6-OHDA-induced dopaminergic neurodegeneration in Caenorhabditis elegans is promoted by the engulfment pathway and inhibited by the transthyretin-related protein TTR-33

Oxidative stress is linked to many pathological conditions including the loss of dopaminergic neurons in Parkinson’s disease. The vast majority of disease cases appear to be caused by a combination of genetic mutations and environmental factors. We screened for genes protecting Caenorhabditis elegans dopaminergic neurons from oxidative stress induced by the neurotoxin 6-hydroxydopamine (6-OHDA) and identified the transthyretin-related gene ttr-33. The only described C. elegans transthyretin-related protein to date, TTR-52, has been shown to mediate corpse engulfment as well as axon repair. We demonstrate that TTR-52 and TTR-33 have distinct roles. TTR-33 is likely produced in the posterior arcade cells in the head of C. elegans larvae and is predicted to be a secreted protein. TTR-33 protects C. elegans from oxidative stress induced by paraquat or H2O2 at an organismal level. The increased oxidative stress sensitivity of ttr-33 mutants is alleviated by mutations affecting the KGB-1 MAPK kinase pathway, whereas it is enhanced by mutation of the JNK-1 MAPK kinase. Finally, we provide genetic evidence that the C. elegans cell corpse engulfment pathway is required for the degeneration of dopaminergic neurons after exposure to 6-OHDA. In summary, we describe a new neuroprotective mechanism and demonstrate that TTR-33 normally functions to protect dopaminergic neurons from oxidative stress-induced degeneration, potentially by acting as a secreted sensor or scavenger of oxidative stress.

Introduction Oxidative stress, caused by an increased occurrence of reactive oxygen species or a decreased cellular defence against these chemicals, is a risk factor for many pathological conditions including Parkinson's disease (for review [1]). The aetiology of Parkinson's disease, which is the second most abundant neurodegenerative disorder, is largely unknown (for review [2]). Although approximately 10% of disease cases can be explained by single gene mutations, the majority are thought to be evoked by the interplay of genetic defects and environmental factors (for review [3,4]). One of the major hallmarks of Parkinson's disease is the degeneration of dopaminergic neurons in the substantia nigra of the midbrain, as well as the abnormal aggregation of intracellular α-synuclein protein known as 'Lewy bodies' in surviving dopaminergic neurons [4]. In both genetic and idiopathic cases of Parkinson's disease, oxidative damage appears to play an important role [4]. Several disease loci are associated with mitochondrial function, and an increased incidence of Parkinson's is associated with rural living, possibly linked to pesticide exposure [4].
The nematode C. elegans is a powerful model system to understand how dopaminergic neurons maintain their integrity when animals are subjected to oxidative stress. Components of dopaminergic neurotransmission are highly conserved between C. elegans and mammals, and dopaminergic neurons can be readily visualised in living animals (for review [5]). C. elegans hermaphrodites possess eight dopaminergic neurons [6]. These are non-essential for viability or locomotion, whereas they function in mechanosensation and are important for adapting movement to nutrient availability [5].
C. elegans dopaminergic neurons have been used to mimic dopaminergic neurodegeneration associated with exposure to known and suspected neurotoxins. One of these toxins is 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP), exposure to which was discovered to cause parkinsonism in a group of drug addicts [7]. In C. elegans, MPTP-induced neurodegeneration is independent of the CED-4 apoptosis protein and seems to be in part affected by necrosis [8]. Exposure to manganese also leads to neurodegeneration in C. elegans [9,10], as does treatment with paraquat and aflatoxin B 1 , agents that were recently implicated in dopaminergic neurodegeneration by causing genotoxic damage in mitochondria [11]. One of the most intriguing agents that induces dopaminergic neurodegeneration in C. elegans and other model organisms is 6-hydroxydopamine (6-OHDA), a neurotoxin related to dopamine (for review [12]). 6-OHDA is specifically taken up into dopaminergic neurons by the dopamine transporter and leads to oxidative stress by blocking complex I of the respiratory chain [12]. When C. elegans are exposed to 6-OHDA, dopaminergic neurons gradually die [13]. It remains underexplored which cell death pathways mediate dopaminergic neurodegeneration after 6-OHDA exposure, and which mechanisms limit their damage upon exposure to oxidative stress.
Key features of apoptotic cell death and cell corpse engulfment mechanisms have been discovered in C. elegans (for review [14]). The vast majority of C. elegans apoptotic cell deaths are mediated by a conserved pathway comprised of a Apaf-1-like protein CED-4 and the CED-3 caspase (for review [14]). Apoptotic corpses and necrotic cells are engulfed and eliminated by the corpse engulfment pathway [14]. In contrast to mammals, C. elegans does not possess specialised phagocytes, but dying cells are engulfed by surrounding cells (for review [15]). In addition, it has been shown that the engulfment pathway cooperates with the apoptotic pathway to ensure efficient apoptosis [16,17]. Furthermore, there are several C. elegans cell deaths during which engulfment is used for apoptosis-independent cell killing [18,19]. This type of cell death, recently termed 'phagoptosis', also appears to occur in mammalian neurons which are engulfed by surrounding glia cells (reviewed in [20]). When C. elegans dopaminergic neurons die in response to 6-OHDA treatment, their neuronal processes exhibit blebbing and are the first to degenerate [13]. Dying cell bodies exhibit condensed chromatin, a common feature of apoptosis; however, blockage of apoptosis does not prevent neuronal death [13]. Typical features of necrotic cell death, such as extensive swelling, do not occur, but dopaminergic neurons were found to be surrounded by glia cells, consistent with a role of phagocytosis in 6-OHDAinduced neurodegeneration [13].
Through an unbiased forward genetic screen, we have discovered that mutations in the C. elegans transthyretin-related gene ttr-33 exacerbate dopaminergic neurodegeneration after 6-OHDA treatment. ttr-33 mutants also exhibit increased susceptibility to oxidative stress at an organismal level. ttr-33 is one of the 59 C. elegans transthyretin-related genes, a family resulting from a nematode-specific expansion [21,22]. Human transthyretin (transporter of thyroxine (T 4 ) and retinol) is synthesised mainly in the liver and the choroid plexus of the brain, and binds and distributes retinol and thyroid hormones (for review [23]). Transthyretin can form protein aggregates and has been connected to amyloid diseases [23]. The only characterised C. elegans transthyretin-related protein, TTR-52, is a secreted protein that facilitates apoptotic corpse engulfment by promoting recognition of apoptotic cells by engulfing cells [24][25][26]. TTR-52 also mediates axonal repair during regeneration of severed neurons [27]. In contrast, we speculate that TTR-33 may act as a sensor or scavenger of reactive oxygen species. Interestingly, we find that mutations affecting the kgb-1 MAPK pathway reduce neurodegeneration in ttr-33 mutant animals. Finally, we provide genetic evidence that the C. elegans engulfment pathway contributes to the removal of 6-OHDA-treated dopaminergic neurons.

