How a new crystallography technique may speed up drug development

For chemists looking to understand molecular structures, crystals play an indispensable role. When a given material is crystallised – the molecules packed together in such a way that a crystal forms – it can be studied using a technique called x-ray crystallography. This is one of the best techniques we have for discovering how the atoms in a material bond together, and can provide a basis for designing new drugs. 

The difficulty here is that not every material can be crystallised, or at least, it can’t be done easily. Often, it requires a kind of trial and error approach, with scientists running many different crystallisation experiments in the hope that one of them will yield results. This can take many weeks to complete, and may require impractically large quantities of the material in question. 

Now a team of chemistry experts, from Newcastle and Durham University and SPT Labtech, have designed a new technique to speed up this process. Known as Encapsulated Nanodroplet Crystallisation (ENaCt), this is a robot-assisted, high-throughput method that requires only micrograms of material per experiment. Within a few minutes, the user is able to set up hundreds of tiny crystallisation experiments. This speeds up the process of getting results.

According to senior lecturer in chemistry at Newcastle University and co-head of the project Dr Michael Hall, chemists can be very conservative in their methods. A lot of the commonly used techniques haven’t really changed in decades.  

“The classical method for making crystals involves dissolving something in a solvent – you evaporate it or add a different solvent and let them mix,” he says. “So you do a lot of these experiments but often on a relatively large scale. If you’re having to use a small spatula full of material for every experiment, that quickly adds up.”

The team wanted to miniaturise the whole process, so as to perform the same screening procedure with a fraction of the material. For every experiment, a tiny amount of the material is dissolved in an organic solvent.

Each droplet is encapsulated in an inert oil, to control the rate of solvent loss through evaporation. The process is automated via liquid handling robots, meaning many unique experiments can get underway at the same time. 

“We took inspiration from recent developments in protein crystallography, in which crystals are made from proteins using very small quantities of material,” says Dr Hall. “Someone did one for the Covid-19 spike protein recently – they were able to make 3D models of the spike protein that came from the crystal structure, using X-ray crystallography to find every atom. We’re doing that for molecules that are much smaller than a protein, using robotics, low volume, and lots of automation.” 

The researchers, who published their findings in the journal Chem, tested their technique on a range of small molecules, including organic dyes, organometallics and pharmaceuticals. Heartened by the very high success rate, they moved onto a supposedly ‘uncrystallisable’ compound, the fungicide dithianon. They set up 394 experiments, 72 of which yielded crystals. 

“As of this point in time there is not a single material we’ve studied where we’ve failed to get a crystalline form using this method,” says Dr Probert, senior lecturer in inorganic chemistry and head of crystallography at Newcastle University and project co-head. “This wasn’t expected. We thought it’d work better than classical methods, but it’s better than we thought it would be.”

Dr Hall adds that part of the success rate lies in the sheer statistical power of being able to do so many experiments. While some molecules take more experiments than other, the success rate is extremely high. 

“Getting a crystal is a bit like playing roulette – if you spin the roulette wheel enough times you’ll eventually get the number you want, if you cheat and use lots of balls at once even better. Similarly if you run enough different crystallisation experiments and you’ll eventually find a crystal,” he says. “With our ENaCt approach we are cheating by running lots of experiments at the same time. The more experiments you can do, the quicker you can map the experimental space and the quicker you can get a result.

Since the technique was developed, the researchers have been talking to a wide range of academic and industrial groups who are interested in trying it out for themselves. While there are several potential uses within the pharma industry, one of the main benefits will be the ability to look at crystal forms early in the drug development process. 

“From a processing point of view, that crystal form is important – it affects how you make tablets, and things like powder flow in a factory,” says Dr Hall. “At the moment most of the crystallography work is done at the process development stage, three or four years into a programme. With our technique, you should be able to start that early on.”

He points out the classic case of AbbVie’s HIV drug, Ritanovir. When the drug first came to market in 1996, the tablets contained a crystal form of the drug that was highly soluble in the body. Two years later, their manufacturing plant suddenly made a different crystal form of the drug that they’d never seen before, a version with much lower bioavailability. They were forced to pull the oral capsule form from the market, setting them back a year or two. 

“With our technique, you fend off those late stage problems,” says Dr Hall. “We’ll end up having very early stage derisking of projects with crystallography, along with lots of other applications.”

Dr Hall and Dr Probert are now looking to start a spin-off company to bring their technique to industry and academia. While they can’t talk about upcoming developments, they are aware of what their next steps will be and are excited about the prospects moving forward. 

“It’s quite rare in the science environment to go from the conceptual idea to a resolved project within this time period,” says Dr Probert. “That’s only been possible via a raft of PhDs, postdocs and industrial collaborators sitting behind the whole process. Paul Thaw from SPT Labtech gave us a huge amount of expertise and guidance in the use of liquid handling robots, while Andrew Taylor, the first author of the paper, contributed a lot of work.” 

As well as opening up further avenues of collaboration, the technique will provide experimental backing for ideas that have only been theoretical up till now. 

“There are many people in the world working on computing what different crystallographic forms should look like for any given molecule,” says Dr Probert. “But until something is realised in an experiment, it remains theory. The method we’ve developed has unlocked some of those additional forms.”

The method, then, could prove an asset to fundamental research and drug discovery alike. Ultimately, it might overturn some long-held preconceptions within the molecular sciences.