The Drosophila BEACH family protein, blue cheese, links lysosomal axon transport with motor neuron degeneration

Abstract

Impaired axon transport is one of the earliest pathological manifestations of several neurodegenerative diseases, and mutations in motor proteins can exacerbate or cause degeneration (Williamson and Cleveland, 1999; Gunawardena and Goldstein, 2004; Stokin and Goldstein, 2006). Compromised function in lysosomes and other degradative organelles that intersect with the lysosomal pathway are also strongly implicated in neurodegenerative disease pathology (Nixon and Cataldo, 2006; Rubinsztein, 2006). However, any functional link between these two phenomena has not as yet been recognized. Drosophila mutants in blue cheese (bchs) undergo progressive brain degeneration as adults and have shortened life span (Finley et al., 2003), but the cellular function of Bchs and the cause of degeneration have not been identified. A role in lysosomal trafficking is suggested by the homology of Bchs with the vesicle trafficking-associated BEACH (Beige and Chediak-Higashi) domain protein family (Wang et al., 2002; De Lozanne, 2003) and by its genetic interaction with a lysosomal transport pathway (Simonsen et al., 2007). Here, we describe the degeneration of a population of identified larval motor neurons in bchs mutants. We present evidence that Bchs is primarily lysosomal in those motor neurons in wild type and, using live fluorescence imaging of individual motor neurons in intact larvae, show that lysosomal vesicles fail to be transported toward motor neuron termini in bchs mutant and Bchs-overexpressing larvae. We suggest therefore that anterograde transport of lysosomes toward synaptic termini is a key factor in preventing motor neuron degeneration and that Bchs reveals a functional link between the lysosomal degradative pathway and transport.

Figure 1.

Bchs is expressed highly in a subset of motor neurons in embryonic CNS. Bchs (red) is seen in longitudinal axon tracts in the CNS and in a subset of motor neuron cell bodies in wide-field fluorescence images of a filleted stage 15/16 embryo. Note that punctate localization (red arrow) of Bchs is prominent in motor neuron cell bodies of aCC, RP2 (white arrows), and in an unidentified cell posterior to RP2. Images were acquired by wide-field fluorescence microscopy. Anterior is up. Scale bar, 20 μm.

Figure 2.

Colocalization between Bchs and vesicular trafficking markers in embryonic aCC and RP2 motor neuron cell bodies. aCC and RP2 motor neurons were labeled with endolysosomal GFP fusion markers expressed via the Eve driver, and immunostained for endogenous Bchs in filleted stage 15/16 embryos. The Costes colocalization thresholding algorithm applied to the Pearson and Manders coefficients (part of the “Colocalization Threshold” ImageJ plugin by T. Collins and W. Rasband) was used to measure colocalization scores between Bchs and the endolysosomal markers LAMP-GFP (A) (n = 72) and spinster-GFP (B) (n = 83); the endosomal sorting marker 2×FYVE-GFP (C) (n = 78); the recycling endosomal marker rab11-GFP (D) (n = 100); and the late endosome marker rab7-GFP (E) (n = 80) in single confocal images acquired with a 63×/1.4 NA objective lens, at 5× zoom. Overlaid red (Bchs) and green (GFP) channels show colocalized pixels in beige with intensity equivalent to √〈ch1 intensity * ch2 intensity〉. F, For each marker, thresholded values for the Pearson correlation coefficient (R-coloc), tM1 (the percentage of colocalized intensity in the red channel relative to total intensity in the red channel), and tM2 (the percentage of colocalized intensity in the green channel relative to total intensity in the green channel) are shown. Scores are plotted for pixels above an automatically determined threshold for both channels, according to the algorithm of Costes et al. (see Materials and Methods). Colocalization values for each parameter are consistently higher with late endolysosomal markers LAMP-GFP and spin-GFP than early endosomal or recycling markers FYVE-, rab7-, and rab11-GFP. Scale bar, 2 μm. Error bars show SEM.

Figure 3.

Abnormalities in ISN nerve bundles of bchs-null larvae. A, B, WT (left) versus bchs (right) ISN motor nerves terminating onto muscles 1, 2, 3, and 9. Third-instar larval fillets were fixed and immunostained with anti-Futsch and FasII (green) and rhodamine–phalloidin for muscles (purple), to show all fibers of the ISN. B, In bchs¹/Df(2L)c17 ISN swelling of the nerve fiber is evident (arrow) in this projected confocal image. C–E, Wide-field fluorescence images of Futsch/FasII-labeled ISN nerves in WT (C), bchs¹/Df(2L)c17 (D), and bchs58/Df(2L)c17 (E), showing thickenings of the nerve bundle (arrows). F, G, WT (left) and bchs¹/Df(2L)c17 (right) third-instar larval fillets expressing eve>CD8-GFP only in aCC and RP2 motor neurons reveal apparent loss and thinning of individual fibers contributing to ISN nerve (arrow). Fillets were immunostained with anti-GFP and rhodamine–phalloidin (purple). H–K, Confocal images of eve>CD8-GFP expressing aCC and RP2 motor axons of third-instar larval fillets, detected with anti-GFP (green; H, J). Axonal varicosities are evident in the remaining aCC process of bchs¹/Df(2L)cl7 (J, arrow). The entire ISN nerve fiber is labeled with anti-Futsch and FasII (red) (I, K). Scale bar, 40 μm (in all images).

