Presentation on Probiotic Bacillus subtilis Protects against alpha synuclein Aggregation in C.elegans

Article

Probiotic Bacillus subtilis Protects against a-Synuclein Aggregation in C. elegans

Graphical Abstract

Highlights d B. subtilis PXN21 inhibits and reverses a-syn aggregation in a

C. elegans model

d Spores and vegetative cells protect through different

mechanisms

d The probiotic inhibits a-syn aggregation by changing the host

sphingolipid metabolism

d Biofilm formation in the gut and bacterial metabolites reduce

a-syn aggregation

Authors

Marı́a Eugenia Goya, Feng Xue,

Cristina Sampedro-Torres-Quevedo, ...,

Kathryn L. Ball, Nicola R. Stanley-Wall,

Maria Doitsidou

Correspondence [email protected]

In Brief How the gut microbiome affects

Parkinson’s disease remains unclear.

Goya et al. show that the probiotic

B. subtilis strain PXN21 inhibits and clears

a-synuclein aggregation in a C. elegans

model. The bacterium acts via

metabolites and biofilm formation to

activate protective pathways in the host,

including DAF-16/FOXO and sphingolipid

metabolism.

Goya et al., 2020, Cell Reports 30, 367–380 January 14, 2020 ª 2019 The Author(s). https://doi.org/10.1016/j.celrep.2019.12.078

Cell Reports

Article

Probiotic Bacillus subtilis Protects against a-Synuclein Aggregation in C. elegans Marı́a Eugenia Goya,1 Feng Xue,1,5 Cristina Sampedro-Torres-Quevedo,1,5 Sofia Arnaouteli,2

Lourdes Riquelme-Dominguez,1 Andrés Romanowski,3 Jack Brydon,4 Kathryn L. Ball,4 Nicola R. Stanley-Wall,2

and Maria Doitsidou1,6,* 1University of Edinburgh, Centre for Discovery Brain Sciences, Edinburgh, Scotland 2University of Dundee, School of Life Sciences, Dundee, Scotland 3University of Edinburgh, School of Biological Sciences, Edinburgh, Scotland 4University of Edinburgh, Institute of Genetics & Molecular Medicine, Edinburgh, Scotland 5These authors contributed equally 6Lead Contact *Correspondence: [email protected] https://doi.org/10.1016/j.celrep.2019.12.078

SUMMARY

Recent discoveries have implicated the gut micro- biome in the progression and severity of Parkinson’s disease; however, how gut bacteria affect such neurodegenerative disorders remains unclear. Here, we report that the Bacillus subtilis probiotic strain PXN21 inhibits a-synuclein aggregation and clears preformed aggregates in an established Caenorhab- ditis elegans model of synucleinopathy. This protec- tion is seen in young and aging animals and is partly mediated by DAF-16. Multiple B. subtilis strains trigger the protective effect via both spores and vegetative cells, partly due to a biofilm formation in the gut of the worms and the release of bacterial me- tabolites. We identify several host metabolic path- ways differentially regulated in response to probiotic exposure, including sphingolipid metabolism. We further demonstrate functional roles of the sphingoli- pidmetabolism genes lagr-1, asm-3, and sptl-3 in the anti-aggregation effect. Our findings provide a basis for exploring the disease-modifying potential of B. subtilis as a dietary supplement.

INTRODUCTION

Protein misfolding and aggregation are key pathological features observed in numerous neurodegenerative diseases, including Alzheimer’s and Parkinson’s disease (PD) (Ross and Poirier, 2004). PD is one of the most prevalent neurodegenerative disor- ders (Pringsheim et al., 2014) and is currently incurable. It is char- acterized by the progressive loss of dopaminergic neurons in the Substantia Nigra area of the brain, leading to the development of progressive motor and non-motor symptoms (Poewe et al., 2017). Central to the condition is the accumulation of a-synuclein (a-syn) aggregates in Lewy bodies (Spillantini et al., 1998), and the extent of this accumulation correlates with disease severity

(Stefanis, 2012). a-syn acquires neurotoxic properties when protein monomers progressively combine to form insoluble am- yloid fibrils via oligomeric intermediates (Poewe et al., 2017). Although Lewy bodies contain mostly fibrillar forms of a-syn, oligomeric intermediates are also toxic and play a central role in PD pathogenesis (Winner et al., 2011). Despite recent progress toward identifying disease-modifying interventions (Savitt and Jankovic, 2019), only symptomatic treatments are available (Fahn, 2015). Thus, therapeutic strategies directed at inhibiting or reversing a-syn aggregation present a clear oppor- tunity for disease-modifying interventions for PD and other synucleinopathies. Although PD is primarily considered to be a central nervous

system disease, there is clear evidence for an involvement of peripheral signals, particularly from the gastrointestinal tract and the gut microbiota, in PD progression. This is supported by observations that PD symptoms and a-syn pathology begin in peripheral tissues, particularly the intestine, and as the disease progresses, a-syn aggregates gradually spread to multiple brain regions (Braak et al., 2003; Rietdijk et al., 2017). Recently, the human gut microbiome has emerged as an important player influencing PD (Scheperjans, 2016). Gut bacteria can affect brain function by producing metabolites that enter the blood- stream, eliciting immune responses in the host or modulating neuronal function (Chow et al., 2010; Fung et al., 2017). Preclin- ical evidence suggests that the gut microbiota and intestinal permeability modulate behavior, mood, and neuropsychiatric disorders (Clapp et al., 2017). Likewise, a large number of recent studies investigating microbiota in patients with PD found notable differences compared to healthy controls (reviewed by Boertien et al., 2019), which correlated with clinical features (Li et al., 2017; Minato et al., 2017; Scheperjans et al., 2015). Remarkably, faecal transplants from PD patients exacerbate symptoms in a mouse model of PD, demonstrating that differ- ences in microbiota are not merely a result of the disease, but also impact its progression (Sampson et al., 2016). Human microbiota consist of trillions of microorganisms and

over 1,000 bacterial species (Lloyd-Price et al., 2016), posing a challenge for understanding the effects of individual species. In

Cell Reports 30, 367–380, January 14, 2020 ª 2019 The Author(s). 367 This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

the bacterivore Caenorhabditis elegans, the gut microbiota can be precisely controlled, making it a powerful model for studying the effects of gut bacteria on physiological processes at a single species-single gene level (Cabreiro and Gems, 2013). Further- more, C. elegans has proven to be a valuable model for studying molecular mechanisms of PD and protein aggregation. Overex- pression of human a-syn in C. elegans results in the formation of aggregates that progressively become amyloid-like (Kaminski Schierle et al., 2011; van Ham et al., 2008), and work in C. elegansmodels has identified conservedgenetic andchemical modifiers of a-syn toxicity (B!uttner et al., 2013; Hamamichi et al., 2008;Kautuet al., 2013;Knight et al., 2014;Kuwaharaet al., 2008; Pujols et al., 2018; Qiao et al., 2008; Roodveldt et al., 2009; Ruan et al., 2010; van Ham et al., 2008; Zhang et al., 2017). Here, we used a C. elegans model of synucleinopathy to investigate the effects of gut bacteria on a-syn aggregation.

