1. The properties of specific Ins(1,4,5)P3- and Ins(1,3,4,5)P4-binding sites have been compared in a crude 'P2' cerebellar membrane fraction. 2. A homogeneous population of [3H]Ins(1,4,5)P3-binding sites was present (KD 23.1 +/- 3.6 nM) at high density (Bmax. 11.9 +/- 1.8 pmol/mg of protein); whereas data obtained for [32P]Ins(1,3,4,5)P4 specific binding were best fitted to a two-site model, the high-affinity binding component (KD 2.6 +/- 0.7 nM) constituted 64.2 +/- 4.3% of the total population and was present at relatively low density (Bmax. 187 +/- 27 fmol/mg of protein). 3. The two high-affinity inositol polyphosphate-binding sites exhibited markedly different pH optima for radioligand binding, allowing the two sites to be independently investigated. At pH 8.0, [3H]Ins(1,4,5)P3 binding was maximal, whereas [32P]Ins(1,3,4,5)P4 specific binding was very low; conversely, at pH 5.0, [32P]Ins(1,3,4,5)P4 binding was maximal, whereas [3H]Ins(1,4,5)P3 binding was undetectably low. 4. Both inositol polyphosphate-binding sites exhibited marked positional and stereo-specificity. Of the analogues studied, only phosphorothioate substitution to form inositol 1,4,5-trisphosphorothioate was tolerated at the Ins(1,4,5)P3-binding site, with only a 2-3-fold loss of binding activity. Addition of a glyceroyl moiety at the 1-phosphate position or addition of further phosphate substituents at the 3- or 6-positions caused dramatic losses in displacing activity. Similarly, complete phosphorothioate substitution of Ins(1,3,4,5)P4 caused an approx. 6-fold loss of binding activity at the [32P]Ins(1,3,4,5)P4-binding site, whereas Ins(1,4,5,6)P4, Ins(1,3,4,6)P4, Ins(1,4,5)P3 and Ins(1,3,4,5,6)P5 were bound at least 100-fold weaker at this site. Therefore, only the phosphorothioate derivatives retained high affinity and selectivity for the two inositol polyphosphate-binding sites. 5. Heparin and pentosan polysulphate were potent but non-selective inhibitors at Ins(1,4,5)P3- and Ins(1,3,4,5)P4-binding sites. N-Desulphation (with or without N-reacetylation) of heparin decreased inhibitory activity at the Ins(1,4,5)P3-, but not at the Ins(1,3,4,5)P4-binding site; however, the selectivity of this effect was only about 4-fold. O- and N-desulphated N-reacetylated heparin was essentially inactive at both sites. 6. The results are discussed with respect to the separate identities of the inositol polyphosphate-binding sites.