Mutations in the C. elegans transthyretin-related gene ttr-33 causes enhanced 6-OHDA-induced dopaminergic neurodegeneration
Mutations in neuroprotective genes are expected to cause increased neuronal loss after oxidative stress exposure. To discover genes that prevent dopaminergic neurodegeneration, we used a C. elegans strain carrying the vtIs1 transgene, in which the promoter of the dopamine transporter dat-1 drives GFP expression and thus allows selectively visualising the dopaminergic neurons ( Fig 1A). These animals (P0) were mutagenised with ethyl methanesulfonate, and L1 stage larvae of the F2 generation were exposed to the neurotoxin 6-hydroxydopamine (6-OHDA) for 1 hour at 10 mM, a treatment that does not induce dopaminergic neurodegeneration in wild-type animals. After 72 hours, the adult F2 population was screened for animals that exhibited excessive loss of dopaminergic head neurons ( Fig 1A). We isolated the gt1983 mutant, which exhibited highly increased 6-OHDA-induced neurodegeneration compared to wildtype animals ( Fig 1B). We found that the gt1983 allele is recessive as gt1983/wild-type(+) heterozygous mutants did not exhibit increased 6-OHDA-induced neurodegeneration ( Fig  1D). To identify the corresponding locus to the gt1983 mutation, we performed wholegenome sequencing and single-nucleotide polymorphism mapping [28,29]. We found that the gt1983 mutant carries a point mutation in ttr-33 (transthyretin-related family domain) that causes a glycine to glutamate conversion (G27E) in the TTR-33 protein (Fig 1C). Another ttr-33 point mutation, gk567379, which induces a leucine to phenylalanine substitution (L72F) (Fig 1C), also caused hypersensitivity to 6-OHDA, albeit to a weaker extent ( Fig 1E). Furthermore, gt1983/ttr-33(gk567379) trans-heterozygous animals still exhibited increased 6-OHDA-induced neurodegeneration ( Fig 1E). Finally, transgenic expression of wild-type ttr-33 under the control of a 1.5 kb endogenous regulatory region partially rescued the 6-OHDA-induced neurodegeneration defect of ttr-33 mutants (Fig 1F). Taken together, our results indicate that loss-of-function mutations in the transthyretin-related gene ttr-33 lead to increased 6-OHDA-induced dopaminergic neurodegeneration.