Figure 4.

Microtubule bundles show abnormalities in bchs motor axons. A, B, Third-instar WT (A) and bchs⁵⁸/Df(2L)c17 (B) third-instar larval fillets expressing eve>GFP-α-tubulin in aCC and RP2 motor neurons and immunostained with anti-GFP. Discontinuity of GFP expression is seen in the bchs axon fibers (arrow) and termini (asterisk), enlarged in C–F. G, H, WT (left) and bchs¹/Df(2L)cl7 (right) ISN nerves at the level of the second branchpoint, showing abnormal organization of microtubules labeled with MAbE7 against β-tubulin.

Figure 5.

Loss of motor neurons in bchs-deficient ISN. A–F, Third-instar WT (A) and bchs1/Df(2L)cl7 (B–F) larvae expressing eve>CD8-GFP (green) in RP2 (arrows) and aCC and immunostained with anti-GFP. All motor neurons were also labeled with anti-Futsch/anti-FasII (red). RP2 atrophy, as judged by loss of GFP and/or loss of Futsch/FasII, has progressed to differing extents, from none evident (B) to severe (E, F; arrows). In some cases, GFP levels were lowered, whereas Futsch/FasII levels remained apparently normal. To see fibers, images should be viewed at high resolution (see supplemental material, available at www.jneurosci.org). G, Graph shows percentages of RP2 loss, scored as complete absence of detectable GFP, as in E and F, in hemisegments of the genotypes shown. Error bars represent SD between three experiments. Asterisks point to morphological abnormalities in synaptic arbors. Scale bar, 40 μm (in all images). H–J, Loss of GFP in motor neuron cell bodies of bchs-deficient larval CNS is often accompanied by loss of Even-skipped. CNSs from eve>CD8-GFP (green) third-instar larvae stained with antibody to endogenous eve (red), which localizes to cell nuclei. Wild-type (H) third-instar CNS shows three eve-positive nuclei per hemisegment, two of which are GFP-positive aCC and RP2 (segments separated by drawn lines). I, bchs1/Df(2L)c17 CNS shows some missing eve nuclei (arrows point to segments in which eve nuclei are missing), in addition to loss of GFP. J, bchs58/Df(2L)c17 CNS shows similar, but more frequent loss of GFP and endogenous even-skipped.

Figure 6.

Enlarged details of motor neuron atrophy. A, Wild-type ISN terminal, showing aCC and RP2 motor neurons labeled by eve>CD8-GFP (green), and other motor neurons stained for Futsch/FasII (red). B–B″, Enlargement of A. A non-GFP-labeled, Futsch/FasII-positive small presumptive type II terminal (red II, arrow) branches posteriorly and terminates close to the RP2 terminal (green Is, arrow). A larger presumptive type Ib fiber (red Ib, arrow) does not express GFP, showing that it does not originate from RP2. C, A bchs¹/Df(2L)c17 terminal in which no motor neuron atrophy has occurred. D–D″, Enlargement of C, showing a non-GFP-labeled presumptive type Ib motor neuron (red Ib; D′, arrow) that is morphologically similar to that in B′. As in B″, RP2 (green Is; D″, arrow) is followed closely by a presumptive type II terminal (red II; D′, D″, arrow). To see fibers, images should be viewed at high resolution (see supplemental material, available at www.jneurosci.org). E–E″, A bchs¹/Df(2L)c17 terminal showing RP2 (downward arrow) in close apposition to a second motor neuron fiber, presumably the type II (upward arrow). F–G″, bchs¹/Df(2L)c17 termini in which diminution of both GFP and Futsch/FasII has occurred (arrows), compared with levels in other fibers. Scale bars: A, 40 μm; enlarged images, 20 μm.

Figure 7.

TUNEL reactivity is increased in bchs larval CNS. A–F, TUNEL (red) in CNS from bchs alleles bchs⁵⁸, bchs¹, WT, bchs^ex22, eve>EP2299, and WT treated with DNase as a positive control for TUNEL reactivity. bchs alleles are each in heterozygosity with Df(2L)c17. All brains express eve>CD8-GFP, labeled with anti-GFP (green). Images in bchs mutant and overexpressing brains (A, B, D, E) are single confocal sections, and wild-type images (C, F) are projected stacks. The red (TUNEL) channel in all images was autoleveled in Adobe Photoshop, and median filtered, for presentation of images. G, Quantitation of CD8-GFP-positive aCC and RP2 nuclei that are also TUNEL positive in bchs¹/Df(2L)c17 CNS. For WT, n = 311 nuclei; for bchs¹/Df, n = 128 nuclei were scored. For quantitation of TUNEL signal, single confocal images of wild-type brain were autoleveled for the red (TUNEL) channel, but not for the bchs/Df images, such that any TUNEL signal in wild-type nuclei would be enhanced. Despite this, many fewer wild-type nuclei scored TUNEL positive (3.5% average intensity >20 in ROI defined as a GFP-positive cell nucleus, vs 51.6% for bchs/Df). Each nucleus from a stack was counted only once, and all images were acquired at the same confocal settings and magnification. Error bars show SEM. H, I, Projections of wild-type and bchs¹/Df(2L)cl7 brains expressing eve>CD8-GFP and labeled with anti-GFP, showing reduced neuronal size in the bchs¹/Df brain, which was acquired at the same magnification. Scale bars: H, I, 40 μm.

