(C) PLOS One
This story was originally published by PLOS One and is unaltered.
 . . . . . . . . . .



Divergent signaling requirements of dSARM in injury-induced degeneration and developmental glial phagocytosis [1]

['Kelsey A. Herrmann', 'Department Of Neurosciences', 'Case Western Reserve University School Of Medicine', 'Cleveland', 'Ohio', 'United States Of America', 'Yizhou Liu', 'Arnau Llobet-Rosell', 'Department Of Fundamental Neurosciences', 'University Of Lausanne']

Date: 2022-08 

Elucidating signal transduction mechanisms of innate immune pathways is essential to defining how they elicit distinct cellular responses. Toll-like receptors (TLR) signal through their cytoplasmic TIR domains which bind other TIR domain-containing adaptors. dSARM/SARM1 is one such TIR domain adaptor best known for its role as the central axon degeneration trigger after injury. In degeneration, SARM1’s domains have been assigned unique functions: the ARM domain is auto-inhibitory, SAM-SAM domain interactions mediate multimerization, and the TIR domain has intrinsic NAD + hydrolase activity that precipitates axonal demise. Whether and how these distinct functions contribute to TLR signaling is unknown. Here we show divergent signaling requirements for dSARM in injury-induced axon degeneration and TLR-mediated developmental glial phagocytosis through analysis of new knock-in domain and point mutations. We demonstrate intragenic complementation between reciprocal pairs of domain mutants during development, providing evidence for separability of dSARM functional domains in TLR signaling. Surprisingly, dSARM’s NAD + hydrolase activity is strictly required for both degenerative and developmental signaling, demonstrating that TLR signal transduction requires dSARM’s enzymatic activity. In contrast, while SAM domain-mediated dSARM multimerization is important for axon degeneration, it is dispensable for TLR signaling. Finally, dSARM functions in a linear genetic pathway with the MAP3K Ask1 during development but not in degenerating axons. Thus, we propose that dSARM exists in distinct signaling states in developmental and pathological contexts.

Following injury, severed axons are actively destroyed by a protein called SARM1, or dSARM in Drosophila. It was recently shown that dSARM/SARM1 accomplishes this feat by degrading the key coenzyme NAD + , leading to energetic collapse in the axon. This was surprising since dSARM/SARM1 is a TIR-domain containing protein, and as such, was thought to act exclusively as an adaptor in Toll receptor-mediated pathways. We recently uncovered a developmental role for dSARM in glia where it promotes engulfment of dying neurons. To understand the relationship between this function for dSARM and its better-known function in axon degeneration, we set out to (1) compare the enzymatic requirement of dSARM in both settings, and (2) determine if signaling mechanisms are the same. To address these questions, we used CRISPR to generate new dSARM knock-in alleles. We find that the NAD + hydrolase activity of dSARM is absolutely required for both its axon degeneration function and its signaling function. Pointing to differences in dSARM-mediated pathways in the two contexts, we provide evidence that Ask1 kinase is required for dSARM signaling in glia but not for axon degeneration. Thus, we propose that the NAD + hydrolase activity of dSARM leads to Ask1 activation and subsequent signal transduction.

Funding: This work was supported by a Swiss National Science Foundation SNSF Assistant Professor award (176855), the International Foundation for Research in Paraplegia (P180), and SNSF Spark (190919) to LJN, and by R21NS110397, R01NS120689, and R01NS095895 to HTB. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

