The malarial life cycle involves repeated rounds of intraerythrocytic replication interspersed by host cell rupture which releases merozoites that rapidly invade fresh erythrocytes. Apical membrane antigen-1 (AMA1) is a merozoite protein that plays a critical role in invasion. Antibodies against AMA1 prevent invasion and can protect against malaria in vivo, so AMA1 is of interest as a malaria vaccine candidate. AMA1 is efficiently shed from the invading parasite surface, predominantly through juxtamembrane cleavage by a membrane-bound protease called SUB2, but also by limited intramembrane cleavage. We have investigated the structural requirements for shedding of Plasmodium falciparum AMA1 (PfAMA1), and the consequences of its inhibition. Mutagenesis of the intramembrane cleavage site by targeted homologous recombination abolished intramembrane cleavage with no effect on parasite viability in vitro. Examination of PfSUB2-mediated shedding of episomally-expressed PfAMA1 revealed that the position of cleavage is determined primarily by its distance from the parasite membrane. Certain mutations at the PfSUB2 cleavage site block shedding, and parasites expressing these non-cleavable forms of PfAMA1 on a background of expression of the wild type gene invade and replicate normally in vitro. The non-cleavable PfAMA1 is also functional in invasion. However - in contrast to the intramembrane cleavage site - mutations that block PfSUB2-mediated shedding could not be stably introduced into the genomic pfama1 locus, indicating that some shedding of PfAMA1 by PfSUB2 is essential. Remarkably, parasites expressing shedding-resistant forms of PfAMA1 exhibit enhanced sensitivity to antibody-mediated inhibition of invasion. Drugs that inhibit PfSUB2 activity should block parasite replication and may also enhance the efficacy of vaccines based on AMA1 and other merozoite surface proteins.