We report that the probiotic bacterium Bacillus subtilis PXN21 (Colenutt and Cutting, 2014), when fed to C. elegans, inhibits, delays, and reverses a-syn aggregation. We characterize these protective effects in both young and old nematodes and investi- gate the contributions of known lifespan-extending pathways. We further show that B. subtilis extracts are able to partially recapitulate the protective effect of live bacteria, indicating that a bacterial metabolite is actively involved. From analysis of gene expression profiles, we find that the protective effect of B. subtilis against a-syn aggregation is mediated through alter- ations in the sphingolipid metabolism pathway. Our findings contribute to the current understanding of how gut bacteria interact with the host to influence physiology in remote tissues, and they will motivate further explorations of the probiotic B. subtilis as a diet-based intervention for PD.

RESULTS

B. subtilis Inhibits and Reverses a-Syn Aggregation in a C. elegans Model of Synucleinopathy To assess the effect of gut bacteria on a-syn aggregation, we used an established C. elegans model (strain NL5901), express- ing human a-syn fused to yellow fluorescent protein (YFP) and driven by a muscle-specific promoter (Punc-54::a-syn::YFP) (van Ham et al., 2008). We fed these worms with different bacte- rial diets and assessed a-syn aggregation in day 1 adult animals (72 h post hatching). Among the bacterial species tested was the B. subtilis strain PXN21 (Colenutt and Cutting, 2014), isolated from the commercially available probiotic product Bio-Kult (by ADM Protexin).

On a regular C. elegans laboratory diet, comprising the non- pathogenic strain of Escherichia coli OP50 (Brenner, 1974), a-syn-expressing animals formed aggregates that can be visual- ized by fluorescence microscopy (van Ham et al., 2008) (Figures 1A and 1B). In contrast, animals fed on B. subtilis strain PXN21 showed a nearly complete absence of aggregates at the day 1 adult stage (Figures 1A and 1B). This striking difference in aggre- gation was not caused by lower expression levels of a-syn in PXN21-fed animals, as unc-54 and a-syn transcript levels were upregulated in day 1 adult animals fed withB. subtilis (Figure 1C). Consistently, there were higher levels of a-syn protein in animals fed on the probiotic (Figures 1D and S1A).

We next tested whether a B. subtilis diet could also clear already-formed aggregates. We grew nematodes on E. coli until the fourth larval (L4) stage when aggregates are evident, then shifted them to aB. subtilis PXN21 diet (Figure 1E) and quantified a-syn aggregation 1 and 3 days later. Most of the aggregates present at the L4 stage cleared 1 day after switching diets, whereas the average size of the foci remained unaffected (Fig- ures 1F, 1G, S1B, and S1C). The clearance of aggregates was not due to reduced levels of a-syn expression (Figures S1D– S1F). Notably, the reduced aggregation levels after the switch to B. subtilis persisted for longer, compared to animals grown continuously on this diet from the first larval (L1) stage (Fig- ure 1G). Similar results were obtained in experiments where the food switch happened on the first day of adulthood (Figures S1G and S1H). We further investigated a-syn native forms using non-denaturing gel electrophoresis. Whereas high molecular mass a-syn forms were detected in extracts from worms ingest- ing either diet, lower molecular weight species that go down to a submonomeric form were primarily detected with the B. subtilis diet (Figure 1H). This indicates alterations of a-syn forms, possibly through cleavage or degradation, by this dietary condition.

B. subtilis Protection Is Effective throughout C. elegans Aging To assess the effects of a B. subtilis diet on a-syn aggregation in aging, we followed animals fed on E. coli OP50 or B. subtilis PXN21 until day 13 of adulthood (corresponding to day 16 of the worm’s life). We assessed aging animals under two different feeding conditions: (1) grown continuously on the specified diet for their entire life or (2) grown on E. coli until the L4 stage and then shifted to a B. subtilis diet. When C. elegans were grown continuously on E. coli after

hatching, aggregates were observed as early as the second larval (L2) stage (data not shown) and progressively increased in number up to day 3 of adulthood (Figure 2A). In contrast, in animals continuously grown on B. subtilis PXN21, there was a near-complete absence of aggregation until day 1 of adulthood, followed by a delayed increase in the number of foci up to day 5 and a subsequent decline. The maximum number of aggregates reached in animals fed with B. subtilis was far lower than that observed on the E. coli diet, indicating that B. subtilis does not simply delay aggregate formation. In the second feeding condition, when worms were switched

from the E. coli to the B. subtilis diet at the L4 stage, aggregation dropped rapidly, reaching a very low, steady level until day 13 of adulthood (Figure 2B). To test whether the reduction of aggrega- tion had an impact on the fitness of a-syn-expressing animals, we performed locomotion assays in a liquid medium. The loco- motion fitness of C. elegans was significantly improved after the switch to the B. subtilis diet, compared to animals continu- ously fed on E. coli, for a time interval that mirrored the time of reduced aggregation (Figure 2C). We conclude that the most striking effect on aggregation is conferred during continuous growth on B. subtilis, with nearly no foci in day 1 of adulthood, whereas the most long-lasting effect is achieved after switching to a B. subtilis diet, with aggregation levels remaining low throughout mid- and late adulthood.

368 Cell Reports 30, 367–380, January 14, 2020

The B. subtilis diet inhibits aggregation during aging without reducing the expression of a-syn, compared to the E. coli diet (Figures 2D and 2E). The decline in aggregation in older E. coli- fed worms correlates with the age-dependent decrease in the expression of the unc-54 promoter (Budovskaya et al., 2008) and, consequently, in a-syn protein levels (Figures 2D and 2E). Remarkably, this decrease is more pronounced in aging worms fed on E. coli (Figure 2E), in agreement with previous reports showing a differential diet-dependent regulation of unc-54 expression (Sánchez-Blanco et al., 2016). Thus, B. subtilis in- hibits aggregation in agingworms despite the consistently higher levels of a-syn in this diet relative to E. coli.

The Protective Effect against a-Syn Aggregation Is a General Property of B. subtilis Species Previous studies report stress resistance and longevity benefits for wild-type animals grown on various laboratory B. subtilis strains (Donato et al., 2017; Garsin et al., 2003; Gusarov et al., 2013; Smolentseva et al., 2017). We therefore asked whether

the observed effect on a-syn aggregation is unique to PXN21 or if it is shared among other strains of the B. subtilis species. We tested a panel of laboratory B. subtilis strains, including 168 (Zeigler et al., 2008), JH642 (Smith et al., 2014), and the un- domesticated strain NCIB 3610 (Branda et al., 2001). All strains showed similar effects on a-syn aggregation to the probiotic strain PXN21 following the continuous or food-switching regime (Figures 2F and 2G), indicating that the anti-aggregation effect is a general property of the B. subtilis species. Furthermore, all tested B. subtilis strains extended the lifespan of a-syn-express- ing transgenic animals (Figure 2H; Table S1).

