Issue Date: July 18, 2005
A DREAM REALIZED
It's not often one finds solid proof of this magazine's impact on the outside world. Here's one: A story on chiral drugs (C&EN, Oct. 9, 1995, page 44) helped bring vibrational circular dichroism (VCD) spectrometers to the market, says Rina K. Dukor, cofounder and president of the company that first offered these instruments off the shelf, in 1997.
How VCD was commercialized, however, is mostly a story about a dream and the single-mindedness to make it come true. The dream came to Dukor in 1994. While at a conference listening to lectures on VCD, she realized that the few people who had devoted their research to this technique were close to retiring and that the science VCD embodies could disappear because no commercial instrument existed. On the plane going home, she formulated the dream: to commercialize VCD. The vehicle would be the company she would form: BioTools.
VCD is a type of vibrational spectroscopy that relies on the difference in a molecule's absorbance of left and right circularly polarized infrared radiation. The technique combines the structural specificity of IR spectroscopy with the stereo chemical sensitivity of circular dichroism. Measurement of VCD generates both IR and VCD spectra. Enantiomers yield identical IR spectra and identical but oppositely signed VCD spectra.
The VCD spectrum of a chiral compound can be calculated by using density functional theory. By comparing the calculated and measured VCD spectra, one can determine a chiral molecule's absolute configuration. As the technique gains wider acceptance, VCD may increasingly be used to obtain this information, which previously came primarily from X-ray crystallography.
VCD's use extends beyond determining absolute configurations, however. It can be applied to conformational analysis of compounds in solution, determination of enantiomeric excess, and monitoring of asymmetric reactions.
DUKOR SHARED her dream with VCD pioneer Laurence A. Nafie, a chemistry professor at Syracuse University and the inventor of Fourier transform VCD, the form that is now embodied in commercial instruments. He reported the first FT-VCD spectrum in 1979. Because he chose not to protect the intellectual property, it moved into the public domain a year later.
In the mid-1990s, Nafie also was thinking about bringing VCD to market. At the time, measuring VCD spectra was mostly an ivory-tower exercise by experts, he says. Instrument makers were unsure of VCD's commercial prospects because the market was not obvious, he recalls.
Enter the C&EN story about how rapidly the market for chiral drugs is growing. That story, Dukor says, showed her that designers and makers of bioactive chiral molecules would be the market for VCD instruments. That market would welcome a tool that can reliably deliver the information from crystallography at less expense and within a shorter time, she reasoned. Armed with the C&EN story, she convinced Nafie to partner with her to fulfill their mutual dream.
After they forged their partnership in 1995, other things fell into place. With Nafie as her partner in BioTools, the C&EN story, and a more detailed market analysis in hand, Dukor persuaded the Canadian instrument maker ABB Bomem to build the instrument. Then, she made her first sales pitch.
In October 1995, Dukor learned that University of Southern California chemistry professor Philip J. Stephens was looking for an instrument maker to help him assemble a VCD spectrometer. Dukor says she asked him not to commit to anything until he heard back from her. Seven months later, she presented to Stephens the plans of BioTools and ABB Bomem to bring VCD instruments to market and clinched her first sale.
Meanwhile, Gaussian, the maker of structure calculation software, was planning to commercialize software to calculate VCD spectra.
To calculate the spectrum, one first has to decide the most likely conformations a molecule will adopt and the relative contribution of each conformer to the equilibrium distribution of molecules in solution. Each conformer is subjected to the computational theory to take it to the lowest possible energy state at equilibrium, and the software generates a corresponding spectrum. Finally, the software takes each spectrum, weights it according to the conformer's relative contribution, and calculates a spectrum for the equilibrium composition of conformers. That final spectrum "reflects what's happening in a measurement," Nafie says. "We are able to get conformations of molecules in solution. That's why this technique is very valuable."
The first VCD spectrometer built for off-the-shelf use came to life on Jan. 26, 1997, according to Dukor and Nafie. Since then, 22 units--under the brand name ChiralIR--have been sold in the U.S., Europe, and Asia. Dukor expects BioTools to sell up to 10 spectrometers this year, mostly to pharmaceutical companies, where the instruments, each costing about $120,000, are gaining wide use in determining absolute configuration.
The absolute configuration of chiral compounds is critical in understanding structure-activity relationships and in developing appropriate chiral separations, resolutions, or syntheses, says Oliver J. McConnell, director of discovery analytical chemistry at Wyeth Research. Discovery chemists always want to know as early as possible the absolute stereochemistry of their syntheses and resolutions, says Jack Z. Gougoutas, a distinguished research fellow at Bristol-Myers Squibb.
