AN INTERESTING STEROID DIMER WITH AN EXTREMELY LONG CELL AXIS: TETRAGONAL, P4₁2₁2, a=b=9.17Å, c=85.90Å, V=7225.06ų. Fusen Han and Roger G. Williams, Pharmcia and Upjohn Inc., Kalamazoo, MI 49001, USA

An impurity in the methylprednisolone precursor was found in August of 1987 from five SOLU-MEDROL lots in the former Upjohn Company. It was the cause of insoluble residues present in the reconstituted solutions. In support of the structure determination, the analogous compound in the hydrocortisone series was synthesized. In experiments on decomposition routes of hydrocortisone, a sample dissolved in deoxygenated methanol was treated with 45% potassium hydroxide. The crude product mixture was purified by chromatography on silica gel with a chloroform eluant. One of the primary fractions proved to be a dimeric condensation product whose structure was unknown.

The X-ray diffraction data of this compound were collected with Siemens P4 diffractometer. The control software, XSCANS, could not determine the cell from its default parameters. After modification, it derived a unusual tetragonal cell: P4₁2₁2, a=b=9.17Å, c=85.90Å, V=7225.06ų. To verify it's extremely long c axis, three axial photos were taken. The cell parameters were confirmed from these photos. 6839 reflections were collected with [[Sigma]] scan at -40deg.C.

The trial solution of this structure could not be easily obtained from the direct methods either. The short a, b and extremely long c axis could be the reason for the difficulties. The structure was finally solved with 10,000 direct methods attempts using SHELXS program package. The structure refinement is underway.