Analysis of the solid-state conformations of six active and two inactive anti-human immunodeficiency virus nucleoside analogues is used to correlate conformational features with the relative activities of the compounds. Ten of the 11 observations of the active compounds (3'-azido-3'-deoxythymidine, 3'-azido-2',3'-dideoxyuridine, 3'-azido-2',3'-dideoxy-5-ethyluridine, 2',3'-dideoxyadenosine, 2',3'-dideoxycytidine, and 3'-deoxythymidine) have C3'-exo sugar ring conformations. The only exception is one of the two molecules of 3'-deoxythymidine. Four have unusual low anti glycosyl link conformations coupled with unusual C3'-exo/C4'-endo sugar ring conformations. The inactive compounds, 2',3'-dideoxyuridine and 3'-(propyl-2-ene)-2',3'-dideoxyuridine, have C3'-endo conformations. The C3'-exo and C3'-endo conformations place C5' in axial and equatorial positions, respectively. This affects the location of the 5'-hydroxyl group in relation to the location of the base. The 5'-hydroxyl group is the site of phosphorylation of the nucleoside and the observation of this conformational preference of the active compounds may be of importance for the development of new nucleosides with activity against human immunodeficiency virus.

The crystal structures of three anti-HIV 3'-azido-3'-deoxynucleosides have been determined to gain conformational information for structure-activity studies. The compounds 3'-azido-3'-deoxythymidine (AZT), 3'-azido-2',3'-dideoxyuridine (CS-87), and 3'-azido-2',3'-dideoxy-5-ethyluridine (CS-85) are all active inhibitors of HIV-1 replication. X-ray diffraction data for all three compounds were measured at 165 K.

No abstract

C16H17O8.1.5H2O, Mr = 364.33, monoclinic, C2, a = 14.314 (3), b = 6.851 (1), c = 18.178 (5) A, beta = 100.90 (2) degrees, V = 1750.46 A3, Z = 4, D chi = 1.382 Mg m-3, lambda(Mo K alpha) = 0.71069 A, mu = 0.933 mm(-1), F(000) = 768, T = 294 K, R = 0.078 for all 2160 reflections. The structure is characterized by the close stacking along the b axis of the planar 4-methylumbelliferyl ring system which is nearly perpendicular to b and the extensive hydrogen bonding scheme in which all hydroxyl groups are within 2.95 A of at least two other O atoms.

Macromolecular crystallization has been identified as the rate-limiting step in the structural biology field. With the advent of genomic-wide approaches to structural biology, efforts to streamline and automate the process of crystal growth are mandatory. In the last eighteen months we have developed a high throughput laboratory capable of executing a million crystallization experiments a year. We will describe the efforts and comment on the first 717,312 crystal growth experiments.