Lethal Factor: Chopping Up MKK's

The first insight into the catalytic activity of LF came from the studies of Klimpel et al. [56], who showed that a small region of the C-terminus contained a motif (686HEFGHAV692) characteristic of the Zn²⁺-binding site of metalloproteases (HEXXH).  Mutating the residues His686 or His690 to alanine or residue Glu687 to cysteine not only inactivated LF, but in the case of the former also reduced Zn²⁺-binding by LF [56].  Moreover, LF toxicity could be blocked by protease inhibitors [56].  Although these data indicated that LF was a metalloprotease, the identity of its substrate(s) remained unknown.  The first clues regarding the nature of the substrate were provided by Hammond and Hanna [57,58], who demonstrated that LF could cleave synthetic peptides and that these substrates usually contained proline residues immediately preceding the cleavage site.  Then, in 1998, two groups independently demonstrated that LF proteolytically cleaved the N-terminus of [59,60] and inactivated [60] mitogen-activated protein kinase kinases (MKK or MEK) 1 and 2.  Subsequently LF was found to cleave not only MEKs 1 and 2, but also MKKs 3, 4, 6, and 7 [61,62,63]. 

A comparison of cleavage sites in each of these proteins indicated that LF preferentially cleaves just before an aliphatic residue which lies 2-3 residues following a stretch of basic amino acids.  This motif is similar to a described generic MAPK binding site, or docking (D) domain, consisting of a basic amino acid center that is flanked by hydrophobic residues on one or both sides [64].  This observation was particularly interesting because it immediately provided a mechanistic explanation for how LF inactivates MEK.  Indeed, Chopra et al. [65] have since demonstrated that proteolysis by LF reduces MEK1’s affinity for its substrate MAPK.

The identification of a consensus cleavage site for LF raises questions regarding protease-substrate specificity.  Despite the fact that D domains are found in a multitude of proteins which interact with MAPKs [66], substrates for LF outside the MEK family have not been identified.  If not the D domain, then what is it that determines LF-substrate specificity?  Interestingly, yeast two-hybrid analyses for binding partners of LF have isolated cDNA for MEK2 which lacks the N-terminal cleavage site [59].  Thus, other regions of MEKs, in addition to the N-terminal cleavage site, must be required for LF-substrate recognition.  In support of this, Chopra et al. [65] have identified a conserved LF-interacting region (LFIR) at the C- terminus (residues 292–318) of MEK1 that is necessary for proteolysis by LF.  Point mutations at conserved residues within this region prevent LF proteolysis of MEK1 without altering MEK1's biochemical activity.  The dual requirement for a D domain and an LFIR may explain the apparent specificity of LF for the MEK protein family.