6-OHDA sensitivity of the ttr-33 mutant occurs in the first larval stages and is inhibited by blocking neuronal drug uptake
TTR-33, like the majority of the C. elegans transthyretin-related proteins [24], is predicted to be a secreted, non-cytoplasmic protein (S2A Fig). Apart from the small alpha-helical signal peptide at the N-terminus, the predicted TTR-33 protein structure is composed predominantly of beta sheets (Fig 2A; S2C Fig). According to this protein structure prediction, ttr-33(gt1983) and ttr-33(gk567379) lead to amino acid changes on two neighbouring beta-strands (Fig 2A).
To determine the developmental stages in which ttr-33 functions to protect C. elegans from 6-OHDA-induced neurodegeneration, we exposed a mixed population of ttr-33 mutants to this drug. The larval stages 1-4 (L1-L4) and adult animals were separated after treatment. We found that L1 and L2 ttr-33 mutant larvae were highly sensitive to 6-OHDA, whereas L3 larvae exhibited less sensitivity (Fig 2B). Mutant L4 larvae and adults did not show increased neurodegeneration ( Fig 2B). We conclude that loss of function of ttr-33 causes 6-OHDA-induced dopaminergic neurodegeneration in the first three larval stages of C. elegans.
Remaining dopaminergic head neurons 24, 48 and 72 hours after treatment with 10 mM 6-OHDA and 72 hours after control treatment with ascorbic acid only ('72h Ctr') for BY200 wild-type and gt1983 mutant animals. Animals possessing all neurons were scored as 'ADE + CEP' (white bar), those with partial loss of CEP but intact ADE neurons as 'ADE + partial CEP' (light grey bar), those with complete loss of CEP but intact ADE neurons as 'only ADE' (dark grey bar) and those with complete loss of dopaminergic head neurons as 'no ADE + CEP' (black bar). Error bars = S.E.M. of 3 biological replicates for all the other strains, each with 60-120 animals per strain. Total number of animals per condition n = 30 for the '72h Ctr' and n = 220-340 for all the other conditions ( ÃÃÃÃ p<0.0001; G-Test comparing BY200 wild-type and mutant data of the same time point). . At present, we speculate that the ttr-33 mutant defect is not be fully rescued by our ttr-33 wild-type plasmids because of a 3 amino acid linker sequence in the constructs. This linker sequence was introduced as part of the cloning strategy between the translational start site and the N-terminal secretion signal of TTR-33 and may compromise protein function. It is also possible that the construct do not contain all promoter elements. As extrachromosomal arrays lead to variable expression between tissues, it is possible that not all tested animals expressed TTR-33 in the arcade cells. Finally, overexpression of TTR-33 might interfere with its function. https://doi.org/10.1371/journal.pgen.1007125.g001 Oxidative stress protection by the C. elegans transthyretin-related TTR-33 ttr-33 mutant animals develop at a slightly slower rate than wild-type animals (S3A Fig), but do not exhibit increased growth delays after exposure to 6-OHDA ( Fig 2C). Thus, ttr-33 mutants exhibit highly increased dopaminergic neurodegeneration at a concentration of 6-OHDA that does not affect their overall development. We also examined whether the ttr-33 mutation affects dopaminergic signalling, yet we found that ttr-33(gt1983) mutants behave like wild-type animals in assays testing dopamine-mediated behaviours (S3B and S3C Fig). Finally, ttr-33 mutants do not exhibit loss of dopaminergic neuron when the uptake of 6-OHDA into these cells is prevented by deletion of the dopamine transporter dat-1 (Fig 2D).

TTR-33 acts independently of TTR-52
We next asked whether TTR-33 and the previously characterised TTR-52 [24][25][26] have analogous roles in neuroprotection and axon repair. To test genetic interactions, we also performed experiments with lower concentrations of 6-OHDA to be able to detect possible enhanced neurodegeneration due to synergic effects (S4A Fig). We found that a ttr-52 lossof-function mutation did not cause increased 6-OHDA-induced neurodegeneration and did not influence the sensitivity of ttr-33 mutants or wild-type animals (Fig 3A and 3B; S4B and S4C Fig). Thus, TTR-52 does not appear to play a role in 6-OHDA-induced dopaminergic cell loss. In addition, we performed laser axotomy on the PLM (posterior lateral microtubule cells) mechanosensory neurons of wild-type and ttr-33 mutant animals, using animals carrying the zdIs5 transgene that express GFP under the control of the mechanosensory neuron-specific mec-4 promoter. We found that PLM axons of ttr-33 mutants did not show overt defects in the rate of axonal regrowth, reconnection or fusion after injury ( Fig 3C and 3D), unlike ttr-52 mutants which have been shown to present a defect in axonal fusion [27]. In summary, TTR-52 and TTR-33 function appear to have specific and independent functions.

