Flavor alchemy

Wines have good years and bad years. And so do oranges. With some crops the oranges mature well and the juice has the right balance of sweetness and tartness. But for most, the sweetness may be off or the variety has delayed biterness. Nothing like the crop from 1999.

The secret to removing the bitterness are plastic beads. For the beads to work, the orange juice needs to be separated into pulp and liquid. The liquid has most of the limonin, the bitter compound in orange juice. When the beads come in contact with the liquid, the limonin molecules get trapped in the beads (along with some other useful chemicals, unfortunately).

After some time the columns of beads need to be cleaned so as not to loose their capacity to trap the limonin. That is done with ethanol, the alcohol in wine or beer. The ethanol is then removed by washing the beads in stages with hydrochloric acid (HCl), sodium hydroxide (NaOH), and water. The idea is that the hydrochloric acid and the sodium hydroxide neutralize each other. After the removal of the limonin (a process called de-bittering), orange juice is reconstituted by adding back some of the pulp and maybe even some orange oil extracted from the peel.

Some of the tricks used to de-bitter citrus juices may work for the cook. From filtering techniques, to adding chemicals or microorganisms, many different ideas have been tried. Most ideas fail in practice because they either remove too many components from the citrus juice, affecting flavor or nutritional value; or because the method does not work in practice for the millions of gallons of juice that need to be processed. But on the smaller scale, these ideas may be practical.

The beads used for orange juice are a made from a styrene divinylbenzene cross-linked co-polymer resin. A long word but not a complicated concept. A polymer is a large molecule made linking together many copies of a few groups of atoms. The groups that get repeated are monomers and often differ from a molecule by the position of a few bonds. Styrene is one of those molecules that by the alteration of one bond can join with another styrene molecule to form a chain of molecules: the polymer polystyrene. Sometimes two different monomers will repeat to form a large molecule, which is then called a co-polymer. Polymers chains can be linked together with the introduction of atoms or small molecules. This freezes the co-polymer chains in a rigid network of molecules, such as when sulfur helps link the polymers of isoprene to form the more durable vulcanized rubber.

Rohm and Haas, a chemical company of Plexiglas and Morton Salt fame, introduced in the 1980s a new kind of cross linked co-polymers. They create the co-polymers in a solution containing a filler material that eventually gets washed away leaving behind large a network of tunnels in the beads. These tunnels are 0.1 to 0.3 micron in diameter, large for the gaps in a cross-linked polymer, but sufficiently small to trap other molecules. With this new polymer it was possible to absorb the limonin from orange juice and many other bitter compounds.

How the oranges get juiced helps with the flavor. The machine of choice is the FMC juice extractor. It has been around long enough to be mentioned in McPhee’s classic book on oranges. Used commercially since 1947, the FMC machine does not slice the orange nor does it mix the peel oil with the juice. Instead, the orange is punctured by a pipe with sharp edges. A plunger then presses the orange, forcing the pulp into the pipe. Within the pipe the pulp is squeezed and the juice is collected in a lower tank. While this is taking place, the oils from the peel have been collected by washing the peel while its being pressed. All this takes place in less than a second.

Orange juice has no added sugar. The not-from-concentrate (NFC) juice sold in the refrigerated sections of supermarkets (typically Tropicana or Minute Maid brands in the United States) has no sugar listed in the list of ingredients, yet the juice tastes sweeter than any juice I have ever squeezed.

Adding to this mystery is the fact that the orange juice from Navel oranges turns bitter after some time. In the United States, the two most common oranges are the Valencia and the Navel. Valencias are harder to peel and have seeds, so when available, Navels are the choice for eating. To taste the difference in flavor between the fresh squeezed juice and the one from the carton, I tried a little experiment.

I found what tasted like ripe and fresh Valencia and Navel oranges. I made two half-cups, each from one variety. I also peeled a Navel orange and puréed it in the blender. I left all three containers tightly covered in the refrigerator overnight. Next day I poured each of the juices in a small cup together with some NFC and after they reached 12ºC I tried them in order: Navel, purée, Valencia, NFC.

Result: the tasting order was from the most bitter to the least. When first tasting the Navel orange juice it is hard to notice it is bitter. One can tell it is not as good as other orange juices. The aromas of orange are still there, together with the sweet and sour tastes, but the bitterness is not overwhelming and only becomes evident in contrast with the next cups of juice. There was no hint of bitterness in the NFC, although it lacked some of the “brightness” of the fresh squeezed juices. The purée tasted as a diluted version of the Navel orange juice.