To date, functions of individual SARM1 domains have been assigned largely via in vitro assays and in vivo overexpression paradigms [ 12 – 15 , 17 – 22 ]. Given caveats associated with overexpression experiments, we interrogated SARM1 signaling requirements by mutating the endogenous locus. We used CRISPR/Cas9-mediated genome engineering to generate a dSARM knockout allele by replacing exons containing the ARM, SAM, and TIR domains with an attP recombination target. We then recombined in a series of domain mutants as well as a catalytically inactive point mutant (E1170A) and assessed the contributions of each domain to injury-induced degeneration and developmental phagocytosis. We find that the E1170A allele exhibits long-lived protection of axons following injury, demonstrating that dSARM’s NAD + hydrolase activity accounts for its pro-degenerative function in vivo. Unexpectedly, the TIR-only allele can drive spontaneous axon degeneration over the course of days, indicating that SAM-mediated multimerization is not essential for dSARM activity in the absence of ARM domain-mediated inhibition. We next analyzed these new dSARM alleles in glial TLR signaling and find that dSARM’s NAD + hydrolase activity is essential for this role. In contrast, the SAM domain is dispensable for signaling in glia. Finally, we explored signaling downstream of dSARM in glia and present evidence that dSARM functions through the MAP3K Ask1 in glia but not in degenerating axons. These findings argue that dSARM has distinct signaling modes in degenerative and non-degenerative signaling pathways.

SARM1 was first identified as an innate immune adaptor protein [ 24 , 25 ] and regulates neurodevelopment [ 16 , 26 – 30 ]. We recently demonstrated that dSARM is a component of a glial Toll receptor pathway required for clearance of neuronal corpses [ 29 ]. Loss of Toll-6 pathway components results in accumulation of apoptotic debris in the developing brain and early-onset neurodegeneration. The identification of this developmental function for dSARM raises important questions. To what extent are dSARM-mediated signaling mechanisms conserved between axon degeneration and glial phagocytic pathways? Specifically, is SAM domain-mediated multimerization and/or the NAD + hydrolase activity of dSARM necessary for TLR-dependent signaling? And are dSARM’s downstream signaling mechanisms conserved in development and degeneration?

The discovery that Wallerian degeneration is an active destructive process prompted forward genetic screens for loss-of-function (LOF) mutants in which axons are protected following injury. Drosophila SARM1 was identified in such a screen as its loss confers robust protection of distal axons following axotomy [ 11 ]. Mice lacking SARM1 exhibit preservation of severed axons for weeks following injury [ 11 , 12 ], demonstrating conservation of function. Underscoring the importance of NAD + , SARM1 drives axonal death via intrinsic NAD + hydrolase activity that is proposed to culminate in metabolic catastrophe [ 13 , 14 ]. SARM1 encodes a protein with an N-terminal ARM domain, two tandem SAM domains, and a C-terminal TIR domain. Biochemical and genetic studies indicate that the TIR domain contains NAD + hydrolase activity, the SAM domains are responsible for multimerization, and the ARM domain mediates auto-inhibition [ 12 – 16 ]. Recent structural studies provide a high-resolution view of SARM1 structure and demonstrate that it assembles into an octamer mediated by SAM domain oligomerization [ 17 – 19 ]. In its inactive conformation, the TIR domain is bound by the inhibitory ARM domain, while SARM1 activation leads to release of this auto-inhibition [ 18 – 22 ]. The TIR domains cleave NAD + once released by the ARM domain in response to an increase in the NMN/NAD + ratio [ 23 ].

Wallerian degeneration is a specific type of axon degeneration in which the axon distal to an axotomy degenerates [ 1 ]. A spontaneous mouse mutant, Wallerian Degeneration Slow (Wld S ), exhibits markedly delayed axon degeneration [ 2 , 3 ]. This phenotype argues that axon degeneration is an active process and not passive wasting of the injured nerve. The Wld S mutation is a tandem triplication of the NAD + synthetic enzyme Nicotinamide mononucleotide adenlyl transferase 1 (Nmnat1) and Ubiquitination factor e4b (Ube4b) [ 4 ]. While NAD + levels normally plummet following injury [ 5 ], NAD + depletion is blocked by Wld S [ 6 – 9 ], hinting at a regulatory role for NAD + in the decision to degenerate across evolution [ 10 ].