B. subtilisBiofilm Formation andNitric Oxide Production Protect from a-Syn Aggregation in Aging B. subtilis was previously shown to increase lifespan and stress tolerance in C. elegans via several partly co-dependent mecha- nisms: the formation of a biofilm, a three-dimensional bacterial community embedded in a self-produced extracellular matrix (Branda et al., 2005), in the gut of day 7 adult worms (Donato

Figure 1. B. subtilis PXN21 Inhibits and Reverses a-Syn Aggregation in the C. elegans Model NL5901 (Punc-54::a-syn::YFP) (A) Representative fluorescent images of a-syn aggregates (foci) in the head of day 1 adult worms fed on E. coliOP50 or B. subtilis PXN21. Higher magnifications

of the highlighted regions are shown.

(B) Quantification of a-syn aggregates larger than 1 mm2 per animal in the head region of day 1 adult worms fed on the indicated diet. ****p < 0.0001; n = 25 worms

per condition.

(C) Expression levels by qRT-PCR of unc-54 and a-syn transcripts in day 1 adult worms normalized to the E. coli diet. Expression level of each gene in worms fed

with E. coli was taken as 1. *p = 0.0245, **p = 0.0029, n = 3 per condition, with three technical replicates each (N represents a population of !4,000 worms).

(D) SDS-PAGE of a-syn transgenic and wild-type (control column) day 1 adult worms grown on the two diets. Arrow and arrow with * indicate a-syn monomeric

and sub-monomeric forms, respectively.

(E) Assay strategy for the food-switch experiment. L1, first larval stage; L4, fourth larval stage; d1ad, adult day 1; d3ad, adult day 3.

(F) Fluorescent images of a-syn aggregates of representative L4 (left) and day 1 adult (upper right) worms grown on E. coli or 24 h after the switch toB. subtilis diet

(lower right).

(G) Average number of a-syn aggregates before and after the worm switching. ****p < 0.0001 versus E. coli; a versus b, ****p < 0.0001; n = 25 worms per time point

per condition.

(H) Immunoblotting of native a-syn conformations of transgenic andwild-type young adult worms. Arrowwith * indicates a-syn sub-monomeric form. Data shown

are mean ± SEM from one representative experiment out of three with similar results.

Cell Reports 30, 367–380, January 14, 2020 369

et al., 2017; Smolentseva et al., 2017); the production of nitric oxide (NO) (Gusarov et al., 2013); and the secretion of colony- stimulating factor (CSF) quorum-sensing pentapeptide (Donato et al., 2017). We first confirmed that the yet-uncharacterized B. subtilis strain PXN21 was very proficient at forming a hydro- phobic biofilm under standard conditions, similar to the well- characterized B. subtilis NCIB 3610 (Figure S2A).

To explore whether any of the above bacterial pathways regu- lating lifespan and stress resistance in C. elegans were also responsible for reducing a-syn aggregation in our model, we applied the food-switching approach with B. subtilis NCBI 3610 alongside the biofilm-deficient derivatives Deps(A-O), DbslA, and DtasA (Figure S2A). Each of these strains lacks a different extracellular matrix component essential for biofilm formation: Deps(A-O) is defective in exopolysaccharide forma- tion (Branda et al., 2001); DtasA lacks protein fibers (Romero et al., 2010); and DbslA is deficient in forming the hydrophobic surface layer that surrounds the biofilm (Hobley et al., 2013; Ko- bayashi and Iwano, 2012). We found no effect of biofilm muta- tions on aggregation in the early days of adulthood (Figures 3A and S2B). In contrast, after day 5 of adulthood, when biofilms form in the gut of the nematodes, a-syn aggregation progres- sively increased when animals were fed the DtasA strain, but

not the Deps(A-O) or DbslA deletion strains (Figures 3A and S2B). The triple-mutant strain combining all three biofilm dele- tions (Deps(A-O), DtasA, and DbslA) did not further increase aggregation. We conclude that the biofilm matrix protein TasA supports the ability of B. subtilis to protect against a-syn aggre- gation later in adulthood. Similar to the DtasA biofilm-deficient strain, deletion strains

for DphrC, defective in the production of the quorum-sensing pentapeptide CSF, and Dnos, defective in NO production, also showed an increase in aggregates later in adulthood but not in earlier stages when compared to the wild-type strain (Figures 3B and S2C). These results are in agreement with previous reports that NO and CSF production is increased by an order of magnitude under biofilm-forming conditions (Donato et al., 2017). Given that the DphrC and Dnos deletion results implicate CSF

and NO in the prolonged protective effect of B. subtilis against a-syn aggregation, we asked whether exogenous supplementa- tion of these metabolites in the absence of biofilm could exert a protective effect earlier in adulthood. Whereas animals grown on E. coli supplemented with CSF showed no changes in aggre- gation under the tested conditions (data not shown), NO directly supplied to the worm’s diet induced a significant reduction of

Figure 2. B. subtilis Protection against a-Syn Aggregation Is Effective throughout C. elegans Aging and Is Triggered by Different Strains (A and B) Time course of a-syn aggregation in

worms continuously grown on the annotated diet

from larval stage L1 (A) or after food switching at

the L4 (B). ****p < 0.0001, ***p = 0.0002. Data

shown are mean ± SEM, n = 25 worms per time

point per condition.

(C and D) Immunoblotting analysis (C) and quan-

tification (D) of a-syn versus b-tubulin levels of

protein extracts from day 1 to day 10 adult worms

grown on the annotated diet from the L1 (left and

middle) or L4 stage (right). Datawere normalized to

a-syn/b-tubulin levels of day 1 adults worms fed

with E. coli.

(E) Locomotion analysis (thrashing rate) of worms

after the food switching at L4 from E. coli to

B. subtilis PXN21. *p = 0.0152, **p = 0.0072, ****p <

0.0001. Mean values ± SEM, n = 50 worms per

condition from two independent experiments are

shown

(F and G). Time course of a-syn aggregation in

worms continuously grown (F) or after the food

switching at L4 (G) onto B. subtilis strains 168,

JH642, NCIB 3610, and PXN21. Black asterisks

indicate comparison with E. coli; green asterisks

denote comparison of PXN21 with NCIB 3610;

****p < 0.0001, ***p < 0.001, **p < 0.01. Data

shown are mean ± SEM, n = 25 worms per time

point per condition.

(H) Longevity of a-syn worms fed on mixed lawns

of the different B. subtilis strains shown in (F).

****p < 0.0001, all strains versus E. coli. n R 200

worms per condition from three independent ex-

periments. Data shown are mean ± SEM from one

representative experiment out of three with similar

results, unless stated otherwise.

370 Cell Reports 30, 367–380, January 14, 2020

aggregation on day 3 of adulthood (Figure 3C). In summary, biofilm-associated bacterial pathways/metabolites responsible for lifespan extension are required for keeping aggregation levels low during aging; however, they do not explain the strong protection observed in early adults.

