L-Glucose and the $50,000 Sweetener

A sugar exists that tastes identical to glucose. It has zero calories. Your body cannot metabolize it. It passes through you unchanged, like water through a sieve. If it were cheap, it could change the daily lives of more than 500 million people with diabetes. It has been known since 1976. It costs between $31,000 and $250,000 per kilogram. Regular glucose costs about fifty cents.

The molecule is L-glucose. Its story is not about chemistry failing. It is about chemistry succeeding in one direction, four billion years ago, and never looking back.

---

The molecule that does nothing

L-glucose is the mirror image of D-glucose. Same atoms. Same bonds. Same molecular weight. Every hydroxyl group points the opposite way. If you placed the two molecules side by side, they would relate to each other the way your left hand relates to your right: identical in composition, impossible to superimpose.

This matters because biology is handed. The molecular machinery that processes sugar in your body was built for D-sugars. SGLT1, the transporter that pulls glucose from your gut into your bloodstream, recognizes the shape of D-glucose and ignores L-glucose entirely. Hexokinase, the enzyme that phosphorylates glucose to begin glycolysis, won't touch it. The molecule enters your mouth, reaches your intestine, and leaves. No absorption. No metabolism. Zero calories, not as a marketing approximation, but as a thermodynamic fact.

The interesting part is what happens before the gut. In 2021, Dubovski and colleagues published a study in which trained taste panels compared D-glucose and L-glucose solutions at matched concentrations. The panels could not distinguish them. The molecular basis is the TAS1R2/TAS1R3 sweet taste receptor, which has two binding sub-pockets arranged so that either enantiomer fits. The receptor does not care about handedness. It responds to both.

Your tongue cannot tell D-glucose from L-glucose. Your metabolism can. That gap between perception and processing is the entire premise of a zero-calorie sweetener that tastes exactly like sugar. Not 70% as sweet. Not sweet with an aftertaste. Identical.

---

The man who tasted Mars

Gilbert V. Levin was not thinking about diabetes when he discovered this. He was thinking about Mars.

Levin was a sanitary engineer turned astrobiologist. In the early 1970s, NASA selected his Labeled Release experiment for the Viking mission: the first attempt to detect life on another planet. The experiment worked by feeding Martian soil a nutrient broth spiked with radioactive carbon. If anything in the soil metabolized the nutrients, it would release labeled carbon dioxide, and the detector would see a signal.

Levin had a problem. Terrestrial life uses D-sugars. If Martian life existed, it might use the opposite handedness. An experiment that only offered D-nutrients could miss half the possibilities. So Levin included L-glucose in his nutrient mix as a control, a sugar that Earth organisms would ignore but hypothetical mirror-image life might consume.

Viking landed on Mars in 1976. The Labeled Release experiment produced a positive signal that remains debated to this day. But Levin noticed something else. Back in his lab at Biospherics Inc. in Beltsville, Maryland, he tasted the L-glucose. It was sweet. Indistinguishably sweet.

He ran a formal taste panel. Confirmed it. Filed US Patent 4,262,032 in 1977, granted 1981, claiming L-glucose as a non-caloric sweetener. Then he spent the next two decades trying to make it cheaply.

He failed. The synthesis was too long, too expensive, too inefficient. Levin eventually pivoted to D-tagatose, a naturally occurring rare sugar that could be manufactured from lactose. He patented that in 1988. Arla Foods licensed it in 1996. The FDA granted it GRAS status in 2001. It reached supermarket shelves in 2003 as Naturlose.

Levin died in July 2021, at 97. L-glucose was still too expensive for anyone to buy in bulk. The compromise sugar he invented as a fallback became a commercial product. The perfect one never did.

---

Fifty percent more than gold

Here is what L-glucose costs today, if you want to buy some.

Sigma-Aldrich sells 500 milligrams of L-glucose for approximately $247. That extrapolates to $494,000 per kilogram at research-grade pricing. Biosynth Carbosynth offers it at a larger scale: roughly $31 per gram at the 25-gram tier, or about $31,000 per kilogram. Gold, as of early 2026, trades near $145,000 per kilogram. Depending on the supplier and quantity, L-glucose costs somewhere between a quarter and twice the price of gold.

Levin used to say his sweetener was worth more than gold by weight. He was not exaggerating.

The cost is not arbitrary. It reflects the synthesis. D-glucose is one of the most abundant organic molecules on Earth. Plants produce it by the hundreds of billions of tons through photosynthesis. L-glucose does not occur in nature. No organism produces it. No enzyme catalyzes its formation. To make it, you must chemically invert every stereocentre of a D-sugar through a multi-step synthetic route.

