Thursday, 24 July 2014

Science Week 2014: Birkbeck and the History of Crystallography

Science Week at Birkbeck in 2014 featured two lectures on Department of Biological Sciences, both presented on 2 July. One of these was a double act from two distinguished emeritus professors and Fellows of the College, Paul Barnes and David Moss. Remarkably, they both started their working lives at Birkbeck on the same day – 1 October 1968 – and so had clocked up over 90 years of service to the college between them by Science Week 2014.

The topic they took was a timely one: the history of the science of crystallography over the past 100 years. UNESCO has declared 2014 to be the International Year of Crystallography in recognition of the seminal discoveries that started the discipline, which were made almost exactly 100 years ago; a number of the most important discoveries of that century were made by scientists with links to Birkbeck.

The presenters divided the “century of crystallography” into two, with Barnes speaking first and covering the first 50 years. In giving his talk the title “A History of Modern Crystallography”, however, he recognised that crystals have been observed, admired and studied for many centuries. What changed at the beginning of the last century was the discovery of X-ray diffraction. Wilhelm Röntgen was awarded the first-ever Nobel Prize for Physics for his discovery of X-rays in 1896, but it was almost two decades before anyone thought of directing them at crystals. The breakthroughs came when Max von Laue showed that a beam of X-rays can be diffracted by a crystal to yield a pattern of spots, and the father-and-son team of William Henry and William Lawrence Bragg showed that it was possible to derive information about the atomic structure of crystals from their diffraction patterns. These discoveries also solved – to some extent – the debate about whether X-rays were particles or waves, as only waves diffract; we now know that all electromagnetic radiation, including X-rays, can be thought of as both particles and waves

Von Laue and the Braggs were awarded Nobel Prizes for Physics in 1914 and 1915 respectively, and between 1916 and 1964 no fewer than 13 more Nobel Prizes were awarded to 18 more scientists for discoveries related to crystallography. Petrus Debye, who won the Chemistry prize in 1936, showed how to quantify the thermal motion of atoms as vibrations within a crystal. He also invented one of the first powder diffraction cameras, used to obtain diffraction patterns from powders of tiny crystallites. Another Nobel Laureate, Percy Bridgman, studied the structures of materials under pressure: it has been said that he would “squeeze anything he could lay his hands on”, often up to intense pressures.

Scientists and scientific commentators often argue about which of their colleagues would have most deserved to win the ultimate accolade. Barnes named three who, he said, could easily have been Nobel Laureates in the field of crystallography. One, Paul Ewald, was a theoretical physicist who had studied for his PhD under von Laue in Munich, and the other two had strong links with Birkbeck. JD “Sage” Bernal was Professor of Physics and then of Crystallography here; he was famous for obtaining, with Dorothy Crowfoot (later Hodgkin) the first diffraction pattern from a protein crystal, but his insights into the atomic basis of the very different properties of carbon as diamond and as graphite were perhaps even more remarkable. He took on Rosalind Franklin, whose diffraction patterns of DNA had led Watson and Crick to deduce its double helical structure, after she left King’s College, and she did pioneering work on virus structure here until her premature death in 1958.

Barnes ended his talk and led into Moss’s second half-century with a discussion of similarities between the earliest crystallography and today, as now, you only need three things to obtain a diffraction pattern: a source of X-rays, a crystalline sample, and a recording device; the differences all lie in the power and precision of the equipment used. He demonstrated this with a “symbolic demo” that ended when he pulled a model structure of a zeolite out of a large cardboard box.

Paul Barnes demonstrates the basic principles of X-ray crystallography using a large cardboard box. Photo © Harish Patel and Ruben Zamora, Department of Psychological Sciences, Birkbeck

David Moss then took over to describe some of the most important crystallographic discoveries from the last half-century. His talk concentrated on the structures of large biological molecules, particularly proteins, and he began by explaining the importance of protein structure. All the chemistry that is necessary for life is controlled by proteins, and knowing the structure of proteins enables us to understand, and potentially also to modify, how they work.

Even the smallest proteins contain thousands of atoms; in order to determine the position of all the atoms in a protein using crystallography you need to make an enormous number of measurements of the positions and intensities of X-ray spots. The process of solving the structure of a protein is no different from that of solving a small molecule crystal structure, but it is more complex and takes much more time. Very briefly, it involves crystallising the protein; shining an intense beam of X-rays on the resulting crystals to produce diffraction patterns, and then doing some extremely complex calculations. The first protein structures, obtained without the benefit of automation and modern computers, took many years and sometimes even decades.

Thanks to Bernal’s genius, energy and pioneering spirit, Birkbeck was one of the first institutes in the UK to have all the equipment that was needed for crystallography. This included some of the country’s first “large” computers. One of the first electronic stored-program computers was developed in Donald Booth’s laboratory here in the 1950s. In the mid-1960s the college had an ATLAS computer with a total memory of 96 kB. It occupied the basements of two houses in Gordon Square, and crystallographers used it to calculate electron density maps of small molecules. Protein crystallography only “took off” in the 1970s with further improvements in computing and automation of much of the experimental technique.

