The shape of all things – the year of crystallography

Crystals have been a source of fascination for thousands of years, and for good reason. Not only are they beautiful, they are the visible evidence that everything in our world has a shape and structure. Everything looks the way it does thanks to the way its atoms fit together in three dimensions.

This year marks the centenary of the award of the 1914 Nobel Prize for Physics to Max von Laue, for his discovery that x-rays could be bent (diffracted) by the tiny crystals that form the basis of all matter.

Von Laue’s discovery founded the science of crystallography which, 100 years later, has grown to influence all branches of science. Its applications are vast, from helping to engineer better nanomaterials, to developing many of today’s cancer drugs.

We’ll get to the cancer drugs later, but first, it’s time for a rapid sprint through a century of crystallographic marvels.

A crystal is for life

Von Laue’s crucial discovery – that x-ray beams bent when they passed through a crystal – was rapidly adopted by the father-and-son team of William and Lawrence Bragg, who won the 1915 Physics Nobel Prize for showing how to work out the structure of any crystal using a process called x-ray diffraction.

An X-ray diffraction pattern of a crystallized protein. The pattern of spots and the intensity of each spot can be used to determine the structure of the protein.

Rather like shining a torch at a chandelier, shining x-rays at crystals bends the beam in different patterns, which can be picked up as arrangements of spots on photographic film.

From these spots, using some complex mathematical jiggery pokery, it’s possible to work out what any crystal looks like. To stretch our analogy further, it’s like working out the shape of a hidden chandelier from the pattern of spots it makes on the walls.

For example, early x-ray crystallographers provided the answer to the puzzle of why diamond and graphite, both made entirely of carbon atoms, look and behave so differently. Graphite is composed of slippery sheets of hexagonally packed carbon atoms, making it a pencil manufacturer’s best friend. But diamond’s carbon atoms are arranged in rigid three-sided pyramids, to give that million-dollar sparkle and strength.

It wasn’t long before scientists wondered whether biological materials might be crystallisable too. Soon enough, researchers were able to work out the complex structure of pretty much any biological molecule, including proteins, and famously, DNA, using crystallography. This gave scientists unprecedented insight into the workings of the tiny molecular machines driving life.

Precision engineering

With the advent of genetic engineering it became possible not only to make the vast quantities of protein required for crystallography experiments, but also to modify the genetic codes of proteins, and then use crystallography to see what the modification did to its 3D shape. It was as if you could alter the blueprint for a car, and then work out from the finished vehicle whether you’d messed up the wheels, or just lost a cup-holder.

And by changing the shape of a protein, and then seeing how it changed that protein’s function, researchers made enormous strides forward.

As scientists learnt more about proteins, it became apparent that tinkering with structures might have a practical medical application: drug development. Instead of the haphazard centuries-old approach to curing diseases – ‘try it and see if it works’ – it became possible to identify and specifically target the crucial parts of disease-causing proteins.

Thanks to a forensic dissection of the workings of both normal and cancerous cells, cancer scientists knew a fair bit about the proteins involved in cancer. The idea that this knowledge could be coupled with crystallography ignited the imaginations of researchers around the world.

And in London, some ambitious and highly talented scientists, funded by Cancer Research UK, set about developing the targeted drugs they hoped could combat some of cancer’s most malevolent molecules.

The dynamic duo of crystallography

The Institute of Cancer Research (ICR), where the work took place, already had a world-class reputation in the fields of cell and molecular biology but, to complement this, they needed to hire new talent in crystallography. They struck gold not just once, but twice, in the forms of David Barford and Laurence Pearl.

Protein crystals

Both applied for the crystallography job, and were so clearly wonderful that the single crystallography job rapidly turned into two positions, initiating a golden era of drug discovery at the ICR’s Fulham Road site in London.

Barford and Pearl are polar opposites in character, with Barford’s quiet intensity a contrast to Pearl’s cheerful flamboyance. But both share attributes that have made them two of the world’s most respected structural biologists.

Funnily enough, the tricky bit of modern crystallography is not the complicated part at the end. The challenge comes long before the crystals are set in enormously powerful x-ray beams, diffraction patterns automatically collected by detectors and structures mostly worked out by plugging the data into computer programmes.

