02. Glutamate: The Oldest Word

In a laboratory at Tokyo Imperial University in 1907, the chemist Kikunae Ikeda is boiling seaweed. The stock is konbu dashi, the kelp broth beneath most of Japanese cooking, and Ikeda has decided its taste is none of the four his textbooks allow. Over months he works through some ninety pounds of kelp, stripping out salts and sugars a fraction at a time, until about an ounce of brown crystals is left. He tastes them: not sweet, sour, salty or bitter, but a fifth thing, savoury and mouth-filling and slow to leave, and he names it umami. The crystals were glutamic acid. A patent followed in April 1908, granted that July; by 1909 the Ajinomoto company was making monosodium glutamate by the ton.

Ikeda's own paper dropped out of the international literature, surviving in English only as a three-quarter-length translation published ninety-three years later, beside Bernd Lindemann's account of its reception. The mechanism took nearly as long. In 2002 Greg Nelson, Nicholas Ryba, Charles Zuker and colleagues reported the receptor in Nature: T1R1+T1R3, a G-protein-coupled receptor — a protein threaded through the membrane that binds something outside the cell and starts a cascade inside — built from two subunits and tuned broadly to L-amino acids, glutamate strongest among them. Ikeda had not invented a taste. He had found a molecule the tongue was already built to notice.

That receptor is the newest thing in this chapter. Glutamate is not, at bottom, a messenger: it is a metabolite, a working chemical of ordinary cell chemistry, made and consumed by cells with no interest in signalling anyone. Glutamate dehydrogenase, present in all three domains of life, runs glutamate ⇌ α-ketoglutarate plus ammonia — the hinge between amino-acid nitrogen and the carbon cycle that burns sugar. Around it sit the aminotransferases, which lift an amino group off one carbon skeleton onto another; comparative genomics places their divergence into four classes before LUCA, roughly four billion years ago. Because they funnel amino groups from most other amino acids through the glutamate/α-ketoglutarate pair, glutamate is not merely abundant. Every cell that handles nitrogen handles glutamate.

Its recognition as a neurotransmitter lagged for decades, and the ubiquity was the reason: a signal was expected to be scarce and dedicated, not the cell's nitrogen currency at five to fifteen millimoles per kilogram of brain tissue. What was missed was most of the brain. Counting cortical excitatory synapses as glutamatergic, David Attwell and Simon Laughlin's 2001 energy budget for grey matter puts roughly ninety percent of synapses in that class, from anatomical counts of 1991 and 1998; a stricter denominator gives thirty to forty percent. Excitation is the default, and the default has no brake of its own.

Traced backward, the receptor leaves the animals almost at once. In 1999 Guo-Qiang Chen, Changhai Cui, Mark Mayer and Eric Gouaux described GluR0 in Nature, from the cyanobacterium Synechocystis PCC 6803 — the first glutamate receptor found in a prokaryote. It is ionotropic, meaning the ligand's binding opens a pore directly rather than starting a chemical relay, and it reads as a bacterial amino-acid-binding protein bolted onto a potassium channel. What it is not is evidence that bacteria talk to one another in glutamate: GluR0 has no known cell-to-cell role; nutrient sensing is likelier. The fold is ancient; the conversation is not.

Plants took the same fold somewhere else. In 2018 Masatsugu Toyota and colleagues reported in Science that wounding one leaf of Arabidopsis releases glutamate from the damaged cells, and a wave of calcium travels through the plant's plumbing to distant, untouched leaves within minutes, switching on defence genes there. The sensors are GLUTAMATE RECEPTOR-LIKE proteins, homologs of the animal channel, running alarm traffic in an organism with no neurons and no synapses. Where that family came from is unsettled: one comparative-genomics line places an ancestral GLR in algae with continuous descent into land plants, another finds land-plant GLRs unrelated to green-algal ones. The mechanism is dated to the year. The lineage is not.

The sharpest result belongs to the other receptor line. In 2006 Donghui Kuang, David Hampson, Belinda Chang and colleagues reconstructed the ancestral ligand-binding pocket of the Family C G-protein-coupled receptors — the class the metabotropic glutamate receptors belong to, metabotropic meaning the receptor starts a cascade rather than opening a pore. Ancestral reconstruction infers an extinct protein's sequence from living descendants, builds it, and tests it. The resurrected pocket already bound L-glutamate with high affinity, at a node long predating any synapse, and that line roots in single-celled eukaryotes. The sensor came before the use.

Inside the animals the ionotropic channels radiate. Àlex Ramos-Vicente and colleagues resolved that radiation in eLife in 2018 into four subfamilies — AKDF, Epsilon, Lambda and NMDA — established within the animal stem, which Vol. I's Bayesian clock brackets between 649.8 and 833.5 million years ago. As Vol. I told through the sponge's empty postsynaptic scaffold and the ctenophore's glutamate-only nervous system, the parts of a synapse do not arrive as a set. Placozoans sharpen that: Trichoplax adhaerens and its relatives have no neurons, no synapses and no nervous system, yet carry working Epsilon- and AKDF-class channels while lacking NMDA outright — and their one characterized Epsilon receptor is not gated by glutamate at all. Glycine, alanine, serine and valine open it, and three substitutions flip its preference to glutamate, as Singh, Bowie and Senatore showed in 2025. Vol. I found scaffold without receptor. This is receptor without a nervous system, and without glutamate.

