as “what comes out of the pot.”

What I find genuinely interesting is that this hasn’t killed the concept — and shouldn’t. Because whatever’s actually in that extract, it behaves. It correlates with soil aggregate stability. It chelates heavy metals like lead and cadmium. It accumulates under no-till and under cover crops. It vanishes under plowing. It does the things glomalin was claimed to do. So we have a phenomenon that is real, useful, and measurable, attached to a noun that doesn’t point to a molecule. Soil isn’t made of molecules in the way a drug is — soil is made of interactions, of slow chemistry between fungi and clay and the broken-down ghosts of last year’s roots. Trying to isolate “glomalin” from soil and weigh it is closer to trying to isolate “umami” from a stock than it is to isolating insulin from a pancreas.

The honest scientific posture here is uncomfortable and I respect it: the operational definition stays, the chemical definition gets demoted, and everyone agrees we will keep using the word “GRSP” while continuing to argue about what it is. There’s a 2020 review titled Glomalin – Truths, Myths, and the Future of this Elusive Soil Glycoprotein that essentially says: we oversold this, we’re not unselling it, and the carbon numbers people quote are probably wrong but the underlying soil-structure story is probably right. That is so much more interesting than “scientist discovers protein, protein saves planet.” It’s a parable for how categories in living-systems science actually mature — first you find a thing, then you realize it isn’t a thing, then you decide whether the not-thing is still useful, and usually it is, just for different reasons than you thought.

My opinion: the hero narrative (carbon-storing fungal glycoprotein!) is wrong; the deflated narrative (oh, GRSP is just humic acid contamination, never mind) is also wrong; the honest narrative is that mycorrhizal fungi are doing something sticky and durable to soil, the assay we invented happens to grab a fingerprint of that something along with a bunch of unrelated stuff, and the fingerprint is still useful even though it isn’t a portrait. What I keep circling back to is this: how much of biology and ecology is actually like this? How many of our nouns — “the immune system,” “the microbiome,” “the gut-brain axis” — are operational categories pretending to be entities? When we say a thing has a name, we are usually claiming we’ve cut nature at a joint. With glomalin, the joint moved. What else am I confidently naming that is really just what comes out when I extract a particular way?

Sources:

— Shelle
Curiosity Lab · ficientdesign.com