5243 Introduction The conversion of epithelial cells to migratory cells plays a central role in development, wound healing and tumor metastasis. Depending on the context, cells may move as integrated sheets, as individual cells or in small groups. During skin repair, large groups of cells migrate collectively as epithelial sheets. On the other hand, neural crest cells and primordial germ cells move in chain-like arrangements. In the latter cases, when migration is initiated, the cells must reduce their adhesion with the epithelial cells that are left behind, while maintaining cohesive contacts with other migrating cells. Changes in expression of specific cell adhesion proteins and intermediate filament proteins have been described that correlate with migration of a variety of epithelial cell types (Ridley et al., 2003). Although some epithelial characteristics are lost when cells become migratory, the fates of the recently characterized epithelial polarity proteins have not been described in migrating cells derived from an epithelium. The migration of the border cells, a small group of follicle cells in the Drosophila ovary, provides a model system to study these events in vivo. The border cells originate from the anterior pole of the follicular epithelium, which surrounds and communicates with an underlying cluster of 16 germline cells to form an egg chamber (Fig. 1A). Migration is initiated when a pair of specialized follicle cells, known as the polar cells, secrete a cytokine signal that activates the Janus Kinase/signal transducer and activator of transcription (JAK/STAT) pathway in the neighboring 6-8 cells, stimulating their motility (Silver and Montell, 2001). Migration is completed once the border cells reach the oocyte. Initially, the border cells are polarized epithelial cells. These epithelial cells have distinct membrane domains. The apical domain contacts the germline. The basal plasma membrane of follicle cells contacts the basal lamina. The lateral sides touch the plasma membranes of neighboring epithelial cells, connecting neighboring follicle cells through cell junctions. These distinct membrane domains harbor different protein complexes that are important for establishing and/or maintaining the epithelium. The Par-3/Bazooka (Baz), Par-6, atypical protein kinase C (aPKC) complex and Crumbs, Stardust (Sdt), Discs lost (Dlt) complex localize to the apical membrane of follicle cells during oogenesis as well as to subapical adherens junctions. Mutations in these genes cause epithelial discontinuity and multilayering (Abdelilah-Seyfried et al., 2003; Tanentzapf et al., 2000). E-cadherin and Armadillo (Arm), which is the Drosophila homolog of β- catenin, localize laterally in follicle cells and are critical for the assembly of adherens junctions. Clones of follicle cells lacking Arm lose adherens junctions, and E-cadherin is no longer detected in the junctional region. In addition, a protein complex composed of Discs large (Dlg), Lethal (2) giant larvae (Lgl), and Scribble (Scrib) localizes to septate junctions in the lateral domain. Mutations in dlg disrupt septate junction structure and cell polarity (Woods et al., 1997). Finally, proteins such as integrins localize to the basal domain. Although spatially separated, these protein complexes communicate with each other and in some cases can affect each other’s proper localization. Although much has been learned concerning the molecules and mechanisms that are responsible for generating and maintaining polarity within the follicular epithelium, little is Polarized epithelial cells convert into migratory invasive cells during a number of developmental processes, as well as when tumors metastasize. Much has been learned recently concerning the molecules and mechanisms that are responsible for generating and maintaining epithelial cell polarity. However, less is known about what becomes of epithelial polarity proteins when various cell types become migratory and invasive. Here, we report the localization of several apical epithelial proteins, Par-6, Par-3/Bazooka and aPKC, during border cell migration in the Drosophila ovary. All of these proteins remained asymmetrically distributed throughout migration. Moreover, depletion of either Par-6 or Par-3/Bazooka by RNAi resulted in disorganization of the border cell cluster and impaired migration. The distributions of several transmembrane proteins required for migration were abnormal following Par-6 or Par-3/Bazooka downregulation, possibly accounting for the migration defects. Taken together, these results indicate that cells need not lose apical/basal polarity in order to invade neighboring tissues and in some cases even require such polarity for proper motility. Key words: Par-6, Bazooka, Border cells, Cell migration, Drosophila Summary Requirement for Par-6 and Bazooka in Drosophila border cell migration Elaine M. Pinheiro and Denise J. Montell* Department of Biological Chemistry, Johns Hopkins School of Medicine, 725 North Wolfe Street, Baltimore, MD 21205-2185, USA *Author for correspondence (e-mail: dmontell@jhmi.edu) Accepted 20 August 2004 Development 131, 5243-5251 Published by The Company of Biologists 2004 doi:10.1242/dev.01412 Research article