Organic chemists, make sure you’re sitting comfortably. The structure of small organic molecules, such as those used in drugs, can be deduced in minutes rather than weeks, thanks to a technique that uses beams of electrons to quickly reveal how atoms are arranged.
The technique, called 3D electron diffraction, has been used by some inorganic chemists and material scientists since the mid-2000s to deduce the structure of molecules. But organic chemists, for whom the implications could be transformative, did not adopt it widely.
In mid-October, two papers appeared online describing a way to use the same technique for drugs, making it much faster and easier to work out the structure of these small organic molecules than current techniques allow.
“I think there are a lot of people smacking their heads, saying, “Why didn’t we think to do this earlier?” says John Rubinstein, a structural biologist at the University of Toronto who uses related techniques to study large molecules such as proteins.
Existing methods for finding the structure of small molecules require scientists to grow crystals for analysis, a laborious process that can take weeks or months. “Something that was a real barrier to their research is now basically removed,” adds Rubinstein.
Knowing how atoms are arranged in a molecule is necessary for understanding that substance’s function. Chemists working to develop new drugs, for example, depend on this structure to understand how the drug acts in the body — and how it could be tweaked to bind more strongly to its therapeutic target or to reduce side effects.
A technique called X-ray crystallography has been used for decades to deduce this arrangement. But it can take weeks of work — and is not always successful.
First, scientists need to coax the molecules to crystallize. They then blast the crystal with an X-ray beam. The crystal’s lattice structure causes the X-rays to diffract, and a special detector records the resulting pattern.
Scientists then use software to analyse the pattern and work out the structure of the molecule.
The challenges arise because X-ray diffraction works only with large crystals, and these can take weeks or even months to form. And some molecules are so hard to crystallize in the first place that it might not be possible to analyse them this way at all.
One alternative is to replace X-rays with electron beams, which can produce diffraction patterns using much smaller crystals. In 2007 and 2008, respectively, crystallographers at the Johannes Gutenberg University in Mainz, Germany, and Stockholm University developed the first methods for detecting the 3D structures of molecules automatically using electron diffraction. Previously, scientists had to laboriously merge multiple 2D diffraction patterns together to get this 3D structure.
Initially, the technique was used mainly with inorganic structures, which aren’t affected by radiation as much as organic molecules are. Then, in 2013, Tamir Gonen, a structural biologist at the University of California, Los Angeles, developed a version of electron diffraction called MicroED, which could be used on large biomolecules such as proteins.
Now, in the two latest papers, Gonen’s team and another group from Switzerland have demonstrated that electron diffraction can also be used to work out the structure of smaller organic molecules. It’s not quite the first time this has been done, says Xiaodong Zou, a structural chemist at Stockholm University. But it is an important demonstration of just how fast and easy this kind of analysis can be.
In the first paper, published on 16 October in Angewandte Chemie International Edition, a team led by crystallographer Tim Grüne at the Paul Scherrer Institute in Switzerland reports the creation of a prototype device for finding the structure of small molecules, using the beam from an electron microscope and a compatible detector1.
The diffraction patterns are analysed by software that is already used in X-ray crystallography. “Everything is composed of parts which have existed before,” says Grüne. “It’s just really the smooth integration of the system.”
The team used its set-up to find the structure of the painkiller paracetamol from tiny crystals formed from the powder inside capsules. These crystals were just a few micrometres long — much smaller than can be analysed using X-ray diffraction.
In the second paper, a preprint uploaded to the ChemRxiv server on 17 October, Gonen’s group adapted the MicroED technique to solve the structure of small molecules instead of proteins2.
Gonen says that making this shift was “trivial”. The main tweaks concerned the preparation of the samples, he says: whereas fragile proteins need to be treated with care, in this case, all he had to do was grind down pharmaceutical powders. The team used this adapted version of MicroED to find the structure of powders of pharmaceuticals including ibuprofen and the anti-epileptic drug carbamazepine.
These crystals were some 100 nanometres wide — a billion times smaller in volume than those used in X-ray crystallography — and their structure could be resolved in under 30 minutes.
Rubinstein says it’s surprising that a technique already used in other fields hasn’t yet been widely adopted by organic chemists. “It’s this great solution that’s been sitting almost in plain sight,” he says.
Gonen puts the oversight down to a lack of communication between disciplines. It was only when he began speaking with chemists, he says, that he became aware that they struggled with growing large crystals to analyse small molecules, leading him to realize he had a solution for them. “As a protein crystallographer, I never really thought very carefully about small molecules,” he says. “For us, small molecules are the things we try to get rid of.”
Excitement and limitations
The technique has caused lots of excitement, but it does have some limitations. The 3D structure of some molecules, for example, results in mirror-image molecules that can have different chemical effects — but it is challenging to distinguish between these structures using electron diffraction.
Differentiation of these mirror-image structures would require further development of the analysis software, says Grüne.
One obvious application for electron crystallography is elucidating potential candidates for drug development. But Gonen says the technique could have other applications, such as forensics, when quick identification of a substance can be vital.
Grüne is optimistic that his work will encourage hardware manufacturers to create new devices built specifically for electron crystallography. At the moment, researchers tend to rely on electron microscopes to generate the electron beams, but these are expensive and include components, such as lenses, that are not necessary for electron diffraction. They are also not optimized to work with the other pieces of equipment used in the analysis. With a purpose-built device, he says, “one could just solve structures with the push of the button”.
Nature 563, 16-17 (2018)