br Conformational properties of DGK
Conformational properties of DGKϵ
Interaction of DGKϵ with its lipid substrate A major component of the interaction between DGKϵ and lipid bilayers is the insertion of the hydrophobic segment comprising residues 20–42 into the membrane, as discussed above. However, this is not the only interaction of this protein with lipids. One of the substrates of DGKϵ, DAG, is also a lipid (the other substrate being ATP). DGKϵ is highly specific towards DAG substrates that contain an arachidonoyl moiety. Because there is a similar pattern of amino cysteine protease residues required for both DGKϵ (Shulga et al., 2011b) and certain lipoxygenases (Neau et al., 2009), we have suggested that part of the interaction of DGKϵ with its lipid substrate occurs with the region of the protein containing this amino acid pattern. We have termed this amino acid pattern as the LOX motif and have defined it as L-X-R-X(2)-L-X(4)-G, in which -X- is n residues of any amino acid. This segment is found within the accessory domain of DGKϵ, corresponding to residues 436–456. It is not found in other mammalian DGK isoforms. This segment is critically required for DGKϵ’s enzymatic activity and acyl chain specificity (Shulga et al., 2011b). Located at the N-terminus of this segment is a conserved LOX motif corresponding to L431-G443. The crystal structure of 8R-lipoxygenase suggests that these essential residues form part of a channel that provides access to the catalytic site (Neau et al., 2009). Mutating any of the conserved residues of the LOX-motif affects both enzymatic activity and arachidonoyl specificity of DGKϵ (Shulga et al., 2011b). Mutations that we made in the LOX motif of DGKϵ resulted in proteins largely devoid of enzymatic activity and hence it was difficult to characterize the residual activity of the mutants. We therefore mutated a segment of DGKϵ adjacent to the LOX-motif that caused a smaller decrease in enzyme activity, and in two cases (Y451F and R457Q) the mutations resulted in an unexpected increase in enzymatic activity (D’Souza and Epand, 2012). These mutations are more hydrophobic, which may facilitate better alignment between the hydroxyl group of the substrate and the catalytic site of DGKϵ. The high residual activity of these and other mutations in the region of DGKϵ between residues 447 and 457 resulted in changes in preference between arachidonoyl and linoleoyl-containing substrates (D\'Souza and Epand, 2012). The native form of DGKϵ has a preference for SAG over 1-stearoyl-2-linoleoyl glycerol (SLG) as substrate. This preference can be either increased or decreased by mutation in this region of the protein. We suggest that because the segment 447–457 is adjacent to the LOX-motif, mutations in this region do not drastically eliminate substrate binding, but they do modify the substrate binding pocket in a manner that affects relative substrate specificity.
DGKϵ and the PI-cycle It should be noted that DGKϵ is not highly expressed in all tissues (it is most prevalent in the brain, retina, and cardiac muscle), yet PI can still be synthesized and enriched with specific acyl chains by other processes (D’Souza et al., 2014, Milne et al., 2008). However, in tissues such as the brain, where there is high DGKϵ expression, there is a particularly large amount of acyl chain enrichment with sn-1 stearoyl and sn-2 arachidonoyl species of PI (D’Souza and Epand, 2014). Studies using DGKϵ KO MEFs demonstrate the importance of DGKϵ in the PI-cycle by showing that there is roughly a three-fold reduction in PA and PI content in the plasma membrane of these cells compared to the plasma membrane of WT MEFs (Shulga et al., 2010). Similar lipodomic studies with DGKϵ KO MEFs show significant differences in the levels of arachidonate-containing glycerophospholipids between WT and KO cells (Milne et al., 2008). The lipid class showing the greatest reduction in arachidonoyl content is the phosphoinositides (PI and PIPn) (Milne et al., 2008). DGKϵ KO cells show the greatest reductions in the levels of 38:4 PI, corresponding to 1-stearoyl-2-arachidonoyl phosphatidylinositol. Other lipid classes also show reduced arachidonoyl content, although the effect on PI is greatest (Milne et al., 2008). As expected, these lipidomic studies show that in DGKϵ KO cells, the enrichment of arachidonoyl content from DAG to PA is decreased as expected from the loss of the arachidonoyl-specific DGKϵ. Unexpectedly, the arachidonoyl enrichment from PA to PI is also decreased, although DGKϵ does not catalyze any reaction in this path. Thus DGKϵ may also be involved in another function to increase enrichment of arachidonoyl content during the conversion of PA to PI (Milne et al., 2008) which we are currently investigating.