50th anniversary of the Stanford SSRL synchrotron radiation and protein crystallography initiative

1.1. On a personal note (JRH)

When Dorothy Hodgkin shared a preprint of that paper (Phillips et al., 1976) with me in 1975, when I was a DPhil student at the University of Oxford (1974 to 1977), it changed my whole research career perspectives. I sum the situation up and summarize the challenges for macromolecular crystallography (MX) at that time in Fig. 1.

Suitably inspired, my first SR experiment commenced in 1976 on the NINA synchrotron at Daresbury; it involved optimizing the anomalous scattering of platinum in my K₂Pt(CN)₄ derivative of my DPhil project `6-phospho­gluconate de­hydrogenase enzyme protein crystallography structure determination' (supervisor in Oxford, Margaret Adams). I was helped greatly by my NINA local contact Dr Joan Bordas.

This experiment was soon followed by my use of the LURE synchrotron MX beamline in Paris (established by Roger Fourme) and the EMBL Hamburg SR facilities at DESY (launched by Ken Holmes and Gerd Rosenbaum). Indeed, the seminal article by Rosenbaum, Holmes & Witz (1971), `Synchrotron Radiation as a Source for X-ray Diffraction', at DESY was cited by Phillips et al. (1976) as the first reference of their references list. That paper, also pivotal, described muscle fibre diffraction with SR but did mention SR protein crystallography in a table. The wide influence of this 1971 paper, and how it led to the EMBL Outstation at DESY Hamburg, is described by Holmes & Rosenbaum (1998). In parallel with the paper by Phillips et al. (1976) was the paper by Harmsen et al. (1976) from DESY, also important, which had a more cautious view of SR protein crystallography, and no coverage of the multiple wavelength phasing method with SR that Hodgson et al. initiated and developed (Phillips et al., 1977, 1979; Templeton et al., 1980; Phillips & Hodgson, 1980). That method was greatly expanded in its applicability by the seleno-me­thio­nine protein production methodology initiated by Wayne Hendrickson (Hendrickson et al., 1990).

In 1981 I joined the UK's dedicated SRS leading the protein crystallography there on a 50%/50% joint appointment with Keele University Department of Physics, and subsequently full time at Daresbury from 1983. We made a strong effort to move to electronic detectors jointly with the Rutherford Appleton Laboratory (a multi-wire proportional chamber) and the Laboratory of Molecular Biology Cambridge and Enraf–Nonius (a TV diffractometer of Uli Arndt). Image plate readers came later, from Rigaku and EMBL Hamburg (Jules Hendrix), then CCDs initiated from Princeton (Sol Gruner) and installed at MacCHESS (see below). The pixel-array detector came considerably later, notably from the Swiss PSI Centre and its spin out company Dectris, which dominates the detector provision at beamlines in the modern era.

Within this period, I led the planning of ESRF's macromolecular crystallography in the mid-1980s and into the 1990s (Aigrain, 1987). Greatly facilitating the detailing of these plans was the `New Rings Workshop' held at SSRL, that led to a direct research and development collaboration between myself at Daresbury SRS and Keith Hodgson and Britt Hedman at SSRL [see Hedman et al. (1985) and Helliwell (1988)]. The eventual extensive development of instrumentation and methods for multi-wavelength anomalous dispersion at the SRS, ESRF, CHESS and Elettra facilities in which we participated was described by Cassetta et al. (1999). As a further measure of the global connectedness of the SR research and facility community, Keith Moffat spent a sabbatical year with me at the Daresbury SRS and at the York University Department of Physics, where I had taken up another joint appointment. We worked on various aspects of the SR Laue protein crystallography method. Then in 1994 I spent four and a half months on sabbatical at MacCHESS, Cornell University, with Steve Ealick, then MacCHESS Director, followed by one month at BioCARS with Keith Moffat, by then at the Advanced Photon Source and University of Chicago.

Overall, I described the evolution of SR and the growth of its importance in crystallography in Helliwell (1992, updated in 2011; see Helliwell, 2011). There are many new developments in the past ten years since 2012. These are encapsulated in all the articles in this special issue. I pay tribute to the 1976 paper by Keith Hodgson, and his many decades of leadership at SSRL, and the other pioneers at DESY Hamburg (Rosenbaum et al., 1971) and LURE Paris (Lemonnier et al., 1978).

1.2. On a personal note (DMS)

The Cornell High Energy Synchrotron Source (CHESS) also became a centre for macromolecular crystallography. Cornell had a long history of projecting the importance of SR in X-ray physics including diffraction (Parratt, 1959). CHESS, constructed in 1978–1980, was one of the first facilities to make use of the parasitic X-rays from an electron–positron collider (CESR, which had been in operation since 1975). Keith Moffat pioneered use of CHESS for macromolecular crystallography, assisting a number of scientists to use the facility beginning in 1981. I recall the excitement of finding that exposures of 24–36 h on a lab source now took only a few minutes at CHESS! In 1984, Moffat secured a grant from NIH to establish the Research Resource MacCHESS (`Macromolecular crystallography at CHESS', or `son of CHESS', according to Moffat's Scottish background), a formalization of the program to support and improve a user facility for structural biologists, and became its first Director. The MacCHESS grant has continued to the present day, and I was part of it from its founding until my retirement in 2022, serving as Director from 2008 to 2022.

The forte of MacCHESS is development of new techniques to facilitate the use of X-rays by structural biologists – largely macromolecular crystallography, but also biological small angle X-ray scattering, BioSAXS. Many of these techniques, developed in the late 20th century, are now standard at synchrotron sources worldwide.

