Recent studies on proteins whose N and C termini are in close proximity have demonstrated that folding of polypeptide chains and assembly of oligomers can be accomplished with circularly permuted chains. As yet no methodical study has been conducted to determine how extensively new termini can be introduced and where such termini cannot (4), gp120 of human immunodeficiency virus 1 (HIV-1) (5), interleukin 113 (6), mouse ornithine decarboxylase (7), bacteriophage T4 lysozyme (8, 9), the catalytic subunit of Escherichia coli aspartate transcarbamoylase (10, 11), ribonuclease Ti (12, 13), Bacillus ,B-glucanase (14), E. coli DHFR (15), interleukin 4 (16-18), the SH3 domain of a-spectrin (19), E. coli outer membrane protein A (20), Bacillus stearothermophilus glyceraldyhyde-3-phosphate dehydrogenase (21), and 'yB-crystallin (22).Most of these enzymes were permuted at one or two positions, and only few were subjected to a systematic permutation analysis (11,13,19). Although evidence has accumulated that new N and C termini in flexible solvent-exposed loops are readily tolerated, the question arises whether free termini can be moved into other areas of a protein; e.g., within secondary structure elements or into the hydrophobic core. However, even with the availabilty of efficient strategies based on tandem gene constructs that are used to produce "sitedirected" circular permutants (6,11,12), a one-by-one exhaustive permutation analysis of a protein of several hundred residues in length is hardly a practicable approach. Therefore, we devised a technique to construct a random collection of circularly permuted DNA molecules containing the complete coding sequence for the desired polypeptide chain along with a linker between the original C-and N-terminal regions. These rearranged genes were incorporated into a suitable plasmid for expression of the randomly permuted polypeptide chains in E. coli. Colonies containing stable and/or active protein formed from these chains were identified by either an immunoblot assay or by the suppression of auxotrophy because of the presence of active enzyme.For these studies we have utilized the catalytic (c) polypeptide chain of E. coli aspartate transcarbamoylase (ATCase; Abbreviations: ATCase, aspartate transcarbamoylase; c chain, catalytic polypeptide chain; C trimer, catalytic trimer or subunit; r subunit, regulatory subunit or dimer.

The crystal structure of an allosteric chorismate mutase, the Thr-226 -B le mutant, from yeast Saccha-romyces cerevisiae has been determined to 2.2-A resolution by using the multiple isomorphous replacement method. Solventflafening and electron-density modification were applied for phase improvement. It has been proposed that both enzymatic and nonenzymatic rearrangement of chorismate to prephenate is a concerted reaction that proceeds via a transition state with chair-like geometry (3-6). The enzymatic reaction is distinctive among known enzymes in that it catalyzes a pericyclic process. The recently published crystal structure of monofunctional chorismate mutase from B. subtilis shows that no functional group from the protein is available for direct involvement in the catalysis (7). The two-million-fold rate enhancement is thus achieved by selection of the active conformation and stabilization of the transition state by the enzyme.The structural data on the catalytic antibody with low chorismate mutase activity (8) also supports the conclusion that the catalysis ofthe isomerization of chorismate proceeds by stabilizing the same pericyclic transition state that occurs in the uncatalyzed thermal reaction, rather than by providing alternative pathways to the product.Yeast chorismate mutase is a monofunctional dimer oftwo 30-kDa polypeptides encoded by the AR07 gene (9). Biochemical data showed that in the dimer there are two substrate-binding sites, two activator-binding sites, and two inhibitor-binding sites that may or may not be distinct from the activator sites. Kinetic data of the wild-type chorismate mutase show positive cooperativity toward the substrate. This cooperativity is lost in the presence of tryptophan. The product of one of the AR07 mutant alleles, with a single substitution of Ile-226 (T226I), has been characterized as a constitutively activated chorismate mutase that does not respond to activation by tryptophan or inhibition by tyrosine. As a relatively small protein in the allosteric family, yeast chorismate mutase provides an ideal model system for exploring the detailed mechanisms of allosteric regulation as well as to examine the unique catalysis. We describe here the crystal structure of the mutant chorismate mutase, T2261, from yeast S. cerevisiae. § In contrast to the available structure of chorismate mutase from B. subtilis which exists as a homotrimer, the yeast protein is a dimer that has a completely different folding topology. In addition, the binding sites for tryptophan in the structure were identified. MATERIALS AND METHODSThe mutant enzyme was isolated, purified, and characterized by R.G., G.S., and G.B., and the structure determination was made by Y.X. and W.N.L., as described below.The crystallization of the protein was done by using the hanging-drop method, as described earlier (10). The space group was established as P6 or P65, and the unit cell parameters are a = b = 95.8 , c = 157.9 A, a =p = 90°, and y = 120°. Heavy-atom derivatives were prepared by soaking the c...

The initiation of DNA replication in eukaryotes requires the loading of the origin recognition complex (ORC), Cdc6, and minichromosome maintenance (MCM) proteins onto chromatin to form the preinitiation complex. In Xenopus egg extract, the proteins Orc1, Orc2, Cdc6, and Mcm4 are underphosphorylated in interphase and hyperphosphorylated in metaphase extract. We find that chromatin binding of ORC, Cdc6, and MCM proteins does not require cyclin-dependent kinase activities. High cyclin A-dependent kinase activity inhibits the binding and promotes the release of Xenopus ORC, Cdc6, and MCM from sperm chromatin, but has no effect on chromatin binding of control proteins. Cyclin A together with ORC, Cdc6 and MCM proteins is bound to sperm chromatin in DNA replicating pseudonuclei. In contrast, high cyclin E/cdk2 was not detected on chromatin, but was found soluble in the nucleoplasm. High cyclin E kinase activity allows the binding of Xenopus ORC and Cdc6, but not MCM, to sperm chromatin, even though the kinase does not phosphorylate MCM directly. We conclude that chromatin-bound cyclin A kinase controls DNA replication by protein phosphorylation and chromatin release of Cdc6 and MCM, whereas soluble cyclin E kinase prevents rereplication during the cell cycle by the inhibition of premature MCM chromatin association.

The initial step of tryptophan biosynthesis is catalyzed by the enzyme anthranilate synthase, which in most microorganisms is subject to feedback inhibition by the end product of the pathway. We have characterized the TRP2 gene from a mutant Saccharomyces cerevisiae strain coding for an anthranilate synthase that is unresponsive to tryptophan. Sequence analysis of this TRP2(Fbr) (feedback-resistant) allele revealed numerous differences from a previously published TRP2 sequence. However, TRP2(Fbr) was found to differ in only one single-point mutation from its own parent wild type, a C-to-T transition resulting in a serine 76-to-leucine 76 amino acid substitution. Therefore, serine 76 is a crucial amino acid for proper regulation of the yeast enzyme. We constructed additional feedback-resistant enzyme forms of the yeast anthranilate synthase by site-directed mutagenesis of the conserved LLES sequence in the TRP2 gene. From analysis of these variants, we propose an extended sequence, LLESX10S, as the regulatory element in tryptophan-responsive anthranilate synthases from prokaryotic and eukaryotic organisms.