We have characterized the self-association of FtsZ in its GDP-bound state (GDP-FtsZ) and the heteroassociation of FtsZ and a soluble recombinant ZipA (sZipA) lacking the N-terminal transmembrane domain by means of composition gradient-static light scattering (CG-SLS) and by measurement of sedimentation equilibrium. CG-SLS experiments at high ionic strengths and in the presence of 5 mM Mg(2+) show that, while FtsZ self-associates in a noncooperative fashion, sZipA acts as a monomer. CG-SLS data obtained from mixtures of FtsZ (A) and sZipA (B) in the presence of Mg(2+) are quantitatively described by an equilibrium model that takes into account significant scattering contributions from B, A(1), A(2), A(3), A(4), A(5), A(6), A(1)B, A(2)B, A(3)B, and A(4)B. However, in the absence of Mg(2+) (with EDTA), the data are best explained by an equilibrium model in which only B, A(1), A(2), A(3), A(1)B, and A(2)B contribute significantly to scattering. The best-fit molecular weights of monomeric A and B are in good agreement with values calculated from amino acid composition and with values obtained from sedimentation equilibrium. The latter technique also confirmed the interaction between sZipA and GDP-FtsZ. Moreover, the association model that best describes the CG-SLS data is in qualitative agreement with the sedimentation data. From these results, it follows that the binding of sZipA to GDP-FtsZ is of moderate affinity and does not significantly affect the interactions between FtsZ monomers. Under the working conditions used, only one sZipA binds to FtsZ oligomers with a length of six at most. The observed behavior would be compatible with FtsZ fibrils being anchored in vivo to the bacterial inner plasma membrane by substoichiometric binding of membrane-bound ZipA.