The cell corpse engulfment pathway is required for 6-OHDA-induced dopaminergic neuron death
Given the reported role of TTR-52 in corpse engulfment, we asked if the cell corpse engulfment machinery influences 6-OHDA-induced neurodegeneration. The phagocytosis pathway consists of two partially redundant branches (for a simplified model see Fig 4A which was adapted from [27]). One branch involves the TTR-52 bridging molecule, the CED-1 receptor and the CED-6 adaptor protein. The second pathway involves the CED-2 SH2 and SH3 domain-containing adaptor protein, and the activation of the CED-10/RAC protein via the bipartite CED-5 and CED-12 guanine nucleotide exchange factor. CED-10 acts as the end point of both pathways ( Fig 4A). We found that mutation in ced-6, in contrast to mutation in ced-2, led to a measurable suppression of dopaminergic neurodegeneration in the ttr-33 mutant background (Fig 4B; S5A Fig). Furthermore, ced-2;ced-6;ttr-33 triple mutants and ced-10;ttr-33 double mutants in which both pathways are disrupted, showed an increased suppression of dopaminergic neuron death compared with ced-6;ttr-33 alone ( Fig 4B). Finally, mutations affecting the cell engulfment pathway also led to reduced dopaminergic neuron loss in a wild-type background ( Fig 4C; S5B Fig) with ced-10 single mutants and ced-6;ced-2 double mutants displaying even less neurodegeneration than ced-2 and ced-6 single mutants ( Fig 4C). Thus, both branches of the cell corpse engulfment pathway are required for 6-OHDA-induced dopaminergic neuron loss. We note that TTR-52, which has a role in engulfing apoptotic cells in the CED-5 pathway, had no discernible role in 6-OHDA-induced dopaminergic neurodegeneration.   [27]). The recognition of the conserved 'eat-me' signal phosphatidylserine (PS) on the surface of apoptotic and necrotic cells triggers two partially redundant pathways in the engulfing cell [73]. The genes tested in this study are highlighted in yellow and the part of the pathway that was shown to be required for axon regeneration [27] is boxed in blue. psr = phosphatidylserine receptor family, ced = cell death abnormality, nrf = nose resistant to fluoxetine Oxidative stress protection by the C. elegans transthyretin-related TTR-33

ttr-33 is likely expressed in pharynx, hypodermis and in posterior arcade cells
TTR-33 is predicted to be a secreted protein and a transcriptional reporter is thus expected to highlight the location of protein production. We examined three independent transgenic strains carrying transcriptional Pttr-33::gfp reporter, and found that six cells residing in the head of L1 larvae were strongly labelled (Fig 5A and 5B). Two of the labelled cells were located next to the first pharyngeal bulb and four were situated halfway between the mouth and the first pharyngeal bulb. All six cells form protrusions towards the mouth of the larva. The cells were labelled as early as in the embryonic pretzel stage (Fig 5C; S6 Fig, S1 Movie). Comparison with the cell atlas of C. elegans [30] and a GFP reporter under the control of the grl-27 arcade cell-specific promoter [31] revealed that these cells represent most likely the posterior arcade cells, which are located between external epidermis and anterior pharynx [32]. In support of this notion, we found that expression of wild-type TTR-33 under the arcade cell-specific promoter of bath-15 (C08C3.2) led to a partial rescue of the defect of ttr-33 mutant animals ( Fig  1F). In addition, in L1 stage larvae, ttr-33 expression was observed in the pharynx and in several unidentified cells immediately posterior to the second pharyngeal bulb. In older animals, the Pttr-33::GFP transcriptional reporter highlights head cells next to and anterior of the first pharyngeal bulb, as well as the pharynx and the hypodermis (S6 Fig, S2-S5 Movies). Given that expression of wild-type TTR-33 under the control of its endogenous regulatory region as well as the Pbath-15 promoter was not sufficient to fully rescue the enhanced neurodegeneration in ttr-33 mutants, it remains possible that the endogenous gene is expressed in further tissue not visible in these transgenic lines. Oxidative stress protection by the C. elegans transthyretin-related TTR-33

ttr-33 mutants are sensitive to oxidative stress
To determine if ttr-33 mutants are also sensitive to oxidative stress at an organismal level, we exposed the L1 stage larvae to different concentrations of paraquat and hydrogen peroxide (H 2 O 2 ) for 1 hour. We found that after treatment with paraquat ( Fig 6A) and H 2 O 2 (Fig 6B), Oxidative stress protection by the C. elegans transthyretin-related TTR-33 the development of ttr-33 mutants is compromised compared with wild-type animals. Since increased oxidative stress has been connected to premature aging (for review [1,33]), we asked whether the lifespan of ttr-33 mutants was altered and found that these animals are indeed short-lived ( Fig 6C, S7 Fig). We next investigated if ttr-33 is transcriptionally regulated in response to oxidative stress using qRT-PCR and found that this does not appear to be the case (S8A and S8I Fig; S9A and S9I Fig). Furthermore, to determine if oxidative stress signalling is altered in ttr-33 mutants as compared to wild-type animals, we measured expression of the oxidative stress reporters gst-1, gst-4 and gcs-1. We found that the ttr-33(gt1983) mutant does not show overt differences in gst-1, gst-4 and gcn-1 mRNA expression after treatment with 6-OHDA and paraquat or under control conditions (S8 Fig and S9 Fig). In summary, TTR-33 appears to provide a constitutive organism-wide oxidative stress protection.