For example, with the intense X-rays from a synchrotron source, it became clear that radiation damage to macromolecular crystals was a serious problem. Cooling the crystals to liquid nitro­gen temperature was the solution, and experiments at CHESS were critical in developing cryocooling techniques. Once it was realized that very rapid cooling was required for success (hence precluding use of a capillary surrounding the crystal), Håkon Hope developed a method of mounting crystals using oil and very thin glass slips (created by blowing a bubble of glass and shattering it) (Hope, 1988). This method was utilized in tests at CHESS but was quickly supplanted by use of a thin free-standing film of mother liquor enclosed in a loop. The loop mount (Teng, 1990) was originally developed by T.-Y. Teng, a scientist in Moffat's lab, as a means of reducing strain on a thin plate crystal, but it proved a boon for cryocooling: easy to make, easy to load with a crystal, and producing very little background scatter (especially after the original gold wire loop was replaced with X-ray transparent nylon). The cryoloop was soon commercialized and became a mainstay of crystallography at synchrotron sources and elsewhere.

Along with other synchrotron facilities, MacCHESS has been active in developing techniques to automate crystal handling, data collection and processing, to the point that now a user need only grow crystals, ship them to the facility, and use a well designed GUI to carry out their experiment. Various beamlines tend to put their own spins on their automation suite, but there has been plenty of `cross-pollination' between them, to which MacCHESS has contributed, particularly in the 1990s and 2000s.

MacCHESS has been key in the development of increasingly sophisticated X-ray detectors for macromolecular crystallography. When Rossmann solved the structure of human rhinovirus at CHESS (Rossmann et al., 1985), it required processing hundreds of pieces of X-ray film. When imaging plates were developed to replace film, first at Kodak and later by Fuji, they were quickly adopted at MacCHESS to provide digitized diffraction patterns. Then followed extremely productive collaborations with Sol Gruner's lab and with Area Detector Systems Corp. to test, and integrate into the crystallographic process, CCD-based detectors (Gruner & Ealick, 1995) and hybrid pixel-array detectors (Rossi et al., 1999).

Although most data taken at CHESS utilize monochromatic X-rays, MacCHESS pioneered the synchrotron Laue method, beginning in 1984 (Moffat et al., 1984), as a means of performing very rapid, i.e. time-resolved, protein crystallography. Using Laue diffraction in conjunction with a borrowed undulator, CESR running in single-bunch mode and a custom high-speed shutter, it was possible to obtain useful data from a 120 ps pulse of X-rays (Szebenyi et al., 1988). The Laue method was promptly taken up by Helliwell (1984) at SRS Daresbury and its users (e.g. Hajdu et al. 1987) and extended to neutrons, firstly at the Institut Laue Langevin in Grenoble in France (Habash et al., 1997).

1.3. Protein crystallography at Daresbury 1987 onwards (CN)

By 1987, the protein crystallography facilities on beamlines 9.6 and 7.2 were well established with a high scientific output and beamline 9.5 was being developed for very rapid Laue and rapidly tunable monochromatic experiments, under a collaboration with the Swedish Research Council. Early Laue experiments clarified the capabilities of the technique and acted as a test bed for Laue software developments. For monochromatic data collection, the SRS was an early adopter of data collection at cryogenic temperatures, a technique which allowed the study of weakly diffracting crystals before radiation damage occurred. Daresbury staff investigated radiation damage in protein crystals at cryogenic temperatures in the early 1990s using beamlines 9.7 and 9.5 in Laue mode. The subject of radiation damage in protein crystallography resulted in a regular series of international workshops which has lasted 26 years thus far.

Scientific highlights using these facilities included structures from Oxford of foot and mouth disease virus and time-resolved studies on glycogen phospho­rylase, the structure determination of light harvesting protein (Glasgow and Daresbury) and, in 1994, the structure of F1-ATPase by a team from the MRC Laboratory of Molecular Biology. This work resulted in a Nobel Prize for Chemistry, awarded to John Walker in 1997. There was also increasing interest in the use of the SRS by the pharmaceutical industry and agreements were set up with a variety of companies to provide both beam time and data collection services. By the beginning of the 21st century the protein crystallography facilities at the ESRF were increasingly being used for the most demanding projects. However, even for these, the SRS was being used for initial investigations including screening crystals to identify the best crystallization conditions. One notable example of this is the work on 30S ribosome which led to the Nobel Prize for Chemistry (to Venki Ramakrishnan in 2009). This Nobel Prize was shared with Ada Yonath (also an SRS user) and Tom Steitz.

The research councils (and later the Wellcome Trust) involved in biological and medical sciences were very keen to see that the biology beamlines on the SRS were maintained at a high level and invested additional funds. This resulted in the construction of three new beamlines – 14.1 and 14.2 [available 2000 (Duke et al., 1998)] and 10.1 [available 2003 (Cianci et al., 2005)] – on multipole wiggler beamlines, two of which were equipped with multi-wavelength anomalous dispersion (MAD) capabilities. These facilities also acted as a focus for automation and incorporated robotic sample changers.

The Collaborative Computational Project for Protein Crystallography (CCP4) was based at Daresbury during the lifetime of the SRS and good links were developed between these two activities. An informal collaboration was set up with other European synchrotrons and the MRC Laboratory of Molecular Biology with the aim of automating data collection for protein crystallography. Several successful joint grant applications extended this further, funded by UK Research Councils or the European Commission.

After the announcement that the new synchrotron Diamond was to be built on the Rutherford Appleton Laboratory site, investment in the SRS nevertheless continued to sustain a key driver for the new facility. The planned transfer of activities from the SRS to Diamond went well for protein crystallography and the developments in automation and data handling which took place on the SRS formed a good basis for further developments at Diamond.