By David N. Leff

"To find a dextro-amino acid in humans," observed neuroscientist Solomon Snyder, "is unprecedented. It's the equivalent of finding a pterodactyl [bat-winged flying dinosaur] in your local pet shop."

Snyder, director of neuroscience at The Johns Hopkins School of Medicine, and his co-workers have in effect done just that. Back in 1993, a Japanese group reported encountering the oddity - a D (dextro)-serine amino acid - in the brains of rats.

"This was against the laws of nature," Snyder said. "There are no D-amino acids in animals. In insects and bacteria, yes; in mammals, no."

When he saw this contrarian report, Snyder told Michael Schell, his grad student at the time, "Here's an interesting challenge about something that, if it's for real, would be extraordinary. Schell," he recalled, "was very skilled at making selective antibodies to small molecules. So he made an antibody so we could map the D-serine amino acid, and confirmed that it really was present in the rat brain.

"The Japanese," Snyder continued, "had reported that the D-serine concentration was highest in the forebrain. We saw the same thing, except that as we were doing it at a microscopic level, we could get very detailed localizations. And we found D-serine right where you have NMDA receptors for glutamate." (NMDA stands for N-methyl D-aspartic acid.)

"Glutamate," he continued, "is the major excitatory transmitter in the brain, and NMDA is the best-known glutamate receptor. When you have a stroke," he related, "there's a massive release of glutamate - like a 50-fold elevation. And that excites neurons to death. Therefore, at least half of the neural damage following a cerebral stroke is due to the excess glutamate-stimulating NMDA receptors.

"NMDA receptor antagonists reduce stroke damage by 50 to 60 percent," Snyder pointed out. "Some of them are in clinical trials. We thought that was particularly interesting because the NMDA receptor is unique, and glutamate is probably one of the most abundant substances in the brain."

It Takes Two To Rein In Glutamate

Snyder asked rhetorically: "How could the NMDA receptor avoid being accidentally stimulated by this glutamate substance just swimming around in the brain? His answer: "Nature was very clever, and made the NMDA receptor a fail-safe. You have to stimulate it not just by glutamate but at a different site as well. And that second site, discovered 10 years ago, people thought was for the amino acid glycine, which does indeed activate NMDA receptors.

"But interestingly, when people had been studying the glycine site of the NMDA receptor, they'd made lots of amino acids, to see the properties of the site. And whenever you do structure-activity analysis, you always use D-amino acids because they're not metabolized; they're just good tools.

"And lo and behold," Snyder recounted, "we found D-serine to be more potent than glycine. So that made us think maybe this D-serine is the endogenous agent that regulates the NMDA receptor. This idea became even more tantalizing when we looked at high-power magnification, and saw that the D-serine was localized right where there are NMDA receptors, which is in gray matter where there's mostly neurons and synapses. It was in a subtype of glia called protoplasmic astrocytes."

Those astrocytes did indeed have glutamate receptors that released D-serine. "So that looked pretty cool," Snyder recalled, "but still could we suggest a D-amino acid? They're not supposed to exist in all animal biology. That was radical. People weren't going to pay any attention.

"So to get evidence for this finding, we took advantage of something very interesting. In 1935, Hans Krebs of the Krebs cycle - the great biochemist - had discovered an enzyme that he called D-amino acid oxidase - because it oxidized only D-amino acids. And nobody paid any attention because everybody knew there aren't any D-amino acids.

"So we stained for it, and its localizations were the exact opposite of D-serine, suggesting that maybe Krebs' enzyme gets rid of D-serine. We looked carefully at its specificity. It's actually the only D-amino acid enzyme on which it works. So we purified a lot of it, added it to rat brain preparations, and showed that it indeed destroyed the D-serine and nothing else. And then the NMDA neurotransmission dropped off. So that convinced me the fail-safe was for real. But it raised the question: How on earth could the body make a D-amino acid anyhow?"

Snyder's present postdoc, Herman Wolosker, "showed it was actually very straightforward. He isolated an enzyme that converted L-serine to D-serine by a process called racemization."

Snyder is senior author of a progress report in the current Proceedings of the National Academy of Sciences (PNAS), dated Nov. 9, 1999. Its title: "Serine racemase: A glial enzyme synthesizing D-serine to regulate glutamate-N-methyl-D-aspartate neurotransmission."

Therapeutic Payoff For Oddball D-Serine

"This is of clinical importance," Snyder told BioWorld Today, "because once you have the enzyme, you can look for enzyme inhibitors, which drug companies will surely do. We already know that drugs that block the glycine site of the NMDA receptor in animal models, block stroke damage. So if you inhibit the enzyme that makes D-serine, you might get the same effects. In general in biology, and in pharmaceutical research, if you have a substance that you want to have less of, you obviously could inhibit the enzyme that makes it, or block its receptor. We'll find out. Now that we have serine racemase, drug companies can develop inhibitors of that enzyme, which we're fairly confident will also block NMDA transmission.

"[Johns] Hopkins," he observed, "is in the process of licensing the patents on serine racemase, and that's not complete yet to my knowledge."