Human erythroid malignancies (polycythemia vera [PV] and erythroleukemia) are associated with erythropoietin (Epo)-independent growth and differentiation. Missense or nonsense mutations in the Epo receptor (Epo-R) have been recently described in experimental erythroleukemia in mice and in cases of erythrocytosis in humans. To search for a similar genetic alteration in erythroleukemia and PV, we entirely sequenced the exons of the Epo-R gene as well as the intron-exon junctions in these disorders using polymerase chain reaction. In 1 of 10 cases of erythroleukemia, a single allele mutation was found in the 8th Epo-R gene exon that changed asparagine 487 into a serine. No Epo-r gene mutation was found in 12 PV cases studied, but the same mutation (N487S) was found in 1 patient with polycythemia that did not fulfill the criteria of PV (polycythemia of unknown origin). We did not detect this mutation after sequencing part of the 8th exon of the Epo-R gene from 21 other patients with polycythemia of unknown origin and 51 normal controls. The Epo-R mutation was also found in Epstein-Barr virus-derived cell lines from both cases, suggesting that it is not related to the malignant clone. Therefore, this mutation does not appear to be somatic, although no familial cases were found. The biologic effect of this mutation was subsequently studied. Erythroid progenitors from the polycythemic patient normally responded to Epo, whereas those from the erythroleukemic patient were Epo-independent due to autocrine stimulation by Epo. The normal and the mutated Epo-R were transfected into the murine Ba/F3 cell line. Both types of cells displayed the same response to Epo for proliferation, differentiation, and inhibition of apoptosis. Although this mutation may destroy a consensus binding site for Grb2, no obvious differences either in the pattern of Epo-induced tyrosine phosphorylated proteins or in the binding of Grb2 to the Epo-R were observed. In conclusion, a somatic Epo-R missense mutation does not appear to be a molecular mechanism involved in the abnormal growth of human erythroleukemia and PV. However, the Epo-R mutation (N487S) that we describe is located in the same tyrosine sequence beginning at AA 485 as the one previously observed (P488S) in as case of polycythemia (Sokol et al, Exp Hematol 22:447, 1994). These results suggest that this phosphopeptide sequence may play an important role in Epo signalling.