Drosophila NMNAT maintains neural integrity independent of its NAD synthesis activity

Abstract

Wallerian degeneration refers to a loss of the distal part of an axon after nerve injury. Wallerian degeneration slow (Wld(s)) mice overexpress a chimeric protein containing the NAD synthase NMNAT (nicotinamide mononucleotide adenylyltransferase 1) and exhibit a delay in axonal degeneration. Currently, conflicting evidence raises questions as to whether NMNAT is the protecting factor and whether its enzymatic activity is required for such a possible function. Importantly, the link between nmnat and axon degeneration is at present solely based on overexpression studies of enzymatically active protein. Here we use the visual system of Drosophila as a model system to address these issues. We have isolated the first nmnat mutations in a multicellular organism in a forward genetic screen for synapse malfunction in Drosophila. Loss of nmnat causes a rapid and severe neurodegeneration that can be attenuated by blocking neuronal activity. Furthermore, in vivo neuronal expression of mutated nmnat shows that enzymatically inactive NMNAT protein retains strong neuroprotective effects and rescues the degeneration phenotype caused by loss of nmnat. Our data indicate an NAD-independent requirement of NMNAT for maintaining neuronal integrity that can be exploited to protect neurons from neuronal activity-induced degeneration by overexpression of the protein.

Figure 1. Mutations in Complementation Group 3R4 Disrupt Synaptic Structures in Photoreceptor Terminals

(A–F) External morphology of the homozygous eyes of 3R4¹ (C) and 3R4² (E) are normal when compared to those of isogenized control (A). Red eye color marks heterozygous patches. (B), (D), and (F) ERG recordings of control and mutant eyes. Note the reduced depolarization and on/off response. Bar above trace in (B) indicates duration of light stimulus. (G–I) TEM micrographs of lamina cartridges containing control, 3R4¹, and 3R4² mutant terminals, respectively. Demarcating glia are colored blue and photoreceptor terminals green to accentuate the structures. Note the photoreceptor terminals are disorganized in mutants. The yellow boxes in (G) and (I) indicate the regions shown in (J) and (K), respectively. The red boxes in (G) and (H) indicate the regions shown in (L) and (M), respectively. Scale bar in (G) for (G–I) indicates 1 μm. (J) and (K) Individual terminals that are boxed in (G) and (I) (yellow boxes). nmnat mutant terminals have amorphic active zone structures (red arrows), aberrant capitate projections (blue arrows), aberrant mitochondria (yellow arrowheads), and abnormal membranes (yellow arrow), as well as an aberrant cytoskeleton (blue arrowheads), which are not observed in wild-type terminals. Scale bar indicates 200 nm. (L) and (M) Individual active zones that are boxed in (G) and (H) (red boxes). Compared to the clearly defined wild-type active zone structure (L), nmnat mutant active zones are amorphic and reduced in size. Both wild-type and mutant T-bars are surrounded by synaptic vesicles. Scale bar indicates 200 nm. (N) and (O) Quantification of synapse number and size. No significant difference was found in the average number of active zones per terminal between control (118 terminals counted), and 3R4¹ (93 terminals counted) or 3R4² (71 terminals counted). Synapse size was measured by the width of T-bar platform profile (L) (insert). The size of T-bars in either 3R4¹ (23 measured) or 3R4² (31 measured) is significantly reduced compared to the control (38 measured). An asterisk (*) indicates p < 0.05.

Figure 2. Complementation Group 3R4 Encodes nmnat, an Adenylyltransferase

(A) Fine mapping using the P element recombination method. The insertion sites of the five P elements used to map the gene are marked by red arrows. The recombination ratio for each P element is listed in centiMorgans (cM). The red box marks the region delineated by fine mapping. (B) The genomic region of CG13645. The point mutations uncovered by sequencing are marked by red arrowheads. The genomic rescue construct and the cDNA rescue construct are both marked in green. The blue arrow indicates the insertion site of P element EY4790 used to generate the excision lines Δ4790–1 and -2. The boundaries of the excised regions are marked by brackets in the blue lines. (C) Drosophila NMNAT protein is homologous to mouse and human NMNAT1, −2, and −3. Percentages of similarity and identity to each mouse and human protein are listed. (D) The organization of the NMN adenylyltransferase activity center. The key amino acids mutated to reduce enzymatic activity are marked by asterisks (*). The amino acid positions of the nonsense mutations in both alleles are marked by red arrowheads. (E) The Drosophila NMNAT protein has similar enzymatic activity as human NMNAT3. The activity is measured by the continuous coupling assay and listed in units per milligram of recombinant protein. The mutant proteins H30A, W98G, R224A, and WR (W98G/R224A double mutant) have 1.4%, 22%, 10.8%, and 0.9% of the activity of wild-type protein, respectively.

