A legumin-like seed protein was purified from the endosperm of coffee (Coff
ea avabica L. cv. Colombia). In contrast to legumes, n here efficient stora
ge globulin extraction requires buffered saline solutions well above the ac
idic pK(1) of the globulins, coffee legumin is readily extracted with acidi
c aqueous buffers. The coffee legumin migrates like other 11S storage globu
lins in sucrose gradients. Subunits of coffee legumin have an apparent mole
cular mass of about 55 kDa after one-dimensional SDS-polyacrylamide gel ele
ctrophoresis in the absence of a reducing agent, In the presence of 2-merca
ptoethanol, two polypeptides appear that have apparent molecular masses of
33 and 24 kDa. Two full-length cDNAs were generated from mRNA of developing
seeds that were more than 98% homologous. They had open reading frames of
1458 and 1467 bp. Each encoded legumin precursors of 486 and 489 amino acid
s, respectively (M-r = 54136 and 54818). Examination of a 5' promoter regio
n from a coffee legumin gene revealed a putative legumin-box. Genomic DNA f
rom C. arabica was digested with six different restriction endonucleases, A
fter separation of the fragments by electrophoresis, single discrete fragme
nts on DIVA blots hybridized strongly to a cDNA probe for the acidic chain.
Other fragments that hybridized weakly with this probe were visible after
hybridization at very low stringency, DNA from other species and commercial
ly important cultivars that comprise the genus Coffea produced similar resu
lts. Immunocytochemical studies revealed that some legumin was detected in
the cytoplasm in mature coffee seeds, but that the majority of it was in la
rge storage vacuoles that accounted for most of the cell volume.