Previous biochemical studies suggested that the human platelet Ca2+ATPase system may be cell-specific. To test this hypothesis, we first undertook the molecular cloning of Ca2+ATPase from human erythroleukaemia (HEL) cells, because this human cell line exhibits megakaryocytic features and expresses a Ca2+ATPase that cross-reacts with platelet Ca(2+)-ATPase. For this cloning, an HEL-cell cDNA library was screened with a rat cardiac Ca2+ATPase cDNA probe. The insert of the longest clone isolated was 3.9 kb and its sequence displayed a 100% identity with that of the non-muscle human Ca2+ATPase 2-b isoform, termed SERCA2-b (sarco-endoplasmic-reticulum Ca2+ATPase). The 3.9 kb cDNA covered a subtotal coding region and part of the 3′ non-coding end of the SERCA2-b mRNA. It cross-hybridized with the 4 kb transcript species of cardiac SERCA2-a and with non-muscle SERCA2-b mRNAs, but not with fast-skeletal-muscle SERCA1 mRNA. We next confirmed that SERCA2-b was a component of the platelet Ca2+ATPase system because (1) the platelet clones isolated from a platelet cDNA library exhibited a 100% homology with HEL-cell cDNA; (2) SERCA2-b mRNA was amplified by PCR on total platelet RNA and (3) platelet Ca2+ATPase cross-reacted with a polyclonal SERCA2-b-specific antiserum. Platelets therefore contain a Ca2+ATPase definitely identified as the SERCA2-b isoform of Ca2+ATPase, thus eliminating the possibility that they only contain a single specific Ca2+ATPase.