JNK stress kinases have a role in 6-OHDA-induced neurodegeneration in ttr-33 mutants
Various MAP kinase pathways are involved in stress signalling (for review [34]), and we therefore tested their role in mediating the response to 6-OHDA. We found that mutations in the p38 kinases PMK-1 (p38 MAP kinase family) and PMK-3 alone did not affect the extent of neurodegeneration (Fig 7, S10 Fig). However, a possible redundant function of these two proteins cannot be excluded. In contrast, we found that a jnk-1 (Jun N-terminal kinase) mutation led to increased dopaminergic neuron loss in the ttr-33 mutant (Fig 7A; S10A Fig), but not in wild-type animals ( Fig 7B). Interestingly, mutations in the JNK kinase KGB-1 (kinase, GLHbinding) strongly alleviated 6-OHDA-induced dopaminergic neurodegeneration in ttr-33 (gt1983) mutants (Fig 8A; S11A Fig) but not in wild-type animals (Fig 8B, S11B Fig). Mutation of MEK-1, the mitogen-activated protein kinase kinase (MAPKK) acting upstream of KGB-1 [34], equally reduced 6-OHDA-induced dopaminergic neuron loss in the ttr-33 mutant ( Fig  8A; S11A Fig), but not in wild-type animals (Fig 8B, S11B Fig). Furthermore, we found that mutations in kgb-1 and mek-1 decreased paraquat-induced developmental delays in ttr-33 mutants (Fig 8C). However, kgb-1 and mek-1 mutations alone did not induce paraquat sensitivity ( Fig 8C). Although mek-1 mutation protects from oxidative stress sensitivity in the ttr-33 mutants, mek-1;ttr-33 double mutants develop slowly and show reduced proliferation (S11C and S11D Fig). In summary, the KGB-1 MAP kinase pathway appears to mediate the hypersensitivity to oxidative stress associated with ttr-33 mutations.

Discussion
We report that the C. elegans transthyretin-related protein TTR-33 protects against 6-OHDAinduced dopaminergic neurodegeneration as well as against oxidative stress at an organismal level. The KGB-1 MAP kinase pathway mediates increased sensitivity to oxidative stress in the ttr-33 mutant, whereas the JNK-1 MAP kinase pathway provides additional protection. In addition, we find that the corpse engulfment pathway is required for 6-OHDA-mediated dopaminergic neurodegeneration.
We speculate that TTR-33 is part of an endocrine defence system against oxidative stress. Transthyretin-related proteins form a nematode-specific expanded protein family that comprises 59 members, 54 of which are predicted to be secreted [24]. Some transthyretin-related genes are induced in response to environmental challenges such as oxidative stress or pathogen exposure: Oxidative stress induced by tert-butyl hydrogen peroxide leads to upregulation of several transthyretin-related genes [35], and treatment with the mitochondrial complex I inhibitor rotenone leads to the upregulation of 42 of the 59 transthyretin-related genes [36]. Furthermore, bacterial infection was reported to cause upregulation of ttr-33 [37] and other transthyretin-related proteins [38]. However, we found that TTR-33 does not appear to be induced in response to oxidative stress and thus might provide constitutive protection. TTR- 33 possesses 4 cysteine residues that are conserved within the C. elegans transthyretin-related protein family (marked in yellow in S2B Fig and S12 Fig). These cysteines could serve as reactive oxygen sensors or scavengers, analogous to those in the antioxidant glutathione. TTR-33 is likely produced in the six posterior arcade cells, which form a syncytium and are located between outer epidermis and the pharynx [32]. At present there is no direct evidence, but we speculate that TTR-33 could be secreted from the arcade cells into the pharynx. This hypothesis is consistent with the partial rescue we observed upon expressing TTR-33 from an arcade cell-specific promoter. In the pharynx, TTR-33 could provide organismal protection against ingested oxidative stress-producing compounds. It will be interesting to experimentally investigate the mechanistic details of TTR-33 function in the future. ttr-33 gene expression correlates with genes that are associated with the gene ontology terms 'defence response to fungus', 'negative regulation of metalloenzyme activity' and 'neuropeptide signalling pathway' (S1 Table). ttr-33 mutants die prematurely and are hypersensitive to oxidative stress at an organismal level, supporting the idea that these animals might exhibit a compromised organismal defence system (reviewed in [39]). Alternatively, mutation of ttr-33 could generally lower the threshold of oxidative damage both for dopaminergic neurons as well as at an organismal level.
We hypothesise that the KGB-1 and the JNK-1 MAPK pathways influence 6-OHDA-mediated neurodegeneration in the ttr-33 mutant based on their role in organismal stress responses. KGB-1 is ubiquitously expressed [40,41] and was reported to transmit a neuroendocrine signal that mediates a food aversion behaviour upon disruption of core cellular processes such as translation, respiration and protein turnover [42]. It is conceivable that KGB-1 acts upstream of TTR-33. However, the kgb-1 mutant on its own does not show enhanced 6-OHDA-induced neuronal loss. It is thus possible that the KGB-1 pathway senses the pathology conferred by the combination of exposure to oxidative stress and TTR-33 deficiency analogous to its described roles in sensing cellular defects [42]. Such sensing mechanisms might promote the degeneration of damaged neurons. In contrast, the JNK orthologues JNK-1 protects against 6-OHDAinduced neurodegeneration in the ttr-33 mutant. This phenotype is consistent with the reported role of JNK-1 in protecting against oxidative stress [43,44].
The C. elegans corpse engulfment pathway could promote 6-OHDA-induced dopaminergic neurodegeneration by driving the phagocytosis of dopaminergic neurons. Morphological evidence for 6-OHDA-induced engulfment of dopaminergic neurons were reported [13,45]. However, it is not known exactly how dopaminergic neurons die. The engulfment pathway typically acts by eliminating apoptotic cell corpses, but has also been shown to promote cell death; mutations in the engulfment pathway enhance C. elegans apoptosis defects [15,17] and a small number of cell deaths are mediated by cell killing conferred by the engulfment machinery [18,19,46]. Consistent with our findings, reduced axonal degeneration in neurons whose axons have been severed has been observed in mutants with a defective engulfment machinery [47]. It is still poorly understood how dopaminergic neurons die in Parkinson's disease patients. Active phagocytic cells in the brain are observed in almost all neurological diseases including Parkinson's disease (for review [48,49]), as well as in 6-OHDA models of Parkinson's disease (for review [50]). The C. elegans 6-OHDA model, in which the conserved engulfment pathway has a role in cell removal, might recapitulate important features of this disease.

