IMAGE UNAVAILABLE FOR COPYRIGHT REASONS © 1995 Nature Publishing Group http://www.nature.com/naturebiotechnology • /BIOTOOLS X-ray Crystallography at Extremely Lovv Temperatures Chilling protein crystals enables the characterization of transient structural intermediates C onventional X-ray crystallography is useful for looking at macromolecular structures that are stable at ambient temperatures. But in enzymic and even in binding processes, in- termediate structures are often more informative. By their nature, of course, intermediate structures do not last long enough under normal conditions to obtain X-ray diffraction data. However, by cooling protein samples to within a few tens of Kelvin of absolute zero, the lifetime of the intermediate states can be extended and their structures determined. It has now become routine for macromolecular crystal structures to be determined at low tempera- tures, around 90-100 K. This is done by mounting the crystal in a thin fiber loop 1 rather than in the more conventional X-ray capillary, and by bathing the crystal and loop in a cold gas stream derived from the boil-off of liquid nitrogen. As with studies at room temperature, this proce- dure yields a time average over all structures present during the X-ray exposure time and a space average over all structures in the crystal. However, key tertiary structural changes that accompany processes such as enzyme catalysis, photocycling, and ligand binding and release are often extremely fast at room temperature, and transient structural intermediates cannot be visualized even by the time-resolved crys- tallographic techniques presently under develop- ment. 2 Under this circumstance, the only recourse may be to work at cryogenic temperatures, in order to slow down these structural processes and prolong the lifetime of the structural intermediates to the point where they can be readily observed. Two groups 3 _. have recently adopted this ap- proach in studying the photostimulated process of carbon monoxide release (and subsequent rebinding in the dark) from the small heme protein myoglobin, in single crystals. In solution at room temperature, the "germinate" component of the rebinding reac- tion, in which the carbon monoxide recombines directly with the heme from which it has just been liberated by light, is extremely rapid; spectroscopic studies show that heme structural changes occur on subnanosecond time scales. As very expensive stud- ies, principally by Frauenfelder and colleagues over the past 20 years, have shown, 5 the rebinding reac- tion and the associated heme and protein structural relaxation are both extremely complicated and ex- quisitely dependent on temperature. The lifetimes of spectroscopically detectable in- termediates-which presumably also differ in their heme and protein structure-can be prolonged by many orders of magnitude, by working at tempera- tures between 10-80 K, to the seconds time scale or even longer. Two groups, one 3 working at 40 K using a new cryostat 6 based on the boil-off liquid helium, and the other4 at 20 K, both succeeded in trapping a normally short-lived intermediate. In this intermediate, the carbon monoxide molecule has been photodissociated from the heme yet remains nearby, and the heme and its surrounding protein have partly 3 or substantially 4 relaxed towards the unliganded, deoxymyoglobin form. The crystallographic results differ in detai I, point- ing to a potential difficulty of cryogenic crystallog- raphy. Results may depend on fine aspects of the cooling protocol (extremely rapid cooling would "freeze in" the room tem- perature structural distri- bution, but in practice this is not achievable, and dif- ferent parts of the crystal freeze at different rates) and on the exact protocols used to illuminate the crys- tal, to photo-initiate the reaction and to acquire the X-ray diffraction data (the energy of all visible and X-ray photons ab- sorbed largely appears as heat and may promote structural relaxation). Finally, the task remains to demonstrate that intermediates which are stabilized at cryogenic temperatures indeed are identical to, or differ in verifiable ways from, those at room tem- perature. The technology to execute such experiments is clearly in place. We now need more experience to assess these questions. References l. Teng, T.Y. 1990. J. Appl. Cryst. 23:387-391. 2. Cruickshank, D.W.J., ct al.1 992. Time-Resolved Macromol- ecul ar Crystallography. Oxford Science Publi cations. 3. Teng. T.Y., et al. 1994. Nature Struct. Biol. 1:701-705. 4. Schlichting, l. , et al. 1994. Nature 371:808-81 2. 5. Austin, R.H., et al.1975. Biochemistry 14:5355-537 l. 6. Teng, T.Y., et a l.1 994. J. Appl. Cryst. 27: 133-139. Keith Moffat is at the department of biochemistry and molecular biology, University of Chicago, 902 E. 58th Street, Chicago, lllinois 60637 (e-mail: moff at@cars 1. uchicago.edu). Keith Moffat FIGURE 1. Conventional X-ray crystallography gives a structure of stable myoglobin (featured), whereas using extremely low temperatures can show the protein in action. Picture courtesy of Oxford Molecular, using atomic coordinates from the Protein Data Bank. BIO/TECHNOLOGY VOL. 13 FEBRUARY 1995 133