Orange juice is bitter because it has limonin in it. Limonin is a very bitter compound from the terpenoid family. The limonin is not present in the orange, but forms in a few hours after the juice has been squeezed. The process can be accelerated by heat. Three quarters of people can detect just 6 milligrams of limonin in one liter of water. That is 6 parts to a million! Around one-tenth of the population is even more sensitive and will detect as little as 0.5 parts per million. Limonin and its molecular cousins naringin and nomilin are responsible for the bitterness of citrus juices. Even Valencias have some limonin.

Limonin was discovered in 1841 in the seeds of oranges by S. Bernay. It wasn’t until 1938 that Ralph Higby isolated limonin from the juice of Navel oranges. Because there is no limonin in a fresh orange there had to be some other compound that was giving rise to limonin. It was only in 1969 that Vincent Maier identified LARL as the precursor to limonin. LARL is a tasteless substance. When the orange is juiced, the juice sacs are broken and the acidic juice and an enzyme promote the conversion of LARL into limonin.

Not all farming regions can produce Valencias year-round. Florida’s weather in the United States is better suited for Valencias than California’s, where Navels are more common. Valencia oranges also take longer to grow, they are a winter harvest. The constraints of farming require that different varieties of oranges be used for the NFC juice. As NFC is not the deep orange of Valencia juice, other varieties must go into the NFC. Around half of these turn bitter with time.

Orange juice is big business. The Agricultural Research Service’s (ARS) lab in California where Maier worked for most of his career has been a hub for citrus science research. Maier’s result is part of a long (and ongoing) program at ARS. That 1969 they also isolated the enzyme that helps convert LARL to limonin, and they continued studying the biochemistry of LARL and related compounds. LARL was present in all major citrus fruits, but it was only in 1989 that Shin Hasegawa, also with the ARS, found the enzyme LGT that helped attach a sugar molecule to LARL transforming it into another tasteless substance, limonin glycoside, that if in the juice would not lead to bitterness. The rate of transformation in Valencia and Navel oranges was probably different, but the experiment took two years to carry out.

In an orange the amount of LARL increases as the fruit grows. For fruit in the United States, the peak for LARL happens in the summer. As the fruit mature on the tree the LGT enzyme transforms the LARL into glycoside. Because the Navel oranges are harvested two months after the peak of LARL, not much of it has been converted, while Valencias are harvested 6 months after the September peak, when much of the LARL has been converted. Valencias get all the time to de-bitter.

An easy guess would be that LGT added to a citrus juice would convert the LARL into the tasteless limonin glycoside, but that is not how NFC looses its bitterness. For how it happens one has to delve into the art and science of citrus de-bitering.

Besides the the six tastes (it’s six because of fat, see a previous post), we also have other sensations in our mouth. We can sense when something is spicy hot, as when food has cayenne or wasabi, and when it has a chemical coolant, such as the menthol in chewing gum. These sensations are different from tastes because they rely on little (molecular little) tubes on the nerve cells. They are not in our taste buds but in our mouth and throat. These little tubes allow certain molecules to enter the cells and change some of the chemical reactions that are going on inside. This is how we perceive spicy hot, temperature, and a few other things no one is quite sure.

These little tubes are known as TRP ion channels. They let ions and some small molecules in and out of cells. Because they are made out of proteins that look like each other, biologists have checked that the human genome codes for at least 28 different TRP ion channels. The fun part for scientists is that most of them have unknown functions. One of them, trpa1, seems to detect low temperatures (under 15ºC) but also the hotness related to mustard. Maybe there is some room for culinary experimentation with mustard or wasabi ice creams. (Exotic ice creams are now all the rage. Heston Blumenthal churned mustard ice cream early on, but I’ve never tried it.) The little ion tubes can be desensitized by certain chemicals, which opens up other culinary possibilities.

Sour and salty seem to use mainly TRP channels, but the other tastes use specialized cell membrane molecules called GPCR. The GPCR communicate to the inside of the cell the presence of some molecule on the outside. Say you have just eaten a nice piece of chocolate cake. The sucrose from the sugar will find its way to a taste bud. A taste bud has many taste receptor cells (often abbreviated TRC) and each cell will have a different collection of GPCR. Sucrose will stick only to one type of GPCR called T1R2+T1R3. The GPCR with the captured sucrose moves along the cell membrane and with the help of other molecules produces camp, which is then used to activate enzymes and start other reactions in the taste receptor cells that will signal the brain that a sucrose was inside your mouth.

We have one GPCR for sweet, one for umami, and it looks like one for fat. But we have over 30 for bitter. We cannot tell one bitter taste from another: the GPCR for bitter tend to be on the same taste receptor cell, so the brain cannot tell them apart.