Research: J.T. Schmidt

Research in my laboratory centers on activity-driven refinement of retinotectal projection in zebrafish:
Visual activity, acting via NMDA receptors, refines developing retinotectal maps by shaping retinal arbors. Initially, retinal axons emit many transient side branches along the shaft (resembling a "bottlebrush"), but some branches are stabilized and branch further to give a mature arbor. When MK801 blocks NMDA receptors, dynamic rates of branch addition and deletion increase twofold, as if this prevents release of a stabilizing signal. Ca ++ entry through NMDARs is known to activate phospholipase A2 (cPLA2, a PSD protein) to release arachidonic acid (AA), a putative retrograde signal (Bliss & Collingridge, Nature 1993). AA released by a second enzyme (DAG lipase) mediates L1, NCAM, N-cadherin and FGF stimulation of axon growth via protein kinase C (PKC) activation and GAP-43 phosphorylation to stabilize F-actin (Meiri et al. J Neurosci.1998). We propose that postsynaptic cPLA2 releases AA as a retrograde signal to tap into this presynaptic growth control mechanism.

We previously reported that blocking either presynaptic PKC or DAG lipase causes increased branch turnover like MK801, but additionally causes arbors to remain immature due to interruption of the growth pathway. We have blocked AA release from cPLA2 by injecting into tectal ventricle either a selective pharmacological inhibitor (AACOCF3) or an antisense oligo to suppress expression. Both methods increased branch turnover without effect on arbor maturation (below), a result more similar to that of blocking NMDA receptors. Scrambled oligos had no effect. Injection of the antisense oligo into eye to suppress only presynaptic cPLA2 produced no effect, suggesting a role for postsynaptic cPLA2. After MK801 treatment, exogenous AA reversed the increase in dynamic rates (bar graph).

Axons exposed to antisense oligos
Graph of results

Finally, we are using the fluorogenic cPLA2 substrate PED6 (Mol. Probes) to show that trains of spikes (driven by strobe illumination) activate tectal cPLA2. The results implicate the cPLA2-AA-PKC-GAP43 pathway as a part of an F-actin based mechanism of synaptic stabilization.

research picture

Next, we used DNA constructs to express GAP-43 in retinal ganglion cells as GFP(green fluorescent protein) fusion proteins and assessing the effects on retinal arbor morphology (below). Injection into fertilized eggs results in scattered GFP expression in neurons, including ganglion cells in the eye (A), and their axons can be seen exiting in the optic nerve head and forming arbors in brain (arrows in B and C). GAP43-GFP fusion proteins can also be expressed in axons and observed in time-lapse at half hour intervals (D1 through 4).

Neurons expressing GFP

These arbors expressing excess GAP43 were abnormally large (see figure below). However, if the protein kinase C phosphorylation site was mutated from serine to alanine (which cannot be phosphorylated), then the arbors did not grow and remained immature. This demonstrated that it was not the level of the protein that mattered, but only the form that was in the phosphorylated state. In addition to growth, the phosphorylated GAP43 cause increased branching.

We are now investigating whether the polarity complex, which contains an atypical PKC, might also be involved in branch formation. The polarity complex, which is necessary for initial axon formation, includes Par3, Par6, cdc42 and atypical protein kinase C (aPKCz) and is recruited and assembled at the growing axon tip by PI3 Kinase (phosphatidyl-inositol-3-kinase). The complex forms a positive feedback loop downstream of PI3 Kinase and organizes both the microtubule and F-actin cytoskeletons.  The activity of the aPKCz is essential for polarity complex activation, and it can be activated by retrograde AA signaling. PI3Kinase, which both recruits and is the upstream activator of the polarity complex, was also tested for involvement in branch formation. The results showed that few branches could be formed when any of these were blocked or expression suppressed. These results suggest a portion of the pathways through which activity can affect the growth and branching of retinotectal arbors.