OASIS and Xe phasing: potential in high-throughput crystallography

1. Introduction

In view of the mounting evidence that one-wavelength anomalous scattering (OAS or otherwise known as SAD) may be sufficient to solve protein structures (Hao, 2000), the OASIS program (Hao et al, 2000) was written to determine phases using one-wavelength anomalous scattering data instead of using additional multiwavelength diffraction data (MAD). This is of particular importance when protein crystals are sensitive to X-ray irradiation or the absorption edges of the anomalous scatterers, such as xenon and sulfur, are difficult to access. A number of minor changes in the new version of OASIS to be released by CCP4 include new keywords to allow resolution and sigma cutoff. The upper limit on the number of reflections has been increased to 150,000 (from 90,000).

In preparing samples for MAD or OAS phasing, the most favored approach is the incorporation of selenium into protein using seleno-methionine during the expression of the protein. However, this is only successful when the gene encoding the particular protein is known and an expression is established and when the substitution does not affect crystalline order. In cases where seleno-methionine substitution is not plausible an attractive method for preparing samples is to incorporate xenon gas into the crystal. Xenon is known to bind to hydrophobic pockets within proteins at modest pressure. The one-wavelength anomalous scattering data (courtesy of Dr Rizkallah and Professor Helliwell, see Table 1 for details) collected from a xenon derivative of the lobster apocrustacyanin A1 protein (Cianci et al., 2001) was used to test the possibility of ab initio phasing.

2. Locating the xenon sites

The Se anomalous scatterers for both structures were located by the conventional direct-methods program SAPI (Fan et al., 1990; http://staff.chess.cornell.edu/~hao/sapi/sapi.html) using magnitudes of anomalous differences, for reflections within 3.0  Å. The solution was selected by a default run of the program. The largest 416 normalized structure factors E's were used in tangent formula phase refinement. The resultant electron density map produced a group of 3 highest peaks; there was a clear gap between this group and other peaks in terms of peak height. A Karle-recycle refinement (an option in SAPI) of these three sites yielded an additional minor site.  The absolute configuration of these sites was determined by the program ABS (http://staff.chess.cornell.edu/~hao/abs/abs.html) based on the P_s-function method (Woolfson & Yao, 1994). These Xe sites agreed well with the published sites (Cianci et al, 2001) and formed the basis for the next phasing step.

3. OASIS and DM phasing

The ab initio phasing of the OAS data was implemented in the computer program OASIS (Hao et al., 2000). All Friedel pairs (including centric reflections) were evaluated using OASIS. The script that was used to run OASIS is shown below:

#oasis.com

oasis HKLIN xea1_19_trn.mtz HKLOUT xea1_oasis.mtz << eof

TITLE DIRECT PHASING OF xea1 xenon OAS DATA

HCO XE 12

FIT

RES 2.3

LCE 7

ANO XE 11.5

POS XE    -0.62360 -0.52335 -0.50331  1  0.318

        XE    -0.87037 -0.55929 -0.99017  2  0.419

        XE    -0.65256 -0.76443 -0.87725  3  0.268

        XE     0.47078  0.20858  0.15206  4  0.080

LABIN F1=F SIGF1=SIGF F2=DANO SIGF2=SIGDANO

LABOUT F1=F SIGF1=SIGF PHI=PHIdp W=Wdp

END

eof

Density modification using the CCP4 program DM (Collaborative Computational Project, Number 4, 1994) was then applied to the resulting phase sets. Phase error analysis and figures of merit before and after DM are given in Table 2. The electron density maps after OASIS and density modification clearly revealed the solvent boundary. The C trace was clearly visible but there were a number of places where the electron density was broken. A correlation coefficient between the OASIS  + DM phased map and the final refined structure was 0.57.

4. Discussion

Here we demonstrate that, by exploiting the anomalous signal at single wavelength, OASIS can be used to determine phases at moderate (2.3  Å) macromolecular crystallographic resolution for a medium-size protein. The total CPU time consumed by SAPI, ABS and OASIS was about 3 minutes on an Alpha XP10000 workstation. As the xenon derivatives can be obtained from native protein crystals using commercially available equipment in a relatively short time frame (a few hours), the method described in this paper may provide an attractive alternative to MAD or MIR phasing, in particular when high-throughput is desirable.