HIGH RESOLUTION DIFFUSE X-RAY SCATTERING BY STRUCTURALLY DEFECTIVE PROTEIN CRYSTALS
Richard J. Matyi
Physics Laboratory , National Institute of Standards and Technology
Gaithersburg , MD 20899
High resolution X-ray methods have a successful history in correlating the growth process of inorganic crystals with the types and densities of grown-in structural defects. As shown in Figure 1, crystal defects will alter the characteristics of a reciprocal lattice vector H hkl (and hence the intensity distribution about a reciprocal space point) in two principal ways. First, strains or composition variations will change the lattice parameter of the sample and will cause a change in the length of the reciprocal lattice vector (equal to 1/d hkl). This will result in a redistribution of the intensity away from the exact Bragg condition in the q/2q direction, or parallel to the reciprocal lattice vector direction in the symmetric geometry. In contrast, mosaic spread or lattice tilts will alter the direction of H hkl and will create extra intensity away from the Bragg peak perpendicular to the reciprocal lattice vector.
Direct observation of the defect-generated deviations of a reciprocal lattice vector is achieved with high resolution triple axis X-ray diffraction and perfect crystal optics. Since the angular breadth of a Bragg reflection from an ideally perfect crystal in a laboratory setting is typically on the order of arcseconds, conditioning the angular spread of the incident (S o/ l) and diffracted (S/ l) wavevectors with crystal optics allows the scattering generated by structural defects can be observed. Of course, the orientation of the reciprocal lattice vector H hkl of the sample must be controlled with arcsecond precision as well. Figure2 shows a series of reciprocal space maps that were recorded from a HEWL crystal as a function of irradiation time. The data are plotted in terms of q x and q z, which are related to the deviations a and b of the sample and analyzer crystals, respectively, from the exact Bragg condition q B by 
and 
.
The evolution of the peak intensities and peak breadths (at 0.5 and 0.01 of the peak maximum) from several lysozyme crystals with irradiation time are shown in Figure 3. The longitudinal peak breadth (constant q/2q value passing through the Bragg peak) remained approximately constant throughout the irradiation study with FWHM values remaining between 20 and 25 arcseconds. The transverse peak breadths at long irradiation times (greater than 45 hours markedly increase, while the Bragg peak intensities decreases, which is expected in a sample with increasing disorder. At shorter times, a slight narrowing of the peak breadth was observed.
Figure 4 shows the off-peak diffuse intensity from one of the crystals on a double-logarithmic plot. Although the large number of data points makes individual scans difficult to pick out, it can be seen that the overall trend in the data is that for short irradiation times away from the Bragg peak, the diffracted intensity as a function of q x falls off with roughly the same functional dependence (i.e. the same slope in a log I – log q plot). At longer irradiation times, however, the trend is for the data to exhibit two different slopes: one at a shallower slope than that for early radiation times, and one steeper slope that dominates at large q x values. As the irradiation time increases, the second, steeper slope starts appearing at larger q x values. The Figure also shows a plot of the transition point for the data collected at long radiation times with dual slopes in terms of the reciprocal of the scattering vector corresponding to this change in slope. Because 1/q x has units of length, we refer to this value as a measure of a “characteristic size” that decreases with increasing irradiation time. It is interesting that these potential defects are on the order of microns, which is quite large. This number may be an indicator of the size of an undamaged region during X-ray irradiation.
The preceding approach presumes, of course, that a protein can be coaxed into a form that is sufficiently ordered to deserve the designation “crystal”. Sadly, in many cases, a crystal that appears by optical evaluation to be promising for subsequent analyses turns out to be so poorly diffracting that no useful structural data can be extracted. Further study is then not possible, leading to a depressing state in which the protein crystal grower may have limited (or no) options available to guide the development of a crystallization regimen. Small angle X-ray scattering (SAXS) may provide an alternate approach, however. SAXS theory shows that a “large” perfect crystal should generate a compact intensity distribution in reciprocal space that should not be observable. In other words, a homogeneous protein crystal of macroscopic size with high levels of structural perfection would be expected to generate zero-order scattering that would be immeasurably small in its extent. In contrast, highly imperfect crystals would be expected to generate a SAXS signal.
Figure 5 compares scattering curves obtained from the two crystals (initially un-irradiated HEWL-N and intentionally radiation damaged HEWL-Q) after each of three synchrotron radiation cycles. Again, it must be remembered that the crystal and its capillary was inserted into the X-ray beam immediately prior to each data collection run and was removed immediately afterwards. The data in the Figure show that at each synchrotron irradiation cycle, the initially damaged crystal exhibits more intense scattering than does its initially un-irradiated colleague. In addition, the increasing ordinate in each plot in the Figure clearly shows the increase in the scattered intensity as exposure time in the synchrotron beam increases. Finally, the scattering curve from the HEWL crystal with presumably the least radiation damage (HEWL-N, first cycle) shows what may be an oscillatory behavior; the first cycle scattering of initially damaged HEWL-Q exhibits less intense oscillations. As synchrotron radiation proceeds, however, the scattering curves seem to “smooth out” even as their net intensity increases.
This work demonstrates that high resolution X-ray scattering methods – in the vicinity of reciprocal lattice points ranging from hkl to 000 – can yield valuable insights into the nature of physical defects generated by radiation damage in protein crystals. In the semiconductor community, it has taken several years for these high resolution methods to elevate their status from that of laboratory curiosity to an established and important analytical technique. Hopefully, with continued progress (particularly in the modeling of the diffuse intensity) a similar progression will occur to the benefit of the protein crystal growth community.
References:
