The type III secretion system (TTSS) mediates the specific translocation of bacterial proteins into the cytoplasm of eukaryotic cells, a process essential for the virulence of many Gram-negative pathogens. The enteropathogenic Escherichia coli TTSS protein EspA forms a hollow extracellular filament believed to be a molecular conduit for type III protein translocation. Structural analysis of EspA has been hampered by its polymeric nature. We show that EspA alone is sufficient to form filamentous

more »

... structures in the absence of other pathogenicity island-encoded proteins. CesA is the recently proposed chaperone of EspA, and we demonstrate that CesA traps EspA in a monomeric state and inhibits its polymerization. Crystallographic analysis of the heterodimeric CesA-EspA complex at a resolution of 2.8 Å reveals that EspA contains two long ␣-helices, which are involved in extensive coiled-coil interactions with CesA. The TTSS is a highly specialized bacterial protein secretory pathway that plays an essential role in the pathogenesis of many Gramnegative pathogens including Shigella, Salmonella, Bordetella, Pseudomonas, and pathogenic E. coli by enabling them to inject virulence proteins (known as effectors) directly into the cytoplasm of the eukaryotic host cells they infect 1 . Many of these type III translocated effectors mimic eukaryotic factors and are capable of subverting key host cellular processes to the benefit of the pathogen during infection 2,3 . Although recent studies have expanded our knowledge on the structure and function of many type III translocated effectors, the precise mechanisms of type III protein secretion and translocation, and the detailed structural characteristics of the secretion apparatus, remain poorly understood. The TTSS secretion apparatus is believed to consist of at least 20 unique but highly conserved proteins that associate into a macromolecular complex. EM analysis of purified TTSS (often termed needle complexes because of their characteristic morphology) shows that they contain a pair of ringlike structures situated in the inner and outer membranes of the bacteria with a protruding extracellular component, a gross morphology conserved across several species 4-6 . The TTSS structure is also similar to the bacterial flagellar assembly apparatus, confirming previous genetic analyses which revealed substantial sequence similarities between several core components of the TTSS and the flagellar system 7 . Although the needle complex allows the bacterium to secrete proteins across the bacterial envelope, the delivery of effectors into the host is mediated by a subset of extracellular proteins that are secreted by the needle complex to the bacterial surface and assemble into the type III translocon 8 . One of the most well-characterized type III translocons is that of enteropathogenic E. coli (EPEC). EPEC is the major cause of infantile diarrhea and child mortality worldwide. This pathogen encodes a TTSS in a 35-kilobase pathogenicity island (known as the locus of enterocyte effacement or LEE) which is absolutely required for its tight adherence to the intestinal surface during infection 9 . The EPEC type III translocon consists of at least three proteins: EspA, EspB and EspD, and nonpolar deletion mutants of these three genes retain the ability to secrete effector proteins into culture supernatant, but cannot translocate these proteins into the host cell cytoplasm 10,11 . In particular, EspA polymerizes into an extracellular filament (with lengths of up to 680 nm) that is associated with the EPEC needle complex 12 , whereas EspB and EspD (the so-called translocators) hetero-oligomerize into pore-forming complexes on the surface of the host cell 10,13 . The EspA filament has been proposed to be the molecular conduit for protein translocation, and this hypothesis is supported by a recent cryo-EM structure of sheared EspA filaments showing that it contains a hollow center of ∼25 Å, adequate for passage of partially or completely unfolded polypeptides 14 . However, a detailed understanding of the mechanisms of filament formation and type III protein translocation is not possible at the resolution of this model (∼26 Å). To address some of these fundamental questions, we set out to determine the highresolution structure of EspA. As with other fibrous proteins, EspA is extremely prone to massive polymerization when overexpressed. We carried out a detailed biochemical analysis of the recently identified EspA secretion chaperone CesA 15 , and discovered that CesA directly binds to EspA, inhibiting its polymerization. This unique function enabled us to determine the structure of EspA in complex with CesA, and this structure provides insights into the EspA filament assembly and secretion process.