Brain homeostasis is maintained by cell-intrinsic and cell-extrinsic surveillance mechanisms. During normal development, a commonly cited estimate is that 50% of neurons die, and injury can precipitate the death of even more. The prevalence of neuron and neurite death during development highlights the importance of defining underlying molecular mechanisms as well as those active in phagocytic glia that engulf and dispose of neuronal corpses.

Results

Genome engineering of the dSARM locus We sought to compare functional requirements of dSARM domains in developmental and degenerative contexts in vivo and so undertook a CRISPR/Cas9-mediated genome engineering approach. Using CRISPR/Cas9, we precisely deleted the ARM, SAM, and TIR domain-encoding exons of dSARM and replaced them with an attP site to create a founder knock-out allele (dSARMKO, Fig 1A–1D) [31,32]. We then utilized phiC31-mediated DNA integration at the attP site to create a series of dSARM alleles (Fig 1E–1G) [33]. These alleles retain endogenous intron-exon structure and differ only in the presence of a 50 nucleotide attR site in the intron preceding exon 17 and a 34 nucleotide loxP site in the intron succeeding exon 21/22 (exon numbering from FlyBase; Fig 1G). We successfully generated the following four domain mutants: dSARMARM-TIR, dSARMARM-SAM, dSARMTIR, and dSARMSAM, which are each named for the domain(s) present in the allele (Fig 1H). To test the function of dSARM’s NAD+ hydrolase activity in signaling, we mutated a key glutamic acid in the active site to alanine, which is equivalent to dSARM E893A in isoform E and human SARM1 E642A (dSARME1170A; [14]; [34]. All of these new dSARM alleles are homozygous lethal. dSARMKO, dSARMARM-TIR, dSARMARM-SAM, dSARMSAM, and dSARME1170A animals die as wandering third-instar larvae (L3). The lethality of dSARME1170A animals indicates that the NAD+ hydrolase activity serves an essential developmental function. Interestingly, the lethal phase of dSARMTIR homozygotes is at the first-instar larval stage (L1), suggesting that this allele has gain-of-function (GOF) activity. We also generated a SARMRescue line by recombining back in wild-type sequences (Fig 1H). dSARMRescue homozygotes are viable and fertile, serving as a control for the overall strategy. Alleles were validated by a combination of PCR and sequencing. Notably, after three rounds of injections into roughly 900 embryos, we were unable to recover dSARMSAM-TIR transformants. This allele likely caused dominant lethality, as suggested by published work indicating that the ARM domain prevents unregulated activation of multimerized TIR domains [12,16]. To assay dSARM transcript abundance in mutant lines, we performed quantitative RT-PCR on first- or second-instar larvae, before the lethal phase of these animals. While dSARM transcripts are undetectable in the dSARMKO founder line, we find normal levels of dSARM transcripts in all knock-in alleles (Fig 1I), indicating that stable transcripts are produced. Moreover, as shown below, we demonstrate intragenic complementation between pairs of alleles containing reciprocal domains demonstrating that the domain mutants are expressed and functional. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 1. Genome engineering of the dSARM locus. (A) dSARM has 9 predicted isoforms. We designed a CRISPR/Cas9 strategy to knock out the ARM, SAM, and TIR domain encoding exons (boxed in green) in all isoforms. (B) A CRISPR/Cas9-induced homology-directed repair gene-targeting approach is used to delete the domain-encoding exons of dSARM. We identified a gRNA target sequence in the 3’ intron immediately following the exons in the green box in A. The homology-directed repair donor DNA plasmid contains 5’ and 3’ homologous arms flanking dSARM domains, a loxP-flanked DsRed selection marker, and an attP site of phiC31. The ARM (orange), SAM (magenta), and TIR (teal) domains are overlayed onto the dSARM locus. (C) In the “dSARM DsRed+ knock-out line”, the domain-encoding exons are replaced by the attP-loxP-DsRed-loxP cassette. (D) In the final dSARM founder knock-out line, the DsRed marker is removed by Cre recombinase, leaving attP and loxP sequences. (E) Genomic DNA is engineered to incorporate desired modifications (“dSARM mutant”) in the pGE-attB-w+ integration vector. The dSARM mutant is integrated into the founder line through phi-C31-mediated DNA integration via attP/attB recombination. (F) The resulting “dSARM w+ mutant integration allele” has the engineered mutant dSARM gene at its original genomic locus together with white+ and vector sequences. (G) w+ and extraneous vector sequences are removed by Cre recombinase to generate the “final engineered dSARM mutant allele” containing the engineered mutant flanked by attR and loxP sites in the adjacent introns. (H) A schematic showing the dSARM alleles that were generated. (I) qRT-PCR analysis of relative dSARM mRNA levels average ± SEM: wild type (Oregon R) (n = 28): 1.04±0.01; dSARMKO (n = 4): 0.02±0.01; dSARMRescue (n = 4): 2.45±0.64; dSARME1170A (n = 4): 1.66±0.47; dSARMARM-SAM (n = 4): 1.46±1.03; dSARMTIR (n = 4): 2.23±0.85; dSARMARM-TIR (n = 4): 0.99±0.18; dSARMSAM (n = 4): 1.03±0.21 Error bars are SEM. https://doi.org/10.1371/journal.pgen.1010257.g001

