Time-Resolved Macromolecular Crystallography at Pulsed X-ray Sources

Abstract

The focus of structural biology is shifting from the determination of static structures to the investigation of dynamical aspects of macromolecular function. With time-resolved macromolecular crystallography (TRX), intermediates that form and decay during the macromolecular reaction can be investigated, as well as their reaction dynamics. Time-resolved crystallographic methods were initially developed at synchrotrons. However, about a decade ago, extremely brilliant, femtosecond-pulsed X-ray sources, the free electron lasers for hard X-rays, became available to a wider community. TRX is now possible with femtosecond temporal resolution. This review provides an overview of methodological aspects of TRX, and at the same time, aims to outline the frontiers of this method at modern pulsed X-ray sources.

Keywords: Monte Carlo integration; bacterial phytochromes; beta-lactamase; serial femtosecond crystallography; time-resolved crystallography.

Figure 1

A time-resolved macromolecular crystallography (TRX) experiment with the pump–probe method. The pump laser pulse (blue bar) starts the reaction, and the ultrashort X-ray pulse (red bar) probes the progress of the reaction after a time delay, ∆t. Here, X-ray data collection in the dark as the reference is interleaved with data collection after laser light activation. The X-ray data collection rate is 120 Hz, as available at the LCLS, and the laser repetition rate is 60 Hz in this example.

Figure 2

Screenshot of a photoactive yellow protein (PYP) crystal (pale yellow) displayed on a monitor at the XPP instrument of the LCLS during laser illumination. The crystal of size 900 × 40 × 40 µm³ is kept in a glass capillary. It is illuminated by a femtosecond laser pulse (laser focus). The crystal apparently acts as a waveguide, with the laser light exiting at both ends.

Figure 3

Ultrafast structural changes in the chromophore pocket of PYP [5] Green represents a positive difference electron density and red a negative difference electron density (on the 3/-3 σ contour level). Yellow structure represents a reference (dark state) structure. The p-coumaric acid (PCA) chromophore as well as some nearby residues are marked. (a) 250 fs after laser excitation (pink structure); the chromophore configuration is still trans. Larger structural changes are denoted by arrows. (b) 3 ps after laser excitation (green structure); the structure is cis. Isomerization occurred about the double bond (curved arrow) at the chromophore tail. Some structural changes are also shown by arrows.

Figure 4

Time-resolved crystallographic photoflash experiment on the L29W mutant of Mb–CO [66]. (a) Overall structure of Mb^L29W–CO in the dark. Dashed box: heme pocket. Some important residues are displayed. The heme iron is shown as a yellow sphere. (b) Close-up of the heme pocket 1 ns after an intense optical laser flash to start photodissociation of the CO from the heme. t. The heme and important residues are marked. Red: negative difference electron density; blue: positive difference electron density (−/+ 3 σ contour levels, respectively). Red arrows show structural relaxations at this time delay. In this mutant, the Trp29 transiently occludes the primary docking site of the CO. CO is found at time delays >1 μs on the proximal side of the heme (red-circled p).

Figure 5

Structural comparison of penicillin-binding protein (PBP, blue, PDB entry 3OCN) and M. tuberculosis β-lactamase (BlaC; brown, PDB entry 6B69). (a) Overall view: BlaC is structurally very similar to the C-terminal domain of the PBP. (b) Detailed view of the active sites with essential amino acids in BlaC/PBP shown according to numbering convention. The cephalosporin antibiotics ceftriaxone (CFO*, red) and ceftazidime (CEFTA, green) are bound to serine residues (arrow) in the active sites of BlaC and PBP, respectively. The leaving group of the cephalosporins (R, arrows) is cleaved off.

Figure 6

Active site structures during the BlaC reaction with CEF [64] as determined by MISC. Simulated annealing ‘omit’ difference density shown in green on the 2.5 σ level. (a) Catalytic cleft of the free enzyme, phosphate (P_i), and water is present. Ser-70 is marked. The position of CEF is shown in gray as a guide to the eye. (b) Catalytic cleft 30 ms after mixing; resolution 2.75 Å. Strong electron density shows the CEF ligand. (c) Catalytic cleft 100 ms after mixing; resolution 2.15 Å. Gap in electron density (red arrow, α) shows that the covalent bond has not yet formed. Black arrows in (b,c) point to the sulfur of the leaving group. (d) Catalytic cleft 500 ms after mixing; resolution 2.2 Å. The gap between the CEF and Ser-70 is closed, indicating a covalent bond (red arrow, β). The sulfur feature (black arrow, behind the hydroxyl) is gone. (e) Catalytic cleft in the steady state. Similarities to the 30-ms and 100-ms time delays are evident.

Figure 7

Bacterial phytochrome structures. (a) Structure of the central chromophore, biliverdin (BV). In dark-adapted BphPs, biliverdin (BV) is found in most cases in the Z syn–syn–anti configuration. Red arrow: isomerization upon red light absorption. (b) Structure of the myxobacterial phytochrome 1 (SaBphP1) photosensory core module; PDB entry 6BAO [32]. PAS, GAF, and PHY domains are colored yellow, green, and magenta, respectively. The sensory tongue, the knot, and the BV are marked. Black arrows: structural displacements of the PHY domains after light absorption; PDB entry 6BAO. (c) Structure of the full-length Idiomarina spp. phytochrome-activated diguanylyl cyclase (IsPadC), pdb entry 5LLW [142].