Figure 3. NMNAT Is Highly Enriched in the Nervous System and the Muscle Nuclei

(A–C) Third instar larval neuromuscular junction (NMJ) immunolabeled for HRP (red), with nc82 to mark the active zones (green), and for NMNAT (blue). In the NMJ, NMNAT appears as punctae that co-localize with nc82. NMNAT is also enriched in the muscle nucleus (marked by N in [A]). (D–F) Optic lobe (P+60%) immunolabeled for Synaptotagmin (syt) to mark synaptic vesicles (blue), with nc82 to mark active zones (green). and for NMNAT (red). NMNAT is highly enriched in the photoreceptor cell body in the eye and in the cell bodies of different types of neurons in the lamina cortex and medulla cortex, but is present at lower levels in lamina. la, lamina; lc, lamina cortex; mc, medulla cortex; me, medulla. (G–I) MARCM analysis of adult lamina. GFP marks the mutant patches. Laminae are immunolabeled with nc82 to mark photoreceptors (blue) and for NMNAT (red). NMNAT labeling appears as a punctate pattern decorating nc82 labeling, suggesting that NMNAT is present in clusters in photoreceptor terminals. (J–L) Adult brain immunolabeled with TOTO3 to mark nuclei (blue), with nc82 to mark active zones (green), and for NMNAT (red). NMNAT is expressed at higher levels in neuronal nuclei (marked by TOTO3) in the brain and lower levels in the neuropil (marked by nc82). AL, antenna lobe; MB, mushroom body. Scale bars indicate 5 μm.

Figure 4. Loss of nmnat Causes Severe and Progressive Age-Dependent Degeneration

(A–E) Retinal sections of control (iso), nmnat mutant eyes of different ages, and a trp^P365 mutant eye. Reduced rhabdomeres and vacuoles are seen in P+96%, and both phenotypes become more severe with age. The nmnat mutant retina has a more severe phenotype than trp^P365 in age-matched animals. (F–J) TEM micrographs of lamina cartridges. Demarcating glia are colored blue and photoreceptor terminals green to accentuate the structures. In the mutant lamina, the number of structurally intact photoreceptor terminals gradually reduces with age. The trp^P365 mutant lamina has a rather organized cartridge structure. The red boxes in (F–J) indicate the regions shown in (K–O), respectively. Scale bar in (F) for (F–J) indicates 1 μm. (K–O) Individual synapses boxed in (F–J). In mutant photoreceptors, active zone structures gradually disintegrate with age (arrows); however, the morphology of the T-bars in trp^P365 is well preserved. Scale bar in (K) for (K)–(O) indicates 200 nm. (P) Quantification of the number of terminals per cartridge of control (iso) or mutant laminae at different ages. The photoreceptor terminals are recognized by the presence of capitate projections. In mutant laminae, the number of structurally intact photoreceptor terminals decrease with age, and dark rearing can delay the decline. The number of cartridges quantified is indicated above each graph. (Q) ERG recordings of control (iso) and mutant photoreceptors at 1 d and 8 d of age. At 8 d, mutant photoreceptors have minimal responses to light.

Figure 5. nmnat Mutant Photoreceptors Develop Normally

MARCM analysis of pupal eye disc at 30 h after puparium formation ( P+30%) (A–D) and 50 h after puparium formation P+50% (E–H). GFP marks the mutant patch in (D) and (H). Anti-Actin antibody labels the rhabdomere structure in (B) and (F). Anti-Armadillo antibody labeling (Arm) marks the adherence junction in (C). Anti-NMNAT antibody shows labeling in wild-type cell bodies, but is dramatically reduced or absent in the mutant patch (G). In both developmental stages, there are no detectable structural differences between wild-type and mutant patches.

Figure 6. Light Enhances Neurodegeneration in nmnat Mutant Photoreceptors

(A–C) TEM micrographs of control (iso) or nmnat mutant ommatidia at 1 d or 10 d of age kept in 12-hr light/dark cycle (12hr D/12hr L). Note the dramatic reduction of rhabodmere size at 1 d of age (B). This phenotype becomes more severe by day 10 (C). Genotypes and ages are marked on the top of each column. (D–F) TEM micrographs of control or nmnat mutant ommatidia at 1 d or 10 d of age reared in constant darkness. Note the dramatic improvement at day 1 ([B] versus [E]) and day 10 ([C] versus [F]). (G–I) TEM micrographs of cartridges containing control and nmnat mutant terminals at 1 d or 10 d of age reared in constant darkness. At 1 d of age, the mutant photoreceptor terminals are well organized, compared to the mutant photoreceptors of the same-aged flies raised in regular light/dark cycle (Figure 1H and 1I). Quantification of the number of terminals per cartridge is shown in Figure 4P. Dark rearing does not block synaptic degeneration, as at 10 d of age, the number of structurally intact photoreceptor terminals is reduced (I). Demarcating glia are colored blue and photoreceptor terminals green to emphasize the structures. Scale bars in (A) for (A–F) and in (G) for (G–I) indicate 1 μm.