dSARMARM-SAM has dominant-negative activity following injury The octameric structure of SARM1 raises the possibility that mutants disrupting domain stoichiometry might display dominant-negative effects by inhibiting TIR domain multimerization [12]. Thus, we tested if reducing the number of TIR domains relative to ARM and SAM domains would slow axon degeneration. Normally, degeneration of distal ORN axons is efficient, with little debris remaining by 24 h post-injury (1 DPI; Fig 3A). Instead of counting axons, we quantified total axonal debris in these experiments to better capture axon fragmentation observed shortly after injury. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 3. dSARMARM-SAM has dominant-negative activity following injury. (A-D, F-I) Representative z-projections of OR22a ORNs of the indicated genotypes labeled with anti-GFP. The control side is on the left and the axotomized side is on the right. These animals were injured 7 DPE and analyzed at 1 day (A-D) and 2 days (F-I) post-injury. (E) Quantification of total fluorescent area of debris on injured side at 1 DPI: dSARMRescue (n = 37): 1.00, dSARMKO (n = 31): 1.16, dSARME1170A (n = 31): 1.27, and dSARMARM-SAM (n = 31): 2.07. (J) Quantification of total fluorescent area of debris on injured side at 2 DPI: dSARMRescue (n = 14): 1.00, dSARMKO (n = 12): 0.90, dSARME1170A (n = 13): 0.62, and dSARMARM-SAM (n = 11): 0.72. Error bars represent min and max data points. n.s. is not significantly different. ****, p < 0.0001. https://doi.org/10.1371/journal.pgen.1010257.g003 Using OR22aGal4 to drive UAS-mCD8::GFP in ORNs, we find that axons in whole animal dSARMKO heterozygotes degenerate as quickly as in controls, indicating that loss of one copy of dSARM does not delay axon degeneration (Fig 3B and 3E). We next tested whether dSARME1170A or dSARMARM-SAM heterozygous axons exhibit delayed degeneration since they alter either TIR domain number (dSARMARM-SAM) relative to ARM-SAM domains or TIR domain enzymatic activity (dSARME1170A). Interestingly, while dSARME1170A heterozygotes degenerate as rapidly as controls, axon degeneration in dSARMARM-SAM heterozygotes is incomplete at 1 DPI (Fig 3A, 3C, 3D and 3E). Axons do degenerate in all backgrounds by 2 DPI (Fig 3F–3J). We propose that dSARMARM-SAM has modest dominant-negative activity because each octamer contains more auto-inhibitory ARM domains than TIR domains, which delays the formation of TIR-TIR dimers. The most parsimonious explanation of the finding that dSARME1170A is not a dominant-negative allele is that loss of NADase activity in one TIR monomer does not interfere with the NADase activity of other TIR monomers.