Methods
We note that experiments shown in this paper were done in parallel to experiments shown in the accompanying paper 'Mutations in Caenorhabditis elegans neuroligin-like glit-1, the apoptosis pathway and the calcium chaperone crt-1 increase dopaminergic neurodegeneration after 6-OHDA treatment' and thus share the wild-type control.

C. elegans strains and maintenance
Strains were grown at 20˚C unless indicated otherwise. Additional information about alleles in this study is provided by WormBase (www.wormbase.org) [51]. Strains from the Caenorhabditis Genetics Center (CGC) or the National BioResource Project (NBRP) were outcrossed to the BY200 strain to eliminate unlinked mutations and to introduce a green fluorescent protein (GFP) marker for dopaminergic neurons. The strain carrying the second ttr-33 allele gk567379 (generated as part of the Million Mutation project [52]) was identified with WormBase [51] and ordered from CGC. Both The isolated gt1983 mutant and the ttr-33(gk567379) mutant were outcrossed a minimum of six times to the BY200 strain.

Mutagenesis and mapping
L4 and young adult staged BY200 animals were mutagenised with 25 mM ethyl methanesulfonate in M9 buffer for 4 hours at 20˚C, washed and grown at 15˚C overnight on NGM plates seeded with OP50 E. coli. The animals were bleached once they started to lay eggs and the F1 progeny left to hatch overnight in M9. This synchronised F1 population was left to develop on seeded plates until reaching early adult stage. The F1 adults were then washed off in M9 to lay a synchronised population of F2 L1 larvae, which were then treated with 10 mM 6-OHDA and screened for dopaminergic neuron loss after 72 hours. Candidates were picked on separate plates, backcrossed at least three times and re-tested for their hypersensitivity to 6-OHDA. The genomic DNA was sent for whole-genome sequencing and SNP mapping was performed as previously described [28,29]. SNP mapping combined with whole-genome sequencing revealed that the location of the phenotype-causing mutation could be narrowed down to an interval between 4 and 9 Mb on chromosome V. This region, besides ttr-33, contained two other exon-altering mutations that were excluded by testing a deletion allele (fbxa-223) and by fosmid injection (F43H9.4). ttr-33 identity was further confirmed by genetic rescue and by complementation testing using the second allele ttr-33(gk567379) which was obtained from the CGC.

