Scientists have worked out how the deadliest malaria parasite is able to "hide" from the body's immune system.
The malaria parasite is carried by mosquitoes
The international team said Plasmodium falciparum constantly changes the appearance of a protein it deposits on infected cells.
This meant the human immune system did not have enough time to begin making antibodies against the protein before the parasite changed its appearance.
The discovery could lead to new avenues for drug research, the team told Cell.
Malaria causes more than 300 million acute illnesses and at least one million deaths each year, most of them in developing countries.
The P. falciparum parasite has evolved to have a long life because it has to be around long enough to live in a mosquito before it is transmitted to a human.
It does this by continuously changing the version of a protein known as PfEMP1 that it deposits on the surface of infected cells.
By the time the human immune system learns to recognise the protein and starts making antibodies against it, the parasite has switched to another form of the protein, and the game of "hide and seek" starts again.
It was already known that a family of genes known as var controls the production of PfEMP1.
The malaria parasite's genome contains at least 50 var genes, but only one is expressed at any given time, giving rise to a single version of the PfEMP1 protein.
Over the course of an infection, expression switches from one var gene to another, and all the rest are silenced.
The research team, led by scientists from the Howard Hughes Medical Institute in Maryland, investigated why that happens.
They found there were differences in the DNA of "silent" and "active" var genes.
In silent genes, a protein called silent information regulator 2 (SIR2) plays a role in "muting" the DNA.
There is also a mechanism in the genome which moves the selected var gene into a "spotlight" so it becomes active.
The researchers say that teasing out more about the mechanism by which the var genes are switched on or off could lead to the development of novel drugs for malaria.
Dr Alan Cowman, who led the research, said: "If you could work out a way of causing the parasite to switch all the var genes on, then the body would see all the variations of var genes, and the immune system would be able to control the infection."
Professor Chris Newbold, of the Weatherall Institute of Molecular Medicine, Oxford, UK, said the research was a useful step in the right direction, but that there was not enough information yet to develop medicines.
"It may be that if one could cause all the var genes to be switched on at one time, or for them to be switched off, then that would be an extremely potent way of controlling disease.
"But we don't yet understand enough about the mechanics."