dSARMTIR mutant ORN clones exhibit injury-independent axon degeneration While imaging 7 DPE animals, we noticed that both the uninjured and injured dSARMTIR clones were less bright than the other genotypes and required more laser power to acquire an equivalent image, suggesting the possibility that the TIR domain alone caused axon degeneration. This prompted us to quantify the intensity of mutant clones for all alleles over time in the absence of injury. First, we analyzed at 1 DPE (Fig 4A–4C) and found that dSARMTIR mutant axons were already approximately 40% less bright than controls. Next, we looked at 7 DPE and found that dSARMTIR mutant clones were approximately 55% less bright than controls (Fig 4D–4F). At 10 DPE the appearance of dSARMTIR axons continued to wane, with an intensity 70% less than that of control genotypes (Fig 4G–4I). None of the other dSARM alleles affected fluorescence intensity at any of these time points. When we compared axon intensity among the three time points for dSARMTIR mutant clones, we found a steady loss of axon integrity (Fig 4J), indicating that without ARM and SAM domains, the TIR domain drives axon loss over the course of 10 days. We speculate that without ARM domain-mediated inhibition, the TIR domains are free to associate with each other, cleave NAD+ and drive degeneration, but do so over a slower time scale then when they are tethered together by SAM domains. Together, these data argue that in ORN axons, SAM domain-mediated TIR multimerization drives high-level NADase activity to drive rapid axon degeneration on the order of hours following an acute injury. On the other hand, TIR monomers, or low frequency formation of TIR dimers, have low-level NADase activity capable of spontaneous axon degeneration over the course of 10 days. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 4. dSARMTIR mutant clones exhibit injury-independent axon degeneration. (A,B) Representative z-projections of OR22a ORNs of the indicated genotypes labeled with anti-GFP at 1DPE. (C) Normalized mean axon intensity at 1 DPE: wild type (FRT2A): 1.00, dSARM896: 0.90, dSARMRescue: 1.00, dSARMKO: 0.92, dSARME1170A: 1.09, dSARMARM-SAM: 1.00, dSARMTIR: 0.59, and dSARMARM-TIR: 0.88. (D,E) Representative z-projections of OR22a ORNs of the indicated genotypes labeled with anti-GFP at 7 DPE. (F) Normalized mean axon intensity at 7 DPE: wild type (FRT2A): 1.00, dSARM896: 0.88, dSARMRescue: 0.85, dSARMKO: 0.77, dSARME1170A: 0.92, dSARMARM-SAM: 1.03, dSARMTIR: 0.45, and dSARMARM-TIR: 0.86. (G,H) Representative z-projections of OR22a ORNs of the indicated genotypes labeled with anti-GFP at 10 DPE. (I) Normalized mean axon intensity at 10 DPE: wild type (FRT2A): 1.00, dSARM896: 0.98, dSARMRescue: 0.93, dSARMKO: 0.93, dSARME1170A: 0.94, dSARMARM-SAM: 0.99, dSARMTIR: 0.30, and dSARMARM-TIR: 0.88. (J) Normalized mean axon intensity over time in dSARMRescue compared to dSARMTIR. dSARMRescue: 1.00 (1DPE), 0.85 (7DPE), and 0.93 (10 DPE) and dSARMTIR: 0.59 (1 DPE), 0.45 (7DPE), and 0.30 (10 DPE). Error bars represent min and max data points. N = 16 for each genotype. n.s. is not significantly different. *, p < 0.05. ****, p < 0.0001. https://doi.org/10.1371/journal.pgen.1010257.g004 We were intrigued by the unexpected finding that dSARMTIR mutant ORN clones exhibit injury-induced degeneration (Fig 2H–2I). To look more carefully at the timing of injury-induced degeneration in these clones, we assessed axon degeneration 12 hours after injury. We performed the experiment at 1 DPE so that the mutant clones still appeared relatively healthy and quantified axons 12 h following axotomy in order to uncover small differences in timing. Surprisingly, we found no difference in degeneration rate in dSARMTIR mutant clones relative to dSARMRescue clones (S1A–S1C Fig), arguing that free TIR domains can support timely injury-induced axon degeneration in neurons that are already undergoing slow and steady spontaneous degeneration.

