Microtubule Organization Determines Axonal Transport Dynamics

Abstract

Axonal microtubule (MT) arrays are the major cytoskeleton substrate for cargo transport. How MT organization, i.e., polymer length, number, and minus-end spacing, is regulated and how it impinges on axonal transport are unclear. We describe a method for analyzing neuronal MT organization using light microscopy. This method circumvents the need for electron microscopy reconstructions and is compatible with live imaging of cargo transport and MT dynamics. Examination of a C. elegans motor neuron revealed how age, MT-associated proteins, and signaling pathways control MT length, minus-end spacing, and coverage. In turn, MT organization determines axonal transport progression: cargoes pause at polymer termini, suggesting that switching MT tracks is rate limiting for efficient transport. Cargo run length is set by MT length, and higher MT coverage correlates with shorter pauses. These results uncover the principles and mechanisms of neuronal MT organization and its regulation of axonal cargo transport.

Keywords: C. elegans; axonal transport; dynein; kinesin; microtubule; microtubule dynamics; microtubule length; microtubule organization.

Figure 1. A fluorescence based method for analyzing MT organization

A. Cartoon of the DA9 neuron (Presynaptic sites in green), and tiled organization of axonal MTs, which have staggered ends and are oriented with the plus-end-out. Minus ends are in red. The boxed area corresponds to the proximal 80µm of the axon in the dorsal cord, where MT organization was analyzed. B. Kymograph from a GFP::TBA-1 movie showing axonal MT dynamics. Distal (anterior) is left in all Figures. Arrows and scheme on the right highlight growth and shrinkage. ΔF indicates a change in fluorescence that is due to the dynamics of a single MT. C. Quantification of MT dynamics. ** indicate p<0.01. Error bars show standard deviations. n=191 growth events and 185 shrinkage events. D. Micrograph showing continuous GFP::TBA-1 fluorescence along with RFP::PTRN-1 puncta that allow to count MTs. Scale bar is 5µm. Insets show a magnification of the region that is magnified in F. E. Correlation plot of single MT fluorescence intensity that was measured by two different methods from each neuron. n=25. F. Workflow: GFP::TBA-1 is imaged and the diffraction limited intensity profile is plotted. The profile is binned using the intensity of a single MT to yield the quantized distribution (F’ overlay of quantized (red) and original (blue) signals) from which average MT length, coverage and spacing are calculated. Note that “spacing” refers to the axial distance between adjacent minus ends, not lateral distance between polymers. The length of individual MTs cannot be determined because it is not possible to pair plus and minus ends. For visualization purposes (F”) steps down on the quantized signal were assigned to the longest MT. This assumption does not alter the average length, coverage or spacing. G. EM micrograph of the DA9 axon showing 5 MT profiles and the reconstruction of MTs from 205 50nm EM sections. H. GFP::TBA-1 and RFP::PTRN-1 were imaged on a confocal spinning disc, then the same worm was fixed for EM, sectioned and imaged. I. Results of the fluorescence/EM correlative analysis: shown is the quantized distribution for each method. The pixel size of the fluorescence method (210nm) was scaled with respect to the thickness of EM sections (50nm). Holes in the EM quantized scan are sections where MTs could not be scored due to tissue damage caused by the confocal laser. The two methods yield similar average coverage (4.95 and 4.96 MTs) and their alignment is statistically significant (p<0.00001 compared to the same scan with a random pixel order of a different section of the same axon). Inset shows two consecutive EM section where MTs coverage increases (arrow) and the corresponding increase in the GFP::TBA-1 signal.