A Bacterial Metabolite from B. subtilis Inhibits a-Syn Aggregation in Early Adults Stress resistance and longevity effects induced by B. subtilis in C. elegans were shown to require live bacteria colonizing the nematode’s gut (Donato et al., 2017; Garsin et al., 2003; Gusarov et al., 2013; Smolentseva et al., 2017). In our case, these mech- anisms seem to be relevant only for the effect of B. subtilis against a-syn aggregation in late adulthood and cannot explain the strong protection seen in early adulthood, when no biofilm is present and only insufficient levels of NO are likely available from ingested B. subtilis. To address whether the effects of B. subtilis in early adulthood required live bacteria, we fed a-syn-expressing worms dead B. subtilis, killed by a combina- tion of UV and antibiotics. Surprisingly, dead B. subtilis were as protective as live bacteria at day 1 of adulthood (Figure 3D). We next considered whether we could recapitulate the

protective effect in the absence of bacteria by supplementing the worms’ diet with B. subtilis extracts. Nematodes grown from the L1 on an E. coli diet supplemented with B. subtilis crude extracts from either the supernatant or pelleted vegetative cells showed a 17%–21% and 21%–33% reduction in aggregation,

respectively (Figures 3E and 3F). Therefore, the effect of B. subtilis on a-syn aggregation in early adults is partially mediated by the action of an active and stable bacterial metab- olite, unlike the short-lived NO, associated with the suppression of aggregation later in life.

B. subtilis Spores and Vegetative Cells Both Protect against a-Syn Aggregation Bacterial metabolic state is affected by environmental conditions and can strongly influence bacteria-host interactions. B. subtilis can exist in two distinct metabolic states: (1) as metabolically active, dividing vegetative cells in nutrient-rich conditions, and (2) as dormant, environmentally resistant spores in nutrient- poor or hostile environments (Nicholson and Setlow, 1990). Un- der our regular experimental conditions, B. subtilis forms lawns that contain a mix of spores and vegetative cells (Figure S3A). Both forms were previously shown to confer longevity and stress-resistance benefits in C. elegans via distinct mechanisms (Donato et al., 2017; Gusarov et al., 2013; Sánchez-Blanco et al., 2016; Smolentseva et al., 2017). To determine whether the effect of B. subtilis PXN21 on a-syn

aggregation depends on the presence of either spores or vege- tative cells, we used selective media to acquire pure cultures of each state (see Method Details and Figure S3A). We found that both B. subtilis vegetative cells and spores fully prevented aggregation in day 1 adult worms, similar to the mixed lawns (Figures 4A and 4B), and both reversed preformed aggregates

Figure 3. Biofilm Formation and Active Me- tabolites Contribute to the B. subtilis Effect (A) Time course of a-syn aggregation of worms fed

with E. coli or switched from E. coli to B. subtilis

wild-isolate NCBI3610 and its isogenic-biofilm

mutant derivatives: Deps(A-O), DbslA, DtasA, and

the triple mutant. Black asterisks show compari-

sons versus E. coli; green asterisks indicate the

differences between B. subtilis NCIB 3610 and its

isogenic mutants; ****p < 0.0001; n R 25 worms

per time point per condition.

(B) Time course of a-syn aggregation of worms fed

with E. coli or switched from E. coli to B. subtilis

wild isolate NCBI3610 or its nitric oxide (NO) and

quorum-sensing peptide (CSF)-deficient mutants

DnosA and DphrC, respectively.; ****p < 0.0001,

**p = 0.0054/0.0098, ***p < 0.001; n R 25 worms

per time point per condition.

(C) Quantification of a-syn aggregates of worms

grown from the L1 on E. coli supplemented with

vehicle (water) or NO donor MAHMA NONOate.

****p < 0.0001, **p < 0. 01; n = 25 worms per time

point per condition.

(D) Quantification of a-syn aggregates in the head

of day 1 adult worms fed with either alive or

UV+antibiotic-killed B. subtilis PXN21 cells. Un-

paired t test; n = 25 worms per condition.

(E and F) Quantification of a-syn aggregates of day

1 adult worms grown from the L1 on E. coli sup-

plemented with crude extracts from the superna-

tant (SN) (E) or pelleted cells (cells) (F) of PXN21

cultures (vehicle: ethyl acetate). ****p < 0.0001, ***p = 0.0001; SN, n = 60 worms per condition from three independent experiments; cells, n = 30 worms per

condition from two independent experiments. Data shown are mean ± SEM from one representative experiment out of three with similar results, unless stated

otherwise. ns, no significant differences.

Cell Reports 30, 367–380, January 14, 2020 371

(Figures 4C and S3B). The reduced aggregation on the vegeta- tive cell diet was not due to lower a-syn expression (Figures S3C–S3E).

We corroborated the protective effect of B. subtilis vegetative cells in early adulthood using the strain 168 carrying a deletion in the spoIIE gene, required for sporulation (York et al., 1992). a-syn expressing animals grown on vegetative cells of 168 DspoIIE strain showed similar levels of aggregation, compared to those grown on vegetative cells of the wild-type B. subtilis strain 168 (Figures 4D and S3F).

Finally, we assessed the effect of vegetative cells of B. subtilis PXN21 on a-syn aggregation during aging. Worms grown continuously on vegetative cells showed a general delay in the formation of aggregates, but the number of aggre- gates eventually reached a maximum comparable to that of worms fed on E. coli (Figure 4E). In contrast, when worms were shifted at the L4 stage from E. coli to a B. subtilis lawn of vegetative cells, the aggregation levels remained low until late in adulthood (Figure 4F), similar to those of a mixed lawn diet (Figure 2B).

As both vegetative cells and spores are protective, we next investigated whether they protect through similar or different mechanisms, focusing first on the known lifespan-extending

pathways, dietary restriction (DR), and the insulin-like signaling (ILS) pathway.

Spores Induce DR and Vegetative Cells Protect via a DR- Independent Mechanism We first considered that DR may underlie the specific protec- tive effects of spores against aggregation, as C. elegans is virtually unable to digest spores (Laaberki and Dworkin, 2008) (Figures 5A and S4A). A B. subtilis spores-only diet poorly sustained growth, inducing severe signs of DR, which is reflected by a strong delay in development to adulthood by 5–7 days and a significantly smaller size in adult worms, compared to those grown on E. coli (Figure S4B). Worms grown on mixed B. subtilis lawns showed mild DR, manifested by a slight developmental delay and smaller body size compared to those grown on E. coli (Figures 5A–5C, S4B, and S4C). In contrast, worms fed only with vegetative cells did not show any signs of DR (Figures 5A–5C, S4B, and S4C). In addition, only small brood size differences between the two diets were observed, which disappeared in the food- switching condition (Figures S4D–S4G). This rules out diet ef- fects on fecundity as a contributing factor to the reduction of aggregation.

Figure 4. B. subtilis Spores and Vegetative Cells Both Protect against a-Syn Aggregation (A) Representative fluorescent images of the head region of day 1 adult worms fed on E. coli or B. subtilis PXN21 vegetative cells or pure-spore cultures. Higher

magnifications of the highlighted regions are shown. NGM, nematode regular growth media; NGM + arginine (+ arg) to inhibit sporulation; NGM no peptone, (-

pep) to prevent spore germination.

(B) Quantification of a-syn aggregates of day 1 adults worms fed with the different diets. ***p < 0.001; n = 25 worms per condition.

(C) Average number of a-syn aggregates of worms before and after the switching, from E. coli to B. subtilis lawn of mixed cells, vegetative cells or spores only.

****p < 0.0001 indicates comparison of each diet versus its respective E. coli control; a versus b, ****p < 0.0001; n = 25 worms per time point per condition.

