Tighter αC-helix-αL16-helix interactions seem to make p38α less prone to activation by autophosphorylation than Hog1.

Many eukaryotic protein kinases (EPKs) are autoactivated through autophosphorylation of their activation loop. Mitogen-activated protein (MAP) kinases do not autophosphorylate spontaneously; relying instead upon mitogen-activated protein kinase (MAPK) kinases (MKKs) for their activation loop phosphorylation. Yet, in previous studies we identified mutations in the yeast MAPK high osmolarity glycerol (Hog1) that render it capable of spontaneous autophosphorylation and consequently intrinsically active (MKK-independent). Four of the mutations occurred in hydrophobic residues, residing in the αC-helix, which is conserved in all EPKs, and in the αL16-helix that is unique to MAPKs. These four residues interact together forming a structural element termed 'hydrophobic core'. A similar element exists in the Hog1's mammalian orthologues p38s. Here we show that the 'hydrophobic core' is a loose suppressor of Hog1's autophosphorylation. We inserted 18 point mutations into this core, 17 of which were able to render Hog1 MKK-independent. In p38s, however, only a very few mutations in the equivalent residues rendered these proteins intrinsically active. Structural analysis revealed that a salt bridge between the αC-helix and the αL16-helix that exists in p38α may not exist in Hog1. This bond further stabilizes the 'hydrophobic core' of p38, making p38 less prone to de-repressing its concealed autophosphorylation. Mutating equivalent hydrophobic residues in Jnk1 and Erk2 has no effect on their autophosphorylation. We propose that specific structural elements developed in the course of evolution to suppress spontaneous autophosphorylation of Hog1/p38. The suppressors were kept wobbly, probably to allow activation by induced autophosphorylation, but became stricter in mammalian p38s than in the yeast Hog1.