Figure 8.

bchs affects axonal transport of endolysosomes. A, Kymographs of spinster-GFP-labeled endolysosomal vesicles undergoing anterograde and retrograde transport in RP2 or aCC motor axons in intact, anesthetized third-instar larvae. Representative kymographs are shown from time-lapse movies of various bchs loss-of-function and gain-of-function conditions. Time periods are shown (arrows) to the left of each frame, and distance encompassed by each movie is shown beneath each frame. Vesicles traveling in the anterograde direction are represented as slopes with negative gradient. Retrograde-directed vesicles are represented as slopes with positive gradient. Stationary vesicles are seen as vertical traces. bchs1, bchs¹/Df(2L)c17; bchs58, bchs⁵⁸/Df(2L)c17; EP, Eve>EP2299/+; EP/Df, Eve>EP2299/Df(2L)c17. Most vesicles in wild-type animals move in a relatively uninterrupted manner, whereas bchs¹ and bch⁵⁸ vesicles frequently pause and change direction. Transport in EP/+ axons is almost completely obliterated. EP/Df animals show partial rescue of the EP gain-of-function phenotype, with restoration of vesicle movement that approaches wild type. B, Direction of transport in kymographs shown. C, bchs reverses net transport direction. After thresholding the kymographs to select vesicles, numbers of anterograde, retrograde, and bidirectional moving vesicles were counted in 20 s segments (n ≥ 28) of the kymograph and represented as a percentage of the total number of moving vesicles. The majority of vesicles in wild-type animals were anterograde directed (black bars). This trend was reversed in the bchs⁵⁸ and bchs¹ mutants, in which retrograde movement dominated (light gray bars). WT: ant, 55%; ret, 31%; bi, 14%; bchs⁵⁸: ant, 32%; ret, 57%; bi, 11%; bchs¹: ant, 39%; ret, 50%; bi, 18%; EP/Df: ant, 41%; ret, 56%; bi, 3%; EP: ant, 38%; ret, 63%. *p < 0.05; **p < 0.014; ***p < 0.0024. D, bchs increases jitter of vesicle movement. The total distance traveled and the actual displacement of vesicles were measured and compared (distance/displacement > or <1). Frequent changes in direction (jitter) give values of distance/displacement >1. In bchs¹ animals, 56% of vesicles had values >1, whereas in WT only 16% were >1. WT (n = 96 vesicles, 6 larvae); bch⁵⁸ (49, 6); bchs¹ (41, 6), EP/Df (39, 4); EP (22, 5). E, bchs lowers overall endolysosomal mobility. Stationary and moving vesicles were counted in each 20 s interval of thresholded kymographs. Vesicle mobility is expressed as a percentage (number of moving vesicles/total vesicles × 100). The majority of vesicles in the wild-type animals move anterograde. However, the trend is reversed in bchs, in which retrograde movement dominates. WT (n = 35 segments counted); bch⁵⁸ (n = 29); bchs¹ (n = 28), EP/Df (n = 33); EP (n = 31). Values of p are as above, denoted by asterisks. Error bars show SEM. Genotypes are denoted as follows: bchs1, bchs¹/Df(2)c17; bch58, bch⁵⁸/Df(2)c17; EP/Df, EP(2)2299/Df(2)c17; EP, EP(2)2299/+. All genotypes are in a background of two chromosomal copies of the eve driver and one copy of the UAS-spinster-GFP reporter. Direction of vesicles are denoted as follows: ant, anterograde; ret, retrograde; bi, bidirectional.

Figure 9.

Effects of bchs on transport parameters. A, Moving spin-GFP-positive vesicles in axons were tracked manually from time-lapse movies and the distance moved was used to calculate the instantaneous velocity of a given vesicle from one frame to the next. Instantaneous velocities were binned into groups in increments of 0.05 μm/s and presented as a frequency distribution histogram. Velocities were distributed over a range; however, there is a clear shift in bchs1 toward higher velocities (rightward) and in eve>EP2299/+ (EP) toward lower velocities (leftward). B (and Table 1), Instantaneous velocities measured over the length of a run for a given vesicle were used to calculate the average velocity of each tracked vesicle. Vesicles moving in the anterograde direction were more affected than those moving retrograde. Anterograde vesicles from bch58 and bchs1 animals move significantly faster, whereas in eve>EP2299 animals, transport velocity in both directions was greatly reduced. Data are presented as means with SEM bars. *p < 0.05, one-way ANOVA test.