(D) Average number of a-syn aggregates of worms fed with E. coli, B. subtilis PXN21, B. subtilis 168 strain, or the sporulation mutant 168 DSpoIIE. ****p < 0.0001;

n = 25 per time point per condition.

(E and F) Time course of a-syn aggregation inworms grown from the L1 (E) or shifted at the L4 stage (F) toE. coli orB. subtilis vegetative cells. ****p < 0.0001, n = 25

worms per time point per condition. Data shown aremean ±SEM from one representative experiment out of threewith similar results, unless stated otherwise. ns,

no significant differences.

372 Cell Reports 30, 367–380, January 14, 2020

We confirmed that B. subtilis mixed lawns induced a state of DR using the marker pha-4, an ortholog of the FoxA transcription factors (Panowski et al., 2007): there was a significant increase in pha-4 levels in animals fed on mixed B. subtilis lawns, but not on vegetative cells, compared to animals grown on E. coli (Fig- ure 5D). Furthermore, when we supplemented B. subtilis mixed lawns with E. coli at a 1:1 ratio, we saw a strong protection against a-syn aggregation (Figure 5E) in the absence of pha-4 upregulation (Figure 5D). These results suggest that DR is not responsible for the anti-aggregation effect of vegetative cells, but it may have an effect when animals are fed on spore-rich lawns. DR was previously shown to suppress proteotoxicity in animal

models of polyglutamine and amyloid beta aggregation (Steink- raus et al., 2008), to modify adverse effects of a-syn on the autonomic nervous system in mice (Griffioen et al., 2013), and to alleviate a-syn toxicity in yeast (Guedes et al., 2017). However, to our knowledge, no direct evidence exists that DR can inhibit a-syn aggregation in animal models. We therefore tested whether loss of function of the nicotinic acetylcholine receptor

subunit eat-2, a genetic mimetic of dietary restriction due to reduced food uptake (Lakowski and Hekimi, 1998; McKay et al., 2004), was able to suppress a-syn aggregate formation. Indeed, eat-2(ad456) animals grown on E. coli showed less ag- gregation in day 1 and day 3 adults, compared to wild-type ani- mals grown on E. coli (Figure 5F). However, this reduction was much weaker than the one seen in worms grown on B. subtilis mixed lawns. In addition, a B. subtilis diet further decreased the number of aggregates of eat-2 mutants (Figure 5F), without further increasing pha-4 expression levels (Figure 5G). Similar effects were obtained with a B. subtilis vegetative cell diet (Fig- ures S4H and S4I). We further confirmed that DR is able to inhibit a-syn aggrega-

tion by feeding worms with limited amounts of E. coli killed by UV, a known experimental way to induce DR in C. elegans (Greer et al., 2007) (Figures 5H, S4J, and S4K). However, shifting worms fed ad libitum on E. coli until the L4 stage (or until day 1 of adulthood) to a DR-inducing UV-killed E. coli condition did not clear preformed aggregates (Figures 5I and S4L), even though it inhibited the formation of new aggregates like the probiotic

Figure 5. B. subtilis Reduces a-Syn Aggregation through Dietary-Restriction-Dependent and Independent Mechanisms (A) Fluorescent images of a-syn worms fed on transgenic NCIB 3610 B. subtilis expressing amyE::Phyper-spank-mKate2. Spores resistant to digestion can be

seen in the entire gut in red (left); vegetative cells are present only before the pharyngeal grinder (right).

(B and C) Developmental stage at 48 h (B) and body size at 72 h (C) of a-syn-expressing worms grown on E. coli or B. subtilismixed-cell lawns or vegetative cells.

***p = 0.0007, ****p < 0.0001; n R 80 worms for developmental stage and n R 80 worms for body length per condition from three independent experiments.

(D) Normalized pha-4 expression levels by qRT-PCR in young adult worms grown on the different diet conditions. pha-4 expression level in worms fed with E. coli

was taken as 1. **p = 0.0059; n = 3 samples per condition, with three technical replicates each (each sample consisting of !4,000 worms).

(E) Quantification of a-syn aggregates in day 1 adult worms fed on E. coli,B. subtilis, or a 1:1mixture (B. subtilis:E. coli). ****p < 0.0001; n = 25worms per condition.

(F) Quantification of a-syn aggregates per animal of wild-type or eat-2(ad456)worms grown on E. coli or B. subtilismixed-cell lawn. ****p < 0.0001; n = 75 worms

per time point per condition from three independent experiments.

(G) Normalized pha-4 expression levels by qRT-PCR of young adult wild-type or eat-2(ad456)worms grown on the diet conditions shown in (F). **p = 0.0096, n = 3

samples per condition, with three technical replicates each (each sample consisting of !4,000 worms).

(H) Quantification of a-syn aggregates of day 1 adult worms fed on low concentrations of freshly alive or UV-killed E. coli. ****p < 0.0001, n = 25 worms per

condition.

(I) Average a-syn aggregates of worms before and after L4 switching to E. coli, B. subtilis mixed lawns, or UV-killed E. coli 48 h after seeding. L4, larval stage 4;

d1ad, day 1 adult; d3ad, day 3 adult. ****p < 0.0001 comparison versus E. coli; a versus b, ns for E. coli to UV-killed E. coli versus E. coli versus, ****p < 0.0001 for

E. coli to B. subtilis versus E. coli; n = 25 worms per time point per condition. Data shown are mean ± SEM from one representative experiment out of three with

similar results, unless stated otherwise. ns, no significant differences.

Cell Reports 30, 367–380, January 14, 2020 373

diet (Figure 5I). Therefore, DR per se does not fully reproduce the sum of B. subtilis effects, which include both inhibition and the clearance of aggregates.

Together, these results reveal that DR has a protective role against a-syn aggregation, and it may underlie part of the protec- tive effect triggered by B. subtilis spores. However, vegetative cells inhibit and dissolve a-syn aggregates through a DR-inde- pendent mechanism.

DAF-16 Contributes to the Protection of B. subtilis Later in Adulthood A C. elegans lifespan extension by B. subtilis was previously linked to the downregulation of the evolutionarily conserved ILS pathway (Donato et al., 2017). Decreased signaling of the insulin growth factor (IGF) receptor DAF-2 (Kenyon et al., 1993; Kimura et al., 1997) extends lifespan by activating two down- stream transcription factors, DAF-16/FOXO (Lin et al., 1997; Ogg et al., 1997) and HSF-1 (Hsu et al., 2003). A reduced ILS also protects worms from stress conditions such as toxic protein aggregation of polyglutamine stretches (Hsu et al., 2003; Morley et al., 2002), amyloid beta (Cohen et al., 2006), and a-syn (Knight et al., 2014). To determine whether the ILS pathway plays a role in theB. subtilis-triggered protection against a-syn aggrega- tion, we used daf-2(e1370) mutant worms with inhibited ILS signaling (Gems et al., 1998). The daf-2(e1370) a-syn-expressing animals grown on E. coli showed a strong suppression of aggre- gates in day 1 and day 3 adults compared to wild-type animals (Figure 6A), confirming previous reports (Knight et al., 2014). However, the daf-2 protective effect was significantly less pronounced than that seen in B. subtilis PXN21-fed wild- type worms, and the B. subtilis diet further reduced aggregation levels of daf-2(e1370) animals (Figure 6A). The additive effect between the B. subtilis diet and daf-2 downregulation indicates that B. subtilis acts through an ILS-independent pathway.

