C OMMENT TRENDS IN MICROBIOLOGY 529 VOL. 8 NO. 12 DECEMBER 2000 0966-842X/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved. PII: S0966-842X(00)01853-9 Viewpoints F or a variety of reasons, patho- genic and non-pathogenic bacteria decorate their cell surfaces with, or secrete into the extracellular environment, many different proteins. The expression of proteinaceous virulence factors directly or indirectly affects the host and thereby increases the pathogen’s ability to survive and multiply; however, it should be re- membered that non-pathogenic organisms also secrete proteins adaptive to their lifestyles. In the case of Gram-positive organisms, surface proteins gener- ally follow the sec-dependent pathway to cross the cell mem- brane and are either released into the extracellular environment or remain anchored by virtue of one of several peptide-anchoring sig- nals 1,2 . Secretion in Gram-negative bacteria, however, is constrained by the presence of the outer mem- brane, which necessitates more complex and specialized protein secretion pathways. Evolutionary pressures and bacterial economics, which dictate that it is easier to adopt and adapt mechanisms than to develop new mechanisms inde- pendently, have resulted in the development of a limited number of Gram-negative bacterial secre- tion systems. To date, five protein secretion pathways have been defined among the Gram-negative bacteria. Type I secretion is exemplified by the secretion pathway defined for Escherichia coli hemolysin (HlyA); here, accessory proteins are needed to construct a channel through the periplasm 3,4 . The se- cretion of pullulanase (PulA) from Klebsiella oxytoca is the best stud- ied example of type II secretion, which requires the action of 14 additional accessory proteins as- sociated with various regions of the cell envelope 5 . The type III secretion pathway requires a complex apparatus of proteins, which forms a tightly regulated oligomeric structure spanning the inner and outer membranes 6,7 . Type IV secretion is perhaps the least understood of the Gram- negative bacterial secretion appa- ratuses 8 . The type IV export process involves the coordinate action of at least nine proteins, which are variously associated with the inner and outer membrane and which are localized within the periplasm and cytoplasm. Autotransporters Given the intricate nature of the other secretion systems described above, the apparent simplicity of the autotransporter (or type V) 9 secretion mechanism is astonish- ing. Proteins secreted by this mechanism possess an amino-ter- minal signal sequence (with fea- tures required for passage through the sec translocon), a passenger domain and a carboxy-terminal -domain 10 . Although proteins se- creted by the type II and type V pathways utilize the sec apparatus to traverse the inner membrane, they differ in how they pass through the outer membrane. Proteins secreted by the type V mechanism mediate their own translocation (hence the apt term autotransporter) across the outer membrane by virtue of their -domains; the -domain forms a pore in the outer membrane through which the passenger do- main of the molecule is translo- cated to the cell surface (Fig. 1) 10 . By contrast, proteins secreted by the type II secretion system require ad- ditional proteins for their secretion across the outer membrane. The nomenclature for secretion systems has not always been as clear as this discussion implies; for many years, debate has continued about the nomenclature used to delimit protein secretion path- ways. As each new extracellular or outer membrane protein is identi- fied, it is expediently slotted into one or another category based on homology to defined systems or on a defined strategy. Yet, until re- cently, little was being done to re- dress the inconsistencies that have arisen by parceling the proteins into such convenient categories. Such is the case with all proteins secreted across the inner mem- brane via the sec-dependent mech- anism. It has been suggested that all such proteins should be classi- fied in the type II pathway and that steps involved in outer membrane secretion should represent differ- ent terminal branches of the path- way. However, the application of this rule presents its own prob- lems. For example, autotrans- porters are generally secreted by the typical sec-dependent system, which would make them a termi- nal branch of the general secretory apparatus 5 . However, a handful of autotransporters have rather unusual extended signal se- quences, and these proteins might utilize pathways other than sec 10 . Furthermore, application of the I.R. Henderson* is in the Dept of Microbiology and Immunobiology, The Queens University of Belfast, Grosvenor Rd, Belfast, UK BT12 6BN; R. Cappello is in the Centro de Investigacion en Ciencia Aplicada y Tecnologia Avanzada del Instituto Politecnico Nacional (CICATA-IPN), Legaria 694, Col. Irrigacion, Mexico, DF 11500; J.P. Nataro is in the Center for Vaccine Development, Dept of Pediatrics, University of Maryland School of Medicine, Baltimore, MD 21201, USA. *tel: +44 28 90 263044, fax: +44 28 90 439181, e-mail: i.henderson@qub.ac.uk Autotransporter proteins, evolution and redefining protein secretion Ian R. Henderson, Renato Cappello and James P. Nataro