dSARM signaling in glia requires its NADase activity, but not its SAM domains We recently uncovered a requirement for a glial dSARM-mediated TLR pathway in clearing neuronal debris during development. In loss-of-function (LOF) mutants of Toll-6, FoxO or dSARM, levels of Dcp-1-positive apoptotic debris are increased in the L3 brain [29]. We demonstrated that this pathway promotes phagocytosis by activating transcription of the key engulfment receptor Draper (Drpr) in a specific population of neuronal cell body-associated glia called cortex glia [39]. The discovery of a function for dSARM as a TLR pathway component raised important questions: (1) is dSARM multimerization required for TLR signaling? (2) does dSARM act solely as a TIR adaptor in this pathway or is its NAD+ hydrolase activity required? And (3) to what extent is signaling downstream of dSARM conserved in development versus degeneration? We previously demonstrated that RNAi-mediated knockdown of dSARM in cortex glia results in an identical increase in Dcp-1 debris as observed in dSARM nulls, while pan-neuronal dSARM knockdown does not affect corpse clearance [29]. Thus, the Dcp-1 phenotype observed in dSARM alleles can be attributed solely to dSARM’s function in cortex glia. To investigate functional requirements of individual dSARM domains in this pathway, we quantified the amount of Dcp-1 debris in L3 brains in all new dSARM alleles. For all alleles except dSARMTIR (see below), we conducted this analysis at early L3 before the lethal phase of these animals. We developed an Imaris imaging pipeline to automate and standardize debris quantification, where to account for differences in brain size, Dcp-1 puncta count is normalized to individual brain lobe volume (see Materials and methods; Dcp-1 puncta visible as white spots in brain lobes in Fig 6). Using this method, we find that neuronal corpses are cleared normally in dSARMRescue animals as evidenced by normal levels of Dcp-1 punta in this background (Fig 6A, 6B and 6H). In contrast, we observe a roughly two-fold increase in apoptotic debris in dSARMKO homozygotes relative to dSARMRescue animals (Fig 6C and 6H). This phenotype is consistent with that observed in the original dSARM LOF alleles and also with RNAi-mediated dSARM knockdown in cortex glia [29]. The phenotypes of the rescue and knockout in apoptotic debris clearance establishes the utility of our dSARM allelic series in dissecting signaling requirements of dSARM domains in this glial pathway. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 6. dSARM signaling in glia requires its NADase activity, but not its SAM domains. (A-G, I-J, and L-N) Representative z-projections of brain lobes of the indicated genotypes labeled with anti-Dcp-1. (H) Quantification of the number of Dcp-1 puncta normalized to brain lobe volume: wild type (Oregon R) (n = 28): 9.82x10-5, dSARMRescue (n = 80): 1.64x10-4, dSARMKO (n = 30): 2.73x10-4, dSARMARM-SAM (n = 20): 2.69x10-4. dSARME1170A (n = 28): 2.91x10-4, dSARMARM-TIR (n = 38): 1.75x10-4, and dSARMSAM (n = 17): 2.35x10-4. (K) Quantification of the number of Dcp-1 puncta normalized to brain lobe volume: dSARMRescue (n = 26): 2.61x10-4, and dSARMTIR (n = 34): 5.22x10-4. (O) Quantification of the number of Dcp-1 puncta normalized to brain lobe volume: dSARMRescue (n = 18): 1.57x10-4, dSARMARM-SAM/TIR (n = 16): 1.92x10-4, and dSARMARM-TIR/SAM (n = 16): 1.32x10-4. n values can be found on each graph. n.s., not significant; *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001. https://doi.org/10.1371/journal.pgen.1010257.g006 We continued by interrogating the contributions of individual dSARM domains to developmental signaling. We find excessive neuronal debris in both dSARMARM-SAM and dSARMSAM homozygotes (Fig 6D, 6G and 6H), indicating a TIR domain requirement in TLR