To further investigate the role of the ILS pathway, we analyzed the role of DAF-16/FOXO transcription factor and found that the daf-16(mu86) loss-of-function mutation (Lin et al., 1997) fully abrogated the daf-2(e1370) protective effect on E. coli (Fig- ure 6B). In contrast, daf-16(mu86) did not affect the efficiency of a B. subtilismixed diet to inhibit aggregation (Figure 6B). Simi- larly, no increase in aggregation levels was observed in day 1 adults in daf-16 mutant worms fed with B. subtilis vegetative cells (Figure 6C). However, in day 3 adult worms fed on vegeta- tive cells, loss of DAF-16 function led to a faster increase in the number of aggregates (Figure 6C), indicating that the later protection triggered by the vegetative cell diet relies partially on the activity of DAF-16. The hsf-1(sy441) mutation, which in- hibits the second major transcription factor downstream of DAF-2, did not increase aggregation levels when grown on any B. subtilis diet (Figure 6D).

In conclusion, the effect of B. subtilis on a-syn aggregation is independent of the ILS pathway in early adults. However, the protective effect later in adulthood induced by vegetative B. subtilis cells is mediated in part by the action of DAF-16. Thus, our results further indicate that B. subtilis spores and vegetative cells act redundantly through distinct protective mechanisms, with spores acting likely via PHA-4/DR and vege- tative cells via DAF-16.

B. subtilis Inhibits a-Syn Aggregation by Altering Sphingolipid Metabolism in the Host To uncover the host response pathways that are modified by B. subtilis to induce the protective effect, we performed comparative global transcriptomics analysis (RNA sequencing [RNA-seq]) to compare young adult animals fed on two different diets: E. coli OP50 and B. subtilis PXN21 mixed state (Figure 7A). In addition, since the mixture of E. coli and B. subtilis retained much of the anti-aggregation effect but did not induce DR (see Figure 5E), we included this condition in the transcriptomics experiment to reveal DR-independent protective mechanisms. We found that 6,510 genes were differentially expressed by

1.5-fold change or higher in animals fed with B. subtilis compared to those fed E. coli (false discovery rate [FDR] < 0.05, p value < 0.05) (Table S2). A summary of the top 50 most differentially expressed genes between B. subtilis- and E. coli- fed animals, ranked by lowest FDR, is shown in Figure 7B. Sam- ple clustering showed that the mix of both bacteria exhibited a gene expression profile closer to that of animals fed on E. coli than on B. subtilis (Figure S5A). In agreement with this, only 2,291 genes were found to be differentially expressed in this case (Figure 7C; Table S3). Of these, 343 genes were commonly upregulated and 359 downregulated in both animals fed with B. subtilis and the mixture of the two bacteria, compared to E. coli (Figure 7C; Tables S4 and S5). The RNaseq results were validated by randomly selecting 10 upregulated and downregulated genes and testing the level of expression by qRT-PCR (Figures S5B and S5C). As expected, pha-4 was significantly upregulated (1.35-fold change) only in animals fed on B. subtilis, compared with those fed on E. coli (Figure S5D), but showed no differences in animals fed on the mix of the two bacteria versus E. coli. No other DR-related transcription factors were differentially regulated at the transcript level in the different diets (Figure S5D). Previous genome-wide screens on C. elegans models have

identified modifiers of a-syn aggregation (Hamamichi et al., 2008; Knight et al., 2014; van Ham et al., 2008). We intersected our transcriptomics datasets with these lists and found that a number of known suppressors of aggregation were upregulated by the B. subtilis diet (Table S6), indicating that B. subtilis may impart its effects on a-syn through the activation of multiple pro- tective pathways. Next, we performed a Gene Ontology (GO) term analysis of

gene sets affected by the two bacterial diets and found that 708 and 506 biological process (BP) terms were differentially regulated by B. subtilis and by the mix of the two bacteria, respectively (Table S7). The summaries of the 50 most signif- icant non-redundant upregulated GO terms in B. subtilis and the mix versus E. coli, processed by REduce & VIsualize Gene Ontology (REVIGO), are shown (Figures 7D and S5E). Among the top upregulated biological pathways by B. subtilis PXN21 are immune system processes, protein localization, redox processes, general metabolism, and, in particular, lipid metabolism. An expanded analysis of lipid- related terms revealed that several lipid-metabolism-related processes are significantly upregulated in both B. subtilis and the mixed diet, compared to E. coli (Figure 7E).

374 Cell Reports 30, 367–380, January 14, 2020

We focused on a specific pathway branch of lipid metabolism, the sphingolipid metabolism pathway, as it has been proposed to modify a-syn pathology in PD (Alecu and Bennett, 2019; Gal- vagnion, 2017; Lin et al., 2019; Plotegher et al., 2019). Ceramide lipid metabolism is the central hub of the sphingolipid metabolic pathway and was upregulated by both B. subtilis and mixed diets, with a p value <0.001 (Figures 7E and S5F). Genes in this pathway that are upregulated by B. subtilis (Figure S5F; Table S8) include lagr-1, a C. elegans ortholog of human ceramide synthase CERS1 (Deng et al., 2008; Jiang et al., 1998), and asm-3 (Kim and Sun, 2012), an ortholog of human acid sphingo- myelinase, SMPD1, which hydrolyses sphingomyelin to cer- amide. Among the downregulated genes, we identified sptl-3, an ortholog of human SPTLC2, a serine palmitoyltransferase that catalyzes the first and rate-limiting step of the ceramide de novo biosynthesis pathway (Miyake et al., 1995). To address the functional significance of the altered expres-

sion of ceramide pathway genes by B. subtilis, we used loss- of-function mutations of lagr-1, asm-3, and sptl-3. Loss of the upregulated genes lagr-1 or asm-3 increased the number of aggregates in worms continuously grown on B. subtilis (Figures 7F–7I). Conversely, disruption of sptl-3, which was downregu- lated by B. subtilis, reduced aggregation on the E. coli diet compared to wild-type worms (Figure 7J). Sphingolipid metabolism genes were previously reported to

be regulated downstream of eat-2-induced DR (Calvert et al., 2016). Our data indicate that several of the sphingolipid meta- bolism genes are regulated also in the B. subtilis feeding condition that does not induce DR. Thus, we conclude that both DR-dependent and DR-independent effects of the B. subtilis diet converge on sphingolipid metabolism. In light of our findings, we propose that alterations in sphingolipid meta- bolism triggered by the B. subtilis diet result in a reduction of a-syn aggregation in C. elegans.