Figure 2. Developmental scaling of axonal MT length

A. MT length in wt animals increases as animals grow and the axon elongates. Error bars are for standard deviation. p<0.01, Spearman rank correlation. B. Axial spacing of MT minus ends does not change significantly with age. C. MT coverage of the axon increases during development (p<0.01). D–F Mutants that affect body size and hence axon length, were examined for alterations in MT organization at various developmental stages and as adults. The difference in body length between wt and the mutants is indicated on the right. No significant differences are observed, suggesting that MT length is not a direct readout of axon length. G–I. MT organization parameters in worms that were cultivated at different temperatures ranging from 12°C to 25°C, which affects their developmental speed. Despite a wide difference in the amount of time it takes to reach L4 stage (from over a week at 12°C to less than two days at 25°C, indicated on graphs), MT length was not statistically different between groups. Error bars are for standard deviation, ** indicates p<0.01.

Figure 3. Molecular control of axonal MT organization

A. Model of axonal MTs from a wildtype L4 worm. B. Distribution of average MT length (Y axis), minus-end axial spacing (X axis) and coverage (diameter) in a panel of mutants and RNAi treatments (unc-33 and tac-1). All points outside the encircled area have at least one parameter that is significantly different from wt at p<0.01. C–F. MT models from individual animals of select genotypes that show abnormal MT length and spacing. C. ebp-1(tm1357). D. cyy-1(wy302); cdk-5(ok626). E. ptl-1(ok621). F. unc-116(e2310). G. Axonal MT length is not severely affected in single mutants for cyy-1 or cdk-5 but is drastically reduced in the double mutant cyy-1(wy302), cdk-5(ok626). Restoring either pathway can rescue the double mutant, and elimination of ebp-1 enhances the phenotype. H. Genetic interactions between genes involved in MT organization. ptl-1(ok621) mutants show increased average MT length, which depends on EBP-1 and the CDK pathways. Error bars indicate standard deviation, * p<0.05, ** <0.01.

Figure 4. Axonal cargo arrests at MT ends

A. Kymograph showing that different transport events often stop at the same locations. Arrow heads of the same color indicate stop sites at the same locations. B. Kymograph from double-color movie near the tip of the axon shows a mobile vesicle stopping at MT plus end (arrowhead). Note that other MT ends are also decorated by immobile cargo (arrows). The polarity of MT ends is indicated in the merged panel. C. A kymograph of SNG-1::GFP from a movie taken after a stack acquisition of RFP::PTRN-1 (intensity profile in blue). Most stops on retrograde runs (left to right) colocalize with RFP::PTRN-1 puncta (green lines). Red lines shows uncorrelated pauses. D. Quantification of colocalization between RFP::PTRN-1 and all immobile cargo (left) or only retrogradely moving cargo (right). Random colocalization was calculated using the frequency of RFP::PTRN-1 (see methods). p<0.01, n=37 axons (all immobile cargo) or 22 axons (retrograde runs only).

Figure 5. MT organization mutants show cargo trafficking defects that correlate with their MT defects

A–D. Kymographs from movies of GFP::RAB-3 taken in the dorsal proximal asynaptic domain of the indicated genotypes. The genotypes are ordered from left to right according to the MT length of the mutants. E. Quantification of GFP::RAB-3 run-length in different genotypes. Error bars denote standard error, n=290–1000 moving events per genotype, ** indicate p<0.01. Red crosses mark the average MT length measured at the same developmental stage (late L4). The increase in GFP::RAB-3 run-length mirrors the increase in MT length. F. Average pause time of the same genotypes. Error bars denote standard error, n=185–489 pauses per genotype, ** indicate p<0.01.

Figure 6. MT organization determines axonal transport dynamics during development

A. Kymographs of SNG-1::GFP in different developmental stages. B. Average run length co-varies with MT length and increases during development (p<0.01). Red + signs indicate average MT length for each developmental stage. Error bars are for standard errors of the mean. n=251–393 events per time point. C. Quantification of pause duration in L3 larvae and 2 days old adults. D. Scheme of DA9 showing the distal area where MT coverage is low, and the proximal area that harbors more MTs. E. Kymographs from movies taken in the low MT density distal axon or the high MT density proximal axon. Photobleaching was used to eliminate fluorescence of preexisting immobile cargo. F. Quantification of pause duration shows that vesicles are immobile for longer periods of times at MT tips when MT coverage is low.