Today, protein crystallography can almost be said to be routine. The first step, crystallising the protein, can still be an important bottleneck, but data collection at powerful synchrotron X-ray sources is extremely rapid and structures can be solved quite easily with user-friendly software that runs on ordinary laptops. There are now over 100,000 protein structures freely available in the Protein Data Bank (PDB), and about 90% of these were obtained using X-ray crystallography. The techniques used to obtain the other 10,000 or so, nuclear magnetic resonance and electron microscopy, are more specialised.

Moss ended his talk by describing one of the proteins solved in his group during his long career at Birkbeck: a bacterial toxin that is responsible for the disease gas gangrene (PDB 1CA1). This destroys muscle cells by punching holes in their membranes, and its victims usually have to have limbs amputated to save their lives. Knowing the structure has allowed scientists to understand how this toxin works, which is the first step towards developing drugs to stop it. But you can learn even more about how proteins work if you also understand how they move. Observing and modelling protein motion in “real time” still poses many challenges for scientists as the second century of crystallography begins.

Structure of alpha-toxin, the key Clostridium perfringen toxin in gas gangrene. Image from the PDB.

Tuesday, 8 July 2014

Mimicking DNA: How a Repressor Meets its Waterloo

Bacillus subtilis is a non-toxic bacterium commonly found in soil, usually in the form of a dormant spore.  It is extremely hardy due to a remarkable list of adaptations to environmental threats.  These include the production of antibiotics and degrading enzymes, amalgamation into biofilms, the formation of an endospore and even the destruction of sibling bacteria.

This impressive range of threat responses is triggered by many different adaptation genes, which are governed by an army of regulators, some specific and some global.

One notable universal regulator protein is AbrB, which represses several adaptation genes when cell conditions are favourable.  There are two mechanisms that block this repressor, when the bacterium faces a threat.

Under stress, an upstream regulator, SpoOA, is phosphorylated and, in an envy inducing display of multi-tasking, is able to bind to and inhibit the gene for AbrB and also activate the gene for anti-repressor AbbA.

AbbA binds to the repressor AbrB and blocks it, thereby allowing adaptation to environmental threat.  AbbA has recently been revealed as a DNA mimic, which competes very effectively for the AbrB binding site by copying key DNA characteristics.

Inspection of the primary structure of AbbA reveals 65 residues of which 20 are polar.  Given that the core of a protein tends to be hydrophobic, this implies an polar surface.

The next step was to establish the oligomerisation of the protein in its native state.  Size-exclusion liquid chromatography (SELC) and native mass spectrometry both showed that the natural size of AbbA was double the mass of the monomer, that is, it dimerises.

NMR was used to solve the dimer structure.  


This figure shows the ten lowest energy structures of the AbbA dimer obtained using NMR.  One monomer has helices coloured blue to green whilst the second is yellow to red.  (PDB 2LZF). 

Figure taken from Tucker, A.T. et al (2014)


The monomer consists of three alpha helices, connected by two loops with a fairly unstructured N terminus.  In the dimer there are substantial interactions between helices two and three of each monomer, which are largely hydrophobic, and further hydrophobic interactions between helices one and two.

To determine the binding site for AbbA and AbrB, SELC studies were performed on AbbA with the C terminus of AbrB and then with the N terminus.  The team found that AbbA binds only to the N terminus, the region of the protein that is responsible for binding to DNA.

Previous mutation studies had shown that four arginine residues are responsible for AbrB binding to DNA, namely R8, R15, R23 and R24.  Each of these was mutated to observe the impact on AbbA binding and three of them, R8, R15 and R23, were found to be critical, potentially indicating a similar binding mechanism.

To determine the strength of the competition offered to AbrB-DNA binding by AbrB-AbbA binding, isothermal titration calorimetry was used to measure the dissociation constants.  This technique is covered in TSMB, one of the courses available after PPS.  The dissociation constants showed that the binding strengths of the two pairs of molecules were similar, such that AbbA offers a significant threat to AbrB-DNA binding.

The last stage of the investigation was to determine the interaction between the Abba homodimer and the AbrB homodimer.   

Molecular docking was used to model this interaction and showed a sizeable interaction area with a complex pattern of 18 hydrogen bonds and 16 salt bridges.  This site is at the highly negatively charged terminal of AbbA and the DNA binding face of N terminal AbrB.

The first helix of each AbbA monomer has to pull back to allow this strong interaction but AbbA's second and third helices maintain their conformation and act as a stabilising anchor.

The picture below gives a striking illustration of the extent of AbbA's mimicry of DNA.  


This figure illustrates the similarities between AbbA and the DNA phosphate backbone.  (a) is the NMR structure of AbbA with positively charged residues shown in blue and negatively charged in red.  This can be compared with (c), a DNA fragment (PDB 1BNA), showing the charge distribution in the same colour scheme.  The length of one turn of the helix and the minor groove are shown with yellow dotted lines.  (b) shows the structure of AbbA in cartoon format with the side chain oxygens of glutamic acid residues 16, 29. 33 and 67 as red spheres and (d) is the same structure with the backbone of the DNA from PDB 1BNA superimposed in yellow.
 Figure taken from Tucker, A.T. et al (2014) .

Few DNA mimics have been discovered but they share the tactic of using negatively charged residues (glutamic acids and aspartic acids) to present similar bonding opportunities to the DNA backbone phosphates. It would seem that, faced with the challenge of competing with the strongly charged backbone of DNA, the mimics are following the logic that if you can't beat 'em, join 'em.