It’s actually the fiddly stuff at the beginning that’s really tough. Crystallography projects stall because proteins refuse to behave themselves: sometimes they fall apart, however much you mollycoddle them; other times they stubbornly refuse to crystallise, whatever tempting conditions you set up.

Good crystallographers sometimes appear to coax crystals from thin air, by a combination of luck, skill and sheer bloody-mindedness. And, fortunately, Barford and Pearl and their teams were very good crystallographers indeed.

A structural showdown: BRAF unplugged

Barford’s lab’s biggest success featured a protein called BRAF, which belongs to a special class of proteins known as kinases.

Kinases attach chemical tags called phosphates onto other proteins. This acts like an on/off switch to make the proteins more or less active, and is a really important way of regulating proteins inside cells.

Chris Marshall, who headed the ICR’s Cancer Research UK-funded Division of Cell and Molecular Biology, had been interested in BRAF for a while.

He suggested to his friend Mike Stratton, who was just starting a revolutionary genomic sequencing project that he should take a look at BRAF, to see if it was changed in human tumours.

In 2002, Chris Marshall and his colleague Richard Marais, a BRAF specialist, got a phone call from a wildly excited Stratton, telling them that Chris’s hunch had been bang on the money – the BRAF kinase was faulty in around 60 per cent of all cancers.

Structure of the BRAF protein

Chris Marshall, Richard Marais, and their close collaborator Caroline Springer, together with David Barford’s team, and many colleagues both at the ICR and in Mike’s new lab at the Wellcome Trust Sanger Institute, went into overdrive to figure out what was going so wrong with BRAF and how we could stop it (you can read more about the BRAF story in the Our Milestones post here).

But it was Barford’s team’s ability to grow crystals – and work out the shapes – of the normal and faulty BRAF molecules that revealed what had gone wrong.

The lab found that the normal protein was kept in the “off’ position by a special loop, and this delicate control mechanism was disrupted in the cancerous versions, where the protein was jammed in the “on” state, sending cells into a growth frenzy.

Together, the BRAF team published a landmark paper in 2004 showing that the cancer-causing faults in BRAF were all clustered in the part of the protein responsible for carrying out its job as a kinase.

BRAF is a challenge for crystallographers. And the 2004 paper, containing the Barford recipe for making and crystallising BRAF fragments, was a boon to the many companies interested in finding drugs that would combat BRAF’s lethal effects in one of the most challenging cancers around, melanoma.

Knowing the conditions for making a BRAF molecule tough enough to survive in a test tube was essential for the drug screening experiments that resulted in vemurafenib (aka Zelboraf), now an approved treatment for late-stage BRAF-mutant melanoma.

And Richard Marais and Caroline Springer have recently gone on to develop experimental drugs that kill the activity of both BRAF and a closely related ‘cousin’, CRAF which are just going into clinical trials.

Tackling HSP90, the cancer chaperone

Drugs that knock out the activity of one crucially important cancer-causing protein – like vemurafenib – have been a big success story in anti-cancer therapy. Unfortunately, tumours frequently acquire resistance to these drugs, regrow and spread.

Structure of two HSP90 molecules stuck together

This led one of our protaganists to ask whether it would be better to knock out the activity simultaneously of many cancer-causing proteins, rather than just one.

When Laurence Pearl, David Barford’s crystallography co-star at the ICR, started pushing this idea in the 1990s, he faced almost universal scepticism, as it was widely believed that such drugs would be extremely toxic.

But Laurence persisted, as the protein he had in mind, HSP90, was rather special.

HSP90 belongs to a protein family called the molecular chaperones, which help newly produced proteins fold up properly.

Like mothers of stubborn cellular toddlers, HSP90 and its cousins help unruly young proteins turn themselves from linear chains of amino acids into complex structures, fully dressed and ready to face the world.

The consequences are more severe for badly folded proteins than underdressed toddlers though. If the chaperones are not around to help, the newly formed proteins are caught and destroyed for not conforming to the proper shape.

But HSP90 went from being an obscure molecule with a cute function, to something of great interest to cancer biologists. It was found that the group of proteins it chaperones includes an awful lot of proteins that when faulty can cause a cell to become cancerous. And without HSP90’s assistance, these cancer-causing nasties were destroyed.

