Chemists at times look to plants, sea life and other natural sources for the basic ingredients needed to develop the next breakthrough medicine. Unfortunately, nature is not always willing to easily part with its secrets, forcing scientists to rely on sophisticated imaging technology—nuclear magnetic resonance (NMR) spectroscopy or mass spectrometry, for example—to decipher the molecular formula of newly discovered organic compounds so they can be replicated in the lab.

Sometimes these new compounds defy even the most powerful lab equipment. Researchers at the University of Aberdeen Marine Biodiscovery Center (MBC) in Scotland found this to be the case last year when studying a bacterial species—Dermacoccus abyssi sp. nov. —found in a mud sample harvested via robotic submarine from the Pacific Ocean's Mariana Trench, the deepest place on Earth, at about 11,000 meters. When the sample produced a chemical compound they were not able to identify, the researchers tried high-resolution mass spectrometry to determine the chemical compound's components but were unable to figure out its exact molecular structure.

The Aberdeen scientists are studying the potential of marine organisms as a source for new chemical compounds, which could be used to develop novel treatments for cancer, inflammation, infection and parasitic diseases. The search for new natural compounds is necessary because nature provides much greater chemical diversity than researchers can come up with, says Rainer Ebel, an Aberdeen lecturer working at the Marine Biodiscovery Center. "Nature is a far more creative chemist," he adds. "We keep finding new templates from nature that synthetic chemicals try to improve upon."

When promising candidates are discovered, the scientists must identify the structure of these compounds in order to determine whether they are viable for use in drug development. This approach has been successful, in particular when researchers in the late 1960s and early 1970s extracted a then-unknown compound from the stem bark of the Pacific yew tree that would turn out to be a key ingredient of the anticancer drug Taxol (pdf).

Each carbon and hydrogen atom in a molecule has a defined frequency in the NMR spectrum that scientists use determine how hydrogen and carbon atoms are connected together. The D. abyssi sp. nov. sample's dearth of hydrogen atoms, however, meant that NMR could not provide Aberdeen researchers with enough information to solve the mystery. Instead, the scientists' efforts left them with four potential structures, none of which could be ruled out by the NMR data alone. The only remaining possibility to find the correct structure would have been to take a chemical synthesis of the proposed structures, a very complex task that can take several months.

"We knew basically what was in the lower left and upper right of the molecule, but we couldn't connect the two sides," Ebel says. It was like solving a jigsaw puzzle without the benefit of a photo of the finished product.

The researchers caught a break late last year when the wife of MBC Director Marcel Jaspars recalled an August 2009 article she had read in the London's Daily Mail about a team of IBM Research—Zurich scientists who had used a modified atomic force microscope (AFM) to create an image revealing all the bonds in an individual molecule. One of the pictures accompanying the Daily Mail article showed the hexagonal shapes of five carbon rings as well as the positions of the hydrogen atoms around these rings. Jaspars got his hands on the IBM scientists' study the following day and soon contacted lead researcher Leo Gross. Jaspars sent Gross a sample of Aberdeen's mystery compound, which IBM started examining in January.

An AFM uses a sharp tip to measure the tiny forces between the tip and a sample, such as a molecule, to create an image. To image a molecule's chemical structure with an AFM, the tip needs to be extremely close—less than a nanometer—to the molecule. To do this, IBM scientists increased the tip's sensitivity by attaching a carbon monoxide (CO) molecule to it. When it was close enough to the sample, the CO-covered tip sensed tiny repulsive forces that help reveal a molecule's atomic-scale chemical structure. "I would liken it to increasing the resolution on a computer screen by making the pixels smaller," Jaspars says.

"You wouldn't normally use AFM because the resolution wouldn't be good enough," Jaspars notes. Before IBM fine-tuned its AFM it was used to look at proteins, which are much larger than the samples that Jaspars and Ebel needed to study.

By the end of February IBM researchers had discovered the identity of the mystery compound to be cephalandole A , an already known, compound that had been originally isolated from a Taiwanese orchid. The process of identifying cephalandole A , which the researchers describe in the August 1 issue of Nature Chemistry , was the first time IBM's modified AFM technology had been used to determine the structure of an unknown molecule. (Scientific American is part of Nature Publishing Group.) The ID'd molecule written about in the Daily Mail article was pentacene, a well-known substance commonly used in solar cells.

Cephalandole A has proved inactive in most of the tests the Aberdeen researchers have carried out thus far, although further testing is still underway, Jaspars says. In the end the testing of IBM's technology may be the most significant aspect of the research.

It will take years before IBM's technique becomes routine, particularly because it is time intensive, requiring researchers to carefully prepare the AFM needle and sample. Placing the sample on the AFM's surface is no trivial matter either, nor is taking the actual measurements. Still, work being done by IBM as well as Culver City, Calif.–based Nanogea, Inc. (which makes a nanoparticle coating to improve the precision of AFM probes and substrates) is crucial to expanding scientists' ability to study molecular structures.