br DAG kinase activity is confined to specific
DAG kinase activity is confined to specific cell compartments A number of reports demonstrating agonist-dependent translocation of DGKs to distinct membrane compartments suggest that DGK activity is restricted to localized DAG pools generated after activation of receptors. Perhaps the best evidence of spatially restricted DAG kinase activity was demonstrated by van der Bend et al. . This group measured DAG kinase activity in cells following receptor activation—which caused physiological DAG production—or after treating the cells with exogenous PLC—which caused global, nonspecific DAG generation. They detected significant DAG kinase activity upon activating a receptor, but found very little DAG kinase activity after treating the cells with exogenous PLC. Their data suggested that DGKs are active only in spatially restricted compartments following physiological generation of DAG. Consistent with this conclusion, Nurrish et al.  found in C. elegans that dgk-1, an ortholog of human DGKθ, regulated DAG signalling that was necessary for DL-Dithiothreitol and release. Their data suggested a model where serotonin signalling—which inhibits locomotion—activated the DGK to reduce DAG accumulation. DGKs appear to be active in a number of cell compartments. For example, Nagaya et al.  demonstrated that overexpressed DGKδ partly localized in the endoplasmic reticulum, while Abramovici et al.  found endogenous DGKζ at the neuromuscular junction. Several groups have noted DGK activity in the cell fractions containing cytoskeleton components. For example, Tolias et al.  noted that DGK activity associated with a complex of proteins including a PIP5K, Rac, Rho, Cdc42, and Rho-GDI, all of which regulate cytoskeleton dynamics. We found that several DGK isotypes co-immunoprecipitated with either Rac, Rho, or Cdc42 when overexpressed in cells (M.K.T. and B.L., unpublished observations), and Houssa et al.  showed that active RhoA associated with DGKθ. Additionally, we found that DGKζ interacted with human PIP5K type Iα and increased its activity by generating phosphatidic acid . The physiological significance of these interactions is not entirely clear, but there are data demonstrating that DGKs can modulate cytoskeleton remodeling. For example, DGK inhibitors—which primarily affect type I enzymes—augmented platelet secretion and aggregation , and Abramovici et al.  recently demonstrated that expression of a DGKζ mutant that localized strongly with the plasma membrane enhanced membrane ruffles and caused the formation of large intracellular vesicles. Consistent with an effect on cytoskeleton dynamics, endogenous DGKζ co-purified with components of the cytoskeleton  and it localized at the leading edge of both glioblastoma cells  and C2 myoblasts . Together, these data suggest that DGKs have a broad role in regulating the cytoskeleton, but at this point, their specific roles are not clear. The nucleus has a phosphatidylinositol cycle that is regulated separately from the plasma membrane PI signalling . Like its extranuclear counterpart, nuclear DAG signalling appears to be compartmentalized. Indeed, D\'Santos et al.  demonstrated independently fluctuating pools of nuclear DAG which had distinct fatty acid compositions. This complexity is not surprising because diverse stimuli can lead to generation of nuclear DAG. For example, different growth factors (e.g. IGF-1 or thrombin) stimulated temporally distinct pulses of nuclear DAG , , and several groups have demonstrated that nuclear DAG fluctuates independently of extranuclear DAG during the cell cycle . Nuclear DAG was shown to peak shortly before S phase, suggesting that it may participate in the G1/S transition . Most data support the conclusion that nuclear DAG promotes cell growth. Consistent with this, and emphasizing the importance of nuclear DAG signalling in the cell cycle, we found that cells overexpressing DGKζ—which partly localized in the nucleus—accumulated at the G0/G1 transition, presumably because the kinase reduced nuclear DAG . Several other DGKs have been detected in the nucleus. DGKs α and ι, translocated to the nucleus , , , while a significant fraction of DGKθ localized there constitutively . By transfecting different DGK isotypes into COS-7 cells, we detected DGKs β, γ, δ, and ε in the nucleus (M.K.T., unpublished observations). Nuclear DGKs appear to be confined to separate, distinct regions of the nucleus: DGKs θ, ζ, and ι were noted in discrete regions within the body of the nucleus , , ,  while DGKα appeared to predominantly localize around its periphery . Combined with the complexity of nuclear DAG signalling, these data suggest that within the nucleus, DGKs regulate distinct pools of signalling DAG. Supporting this conclusion, experimental evidence suggests contrasting roles for the nuclear DGKs α and ζ: overexpression of DGKζ inhibited progression from G1 to S phase of the cell cycle , while PA generated by nuclear DGKα appeared to be necessary for IL-2-mediated progression to S phase of the cell cycle . The opposing function of these DGKs indicates both the importance and complexity of nuclear DGK activity and lipid signalling.