Study Of A Ribonuclease And Its Inhibitor From Bacillus

点击翻译按钮获取中文摘要

Two proteins, barnase, the extracellular ribonuclease of Bacillus amyloliquefaciens, and barstar, its intracellular inhibitor, are used as a model system for the study of protein folding and protein-protein interactions. Barnase is one of a homologous group of ribonucleases occurring in both prokaryotes and eukaryotes. Recombinant DNA techniques are being applied with three major aims: (1) to facilitate production of wild type and mutant proteins; (2) to examine the structural and control sequences of the genes; and (3) to make specific changes in the sequences to test theories of folding and to probe the barnase-barstar interaction. Both proteins can now be obtained from recombinant genes in E. coli where expression of barstar counters the lethal effect of barnase expression. The structures of both proteins and their complex are known, barnase at 1.5 Angstrom resolution. Crystal structures of several barnase-barstar pairs having complementary mutations in the interface, obtained by an in vivo selective technique, have been solved, providing insight into the mechanisms that determine the strength of the bond. Barstar also inhibits a group of RNases from Streptomyces strains. These enzymes are distantly related to barnase with a sequence identity of only 25%. Among the four such enzymes in hand, identities range from 40% to 70%. Cloned and expressed in E.coli with the aid of barstar, several of these have been well characterized, along with three barstar homologs, all of which inhibit barnase. Recently, we have used phage display to find and clone the gene for Sti, a barstar homolog and natural inhibitor of RNase St from S. erythreus. Useful expression of this gene required considerable manipulation of its sequence with silent mutations. It allowed some increase in production of RNase St in E. coli but still not enough to be useful. A phage display system has been developed for selection of varieties or homologs of barstar that bind tightly to barnase or its mutants. Procedures have been developed for total synthesis of the barstar gene with randomization of selected residues and a multiplicity (the number of independently randomized sequences) on the order of 10exp9. We have screened 3 synthetic barstar libraries with spatially compact 8, 11 and 11 residue portions of the hydrophobic core randomized and another with randomization of all 22 residues of the core. The possible residues in each substituted position were Leu, Ile, Val, Met or Phe. Unselected genes transferred to a plasmid barstar vector showed no indication of functional barstar, while all selected genes appear to be functional in vivo, allowing transformation of their host by a compatible plasmid carrying barnase. All of the latter also produce measurable barstar activity in vitro, most at a very low level. Some of each of the partially randomized sets produce enough for physical studies. A much larger library of genes for completely randomized core barstars is being prepared. Sequences of functional barstars with all 22 core residues randomized indicate that no particular amino acid is indispensable, although in some positions the wild type residue is clearly preferred. The volume occupied by the core sidechains, as represented by the number of non-hydrogen atoms, is much more variable than expected, especially on the low side, suggesting that the secondary structure framework can collapse considerably without destroying the barnase-binding site. Development of a continuous flow electroporation technique has allowed direct screening of fairly large E. coli libraries for resistance to barnase expression, allowing us to estimate the yield of active barstars with fully randomized cores at about one in 10exp4 or 5.