C. elegans assays
All assays were performed in a blinded manner.
For 6-OHDA assays, 1-10 adult animals were incubated to lay eggs in 70 μL M9 without food on a shaker at 20˚C, 500 rpm for 24-30 h. 10 μL 200 mM ascorbic acid and 10 μL of the respective 6-OHDA 5x stock concentration were freshly prepared in H 2 O and added to 30 μL of L1-stage larvae in M9. 6-OHDA concentrations were chosen such that differential sensitives of various strains could be efficiently analysed. After 1 hour incubation at 20˚C and shaking at 500 rpm, 150 μL M9 buffer was added to oxidise and inactivate the 6-OHDA. The animals were pipetted to one half of an NGM plate containing a stripe of OP50 bacteria on the opposite half of the plate and adult animals and eggs were picked off to prevent growth of those that were not treated at L1 stage. Plates were incubated at 20˚C before examining the 6-OHDAinduced degeneration using a Leica fluorescent dissecting microscope. Unless otherwise indicated, head dopaminergic neurons (ADE and CEP neurons) were scored 72 hours after treatment in technical duplicates. Posterior deirid (PDE) dopaminergic neurons, which are formed later in L2 stage larvae, did not appear to be affected by 6-OHDA treatment, as previously observed [13]. For treatment of different developmental stages, the respective stages were picked after 6-OHDA treatment from a mixed population of animals.
For paraquat and hydrogen peroxide (H 2 O 2 ) assays, 10 μL of the 5x paraquat or H 2 O 2 stock concentration was freshly prepared and added to 40 μL of filtered L1-stage larvae in M9. After 1 hour incubation in the shaker, 150 μL M9 buffer was added and the L1 larvae were counted after pipetting approximately 100 of them to a seeded NGM plate. After 24 hours, animals that reached L3 larval stage were examined and scored.
For laser axotomy and microscopy, animals were immobilised in 0.05% tetramisole hydrochloride on 3.5% agar pads. Axotomies were performed as previously described [54], using a MicroPoint Laser System Basic Unit attached to a Zeiss Axio Imager A1 (Objective EC Plan-Neofluar 100X/1.30 Oil M27). Axons were severed in L4 stage animals 50 μm anterior to the PLM cell body, visualised on a Zeiss Axio Imager Z1 microscope equipped with a Photometrics camera (Cool Snap HQ2; Tucson, AZ) and analysed with MetaMorph software (Molecular Devices, Sunnyvale, CA) or Image J. Animals were analysed 24 hours post-axotomy for regrowth, reconnection and fusion. Regrowth was recorded as the length of regenerative growth from the proximal side of the cut site: the proximal axon was traced from the start of regrowth to the tip of the longest regenerative branch using Metamorph software; neurons that underwent axonal fusion were excluded from these quantifications. Axons were deemed to be reconnected when the proximal and distal axons appeared connected and within the same focal plane when observed with a 63X objective. Axonal fusion was defined as proximaldistal reconnection that prevented the onset of degeneration in the distal axon. For clarification of successful fusion, animals were re-analysed at 48 hours and/or 72 hours post-axotomy as necessary.
For lifespan assays, 50 L4 stage animals were picked for each strain and moved to new plates every day during the first 6 days and again at later days to avoid mixing with the progeny. Animals that did not move after being touched with a platinum pick were scored as dead. Dead animals were counted and removed every day. 'Bag of worms' and dry and burst animals were not included in the lifespan analysis. The data was analysed using OASIS (Online application for Survival Analysis, http://sbi.postech.ac.kr/oasis/) [55].
For egg-laying and development assays, over 20 L4 stage animals per strain were picked separately 24 hours before the assay. On the day of the assay, two plates per strain were seeded with concentrated OP50 E. coli and ten adults each deposited onto the plates. The adults were left to lay eggs for 1.5 hours, removed from the plate and the eggs counted. Developmental stages were determined at 48 and 61 hours, respectively, after egg-laying.
Dopamine paralysis assays were carried out as previously described [56]. Well-fed adults 24 hours post the L4 stage were incubated for 20 min on a plate containing different concentrations of dopamine. Animals were scored as paralysed when they failed to complete a full sinusoidal body movement during a 5 second observation period. Two plates containing 25 animals each were assessed for each condition.
Basal slowing assays were performed as previously described [57]. NGM plates were prepared and half of them seeded with HB101 E. coli before incubating them at 37˚C overnight. Also, mid-L4 larvae were picked separately on seeded plates. The next day, these staged, wellfed young adult animals were placed in 40 μL of M9 for 2 minutes to clean them from bacteria, and then picked to the centre of a plate. 6 animals each were put on a plate with and without bacteria. After 2 minutes adaptation time, the locomotion rate of each animal was quantified by counting the number of body bends completed in five consecutive 20 s intervals. The experiment was performed twice on different days.

qRT-PCR
L1 stage animals were filtered using an 11 μm filter from two 9 cm plates containing mixed stage animals. Animals were washed twice and the population was split in half for control treatment and drug treatment. For the 6-OHDA assays, ascorbic acid and H 2 O was added to the control wells and 10 mM 6-OHDA to the treated animals. For paraquat assays, H 2 O was added to the control wells and 25 mM paraquat to the treated animals. After 1 hour of treatment, RNA extraction was performed based on [58] using Trizol (Thermo Fisher Scientific). The following changes were introduced. The chloroform extraction was repeated once more and 0.