signaling. Moreover, dSARME1170A homozygotes display an equivalent increase in Dcp-1 debris (Fig 6E and 6H), indicating that the NAD+ hydrolase activity is essential for the signaling role of dSARM in this setting. We next tested whether SAM-mediated dSARM multimerization is required by quantifying Dcp-1 puncta in dSARMARM-TIR homozygotes. Strikingly, apoptotic debris remains at control levels in dSARMARM-TIR mutants (Fig 6F and 6H), indicating that the SAM domain is dispensable for dSARM’s signaling role. We wanted to test whether dSARMTIR homozygotes display a glial phagocytosis phenotype at the L3 stage, but these mutants do not live this stage of development. Thus, we quantified apoptotic debris at L1. dSARMTIR homozygotes display a roughly two-fold increase in neuronal debris relative to dSARMRescue animals at this stage (Fig 6I–6K). Given the likely GOF activity observed in dSARMTIR mutants, it is unclear whether the increased debris in these animals reflects dSARM function in cortex glia or increased neuronal death caused by a different mechanism. Regardless, these data imply that isolated TIR domains are insufficient to carry out dSARM’s function in cortex glia. To test if dSARMTIR behaves as a dominant allele in this assay, we tested if dSARMTIR heterozyogtes display increased levels of Dcp-1 debris at the L3 stage. We find normal levels of Dcp-1 debris in dSARMTIR heterozygotes and in the rest of our new dSARM alleles (S3 Fig), demonstrating that none of the alleles have dominant activity in this context. These experiments suggest two main conclusions. (1) The finding that dSARMARM-TIR behaves as a null in injury-induced axon degeneration yet supports developmental signaling indicates a differential requirement for the SAM domains these two contexts. (2) The enzymatic activity of dSARM is essential for signaling, thus extending the known roles of the NADase activity of dSARM to signal transduction. In pathological axon degeneration, the ARM, SAM, and TIR domains have all been assigned unique, separable functions. The extent to which these domains are distinct functional elements in TLR signal transduction has not been investigated. Intragenic complementation provides a classic genetic test of domain separability [40], and the generation of a series of complementary domain mutants of dSARM (Fig 1H) provides a unique opportunity to investigate this question. We generated dSARMARM-SAM/dSARMTIR and dSARMARM-TIR/dSARMSAM heteroallelic animals to test whether the domains can complement each other and restore wild-type dSARM function. We find that heteroallelic combinations of both pairs of reciprocal mutants are viable until the mid-pupal stage, while all homozygous mutants die as wandering third-instar larvae. The finding that ARM-SAM suppresses the early lethality observed in TIR-only homozygotes argues that ARM-SAM inhibits the GOF activity observed in this allele in trans. To look more specifically at the function of these reciprocal pairs of mutants in signaling, we quantified Dcp-1 apoptotic debris. We find that Dcp-1 counts are at wild-type levels in both heteroallelic combinations (Fig 6L–6O). This result is not surprising in the case of dSARMARM-TIR/dSARMSAM, since dSARMARM-TIR homozygotes do not display a debris clearance phenotype. However, the rescue of debris clearance in dSARMARM-SAM/dSARMTIR animals indicates that the ARM and TIR domains need not be covalently bound to restore wild-type dSARM function in a TLR pathway. Together, these findings indicate that the ARM, SAM, and TIR domains have separable functions during development and that the ARM domain can restrain the activity of the TIR domain in trans.

[END]
---
[1] Url: https://journals.plos.org/plosgenetics/article?id=10.1371/journal.pgen.1010257

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/