In vivo assay of presynaptic microtubule cytoskeleton dynamics in Drosophila

Abstract

Disrupted microtubule dynamics in neuronal synapses has been suggested as an underlying cause for several devastating neurological diseases, including Hereditary Spastic Paraplegia (HSP) and Fragile X Syndrome (FXS). However, previous studies have been restricted to indirect assays of synaptic microtubules, i.e. immunocytochemistry of microtubule-associated proteins and post-translationally modified tubulins characteristic of microtubules with different stabilities. Very little is known about synaptic microtubule dynamics in vivo, or how microtubule dynamics may be disrupted in disease states. In this study, we develop methods to analyze microtubule dynamics directly in living synaptic boutons in situ. We use fluorescence recovery after photobleaching (FRAP) of transgenic green fluorescent protein (GFP) tagged tubulin at the well-characterized Drosophila neuromuscular junction (NMJ) synapse. FRAP measurements of tubulin-GFP demonstrate biphasic recovery kinetics. Treatment with taxol to stabilize microtubules and promote microtubule assembly reduces both recovery phases. Treatment with vinblastine to disassemble microtubules increases the fast recovery phase and decreases the slow recovery phase. These data indicate that the fast recovery phase is generated by rapid diffusion of tubulin subunits and the slow phase is generated by the relatively slow turnover of microtubules. This study demonstrates that tubulin-GFP fluorescence recovery after photobleaching can be used to assay microtubule dynamics directly in living synapses.

Figure 1. Comparison between transgenic tubulin-GFP expression and native tubulin immunocytochemistry in presynaptic boutons of the Drosophila NMJ

(A) Representative NMJ on muscle 13 (segment A4) expressing UAS-tubulin-GFP driven by elav-GAL4. The terminal is shown at low magnification and the arrow points to a typical type Ib synaptic bouton. Scale bar = 5 μm. (B) Representative tubulin-GFP fluorescence patterns in individual presynaptic boutons at high magnification. Scale bar = 2 μm. (C) Double staining of the fixed NMJ with antibodies specific for α-tubulin (green) and HRP, a presynaptic membrane marker (red). Scale bar = 2 μm.

Figure 2. Characterization of tubulin-GFP distribution between soluble subunit and polymer pools in the presynaptic bouton under different drug conditions

In situ NMJ synapses were permeabilized with 0.5% saponin in the presence of microtubule stabilizing buffer. The fluorescence immediately before permeabilization represents total tubulin-GFP (before) and fluorescence immediately after permeabilization represents tubulin-GFP assembled into microtubule polymers (after). A) Representative images of synaptic boutons without treatment (top), with 50 μM taxol for 30 minutes before permeabilization (middle) and with 1 μM vinblastine for 60 minutes before permeabilization (bottom). All images scaled identically for comparison. Scale bar = 5 μm. B) Histograms showing the quantification of the proportion of tubulin proteins in polymer and soluble subunit pools under the three different conditions.

Figure 3. FRAP of tubulin-GFP under conditions that alter microtubule dynamics

(A) Representative time-course images of tubulin-GFP in a control bouton (top), treated with taxol (middle) and treated with vinblastine (bottom). Red dotted lines indicate the bouton regions that were photobleached and subsequently monitored for fluorescence recovery. The yellow dotted lines indicate the background regions that were used in the fluorescence analysis (see details in Methods and Materials). Each panel in the FRAP series shows before (0 sec), immediately after photobleaching (0.5 sec), ten seconds and nine and half minutes after photobleaching. All images in each condition were scaled identically for comparison. Scale bar = 5 μm. (B) FRAP recover curves of tubulin-GFP under the three conditions. The black solid lines connect the mean value and the bars are s.e.m at each time point. The colored lines are the best two exponential fit to all time points. The equation that was used for the fitting is A = 1-A1*exp(-time/alpha)-(1-A1)*exp(-time/beta), in which A is the normalized fluorescence intensity (dependent variable), time is the actual time (independent variable), and A1, alpha and beta are the three parameters to be determined. Table I list all the fitting parameters that were used for FRAP under the different conditions.