Structures of naturally occurring circular proteins from bacteria.

Bacteria produce a wide variety of both proteinaceous and nonproteinaceous molecules for defense, mediation of microbial competitions, and signaling purposes. Of the proteinbased molecules, there are examples of large folded polypeptides that are translated conventionally from genes as well as smaller peptides that are assembled nonribosomally by peptide synthetases (41). The latter often contain modified amino acids, including those with altered chirality, N-methylation, and nonamide backbone bonds. Such modifications may give the producing organism an advantage, as the modified molecules are less susceptible to the normal proteolytic cleavage reactions of proteins. Cyclization is another strategy that has been used, exemplified by well-known natural products such as cyclosporin A and gramicidin S. The advantages of such a strategy for stabilizing peptides may be seen by the fact that a significant number of macrocyclic natural products have found pharmaceutical applications, including, for example, the widespread use of cyclosporin A as an immunosuppressive agent. Synthetic cyclic peptides are also widely used as lead molecules in the pharmaceutical industry. The biosynthesis of cyclic nonribosomal peptides such as cyclosporin A and polyketides such as the antibiotic erythromycin, as well as hybrid peptide/polyketide drugs such as rapamycin, has recently been reviewed (41). Briefly, it involves the ordered condensation of monomer building blocks by an enzyme-driven process to produce a linear acyl chain that is cyclized by a thioester domain at the C-terminal end of the biosynthetic assembly line (41). Over recent years, several examples of naturally occurring circular proteins fundamentally different from the nonribosomal cyclic peptides have been discovered (58). These molecules are true proteins in that they have a well-folded threedimensional structure and are produced via translation of genes. Their only difference from conventional proteins is that their gene-coded precursor proteins are posttranslationally modified to join the N and C termini to produce a seamless circle of peptide bonds. Such circular proteins occur in a diverse range of organisms, from bacteria to plants and animals, but the focus here is on circular proteins produced by bacteria. In this review we describe the sequences and structures of these proteins and examine what is known about their biosynthesis. We compare them to other recently discovered circular proteins from higher organisms and speculate on the possible roles of backbone cyclization. Circular proteins were unknown a decade ago, and the field is still in its infancy, but there are now enough examples known to make it timely to examine the structures and properties of bacterially produced circular proteins. Bacterial protein expression has also been used to facilitate the production of synthetic circular variants of noncyclic proteins, including -lactamase (31) and green fluorescent protein (30). These studies have adapted intein-based methods to enable protein ligations that result in circular proteins. While the focus of this review is on naturally occurring circular proteins, the studies on artificially produced circular proteins highlight the importance and interest in this area. We note at the outset that we generally use the term circular rather than cyclic to emphasize the fact that the molecules that we are focusing on have a head-to-tail cyclized backbone rather than other cross-links, such as disulfide bonds, that might make just part of the structure cyclic. While the molecules that we examine are thus topologically circular, as we shall see, they fold into complex three-dimensional shapes.