Preparing heavy-atom derivatives

Several diffraction criteria define a promising heavy-atom derivative. First, the derivative crystals must be isomorphic with native crystals. At the molecular level, this means that the heavy atom must not disturb crystal packing or the conformation of the protein. Unit-cell dimensions are quite sensitive to such disturbances, so heavy-atom derivatives whose unit-cell dimensions are the same as native crystals are probably isomorphous. The term isomorphous replacement comes from this criterion. 

The second criterion for useful heavy-atom derivatives is that there must be measurable changes in at least a modest number of reflection intensities. These changes are the handle by which phase estimates are pulled from the data, so they must be clearly detectable, and large enough to measure accurately. 

Finally, the derivative crystal must diffract to reasonably high resolution, although the resolution of derivative data need not be as high as that of native data. Methods of phase extension (Chapter 7) can produce phases for higher-angle reflections from good phases of reflections at lower angles. 

Having obtained a suitable derivative, the crystallographer faces data collection again. Because derivatives must be isomorphous with native crystals, the strategy is the same as that for collecting native data. You can see that the phase problem effectively multiplies the magnitude of the crystallographic project by the number of derivative data sets needed. As I will show, at least two, and often more, derivatives are required. 

Obtaining phases from heavy-atom data 

---

Prepare heavy-atom derivatives of the protein with heavy atoms in different positions in the unit celt. Measure an intensity data set for each heavy-atom derivative. 

In particular the structure determination of cucumber basic protein has been made using MAD data recorded on this instrument around the CuK edge (Guss et al 1988, section 9.7.5). This protein crystal structure had previously defied solution for many years because of problems associated with preparing heavy atom derivatives. 

The electron density maps published so far have been based on the orthorhombic crystal form. The anions Pt(CN)4 - and Au (CNlz and the mercury complex ethyl mercurythiosalicylate were used to prepare heavy atom derivatives (106,107). The overall dimensions of the molecule... 

NCP crystals. There were two facets to this approach. First, it was necessary to reconstitute NCPs from a defined sequence DNA that phased precisely on the histone core to circumvent the random sequence disorder. It was obvious that the DNA was important for the quality of the diffraction from NCP crystals but the role of histone heterogeneity was not so clear. Heavy atom derivatives (i.e., electron rich elements bound in specific positions on the proteins) were not readily prepared by standard soaking experiments, due to a paucity of binding sites. Hence, it was necessary to selectively mutate amino acid residues in the histones to create binding sites for heavy atoms. 

Although the soaking method for heavy-atom derivative preparation is by far the simplest and most common, it is not the only method used. One can first derivatize the macromolecule, and then crystallize. This procedure is less frequently used because of drawbacks such as the inabihty to produce isomorphous crystals due to the disruption of intermolecular contacts by the heavy atoms. Other frequent problems are the introduction of additional heavy-atom sites (a potential complicating factor in phasing) by exposing sites hidden by crystal contacts, and changing the solubility of the derivatized macromolecule. 

Hatfull, G. R, Sanderson, M. R., Freemont, P. S., Raccuia, P. R., Grindley, N. D. F. and Steitz, T. A. (1989). Preparation of heavy-atom derivatives using site-directed mutagenesis introduction of cysteine residues into yS resolvase. /. Mol. Biol. 208,661-667. 

Petsko, G. A. (1985). Preparation of isomorphous heavy-atom derivatives. Method Enzymol. 114,147-156. 

Heavy-atom derivation of an object as large as a ribosomal particle requires the use of extremely dense and ultraheavy compounds. Examples of such compounds are a) tetrakis(acetoxy-mercuri)methane (TAMM) which was the key heavy atom derivative in the structure determination of nucleosomes and the membrane reaction center and b) an undecagold cluster in which the gold core has a diameter of 8.2 A (Fig. 14 and in and ). Several variations of this cluster, modified with different ligands, have been prepared The cluster compounds, in which all the moieties R (Fig. 14) are amine or alcohol, are soluble in the crystallization solution of SOS subunits from H. marismortui. Thus, they could be used for soaking. Crystallographic data (to 18 A resolution) show isomorphous unit cell constants with observable differences in the intensity (Fig. 15). 

X 77.8 X 51.4 A, a space group of P2i2i2j, and a Z value of 4 (molecules per unit cell) (46). Several heavy atom derivatives, whose crystalline structure are isomorphous with that of the native enzyme, have been prepared for the X-ray diffraction analysis of crystalline cytochrome c peroxidase. 

A number of studies have been performed with methyl picolinimidate (Benisek and Richards 1968 Plapp et al. 1971) aimed at exploring the usefulness of the metal-chelating properties of such derivatives in the preparation of isomorphous heavy atom derivatives of proteins for X-ray diffraction studies. 

In our laboratory investigation a phase ambiguity arose when the experiment was performed only once. It was resolved simply by moving the oscillator to a new location and changing the phase of the reference wave. The analogous operation in X-ray crystallography is not quite so simple but still within reach. We must prepare a second (or even more, to be sure) heavy atom derivative bound at some other location in the unit cell. This second oscillator position generates then a second reference wave that yields, in the phase calculation, one phase solution in common with the first experiment. 

---