Figure 7. Loss of NMNAT-Induced Degeneration Can Be Attenuated by Blocking Phototransduction and Is Independent of Apoptosis

(A) and (B) Ommatidial morphology of 1-d-old norpA mutant (A) or norp; nmnat double mutant (B) retina. norpA mutant retina appears normal at this age. norpA; nmnat double mutant retina has greatly improved rhabdomere structure when compared to nmnat single mutant (C). (C) and (D) Ommatidial morphology of nmnat mutant retina (C) or mutant retina overexpressing P35 (D). There are no detectable differences between (C) and (D), suggesting that P35 overexpression does not rescue the degeneration induced by loss of nmnat. (E) Quantification of the number of rhabdomeres and vacuole size for each genotype. Blue columns are the number of rhabdomeres per ommatidium, and the red columns are the vacuole size per 10 μm² of ommatidia. Six animals per genotype and 400 μm² of ommatidia per animal were quantified. Double asterisks (**) indicate p < 0.005; and triple asterisks (***) indicate p < 0.0005 (Student t-test).

Figure 8. Enzymatically Inactive NMNAT Can Rescue the Neurodegeneration Phenotype Caused by Loss of nmnat

(A–F) ERG recordings of mutant photoreceptors overexpressing human NMNAT3 (A), or inactive Drosophila NMNAT forms H30A (C) or WR (E) in nmnat mutant photoreceptors. The genotypes are marked on top of each column. Note that the magnitudes of both depolarization and on/off transients are partially restored. (B), (D), and (F) Retinal structures are partially restored in each genotype. (G–I) TEM micrographs of lamina cartridges. Photoreceptor terminals are well organized in cartridges, similar to wild type. Demarcating glia are colored blue and photoreceptor terminals are colored in green to identify the structures. The red boxes in (G) indicate the regions shown in (J) and (K); the boxes in (H) indicate the regions shown in (L) and (M); and the box in (I) indicates the region shown in (N). Scale bar in (G) for (G–I) indicates 1 μm. (J–N) Individual terminals boxed in (G–I). All active zones have defined platform (arrowheads) and pedestal structures. Scale bar in (J) for (J–N) indicates 200 nm.

Figure 9. Overexpression of Enzymatically Active or Inactive NMNAT Can Rescue the Neurodegeneration Phenotype Caused by rdgA, trp^P365, and Constant Light Exposure

(A–D) Ommatidial morphology of rdgA mutant retinae (A) expressing wild-type NMNAT (B) or enzymatically inactive NMNAT (C) at 2 d of age. The quantification of the number of rhabdomeres per ommatidium (blue columns) and the size of vacuoles per surface area (red columns) are displayed (D). The number of rhabdomeres per ommatidium is significantly rescued by expression of wild-type or enzymatically inactive NMNAT. The vacuole size is significantly reduced with the expression of inactive NMNAT but not wild-type NMNAT. (E–H) Ommatidial morphology of trp^P365 mutant retinae (E) expressing wild-type NMNAT (F) or enzymatically inactive NMNAT (G) at 2 d of age. The quantification of the number of rhabdomeres per ommatidium (blue columns) and the size of vacuoles per surface area (red columns) are displayed in (H). The number of rhabdomeres per ommatidium is significantly rescued by expressing enzymatically active or inactive NMNAT, although the vacuole size remains unchanged. (I–M) Ommatidial morphology of 30-d-old wild-type flies either kept under ambient light in a 12-h light/dark cycle (I), or under constant intense light (2.2 kLux) (J–L). Overexpression of wild-type NMNAT (K) and the inactive protein NMNAT-WR (L) significantly protect ommatidial morphology compared to wild type (J). Quantification (M) shows an increased number of rhabdomeres per ommatidium (blue columns) and a reduction in vacuole size (red columns). Five animals per genotype and 400 μm² of ommatidia per animal were quantified. A single asterisk (*) indicates p < 0.05; double asterisks (**) indicate p < 0.005; and triple asterisks (***) indicate p < 0.0005 (Student t-test).