Trypanosoma brucei is the protozoan parasite that causes African sleeping sickness. Its surface is packed with 10(7) copies of the glycosyl phosphatidylinositol (GPI)-anchored variant surface glycoprotein (VSG). This GPI anchor is unusual in that it contains two myristates (14:0) in its lipid moiety. This fatty acid specificity is achieved through myristoylation of the GPI precursor, and the acyltransferases involved in the GPI remodeling were presumed to be specific for myristate. However, their specificity had never been fully evaluated. Here we found as expected that the remodeling acyltransferases completely excluded palmitate (16:0) and stearate (18:0) in a cell-free fatty acid remodeling system. In contrast, we found surprisingly that one of these enzymes was permissive to shorter fatty acids such as laurate (12:0) and octanoate (8:0). However, the rates of incorporation of shorter fatty acids were lower than that of myristate at low substrate concentration. Since shorter fatty acids are virtually absent in the parasite and in the host bloodstream, it is unlikely that shorter fatty acids compete effectively with myristate as remodeling substrates under physiological conditions. Even if they were present in small quantities, a recently identified specialized fatty acid synthetase efficiently elongates shorter fatty acids to myristate prior to incorporation into GPIs (Morita et al., Science 288 (2000) 140-3.). Therefore, even though a remodeling acyltransferase is permissive with regard to substrate chain length, the myristate specificity in GPI anchors is very high.