Tetracyclines are effective antimalarials, but their mechanism of action is unclear. Since these agents block prokaryotic protein synthesis, it has been proposed that they disrupt the mitochondrion or the apicoplast, both of which include prokaryotic ribosomal subunits in their genomes. We have shown that both the mitochondrion and apicoplast appear normal through a cycle of treatment with doxycycline, that these organelles are successfully segregated into daughter parasites, and that they remain intact in the progeny of treated parasites. However, doxycycline specifically blocks the expression of apicoplast genes, leading to the distribution of nonfunctional apicoplasts into daughter parasites and a subsequent block in parasite development. These results indicate that the site of action of tetracyclines is the apicoplast but that loss of apicoplast function is not apparent until late in the cycle following treatment, explaining the slow action of these drugs.
Previous work on the antimalarial effects of tetracyclines demonstrated increased efficacy with prolonged treatment (9
). Doxycycline inhibited global protein synthesis only at suprapharmacological concentrations (6
), arguing against cytosolic ribosomes as a target for tetracyclines in vivo. Early work examining the mitochondrion as a potential target of tetracyclines demonstrated decreased mitochondrial uptake of rhodamine 123 after 72 h (18
). However, this observation may have reflected secondary toxicity to the mitochondrion following primary effects of the drug. Other studies demonstrating depression of mitochondrial enzyme activity (26
) and a block in both apicoplast and mitochondrial transcription (22
) assessed tetracyclines at concentrations well above those that are clinically achievable.
The slow action of tetracyclines against P. falciparum
was observed even if trophozoite and early schizont stage parasites were treated for as little as 12 h. Apicoplasts in treated parasites initially appeared morphologically normal, replicated their genomes, processed imported proteins, and segregated into developing merozoites. Doxycycline specifically disrupted the expression of apicoplast genes. However, most of the proteins predicted to be required for apicoplast function are encoded by the nuclear genome (28
), so the apicoplast would be expected to perform most functions normally as long as its import machinery remained intact. We propose that, though doxycycline does not prevent apicoplast function initially, the apicoplasts inherited by the progeny of doxycycline-treated parasites contain insufficient levels of apicoplast-encoded proteins required for the importation and processing of the several hundred nuclear genes needed for normal function. This loss of apicoplast function ultimately results in parasite death.
Our data do not support primary action of doxycycline against the mitochondrion, as transcription within this organelle and replication of the mitochondrial genome were not obviously altered over two parasite life cycles. Mitochondria also appeared to segregate normally in the presence of doxycycline, and in the progeny of doxycycline-treated parasites, mitochondria elongated and formed elaborately branched structures similar to those of untreated parasites. We did not observe segregation of these mitochondria at the end of the second cycle, consistent with previous observations that apicoplast segregation always precedes mitochondrial segregation in healthy parasites (37
). However, parasites at this stage displayed gross morphological abnormalities, so lack of mitochondrial segregation was likely secondary to loss of apicoplast function.
In addition to data from plasmodia, there is evidence that prokaryotic protein synthesis inhibitors target the apicoplast of the related apicomplexan parasite Toxoplasma gondii
). In T. gondii
, treatment with clindamycin causes a “delayed-death” phenotype in which progeny are able to invade new host cells but die shortly thereafter. A similar phenotype was observed in T. gondii
parasites that were unable to inherit an apicoplast due to a genetic-segregation defect, leading to the proposal that the apicoplast is required for the establishment of the parasitophorous vacuole. We have observed that P. falciparum
parasites containing defective apicoplasts survive until the end of their cycle, arguing against a role in establishing the parasitophorous vacuole. This observation is in agreement with ultrastructural studies on the progeny of clindamycin-treated T. gondii
parasites, which also contain multiple nuclei and appear unable to complete cell division (7
). In both species, death coincides with the initiation of cell division, which occurs early in the T. gondii
cycle and late in the P. falciparum
cycle. We propose that the apicoplast is required for the formation of daughter cell plasma membranes, since fatty acid biosynthesis is a likely function of apicoplasts and since our ultrastructural studies indicated that these structures were lacking following doxycycline treatment. Since doxycycline needs to be administered only transiently to disrupt apicoplast function in the following cycle and since parasites containing nonfunctional apicoplasts can survive for nearly 48 h, doxycycline-treated plasmodia are an ideal model system for probing specific functions of the apicoplast.