Serotonin binding protein (SBP) is found in synaptic vesicles of mammalian central and peripheral serotonergic neurons. 5-Hydroxytryptamine (5-HT, serotonin) is physiologically stored as a complex with SBP in vivo. Two forms of SBP have been detected with apparent molecular weights of 45,000 and 56,000 (45K and 56K). To understand the relationship between the two forms more fully, we purified the two proteins to homogeneity and partially characterized them. Purification steps included (NH4)2SO4 fractionation and chromatography on Sepharose 4-B, Affi-Gel-Blue, hydroxylapatite, and phosphocellulose. The 45K from of SBP was obtained pure, whereas the 56K form of SBP was obtained about 90% pure by these methods. To isolate pure 56K SBP for induction of antibodies, the protein was further purified by sodium dodecyl sulfate-gel electrophoresis followed by electroelution. The 56K form of SBP was thus isolated, but in a denatured state; its purity was established by two-dimensional gel electrophoresis. The two forms of SBP (pure 45K and 90% pure undenatured 56K SBP) were similar in their 5-HT binding capacity; the enhancement of 5-HT binding by Fe2+; and inhibition by--SH reagents, chelators, and sodium salts. Antibodies raised against the pure 56K form of SBP cross-reacted with the 45K SBP. The two forms of SBP differed in the following properties: (1) dissociation constants--56K form showed higher affinity for 5-HT (KD1 = 0.4 nM; KD2 = 32 nM), whereas the 45K form showed lower affinity (KD1 = 9.7 nM; KD2 = 120 nM); (2) ratio of number of 5-HT binding sites with low affinity to those with high affinity--56K (19:1), 45K (10:1); (3) isoelectric point--the 56K form of SBP is more acidic (5.6 and 5.9) than the 45K form (6.1); (4) binding enhancement by gangliosides and bicarbonate. To establish whether the 45K form of SBP is found in vivo or is produced by proteolysis during isolation, two additional experiments were carried out. (1) We added a mixture of proteolytic enzyme inhibitors to our homogenization buffer; this addition did not change the ratio of the two forms of SBP. (2) We mixed regions of the CNS enriched in the 45K form of SBP (spinal cord) with regions rich in the 56K form of SBP (raphe nuclei) and homogenized them together. Again, this procedure failed to change the ratio of the two forms of SBP as judged by polyacrylamide gel electrophoresis.(ABSTRACT TRUNCATED AT 400 WORDS)