IS ANNEXIN III AN ENZYME? Anita Lewit-Bentley*, Béatrice Perron[[section]]* & Françoise Russo-Marie[[section]] *LURE, Univ. Paris-Sud, 91405 Orsay, [[section]]ICGM, 22 rue Méchain, 75014 Paris, France

Annexin III is the only member of the annexin family for which an enzymatic activity has been proposed: it was identified as an inositol 1,2-cyclic phosphate 2-phosphohydrolase (1).

Annexins are a family of calcium and phospholipid binding proteins whose biological function is not well defined yet. They consist of a highly conserved C-terminal core of four (exceptionally eight) repeats of about 70 amino acids, and a more variable N-terminus (2). Unlike most members of the family, annexin III is expressed preferentially in neutrophiles, in their final stage of differentiation (3). The 3D structure of the conserved core of annexins V and I shows a protein formed of 4 structural domains, with a hydrophilic channel between modules formed of domains 1+4 and 2+3. The calcium binding sites are all located on one side of the molecule that has thus been defined as the membrane-binding surface.

The overall structure of annexin III is similar to that of annexin V, but with a well defined calcium site in domain 3, which is observed only under high calcium conditions for annexin V. The N-terminus of annexin III carries a Trp which lies with its side-chain inserted into the hydrophilic space between the modules described above, which may affect any potential ion channel activity. The relative position of domain 3 with respect to domain 2 is different from annexins I and V, suggesting a molecular flexibility that may be important for the attachment of annexins on membrane surfaces (4).

The crystals grown in the presence of inositol-2-phosphate (InP) are not isomorphous to native ones, yet we find only one molecule of InP in the structure. It lies on the calcium-binding surface of the protein, interacting with a secondary calcium site in domain 3. The presence of its charged phosphate group favours the coordination of a further metal ion on this surface. Its location does not suggest an enzyme active site, but on the basis of its presence on the membrane- binding surface of annexin we can model the interaction of annexins with phospholipids embedded in a membrane.

(1) Ross, T. et al (1990) Science 248, 605-607.

(2) Raynal, P. & Pollard, H. B. (1994) Biochim. Biophys. Acta 1197, 63-93

(3) Comera, C. et al (1989) J. Cell Biochem. 40, 361-370.

(4) Favier Perron, B. et al (1996) Biochemistry, in the press.