The primary tool for determining absolute configuration is single-crystal X-ray crystallography, which requires high-quality crystalline material. When that is not available, nuclear magnetic resonance spectrometry methods or stereospecific syntheses are alternatives, says Amid Salari, research manager for analytical chemistry at Roche Palo Alto.
VCD is slowly joining these well-established techniques, Dukor says.
"The main advantage of VCD is that it does not impose any limit on the physical state of the sample," Salari says. "One can obtain information from crystalline, amorphous, liquid, or oily samples. VCD is also more cost-effective and, in the hands of an experienced user, provides faster turnaround."
"At Wyeth, VCD complements other techniques for determining absolute stereochemistry, as it does not require either derivatization or crystallization," McConnell says.
For Gougoutas, it is important that VCD is a bulk measurement. With single-crystal X-ray crystallography, all the information comes from only one crystal, he points out. Even if the crystal is from a sample that is 99+% pure, there's no way of knowing that it's not the one in a hundred that is aberrant in that sample. So "you need some other method to correlate conclusions based on one crystal to the properties of a bulk sample, and that's easier said than done," he observes.
WITH VCD, "the biggest pitfall is the possibility of missing an important conformer," Nafie says. When that happens, the calculated and measured spectra will not show a good fit, and predictions of absolute configuration cannot be made with high confidence.
Nafie and others have been working on criteria to assess whether a prediction can be made with confidence. Douglas J. Minick, a researcher at GlaxoSmithKline, is preparing a paper on the subject based on hundreds of analyses and predictions he has made in the past five years.
The key to confident predictions at GSK is the value of the statistical parameter R2, in which R is the coefficient of correlation between the intensities of 1015 corresponding bands in the calculated and measured VCD spectra, Minick explains. "Our assignments are considered reliable" if R2 is at least 0.90, he says. If not, "you have to go back" and do more computational modeling.
At GSK, where most of the compounds Minick analyzes are fairly complex (molecular weights of at least 350), the first-pass calculation includes at least six conformations, Minick says. Also for the first pass, the conformers are weighted according to a Boltzmann distribution, a probability distribution based on gas-phase energies. A calculated spectrum based only on these parameters, however, rarely passes GSK's confidence limit, he says. On top of refining the conformations in the model, Minick adds, he often tweaks the distributions manually until the confidence standard, R2 = 0.90, is met.
VCD data are now being used to support New Drug Applications and patents, Dukor says. "We have not heard of any case of the VCD proof being rejected."
"I'm not personally aware of any compound filed with the FDA with its absolute configuration established by VCD alone," Salari says. FDA, however, generally responds positively toward new technologies if validated against established techniques, he adds.
A recent search of FDA review documents has turned up at least two submissions that include VCD data, according to Stephen P. Miller of the Office of New Drug Chemistry in FDA's Center for Drug Evaluation & Research. Both cases were for Phase I clinical studies.
In one case, the absolute configuration was determined by other means, but VCD data provided information on the conformation that the active ingredient adopts in solution, Miller says. In the second case, the active ingredient is a single enantiomer produced by resolution of a racemate, and its absolute configuration was assigned on the basis of VCD data. The FDA reviewers believed that this manner of assigning absolute stereochemistry was appropriate for the development stage of the drug, he tells C&EN. "The bottom line with VCD or any new analytical method is if the data are convincing to the scientists developing the drug, then that data should be convincing to the scientists here at FDA."
Three years after ChiralIR was launched, Dukor quit her full-time job with a medical diagnostics company to focus on turning BioTools into a bona fide instrument company. She sold her house, moved her family to rent-free housing provided by her parents, and maxed out several credit cards to raise about $150,000. That money and about $50,000 pitched in by Nafie went into commercializing VCD's sister technique, Raman optical activity. The ROA instrument, called ChiralRaman, is BioTools' newest product. In 2004, it received an R&D 100 Award for analytical instruments.
The dream is fulfilled. When the first BioTools VCD spectrometer sprang to life in the midst of the 1997 winter, Nafie's emotions welled up, Dukor recalls. "Looking at this spectrum coming out was very moving," Nafie tells C&EN. "In 10 seconds, I could see that everything was working. I just started crying. I couldn't believe it."
Other instrument companies now also offer VCD spectrometers off-the-shelf, but only BioTools' has a feature that allows reaction monitoring, Dukor says. "We are thrilled with the competition: They spread the word. They solidified what we did."
Meanwhile, Dukor has yet to make a living from her work with BioTools. "I have been paying off debts instead of giving myself a salary," she says, adding that she might start writing herself checks by the end of this year.
- Chemical & Engineering News
- ISSN 0009-2347
- Copyright © American Chemical Society