The analysis of X-ray diffraction intensities is complicated by the anisotropy of anomalous scattering (AAS) that can occur due to resonance associated with transitions between core electrons and valence molecular orbitals. Substantial AAS has been observed directly in diffraction data near the K edge of selenium in selenolanthionine [Templeton & Templeton (1988). Acta Cryst. A44, 1045-1051] and in pleiochroism of X-ray absorption in selenobiotinyl streptavidin [Hendrickson, Pähler, Smith, Satow, Merritt & Phizackerley (1989). Proc. Natl Acad. Sci. USA, 86, 2190-2194]. The impact of AAS on the multiple-wavelength anomalous diffraction (MAD) method for phase determination is of particular interest in the context of this chemical state of selenium in the light of a general method that has been developed to incorporate selenomethionine into proteins for use in MAD phasing [Hendrickson, Horton & LeMaster (1990). EMBO J. 9, 1665-1672]. The first step of the MAD phasing method necessarily assumes that the anomalous-scattering factors are isotropic and our first aim here is to evaluate the effect of this approximation on initially determined phases. To obtain ultimate phases free from the effects of anisotropy, a least-squares procedure has been written in which global parameters (i.e. pertaining to the whole data set) are refined simultaneously with local parameters (i.e. pertaining to a given node h). The AAS is taken explicitly into account by considering f' and f” as tensors instead of scalars [Templeton & Templeton (1982). Acta Cryst. A38, 62-67], and the components of the f' and f'' tensors are among the refinable global parameters. The effectiveness of this procedure is tested with data simulated from the refined atomic model of selenobiotinyl streptavidin. The application of this procedure to actual Photon Factory measurements is also described. The results show that AAS does not cripple the MAD method, and that phases uncorrupted by these effects can be recovered.

We present the final design of the x-ray optical systems and experimental stations of the two macromolecular crystallography (MX) beamlines, FMX and AMX, at the National Synchrotron Light Source-II (NSLS-II). Along with its companion x-ray scattering beamline, LIX, this suite of Advanced Beamlines for Biological Investigations with X-rays (ABBIX, [1]) will begin user operation in 2016. The pair of MX beamlines with complementary and overlapping capabilities is located at canted undulators (IVU21) in sector 17-ID. The Frontier Microfocusing Macromolecular Crystallography beamline (FMX) will deliver a photon flux of ~5x10^12 ph/s at a wavelength of 1 Å into a spot of 1 - 50 µm size. It will cover a broad energy range from 5 - 30 keV, corresponding to wavelengths from 0.4 - 2.5 Å. The highly Automated Macromolecular Crystallography beamline (AMX) will be optimized for high throughput applications, with beam sizes from 4 - 100 µm, an energy range of 5 - 18 keV (0.7 - 2.5 Å), and a flux at 1 Å of ~10^13 ph/s. Central components of the in-house-developed experimental stations are a 100 nm sphere of confusion goniometer with a horizontal axis, piezo-slits to provide dynamic beam size changes during diffraction experiments, a dedicated secondary goniometer for crystallization plates, and sample- and plate-changing robots. FMX and AMX will support a broad range of biomedical structure determination methods from serial crystallography on micron-sized crystals, to structure determination of complexes in large unit cells, to rapid sample screening and data collection of crystals in trays, for instance to characterize membrane protein crystals and to conduct ligand-binding studies. Together with the solution scattering program at LIX, the new beamlines will offer unique opportunities for advanced diffraction experiments with micro- and mini-beams, with next generation hybrid pixel array detectors and emerging crystal delivery methods such as acoustic droplet ejection. This work is supported by the US National Institutes of Health.

Native metalloproteins have been routine subjects for mutli- and single-wavelength anomalous diffraction (respectively MAD and SAD) analyses for some time; however, despite notable early successes, native molecules without heavy atoms (Z ≥ 25) have only recently become routinely accessible. Crystals of native proteins and nucleic acids have substantial contents of light elements (P, S, Cl, K, Ca) of potential for use in SAD phasing. Anomalous signals from such elements can be enhanced by using a lower than usual x-ray energy; nevertheless, typical Bijvoet differences usually still remain at a level comparable to noise. We have devised robust SAD procedures to study native, light-atom-only biological structures. We have so far used a modestly low energy (6 - 7 keV), but we further enhance the signal-to-noise in anomalous diffraction by combining data from multiple crystals chosen to be statistically equivalent. We have applied our multi-crystal native SAD approach in several structure determinations (1,2) at sizes up to 1200 ordered residues per asymmetric unit and at resolutions so far as low as 3.2 Å. Our tested practices can be replicated readily, and we plan further improvements in computing protocols and in instrumentation.