The commonly accepted view that enteroinvasive bacteria enter cells of the intestinal epithelial lining through the apical surface can be challenged in the case of shigellosis. This study is based on in vitro experiments that showed that the invasion of human colonic Caco-2 cells by Shigella flexneri occurred through the basolateral pole of these cells. In these experiments, the few bacteria that interacted with the apical surface either bound to microvilli of the cell dome without causing detectable alteration or bound at the level of intercellular junctions at which they demonstrated a limited capacity for paracellular invasion, which permitted subsequent entry through the lateral domain of the cells. Treatment of Caco-2 cell monolayers with ethylene glycol-bis(beta-aminoethyl ether)-N,N,N',N'-tetraacetic acid (EGTA), which disrupts intercellular junctions, greatly enhanced the rate of cell infection. These observations suggest a physiopathological paradox that may have important consequences for the understanding of the process of colonic invasion in vivo during shigellosis.