The crystal structure of Escherchia coli asparaginase II (EC 3.5.1.1), a drug (Elspar) used for the treatment of acute lymphoblastic leukemia, has been determined at 2.3 A resolution by using data from a single heavy atom derivative in combination with molecular replacement. The atomic model was refined to an R factor of 0.143. This enzyme, active as a homotetramer with 222 symmetry, belongs to the class of a/P proteins. Each subunit has two domains with unique topological features. On the basis of present structural evidence consistent with previous biochemical studies, we propose locations for the active sites between the N-and C-terminal domains belonging to different subunits and postulate a catalytic role for Thr-89.

We have isolated 16 peptides from a cyanogen bromide digest of rhodopsin. These cyanogen bromide peptides account for the complete composition of the protein. Methionine-containing peptides from other chemical and enzymatic digests of rhodopsin have allowed us to place the cyanogen bromide peptides in order, yielding the sequence of the protein. We have completed the sequence of most of the cyanogen bromide peptides. This information, in conjunction with that from other laboratories, forms the basis for our prediction of the secondary structure of the protein and how it may be arranged in the disk membrane.

Retroviral gag, pol and env gene products are translated as precursor polyproteins, which are cleaved by virus-encoded proteases to produce the mature proteins found in virions. On the basis of the conserved Asp-Thr/Ser-Gly sequence at the putative protease active sites, and other biochemical evidence, retroviral proteases have been predicted to be in the family of pepsin-like aspartic proteases. It has been suggested that aspartic proteases evolved from a smaller, dimeric ancestral protein, and a recent model of the human immunodeficiency virus (HIV) protease postulated that a symmetric dimer of this enzyme is equivalent to a pepsin-like aspartic protease. We have now determined the crystal structure of Rous sarcoma virus (RSV) protease at 3-A resolution and find it is dimeric and has a structure similar to aspartic proteases. This structure should provide a useful basis for the modelling of the structures of other retroviral proteases, such as that of HIV, and also for the rational design of protease inhibitors as potential antiviral drugs.

A prediction algorithm based on physical characteristis of the twenty amino acids and refined by comparison to the proposed bacteriorhodopsin structure was devised to delineate likely membrane-buried regions in the primary sequences of proteins known to interact with the lipid bilayer. Application of the method to the sequence of the carboxyl terminal one-third of bovine rhodopsin predicted a membrane-buried helical hairpin structure. With the use of lipid-buried segments in bacteriorhodopsin as well as regions predicted by the algorithm in other membrane-bound proteins, a hierarchical ranking of the twenty amino acids in their preferences to be in lipid contact was calculated. A helical wheel analysis of the predicted regions suggests which helical faces are within the protein interior and which are in contact with the lipid bilayer.X-ray diffraction studies of crystalline soluble proteins have resulted in nearly one-hundred known tertiary structures [l, 21. With their advent have come many secondary-structure prediction methods which require only a knowledge of the amino acid sequence (cf. 13 -51). These techniques generally rely on a statistical or informational analysis of the frequency with which the 20 amino acids appear within the observed secondary structures (a-helices, p-strands, and reverse turns). Since their data base is composed of only soluble protein structures, the prediction algorithms may not be applicable to primary sequences of proteins that are in contact with a membrane. The only membrane-bound polypeptide topology known to a reasonably high resolution has been determined by electron-scattering techniques for the purple-membrane protein of Halobacterium halobium [6,7] whose seven rods of electron density are believed to be a-helical segments that traverse the membrane [8]. Such a structure should be predictable from the amino acid sequence alone through utilization of the physico-chemical characteristics of the residues interacting with the uniquely apolar membrane environment. A prediction algorithm based on the physical characters was developed in the present work and refined by application to the bacteriorhodopsin primary structure. The correlation between the predicted helical and turn regions and those of the model structure suggested by Engelman et al.[8] was 0.69, a value at least comparable to the prediction accuracy of techniques devised for soluble proteins [5,9,10]. The prediction method was then applied to the known C-terminal sequence of bovine rhodopsin and to the primary sequences of other membrane-bound proteins. Two turn segments and two helical regions were predicted for bovine rhodopsin. The algorithm clearly delineated likely membrane-bound regions within the amino acid sequence of the other proteins. A hierarchical ranking of the twenty amino acids in their preference to membrane buried was then determined. A helical 'wheel' analysis of the predicted regions indicated the helical faces within the protein interior and in contact with the lipid bilayer.