Many, and possibly all, plants contain ribosomeinactivating proteins (RIPs) [1]. RIPs cleave the N-glycosidic bond at adenine 4324 of rat liver 28S rRNA to inactivate the ribosomes, resulting in the inhibition of protein synthesis [2,12]. RIPs attract the interest of many researchers for their possible usage as therapeutic agents, especially against human immunodeficiency virus [ 11,13]. Luffin-a and luffin-b are single-chain RIPs isolated from the seeds of Luffa cylindrica (sponge gourd) [6,9], and their amino acid sequences have been reported [4,5].We have isolated two types of cDNAs for luffin from the cDNA library of the immature seeds of L. cylindrica using PCR-derived probe synthesized on the basis of the amino acid sequence of luffin-a [7]. One contained the same restriction fragment patterns with the probe and encoded c~-luffin, a protein highly homologous to luffin-a [7]. Here we report a nucleotide sequence of another cDNA which hybridized to the probe but contained Eco RI site not found in the probe. The cDNA was composed of 914 nucleotides containing an open reading frame of 278 amino acid residues ( Fig. 1). Its 5'-non-coding region, 5'-GGGAAAA-3', was almost identical to that of e-luffin, 5'-GGGAAAG-3'. Comparing its 3'-non-coding region to that of c~-luffin, it seems to be truncated by a polyadenylation site different from e-luffin. The protein deduced from the cDNA was designated as fi-luffin. From the alignment of the amino acid sequences of fi-luffin, c~-luffin and luffin-b, fi-luffin seems to be a preproprotein with a secretory signal peptide of 21 residues and an extrapeptide of 10 residues at the NH2 and COOH terminals, respectively (Fig. 1).The amino acid sequence offi-luffin was aligned with those of luffin-a [4], luffin-b [5], and 0~-luffin [7] (Fig. 2). It has been reported that several conserved amino acid residues among many RIPs are involved in forming or stabilizing a putative active cleft of ricin A-chain [8]. Most of them are also conserved among all the luffins (Fig. 2). Mature fl-luffin is more homologous to luffin-b (91 ~o homology) than luffin-a (80 ~o ) and mature c~-luffin (83 ~o), while mature e-luffin is more homologous to luffin-a (96~o) than luffin-b (80~o). These data suggest that the small differences between ~-luffin The nucleotide sequence data reported will appear in the EMBL, GenBank and DDBJ Nucleotide Sequence Databases under the accession number X62372. ~. 1. Nucleotide sequence of the cDNA encoding fi-luffin ~d its deduced amino acid sequence. Isolation of the poly(A) + RNA ~d construction of the cDNA librau ~e described in our previous paper [7]. The cDNAs were isolated from the librwy by plaque hybridization using the PCR-derived probe [7]. Their nucleotide sequences were determined using T7 sequencing kit (Phwmacia). Closed and open ~rowheads indicate putative processing sites at NH 2 and COOH terminals from the Nignment of the amino acid sequences of fi-l~n, ~-luffin mid luffin-b. and luffin-a and those between/Muffin and luffinb in their amino acid sequences...