DISCUSSION

The accumulation of misfolded a-syn into pathological aggre- gates plays a central role in the pathogenesis of PD and other synucleinopathies (Alafuzoff and Hartikainen, 2017). Significant effort has been invested into finding ways to suppress the forma- tion or enhance the clearance of toxic a-syn aggregates as a treatment for PD (Savitt and Jankovic, 2019), though no such therapies are available yet. Previous studies suggest that the presence of distinct groups of bacteria in the gut microbiome modulate PD pathology (Minato et al., 2017; Sampson et al., 2016; Scheperjans et al., 2015). However, deciphering the pre- cise effect of individual bacterial species remains challenging. In this study, we show that B. subtilis PXN21, a probiotic strain that is available for human consumption, both inhibits aggrega- tion and efficiently removes preformed aggregates in a C. elegans model with ectopic expression of human a-syn. It was previously reported that biofilm formation and NO

production by B. subtilis confers C. elegans with stress resis- tance and enhanced longevity (Donato et al., 2017; Smolentseva et al., 2017). Our results reveal that while these pathways contribute to the suppression of a-syn later in life, the protective effect seen earlier in life is independent of these mechanisms. In young adults, the probiotic acts independently of gut coloniza- tion and triggers its protective effects partly via the production of bacterial metabolites other than NO. We provide evidence that distinct metabolic states of the

bacteria affect the physiology of the host as well as a-syn aggre- gation in different ways. B. subtilis spores, which are resistant to digestion and are metabolically inert, induce DR. DR condi- tions are known to activate the lysosomal autophagy pathway (Levine and Kroemer, 2008), one of the main systems of a-syn clearance in cells (Poewe et al., 2017). We find that DR is an effective mechanism to inhibit the accumulation of a-syn in

Figure 6. DAF-16 Contributes to the Protection of B. subtilis in Aging (A) Quantification of a-syn aggregates in the head of wild-type or daf-2(e1370)worms grown on E. coli or B. subtilis PXN21 mixed-cell lawn. ****p < 0.0001, ***p <

0.001, **p < 0. 01; n R 25 per time point per condition.

(B) Average a-syn aggregates in wild-type, daf-2(e1370), daf-16(mu86), and daf-2;daf-16(mu86) double-mutant worms grown on E. coli or B. subtilis PXN21

mixed-cell lawn (spore-rich). ****p < 0.0001, *p = 0.0358; n R 25 per time point per condition.

(C) Average a-syn aggregates in wild-type and daf-2;daf-16(mu86) worms grown on E. coli or vegetative B. subtilis PXN21 lawn. ****p < 0.0001; n = 25 per time

point per condition.

(D) Average a-syn aggregates in wild-type and hsf-1(sy441)mutant worms grown on E. coli or mixed-cell B. subtilis PXN21 lawns or vegetative-only diet (+ arg).

****p < 0.0001; n = 25 per time point per condition. Data shown are mean ± SEM from one representative experiment out of three with similar results, unless stated

otherwise. ns, no significant differences.

Cell Reports 30, 367–380, January 14, 2020 375

Figure 7. B. subtilis Protects against a-Syn Aggregation by Changing the Sphingolipid Metabolism in the Host (A) Assay strategy for the comparative transcriptomics experiment.

(B) Heatmap showing the top 50most differentially expressed genes by false discovery rate (FDR) between E. coli andB. subtilisPXN21 or E. coli andB. subtilis:E.

coli mix. A fold change R 1.5, p < 0.05, and FDR < 0.05 were considered for statistical significance.

(C) Venn diagrams showing the overlap between the statistically significant upregulated and downregulated genes inB. subtilisPXN21 versus themix ofB. subtilis

and E. coli diet.

(D) Summary of the top 50 statistically significant non-redundant BP GO terms of B. subtilis PXN21 versus E. coli by log10 p value.

(E) Lipid-metabolism-related BP GO terms upregulated by B. subtilis PXN21 and the mix versus E. coli diets by log10 p value. Commonly upregulated lipid GO

terms (top), B. subtilis exclusive (middle), and exclusive for the mix of B. subtilis and E. coli diet (bottom) are shown. Gray indicates processes not differentially

regulated.

(F–J) Average a-syn aggregates of wild-type or mutant animals for sphingolipid metabolism genes: lagr-1(gk331) fed from the L1 with B. subtilis PXN21 mixed

lawn diet (F) or vegetative cells (G); asm-3(ok1744)mutant animals fed from the L1 with B. subtilis PXN21 mixed lawn diet (H) or vegetative cells (I); sptl-3(ok1927)

mutant animals fed from the L1 with B. subtilis PXN21 mixed lawn diet (J), compared to E. coli. ****p < 0.0001, **p < 0.01, *p < 0.01; ns, no significant differences.

Mean values ± SEM, n = 50 worms per time point per condition from two independent experiments are shown.

376 Cell Reports 30, 367–380, January 14, 2020

C. elegans and is therefore a likely partial mechanism of action of B. subtilis spores. In contrast, B. subtilis vegetative cells protect via a DR-independent mechanism that partly depends on the ac- tion of DAF-16 in older animals. Downregulation of the ILS pathway, although implicated in the lifespan-extending effects of B. subtilis, is not required for the early protection against a-syn. Therefore, the anti-aggregation properties of B. subtilis remain, to a large extent, distinct from its anti-aging effects. Our transcriptomics analysis revealed that part of the probiot-

ic’s effect is mediated by alterations in the sphingolipid meta- bolism pathway, particularly the regulation of the enzymes LAGR-1/CERS1 (ceramide synthase), ASM-3/SMPD1 (acid sphingomyelinase), and SPTL-3/SPTLC2 (serine palmitoyltrans- ferase). Previous studies suggest that an imbalance of lipids, including ceramides and sphingolipid intermediates, may contribute to the pathology of PD. For example, reduced levels of ceramides occur selectively in brain regions affected by PD pathology (Abbott et al., 2014). Several genetic risk loci in PD affect ceramide metabolism and cellular sphingolipid content (Ferrazza et al., 2016; Gan-Or et al., 2013; Henry et al., 2015; Lin et al., 2019; Plotegher et al., 2019), including mutations in ASM-3/SMPD1 (Foo et al., 2013; Gan-Or et al., 2015, 2013; Ylö- nen et al., 2017) and GBA (the lysosomal glucocerebrosidase). Furthermore, ASM-3/SMPD1 deficiency in cell-based models was shown to lead to a-syn accumulation (Alcalay et al., 2019), whereas inhibition of the Drosophila melanogaster ortholog of SPTL-3/SPTLC2 was found to suppress a-syn-associated neurodegenerative phenotypes (Lin et al., 2018). Furthermore, direct interactions between a-syn and lipids are known to modu- late the aggregation propensity of this protein both in vitro and in vivo (Galvagnion, 2017). We propose that the B. subtilis probi- otic diet in the C. elegans model alters the lipid composition of the cell, directly affecting a-syn aggregation. Our data further demonstrate that a simple dietary intervention can concurrently affect several branches of the sphingolipid pathway, to beneficial effect. PD is typified by the presence of intraneuronal a-syn aggrega-

tion and dopaminergic degeneration (Poewe et al., 2017). Our current study is based on an established C. elegans model that expresses human a-syn in muscle cells, which allows us to assess aggregation in vivo. The effects of B. subtilis on the ner- vous system, as well as its efficacy in mouse models of PD, pre- sent promising avenues of future investigation. The prospect of B. subtilis modifying a-syn aggregation in humans could open exciting possibilities for diet-based, disease-modifying interven- tions through the manipulation of microbiome composition in the gastrointestinal tract or the development of drug therapies based on protective bacterial metabolites.