Statistical analysis
The two-tailed t-test was performed with the Analysis ToolPak Add-In in Microsoft Excel 2010. To choose between a t-test assuming equal or unequal variances, an F-test was conducted with the Analysis ToolPak Add-In. A G-Test was performed with the R Studio 1.0.44 Desc-Tools descriptive statistics tools package to compare 6-OHDA treatment data. p-values were calculated for each biological replicate and for the pooled biological replicate data, and the least significant of these p-values is indicated in the graphs. Data is only marked as statistically significant if all replicates and the pooled replicate data were found to be significant. Statistical analysis for axonal regeneration was performed using GraphPad Prism and Microsoft Excel. Two-way comparison was performed using the t-test.

DNA cloning and generation of transgenic lines
For generation of transgenic constructs, genomic DNA was amplified using primers containing additional 8 base pair restriction sites (marked in italics) for insertion into the pCFJ151 and the pCFJ212 vector for insertions on chromosome II and IV, respectively [70]. In the following primer sequences, restriction sites are highlighted in italics, translation start and translation stop codons in bold. Primers for the transcriptional reporter construct AscI-Pttr-33 -NotI-GFP-FseI-3'ttr-33 -PacI: AscI_Pttr-33 -GTACTTggcgcgccCAAAGAAACTTGCGTG TTTC, Pttr-33_NotI-AAGGAAAAAAgcggccgcaCATTATTTTTGTCTGAAAATACAAAC AC, NotI_GFP-GCTAgcggccgcAGTAAAGGAGAAGAACTTTTCACTGG, GFP_FseI-TT ATTAggccggccCCTTGTATGGCCGGCTAG, FseI_3'ttr-33-ACAAGGggccggccTAATAAT TATACTTAAAAGTATTTTCAACTG, 3'ttr-33_PacI-GCTAttaattaaGCATGCATCTTCCC ATAAAAG. Additional primers for the translational construct AscI-Pttr-33 -NotI-TTR-33 -FseI-3'ttr-33 -PacI: NotI_TTR-33 -AAGGAAAAAAgcggccgcTCACGACTCGCGTGCATT, TTR-33_FseI-AGAAAAggccggcctTCTGTGAAGACAGTCACGTGACTC. The Pttr-33 promoter sequence in these constructs ranges from -1500 to the translational start site. The predicted promoter sequence of ttr-33 ranges from -499 to 100 bp with respect to the translational start site [71]. Additional primers for the translational construct AscI-Pbath-15 -NotI-TTR-33 -FseI-3'bath-15 -PacI: AscI_Pbath-15: GTACTTggcgcgccTTTTCACTTTTTTTGGTCAGA ATC, Pbath-15_NotI: AAGGAAAAAAgcggccgccCATTTTGAAGAGTTTGCTG, FseI_3'bath-15: CTCATAggccggccTAATCATTTTTTTCCTGTTTTAGC, 3'bath-15_PacI: GAGACCttaat-taaCTGGAGTTCACCACTACAAC. For single-copy integration of transgene according to [70], 5 ng/μL of the constructs were injected into unc-119 animals, together with one or two fluorescent co-injection marker(s) (2.5 ng/μL Pmyo-2::mCherry to label the pharynx and 5 ng/ μL Pmyo-3::mCherry to label the bodywall), and the Mos transposase under the control of the eft-3 promoter as described [72]. Wild-type moving animals that did not show expression of the fluorescent co-injection markers were singled out and tested for transgene integration by PCR. Extrachromosomal lines in contrast were identified by expression the fluorescent coinjection markers as well as by the appearance of UNC animals. At least three independent transgenic lines were analysed for all translational and transcriptional constructs. su.se/). No transmembrane domains (in grey) and no cytoplasmic parts (in green) were predicted. (B) TTR-33 secondary structure with helices (red tubes) and beta strands (green arrows) as predicted by JPred4 [67]. (C) Predicted TTR-33 structure (in grey with indicated mutated sites in red and cysteine bridges in yellow) with superimposed TTR-52 structure (in black). The small TTR-52 α-helix (top right part of structure) is likely an artefact caused by mutations that were introduced for crystallisation [ Fig. Transcript levels of ttr-33, gst-4, gcs-1 and gst-1 Table. Gene Ontology (GO) term analysis for genes for which the expression pattern correlated with ttr-33. Genes with expression patterns correlating to the ttr-33 expression pattern were predicted using SPELL [69]-The GO terms and the according genes are sorted by pvalue. The number of genes that are associated with this GO term out of all genes with a similar expression pattern to ttr-33, and out of all genes in the genome, are shown in column '% Query' and '% Genome', respectively. Very specific processes that are associated with few genes in the genome are marked in blue. Ã Similar GO terms based on the same group of genes were merged.