Purification, characterization, and cDNA cloning of a 27-kDa lectin (L10) from horseshoe crab hemocytes.

We separated granular components of horseshoe crab hemocytes by a combination of centrifugation on sucrose density gradient and high performance liquid chromatography, and a 27-kDa protein termed L10 was found to be a major component in the large granules (Shigenaga, T., Takayenoki, Y., Kawasaki, S., Seki, N., Muta, T., Toh, Y., Ito, A., and Iwanaga, S. (1993) J. Biochem. (Tokyo) 114, 307-316). In the present work, lectin activity of this protein and its unique primary structure were elucidated. L10 was purified by four steps of chromatography, including dextran sulfate-Sepharose CL-6B, CM-Sepharose CL-6B, Sephacryl S-200, and Mono S. At least three 27-kDa isoproteins, named L10a, L10b, and L10c, were isolated. Their amino acid compositions were almost indistinguishable, and there were no amino sugars. All the isoforms had hemagglutinating activity against human A-type erythrocytes, in a Ca(2+)-independent manner with L10b showing the highest activity. The L10b-mediated hemagglutination was inhibited in the presence of N-acetylglucosamine or N-acetylallolactosamine, and the association constant (Ka) between L10b and N-acetylglucosamine was 1.95 x 10(4) M-1. Furthermore, L10b specifically agglutinated Staphylococcus saprophyticus KD. Ultracentrifugation analysis revealed that L10b is present in monomer form in solution. A cDNA coding for an isoform of L10 was isolated from a hemocyte cDNA library. The open reading frame of the 768-base pair cDNA coded for the signal sequence of 19 residues. The mature protein had 236 residues with the calculated molecular weight of 26,757. Amino acid sequences of the peptides derived from L10c exactly corresponded to the predicted sequence of the cDNA, whereas amino acid replacements of Ile-129 to Val and His-213 to Tyr existed both in L10a and L10b, suggesting that the cDNA codes for L10c. Cysteine was absent and there were five tandem repeats with 47 amino acids in each segment with internal sequence identities of 49-68%. The entire amino acid sequences had no significant sequence similarity with other known proteins.