The S100PX (PX50) system for protein crystallography

S100PX Facts

The S100PX system is the first commercially available system for protein crystallography to use an ellipsoidal monocapillary optic as the primary optical element for beam intensification and focusing. The system is now used in an increasing number of laboratories where it demonstrates the most intense beam available outside synchrotron facilities for the study of small (<0.1mm) protein crystals.


The S100PX monocapillary optic was originally designed to demonstrate the most intense beam focus possible using a high-brightness rotating anode source for the analysis of protein crystals previously considered too small to analyse in the laboratory. Laboratory trials of the system conducted in conjunction with the Biomolecular Research Institute showed that the optic not only produced unprecedented data from small crystals, but also produced better data from large crystals than had been previously collected using conventional mirror systems (see below). The collection of a structure-solution dataset to 2.2Å from a 184Å cell-edge exoglucanase crystal of only 30 microns diameter - collected using an ageing 600W generator - was reported at IUCR XX (Varghese et al, Abstracts, IUCR '99).

Optical Principles

The design of the S100PX optic called for the most intense possible source, to be re-imaged at the specimen position to maximise intercepted flux by a small specimen crystal. This requires the use of a 0.1 x 1 mm focus cup in the generator, limiting total power to ~1kW, but producing an anode loading of 10 kW/mm - about four times the source brightness of the larger target sizes typically used with mirror optics. Current mirror technologies cannot best utilise high-brightness sources due to the high aberration introduced by long beam paths and multiple reflections. The source image produced at the specimen by the S100PX is about 0.1 mm in diameter, and the length of the optic is selected to limit total (i.e. above background, not FWHM) beam divergence to 0.2 degree. The optic is a "low-pass" filter, efficiently suppressing white radiation and attenuating Kb wavelengths.


"Don't capillary optics produce too much divergence for protein crystallography?"

This is certainly true of multiple-reflection capillary optics such as polycapillaries and nonimaging (or nonfocusing) monocapillaries where the output divergence is always twice the critical angle for reflection from the glass surface, or about 0.6 degree total for CuKa. The S100PX produces a true geometrical focus of singly-reflected X-rays at a point in space about 100 mm from the optic with a total beam divergence above background of 0.2 degree, or about 0.16 degree FWHM. The small incident beam diameter at the specimen permits a slightly greater beam divergence for additional diffracted flux without sacrificing max. unit-cell resolution. Beware of optics specifying only FWHM divergence - your specimen crystal proportionally diffracts everything incident on it within its oscillation range and mosaic spread, not just that between arbitrary FW limits! Some optics produce low-intensity divergence "tails", often asymmetrically distributed between the horizontal / vertical / diagonal if multiple mirror-reflections are employed.

"Shoudn't you have a beam bigger than the crystal?"

The valid comparison is with the typical synchrotron PX beamline. For intense beams, a small-diameter beam reduces scatter and samples a smaller volume of crystal. For frozen specimens, this often means sampling a smaller mosaic region of crystallite - which can mean the difference between good data and no data. With poor-quality crystals, better data are obtained due to the reduced long-range disorder seen by the smaller beam - often resulting in a dataset where none was obtainable using a large beam. Large, perfect crystals that diffract well will produce excellent data using the S100PX - but these are not the reason for choosing an optical system. In this case, lysozyme refinements with the S100PX demonstrate better R-merge and max. structure resolution than that seen with mirrors, mostly due to reduced scatter giving better I/sigma<I>.

"Does the monocapillary beam require corrections in our structure package?"

The monocapillary-focused beam is unpolarised and in fact more symmetrical about the beam axis than any other optical system, due to the beam being toroidally (cylindrically) focused with a single reflection between source and specimen. Our users report better data integration between azimuthal and lateral regions within each diffraction image than previously seen using mirror systems. No refinement changes are needed apart from the removal of the polarisation correction if a monochromator was previously used. Some older packages may require the use of the center-of-mass (as opposed to peak pixel value) method for finding reflection centres.

"Should I use the same camera length as I did with mirrors?"

The small beam diameter at the specimen often results in the ability to collect data at shorter camera lengths than previously possible with a larger-diameter beam. This provides two major advantages: Smaller spots on the camera give more counts/pixel, smaller integration boxes and hence better I/sigma<I>, and more two-theta coverage by the camera. Smaller spots also seem to greatly assist integration statistics in spiral-scan camera systems. Once again, this is only possible due to the reduced incoherent scatter produced by the small-diameter monocapillary-focused beam.

"What about Kb?"

The S100PX optic is designed to preferentially reflect Ka and increasingly suppress higher energies. In many cases, residual Kb can either be ignored - if the unit cell is small, or the crystal small or weakly diffracting - or filtered by a continuously-variable Ni filter integral to the S100PX system. Experienced users often find that they often can still routinely collect unfiltered data from large unit cells in the presence of parasitic Kb reflections as the extremely low background produced by the S100PX system results in only the loss of a relatively small amount of low angle, low structure-resolution data. Note - Many of the claims made against nondispersive mirror optics regarding non-characteristic "background" and I/sigma<I> have more to do with incoherent scatter of Ka from the sample and its environment, especially with large-diameter beams.

"How long do they last?"

We haven't had one fail yet, in over ten years of use. Not even our synchrotron optics demonstrate measurable reflectivity degradation after many beam-hours, often in white-beam. This is due to the use of a proprietary mirror surface which cannot tarnish and contains no heavy elements to "boost" reflectivity - or add fluorescence spectra to the output beam!

"Are there major modifications required to our existing system?"

AXCO has endeavoured to produce a range of "bolt-on and plug-in" retrofit S100PX system profiles for the majority of rotating-anode / camera combinations currently in use. Refits can take as little as a few hours and can be done by any qualified technician. For nonstandard or custom refits, AXCO's workshop can supply a ready-to-run system given a set of engineering details of the existing setup.

"Is the optic hard to align?"

The monocapillary optic has to be the easiest optical system to align on the market. Only one reflection to get right, around one axis, with no left/right or up/down bias. Max. intensity at the specimen position coincides with optimum alignment and ideal beam symmetry. The alignment technique is identical to that for a collimator.

" My question isn't in this list!"

Please ask! Contact us

(c) copyright AXCO 2000