Monday 29 January 2018

The Joy and Pain of Structural Biology

The British Crystallographic Association is, as its name implies, the main organisation supporting crystallograpy and crystallographers in the UK. Theirs is a multi-disciplinary science, and the different needs of the eclectic group of people who call themselves crystallographers are met by the Association's special interest groups: among them, the Biological Structures Group for structural biology. The annual meeting of this group, known as the Winter Meeting, takes place in December, often just a few working days before Christmas. The BSG Winter Meeting has featured on this blog on several occasions: in 2016, it was held at Birkbeck and celebrated the work of one of our distinguished emeritus professors, Steve Wood.

The 2017 meeting, held in the University of Cambridge's famous Cavendish Laboratory, had a rather unusual theme. The organisers asked each of the invited speakers to talk about the ups and downs of their scientific career - the 'joy and pain' of the meeting title - by focusing on one challenging or important piece of work, perhaps described in a single published paper. Not every speaker managed to keep to just one paper, but all the talks gave useful and at times inspiring insights into how structural biology is done.

First of all, however, Malcolm Longair, head of the Cavendish Laboratory from 1997 to 2005 and perhaps the only astrophysicist to address the Biological Structures Group, gave a short history of the university's Physics department that was based there and its links to structural and molecular biology. That early history was quite extraordinary; many of the most important advances in atomic and nuclear physics, including the discoveries of the electron and the neutron and the first controlled nuclear disintegrations, were made there. A lab photo taken in 1932 includes no fewer than nine Nobel laureates.

Crystallography, in those early years, was thought of as part of physics; J.D. Bernal and his group joined the lab in 1931, and the younger Bragg became head of the department in 1938. The rest, as Longair said, was history: seeds of the discipline we now know as structural molecular biology were sown in Bragg's time with Perutz and Kendrew's work on globin structure as well as Watson and Crick's on that of DNA. By the time those studies reached their triumphant conclusion, however, the crystallographers were no longer strictly part of the Cavendish. The 'Unit for Research on the Molecular Structure of Biological Systems’, set up by the Medical Research Council, moved out of the main lab in 1957 into a building known as the 'MRC Hut'. This was the first home of the MRC Laboratory for Molecular Biology (MRC-LMB) at Cambridge with its enduring reputation for excellent structural biology research.

The next speaker, Cambridge University's Tom Blundell, began by describing his early career in 'the Other Place': Dorothy Hodgkin's lab at Oxford, where he had shared some of the glory of the insulin structure. He had considered talking about insulin at this meeting, but, he explained, "Dorothy had had the pain of trying to solve the structure for 30+ years... I had the joy of a paper in Nature!" The story he told instead was his group's own: solving the structures of proteins involved in DNA repair. This was a long story, taking in 15 years' worth of papers in Nature (2002, 2010) and Science (2017) and culminating in the 'great joy' of discovering inhibitors validated against an important protein target for oncology.

DNA damage taking the form of simultaneous breaks in both DNA strands (double-strand breaks) are common but can lead to cell death or cancer. Fortunately, they are easily repaired in healthy cells, mainly through the mechanism of non-homologous end joining (NHEJ). Blundell's group have studied the proteins involved in this complex mechanism for many years. It is a three-stage process, in which the component proteins assemble on the DNA molecule either side of the break; the ends are 'pruned' by adding or removing nucleotides to restore the original sequence and finally, the ends are joined through DNA ligation. One of the proteins involved is a kinase, DNA-PKcs, that exists as a single polypeptide chain of 4128 amino acids. Blundell's group published the structure of this huge molecule in 2010 (PDB 3KGV) and it is still the longest single-chain protein to have been solved by X-ray crystallography. Blundell explained that the chain folds into a flexible, circular 'cradle' like structure that can support the DNA double helix, with the ligation taking place inside. The mechanism requires proteins to work as 'stages, scaffolds and steps' to hold the complex together for repair, and his group has solved structures of many other components including the Ku70-Ku80 heterodimer that recognises and binds to the break, initiating the repair, and a nuclease named Artemis with 'a nice pocket for drug discovery'.

Two talks on structural biology as applied to drug discovery followed. The first was by Pamela Williams from Astex Pharmaceuticals, a company founded by Blundell with Harren Jhoti in 1999 that has just registered its first drug - a kinase inhibitor, Kisqali® (ribociclib) - for clinical use in breast cancer. Williams' talk highlighted another protein family that is just as important in pharmacology as the kinases: cytochromes P450. We have about 50 different P450 subtypes in our livers, and they catalyse reactions that modify drug molecules so they can be more easily removed from our bodies. A handful of these - the subtypes known as 1A9, 2C9, 2C19 and 3D6 - metabolise most prescription drugs. Human (and all eukaryotic) P450s are monotopic membrane proteins with flexible active sites, which allow them to bind a wide variety of substrates but which make the structures hard to solve. Williams' involvement with P450 structural biology began with the first mammalian structure, rabbit cytochrome 2C5, and she joined Astex from California to work on the first human structure, the subtype 2C9. This was published in 2003 (PDB 1OG2); a large number of other human structures have followed, yielding useful insights into drug metabolism.

Ben Bax, who studied for his PhD under Tom Blundell at Birkbeck, has just moved to the University of York after eighteen years at the pharma company GlaxoSmithKline (GSK). His talk described work at GSK to determine the structures of bacterial DNA gyrases. These are members of a large class of enzymes called topoisomerases that catalyse topological transitions in DNA; the gyrase, which catalyses DNA supercoiling, is the target of the widely used quinolone family of antibiotics (e.g. ciprofloxacin). However, quinolone resistance is increasing, mainly through mutations at specific amino acid positions of the target gyrase. GlaxoSmithKline is investing heavily in the development of novel gyrase inhibitors based on oligonucleotides, and Bax' structural biology group has contributed a large number of still unpublished structures of the enzyme with and without inhibitors or DNA bound to this work.

Janet Thornton, emeritus director of the European Bioinformatics Institute, is one of the best known figures in British bioinformatics. Her talk, on what she termed an 'accidental' paper, took the audience back to the basic principles of protein structure. In the late 70s, when she started her career, there were only about fifteen protein structures known but scientists were already examining those structures to determine characteristic patterns. Many of these first structures determined had major inaccuracies, and discovering and correcting these was a major task for early structural biologists. The Ramachandran Plot, now half a century old, was one of the first tools to be developed to gauge the quality of a protein structure, and it is still widely used. Thornton's 'accidental' (and very highly cited) paper described the program PROCHECK, which runs this and other checks on a structure to give a comprehensive assessment of its quality. A PROCHECK record for each structure in the PDB is linked from the database PDBsum.

The final talk provided delegates with a rare opportunity to hear a new Nobel Laureate - in this case, the Laboratory of Molecular Biology's own Richard Henderson - tell the story behind some of his ground-breaking research. Henderson shared the 2017 Chemistry Nobel, for "developing cryo-electron microscopy for the high-resolution structure determination of biomolecules in solution", with Joachim Frank and Jacques Dubochet. He chose to talk about one structure that he had in some senses made his own: that of bacteriorhodopsin, a proton pump found in Archaea that captures light energy as photons and that has many structural and mechanistic similarities with the G-protein coupled receptors, although the exact evolutionary relationship is unclear. Henderson's studies of this important molecule started in the 1970s with structures that were just about detailed enough to show the cylindrical helices. It took him over 15 years'effort with collaborators in Berlin, Berkeley and elsewhere to improve the technology enough to solve the so-called 'phase problem' and obtain an atomic-resolution structure by electron diffraction. The rest, again, is history.