The Word the Family Did Not Start With

Which of those events, then, is the first glutamate signal? The question is harder than it looks, because the family's name came from its founding vertebrate members rather than from its ancestral ligand. In 2010 Vincent Croset, Richard Benton and colleagues showed in PLOS Genetics that a whole divergent branch of it, the protostome Ionotropic Receptors, was repurposed as the molecular basis of insect smell and taste: same fold, synaptic job abandoned, scaffold turned outward. Ramos-Vicente's own tree holds an amphioxus receptor gated by glycine. Kuang and Chang's resurrection reads as a pocket glutamate-shaped from the beginning; the redeployments read as a generic amino-acid detector on which glutamate is one setting among several. Neither frames its work as a rebuttal of the other; the disagreement sits in the data, not between the people.

The same discipline applies to function. Vertebrate NMDA receptors are coincidence detectors: a magnesium ion plugs the pore and is expelled only when the cell is already depolarized and glutamate is bound, so current passes only when two conditions coincide — the substrate for Hebbian learning, the rule that neurons firing together strengthen their connection. That property is not the family's birthright. In the fruit fly's dNR2 subunit the asparagine residue governing the block is a glutamine instead, and C. elegans NMDA-type receptors may lack it altogether, plausibly tracking the higher magnesium those animals carry. Coincidence detection looks like a vertebrate-side refinement laid over an older, blunter channel, and no ancestral reconstruction of that site exists to date it. The class runs at least as deep as the cnidarian–bilaterian ancestor, yet no vesicular, synapse-specific release of glutamate has ever been confirmed in a cnidarian. Receptor existence and synaptic use are two different commits.

The oldest rows in this log are its firmest and the most familiar one is its weakest. Glutamate's place at the centre of nitrogen metabolism rests on comparative genomics across all three domains; the receptor radiation on a published clock with stated bounds; the property that makes a vertebrate synapse capable of learning carries no date at all, filed at grade D for that reason. What the trace agrees on is the ordering — before the synapse a receptor, before the receptor a metabolite, and the receptor, when it appeared, not yet particular about which amino acid it answered to. What Ikeda tasted in 1907 was the oldest word in the vocabulary, and a vocabulary whose only word means more is not yet a language: the next commit is the one that learned to say no.

Commit Ledger

Overturned if: an ancestral-sequence reconstruction across the animal ionotropic tree shows the ancestral channel was already glutamate-selective, making the placozoan receptor's taste for glycine, alanine, serine and valine a late local quirk rather than a survival of the original state.

Sources

• 1999 · Chen, Cui, Mayer & Gouaux, "Functional characterization of a potassium-selective prokaryotic glutamate receptor" — Nature 402:817–821
• 2001 · Attwell & Laughlin, "An Energy Budget for Signaling in the Grey Matter of the Brain" — J. Cereb. Blood Flow Metab. 21:1133–1145
• 2002 · Nelson, Chandrashekar, Hoon, Feng, Zhao, Ryba & Zuker, "An amino-acid taste receptor" — Nature 416:199–202
• 2002 · Ikeda, "New Seasonings" — Chemical Senses 27(9):847–849
• 2002 · Lindemann, Ogiwara & Ninomiya, "The Discovery of Umami" — Chemical Senses 27(9):843–844
• 2006 · Kuang, Yao, MacLean, Wang, Hampson & Chang, "Ancestral reconstruction of the ligand-binding pocket of Family C G protein-coupled receptors" — PNAS 103:14050–14055
• 2010 · Croset et al. (incl. Kaessmann, Gibson & Benton), "Ancient Protostome Origin of Chemosensory Ionotropic Glutamate Receptors and the Evolution of Insect Taste and Olfaction" — PLoS Genetics 6(8):e1001064
• 2015 · dos Reis, Thawornwattana, Angelis, Telford, Donoghue & Yang, "Uncertainty in the Timing of Origin of Animals and the Limits of Precision in Molecular Timescales" — Current Biology 25(22):2939–2950
• 2017 · Smith, Li, Stanley & Smith, "Glutamate Dehydrogenase, a Complex Enzyme at a Crucial Metabolic Branch Point" — Neurochemical Research 42(10)
• 2018 · Ramos-Vicente et al. (incl. Soto & Bayés), "Metazoan evolution of glutamate receptors reveals unreported phylogenetic groups and divergent lineage-specific events" — eLife 7:e35774
• 2018 · Toyota et al., "Glutamate triggers long-distance, calcium-based plant defense signaling" — Science 361(6407):1112–1115
• 2020 · Serrano-Saiz et al. (incl. Hobert), "Modular Organization of Cis-regulatory Control Information of Neurotransmitter Pathway Genes in Caenorhabditis elegans" — Genetics 215:665–681
• 2022 · Koper, Han, Casas Pastor, Yoshikuni & Maeda, "Evolutionary origin and functional diversification of aminotransferases" — J. Biol. Chem. 298:102122
• 2025 · Singh et al. (incl. Bowie & Senatore), "Evolution of iGluR ligand specificity, polyamine regulation, and ion selectivity inferred from a placozoan epsilon receptor" — Communications Biology 8:994
• 2025 · dos Santos, Melgarejo, Ligabue-Braun & de Oliveira, "Phylogenetic and Structural Analyses of Vesicular Glutamate Transporters" — Molecular Neurobiology 62:11962–11971