Mitogen-activated protein (MAP) kinases are serine/threonine protein kinases activated by dual phosphorylation on threonine and tyrosine residues. A MAP kinase kinase (MKK1 or MEK1) has been identified as a dual-specificity protein kinase that is sufficient to phosphorylate MAP kinases p42mapk and p44mapk on the regulatory threonine and tyrosine residues. Because of the multiplicity of MAP kinase isoforms and the diverse circumstances and agonists leading to their activation, we thought it unlikely that a single MKK could accommodate this complexity. Indeed, two protein bands with MKK activity have previously been identified after renaturation following sodium dodecyl sulfate-polyacrylamide gel electrophoresis. We now report the molecular cloning and characterization of a second rat MAP kinase kinase cDNA, MKK2. MKK2 cDNA contains an open reading frame encoding a protein of 400 amino acids, 7 residues longer than MKK1 (MEK1). The amino acid sequence of MKK2 is 81% identical to that of MKK1, but nucleotide sequence differences occur throughout the aligned MKK2 and MKK1 cDNAs, indicating that MKK2 is the product of a distinct gene. MKK1 and MKK2 mRNAs are expressed differently in rat tissues. Both cDNAs when expressed in COS cells displayed the ability to phosphorylate and activate p42mapk and p44mapk, both MKK1 and MKK2 were activated in vivo in response to serum, and both could be phosphorylated and activated by the v-Raf protein in vitro. However, differences between MKK1 and MKK2 in sites of phosphorylation by proline-directed protein kinases predict differences in feedback regulation.