So could switching off HSP90 help push these dangerous proteins towards the dustbin?

Finding the right partner

Most people thought that disrupting HSP90 would simply kill all cells, rather than just cancerous ones. Laurence, whose crystallography talents had already revealed more about the structure and function of HSP90 than anyone else in the world, disagreed vehemently. He knew from his work that there was an easy, obvious way to stop HSP90 from working – test thousands of molecules and find one that does the job. Something called a drug screen.

He just needed a partner to help him do it.

Fortunately for cancer research, almost the only other person in the world who agreed with Laurence about HSP90’s potential had, by a complete coincidence, just started work at the ICR.

Paul Workman, then the new boss of the ICR’s Cancer Research UK-funded Centre for Drug Discovery, was just as keen as Laurence to run a drug screen featuring HSP90, and – by coincidence – had been looking for a crystallographer to help him.

Drug screening requires a lot of samples

Despite considerable scepticism, Pearl and Workman’s teams started working together and, using HSP90 crystals provided by Pearl’s team, began looking for potential drugs that would bind HSP90.

Their first attempt, using around 50,000 compounds begged and borrowed from the backs of various cupboards at the ICR, produced just one positive hit. A chemical called CCT018159 not only stuck to HSP90, but prevented it from interacting with other proteins. In other words, it had potential as an anti-cancer drug.

CCT018159 was further modified to turn it into a usable drug, and in its present incarnation as NVP-AUY922, is now in clinical trials. The objections that it would be too toxic to use have proved unfounded. It turns out that cancer cells are unstable beasts with far higher dependence on the proteins that HSP90 chaperones than normal cells. So NVP-AUY922 can be administered at a dose that kills cancer cells while leaving the rest of the body relatively unscathed. NVP-AUY922 even seems to work against tumours that are resistant to cancer drugs targeting a single protein, meaning that it is a promising candidate as a companion drug in many cancer types.

Pearl and Workman’s big idea has been more than vindicated.

For this work and much more, David Barford and Laurence Pearl were elected, in 2006 and 2008 respectively, to the Fellowship of the Royal Society, British science’s greatest accolade. Both still have a huge interest in cancer drug development, Barford in the field of cell cycle control, and Pearl in that of DNA damage, with a side order of chaperones, his first love.

Paul Workman, who so successfully pioneered the concept of academic laboratories doing drug development, is now Chief Executive and President of the Institute of Cancer Research.

In 2013, for their HSP90 work, Pearl and Workman’s teams shared the Cancer Research UK Translational Cancer Research Prize. The award was, in Laurence Pearl’s words, “recognition for a truly multidisciplinary scientific approach, which brought together basic, translational and clinical research in a pioneering collaboration. It was structural biology that gave us the insights into the 3D structure and function of HSP90, which in turn helped us to design prototype drugs which could disrupt its action.”

And on that note, we wish x-ray crystallography a very happy 100^th birthday, and look forward to the wonders it will reveal during its next century.

Kathy Weston is a senior science communications advisor at Cancer Research UK

References

• Wan, P., et al. (2004). Mechanism of Activation of the RAF-ERK Signaling Pathway by Oncogenic Mutations of B-RAF Cell, 116 (6), 855-867 DOI: 10.1016/S0092-8674(04)00215-6
• Eccles, S., et al. (2008). NVP-AUY922: A Novel Heat Shock Protein 90 Inhibitor Active against Xenograft Tumor Growth, Angiogenesis, and Metastasis Cancer Research, 68 (8), 2850-2860 DOI: 10.1158/0008-5472.CAN-07-5256

Image credits

• Protein crystals by orinoco14, via Flickr under CC-BY-NC 2.0
• X-ray diffraction image by Jeff Dahl (Own work) under CC BY-SA 3.0, via Wikimedia Commons
• Protein crystals by Areej Abuhammad, via Flickr under CC BY-NC-SA 2.0
• BRAF structure by Emw (Own work) under CC BY-SA 3.0, via Wikimedia Commons
• HSP90 structure by Tirkfl at en.wikipedia [Public domain], from Wikimedia Commons
• Science image by Amy, via Flickr under CC BY-NC-ND 2.0