In a development that may lead to less expensive drug production processes mediated by enzymes, Berkeley Lab chemists have successfully replaced a natural compound used in biocatalytic reactions with a simpler and more stable analog.

The team stripped a naturally occurring molecule used as a cofactor in enzymatic reactions — one with the jawbreaking name 1,4-dihydronicotinamide adenine dinucleotide, more often called 1,4-NADH — of its sugar, phosphate, and adenosine groups. The simpler compound, called a biomimetic because it mimics the more complex molecule, was then used instead of its naturally occurring counterpart in a common catalytic cycle in which an enzyme produces a pharmaceutically important molecule called a chiral compound.

	Enantiomers are mirror-image molecules that are structurally equivalent but cannot be superimposed because of their "handedness," or chirality. Also called optical isomers, in pure solutions one of a pair rotates the plane of polarized light to the right, the other to the left.

Chirality refers to the "handedness" (right or left handed) of molecules which occur in two forms, mirror images which cannot be superimposed upon each other. Such molecules are called enantiomers. Chiral compounds used in the pharmaceutical industry are characterized by their optical purity as single enantiomers — overwhelmingly right handed or left handed — a critical requirement in large-scale drug production processes.

"We fooled the enzyme into thinking our stripped-down, reduced cofactor mimic is the natural compound," says Richard H. Fish of Berkeley Lab's Environmental Energy Technologies Division. "This marks the first time, as far as we can tell, a biomimic of 1,4-NADH has been used to produce a chiral compound."

The team's success, although still in the experimental stage, promises to streamline a sometimes expensive drug production method using enzymes. In this process, enzymatic reactions produce chiral molecules of one optical form. Drug manufacturers favor this enzyme-based method over processes that rely on synthetically formulated chemicals because enzymes require less energy to perform reactions. Millions of years of refinement ensure that an enzyme's binding sites are as molecule-specific and energy efficient as possible. This translates to faster reactions and potentially less expensive chiral synthesis, an advantage not easily matched by most synthetic processes.

But there's one drawback. In order for an enzyme to produce a chiral product, it must obtain electrons from a cofactor. Traditional cofactors, such as the phosphate-and-sugar-equipped 1,4-NADH molecule used in Fish's research, supply these electrons in the form of hydride ions. Unfortunately, the complex 1,4-NADH molecule, like many cofactors, is unstable in air and water and therefore has a shorter lifetime. Specifically, it is susceptible to hydrolytic reactions that degrade the compound. This means that an expensive 1,4-NADH cofactor may be required to make a less expensive pharmaceutical product, meaning in some processes the cofactor is more expensive than the product.

What's needed is a cofactor that's stable and fits seamlessly into complex enzymatic reactions. Fish's team wondered if a biomimetic cofactor offered the solution. In other words, replace the bulky, natural 1,4-NADH molecule with a leaner, more stable mimic. Before setting their sights on chiral synthesis, however, the team first had to ascertain whether this simpler version works in a fundamental catalytic cycle that regenerates the needed biomimetic cofactor, 1,4-NADH — a process that essentially provides the fuel for biocatalysis. In this catalytic cycle, the oxidized biomimetic nicotinamide adenine dinucleotide (NAD+) cofactor grabs a hydride ion from an organometallic rhodium catalyst. This hydride ion must always attack the oxidized cofactor's carbon in the number four position, even though the hydride ion has two other positions (two and six) to choose from. This site-specific targeting, called regioselectivity, must occur over and over again in order to sustain cofactor regeneration.

But does the simplified molecule exhibit the same regioselectivity as its more complex cousin? To find out, Fish's team pared down the NAD+ molecule to its simplest form, replacing its sugar, phosphate and adenosine groups with a simple benzyl group. They introduced this biomimetic version into the catalytic cycle, and found that the regioselectivity remains unchanged. The hydride ion attacks the simpler version's number four carbon site to produce the reduced cofactor, 1,4-NADH, just as it does in the naturally occurring compound.

"We can use the simplified NAD+ analog or the natural product; either way it's the same," Fish says.

This solved the first piece of the puzzle. The hydride ion successfully attacks the correct position on the stripped-down NAD+ molecule, a critical intermediate step that enables the molecule to present the hydride ion of the reduced cofactor to an enzyme.

Another, bigger roadblock loomed, however. Would the enzyme even recognize the reduced biomimetic cofactor molecule? Enzymes are efficient because they are engineered to interact only with specific molecules. This also makes them finicky. Change a molecule even slightly, and the enzyme may not accept it. Fish's team had to determine whether the stripped-down, biomimetic 1,4-NADH is convincing enough for the enzyme's finely tuned selection process.

They melded a catalytic cycle that regenerates the 1,4-NADH cofactor with an enzymatic cycle in which a horse-liver alcohol dehydrogenase enzyme accepts hydride ions from the 1,4-NADH to produce a chiral alcohol. Instead of using a naturally occurring 1,4-NADH molecule that readily interacts with the enzyme, they introduced their biomimetic version, stripped of its sugar, phosphate, and adenosine groups. After running the reaction, they determined that biocatalysis was successful: the biomimetic 1,4-NADH cofactor offered its hydride ion to the enzyme, and the enzyme easily accepted the hydride ion as if it were interacting with the natural compound. This, in turn, precipitated an enzymatic process that resulted in a chiral product. In addition, the reaction consistently produced a single chiral product, which may suggest that biomimetic 1,4-NADH compounds can be used in large-scale pharmaceutical and industrial processes.

Ultimately, the more such biomimetic cofactors are found to mesh with a variety of enzymes to produce pharmaceutically valuable products, the cheaper drug production might become, Fish says.

	Richard H. Fish of Berkeley Lab's Environmental Energy Technologies Division.

"If you can use a cheaper biomimetic cofactor over and over again without it decomposing, then you don't have to constantly reinvest in replenishing the cofactor — this saves money," Fish says. "This also opens the door for researchers to determine whether other enzymes can produce pharmaceutical products using biomimetic cofactors."

Berkeley Lab's John B. Kerr and former postdoctoral fellows H. Christine Lo, Carmen Leiva, and Olivier Buriez also contributed to this research.

"Biomimetic NAD+ models for tandem cofactor regeneration, horse liver alcohol dehydrogenase recognition of 1,4-NADH derivatives, and chiral synthesis," by H. Christine Lo and Richard H. Fish, appeared in Angewandte Chemie International Edition, 2002, 41, No 3.