Mouse erythroleukaemia (MEL) cells, which have not been induced into erythroid development, contain a protein kinase (MKJ which phosphorylates the a subunit of protein-synthesis-initiation factor 2 (eIF-2a). In this paper, we show that this kinase phosphorylates both eIF-2a and a synthetic peptide based on the phosphorylation site in eIF-2a at Ser51, the target residue for other eIF-2a kinases. Consistent with this, prior treatment of eIF-2 with MK,, impaired the exchange of bound GDP for GTP which is catalysed by the exchange factor eIF-2B.Using a modified cell-free translation system, we have shown that MK, inhibits translation, consistent with the above observations concerning the site of phosphorylation and the effect of phosphorylation on eIF-2B-mediated guanine-nucleotide exchange.MK, has been purified and its properties have been compared with those of the haem-controlled repressor eIF-2a kinase (HCR) from rabbit reticulocytes. Its behaviour on gel filtration is similar to that of HCR, while its behaviour on anion exchange resembles that of certain phosphurylated species of HCR. Highly purified preparations of MK,, contain a protein with an apparent molecular mass of 98 kDa which comigrates with HCR on SDSDAGE. This protein undergoes phosphorylation when incubated in the presence of Mg2+-ATP, and both this apparent autophosphorylation and the activity of the kinase against eIF-2a are inhibited by the same, low, (10pM) concentrations of haemin. Phosphorylation of the 98-kDa components present in the MEL-cell kinase preparation and in purified rabbit reticulocyte HCR occurs on serine and threonine residues. Analysis of these phosphoproteins by peptide mapping reveals significant differences in their structures, indicating that they may be closely related, but are certainly not identical.Eukaryotic initiation factor 2 (eIF-2) mediates the binding of the initiator tRNA (fMet-tRNAJ to the 40s ribosomal subunit in eukaryotic cells [l]. It does so in the form of a complex which also contains GTP, which is subsequently hydrolysed to GDP prior to the binding of the 60s subunit. eIF-2 is therefore released from the ribosome as a complex with GDP. Since this is inactive in fMet-tRNA, binding, the GDP must be exchanged for GTP. This process has been termed recycling and is catalysed by a further protein factor, the guanine-nucleotide-exchange factor (eIF-2B) [2].It is well-established that phosphorylation of the a subunit of eIF-2 (eIF-2a) at a single serine residue (Ser51) impairs its recycling by eIF-2B, and hence leads to inhibition of peptide-chain initiation (reviewed in [3]). This regulatory mech- Abbreviations. dsI, double-stranded RNA-activated inhibitor, an eIF-2a kinase ; eIF-2, eukaryotic protein-synthesis-initiation factor 2 ; eIF-2a, a subunit of eIF-2; eIF-2B, eukaryotic initiation factor 2B ; HCR, the haem-controlled repressor, an eIF-2a kinase ; hsp90, 90-kDa heat-shock protein ; MEL, mouse erythroleukaemia ; MK,, eIF-2a kinase from induced MEL cells ; M K , eIF-2a kinase from uninduced MEL cells ; f ...