The classic approach is Szarek's 1984 synthesis from D-galactose, involving protection, inversion, and deprotection steps with chromatographic purification at each stage. In 2013, a streamlined route was published. A 2014 patent eliminated the need for column chromatography, a significant improvement. Each advance has trimmed the cost. None has achieved the order-of-magnitude reduction needed for a food-grade product.

The fundamental constraint is that L-glucose sits on the wrong side of biology's mirror. Every shortcut that makes D-sugars cheap (enzymatic conversion, microbial fermentation, agricultural extraction) relies on the fact that life already makes D-sugars. For L-glucose, there is no biological infrastructure to exploit.

---

The sugars that settled

While L-glucose remained locked behind synthesis costs, two other rare sugars made it to market.

D-allulose is about 70% as sweet as sucrose and provides somewhere between 0.2 and 0.4 kilocalories per gram (sucrose provides 4). D-tagatose, Levin's consolation prize, is roughly 92% as sweet and provides about 1.5 kilocalories per gram. Both are real products. You can buy D-allulose on Amazon. It is in commercial food formulations. The FDA recognizes it as GRAS and, since 2019, has excluded it from total and added sugar counts on nutrition labels.

The reason these two succeeded where L-glucose failed is enzymatic conversion. D-allulose is made from D-fructose using a single enzyme: D-psicose 3-epimerase. D-tagatose is made from D-galactose using L-arabinose isomerase. One enzyme, one step, one cheap feedstock. That is the formula for an affordable rare sugar.

The intellectual framework behind this approach came from Ken Izumori at Kagawa University, who mapped what he called the Izumoring: the complete network of interconversions among all 34 hexose sugars. The strategy is elegant. Start from a cheap, abundant sugar. Find or engineer an enzyme that converts it to the target in one or two steps. The Izumoring tells you which conversions are possible and which enzymes you need.

It works beautifully for D-series sugars. D-fructose to D-allulose is a single epimerization. D-galactose to D-tagatose is a single isomerization. But the Izumoring has a boundary. D-sugars and L-sugars sit on separate halves of the network. Crossing from D to L requires inverting every stereocentre simultaneously. No known enzyme does this. The strategy that solved D-allulose breaks completely at the D/L boundary.

There is one more complication. Not all L-sugars are metabolically inert. Batt and Quimby showed in 1995 that L-fructose and L-gulose are partially fermented by human gut bacteria, producing short-chain fatty acids with caloric value. L-glucose is unusual even among L-sugars: it is the one that passes through the human body completely untouched. The perfect candidate is also the hardest to make.

---

A pathfinding problem

Here is the reframe that changes how you think about this.

Making L-glucose is a graph problem. There are hundreds of monosaccharide stereoisomers. There are hundreds of known chemical and enzymatic reactions that interconvert them. The evidence for these reactions is scattered across decades of literature, ranging from well-validated industrial processes to single published reports to reactions that should work in principle but have never been tested.

The question is not "can we make L-glucose?" The question is: what is the shortest route from a cheap starting material to L-glucose, with reliable evidence at every step?

This is a pathfinding problem, the same class of problem that powers GPS routing and network optimization. Except the graph has never been fully assembled. No single database covers all monosaccharide interconversions. The evidence quality varies enormously. And the nodes and edges span organic chemistry, enzymology, and microbial metabolism.

This is what SUGAR was built to do. It enumerates 135 monosaccharide compounds and 696 reactions generated from first principles of stereochemistry. Each reaction is assigned one of four evidence tiers: validated (demonstrated at scale), published (reported in literature), inferred (analogous to a known reaction), or hypothetical (chemically plausible, never tested). The tool runs client-side pathfinding across this network.

What the graph reveals is the shape of the problem. D-glucose sits in a dense, well-connected region of the network. Multiple validated routes connect it to other D-sugars through short paths. But crossing to L-glucose requires traversing inferred or hypothetical edges. There is no short, fully validated path from any cheap starting material to L-glucose. The validated chemistry runs out before you get there.

This is not a proof that the route doesn't exist. It is a map showing where the known roads end and the unmapped territory begins. The difference matters. A blank map and a map with "unknown" marked on it are not the same thing. The second one tells you where to explore.

---

Almost fifty years have passed since Gilbert Levin tasted L-glucose in a laboratory in Beltsville, Maryland, and realized what he had. The molecule is still too expensive for anyone to put in food. The enzymatic route that would make it cheap does not exist. D-allulose sits on supermarket shelves as a compromise: pretty good, not perfect, good enough for now.

More than 500 million people manage their blood sugar every day. The perfect sweetener for them, the one that tastes exactly like glucose and provides exactly zero calories, is locked behind a synthesis problem that biology never solved because biology never needed to. Life chose D-sugars roughly four billion years ago. Everything since has been built on that choice.

Whether someone finds the right enzyme, engineers the cascade, or discovers a microbial route that crosses the D/L boundary is an open question. The landscape is mapped. The path through it is still missing.