STAR+METHODS

Detailed methods are provided in the online version of this paper and include the following:

d KEY RESOURCES TABLE d LEAD CONTACT AND MATERIALS AVAILABILITY d EXPERIMENTAL MODEL AND SUBJECT DETAILS

B Nematode and bacterial strains

d METHOD DETAILS B C. elegans growth conditions B Bacterial growth conditions B Characterization of biofilm formation by B. subtilis

strains B Experiments with killed bacteria B Experiments with bacterial extracts B Nitric Oxide (NO) experiments B Quantification of aggregation B Locomotion analysis B Lifespan assays B Quantification of life-traits B Immunoblot analysis B Nematode RNA Sequencing B Reverse transcription and quantitative real-time PCR

(qRT-PCR) d QUANTIFICATION AND STATISTICAL ANALYSIS d DATA AND CODE AVAILABILITY

SUPPLEMENTAL INFORMATION

Supplemental Information can be found online at https://doi.org/10.1016/j.

celrep.2019.12.078.

ACKNOWLEDGMENTS

We thank Claire Bénard, Sebastian Greiss, Oliver Hobert, Anton Gartner,

Luisa Cochella, Emanuel Busch, Diego Golombek, Ailish Tynan and

Claudio Valverde for helpful discussions and feedback on the manuscript;

EdinOmics and Tessa Moses for expert assistance; Viktoria Bajuszova for

technical support; Francis Turner and Edinburgh Genomics for RNA-seq

analysis; José C. Riquelme, Crispin Jordan, and Ignacio Spiousas for input

on statistical analysis; Ellen Nollen for the NL5901 strain; Life Science

Editors for editing assistance; and Ms. Vickie Cunnane for her generous

donation. Funding sources include EMBO ALTF 529-2017 and MSCA-IF

798650 to M.E.G.; Parkinson’s UK, United Kingdom, project grant PRO-

17-21 to M.D. and F.X.; BBSRC, United Kingdom, BB/P001335/1 grant to

N.R.S.-W. and BB/R50614X/1 grant to K.L.B. which supports J.B.; and

the Wellcome Trust-University of Edinburgh Institutional Strategic Support

Fund ISSF3 award to M.D. Some strains were provided by the CGC, funded

by NIH Office of Research Infrastructure Programs (P40 OD010440), the Ba-

cillus Genetic Stock Centre funded by the National Science Foundation

(1756219), and the Gene Knockout Project at the Oklahoma Medical

Research Foundation. The Wormlab was acquired with funding from the

Muir Maxwell Epilepsy Centre.

AUTHOR CONTRIBUTIONS

M.D., M.E.G., F.X., C.S.-T.-Q., and L.R.-D. designed, performed, and analyzed

experiments with C. elegans; J.B. and K.L.B. designed, performed, and

analyzed western blot experiments; N.R.S.-W. and S.A. designed, performed,

and analyzed B. subtilis biofilm experiments and generated bacterial strains;

A.R. contributed with the RNA-seq and qRT-PCR analysis; and M.D. provided

oversight for the project and wrote the paper. All authors contributed to editing

the manuscript.

DECLARATION OF INTERESTS

The authors declare no competing interests.

Received: July 15, 2019

Revised: October 23, 2019

Accepted: December 19, 2019

Published: January 14, 2020

Cell Reports 30, 367–380, January 14, 2020 377

REFERENCES

Abbott, S.K., Li, H., Muñoz, S.S., Knoch, B., Batterham, M., Murphy, K.E., Hal-

liday, G.M., and Garner, B. (2014). Altered ceramide acyl chain length and cer-

amide synthase gene expression in Parkinson’s disease. Mov. Disord. 29,

518–526.

Alafuzoff, I., and Hartikainen, P. (2017). Alpha-synucleinopathies. Handb. Clin.

Neurol. 145, 339–353.

Alcalay, R.N., Mallett, V., Vanderperre, B., Tavassoly, O., Dauvilliers, Y., Wu,

R.Y.J., Ruskey, J.A., Leblond, C.S., Ambalavanan, A., Laurent, S.B., et al.

(2019). SMPD1 mutations, activity, and a-synuclein accumulation in Parkin-

son’s disease. Mov. Disord. 34, 526–535.

Alecu, I., and Bennett, S.A.L. (2019). Dysregulated Lipid Metabolism and Its

Role in a-Synucleinopathy in Parkinson’s Disease. Front. Neurosci. 13, 328.

Ashburner, M., Ball, C.A., Blake, J.A., Botstein, D., Butler, H., Cherry, J.M., Da-

vis, A.P., Dolinski, K., Dwight, S.S., Eppig, J.T., et al.; The Gene Ontology Con-

sortium (2000). Gene ontology: tool for the unification of biology. Nat. Genet.

25, 25–29.

Boertien, J.M., Pereira, P.A.B., Aho, V.T.E., and Scheperjans, F. (2019).

Increasing Comparability and Utility of Gut Microbiome Studies in Parkinson’s

Disease: A Systematic Review. J. Parkinsons Dis. 9, S297–S312.

Braak, H., R!ub, U., Gai, W.P., and Del Tredici, K. (2003). Idiopathic Parkinson’s

disease: possible routes by which vulnerable neuronal typesmay be subject to

neuroinvasion by an unknown pathogen. J. Neural Transm. (Vienna) 110,

517–536.

Branda, S.S., González-Pastor, J.E., Ben-Yehuda, S., Losick, R., and Kolter,

R. (2001). Fruiting body formation by Bacillus subtilis. Proc. Natl. Acad. Sci.

USA 98, 11621–11626.

Branda, S.S., Vik, S., Friedman, L., and Kolter, R. (2005). Biofilms: the matrix

revisited. Trends Microbiol. 13, 20–26.

Brenner, S. (1974). The genetics of Caenorhabditis elegans. Genetics 77,

71–94.

Budovskaya, Y.V., Wu, K., Southworth, L.K., Jiang, M., Tedesco, P., Johnson,

T.E., and Kim, S.K. (2008). An elt-3/elt-5/elt-6 GATA transcription circuit guides

aging in C. elegans. Cell 134, 291–303.

B!uttner, S., Habernig, L., Broeskamp, F., Ruli, D., Vögtle, F.N., Vlachos, M.,

Macchi, F., K!uttner, V., Carmona-Gutierrez, D., Eisenberg, T., et al. (2013).

Endonuclease G mediates a-synuclein cytotoxicity during Parkinson’s dis-

ease. EMBO J. 32, 3041–3054.

Cabreiro, F., and Gems, D. (2013). Worms need microbes too: microbiota,

health and aging in Caenorhabditis elegans. EMBO Mol. Med. 5, 1300–1310.

Calvert, S., Tacutu, R., Sharifi, S., Teixeira, R., Ghosh, P., and de Magalh~aes,

J.P. (2016). A network pharmacology approach reveals new candidate caloric

restriction mimetics in C. elegans. Aging Cell 15, 256–266.

Chow, J., Lee, S.M., Shen, Y., Khosravi, A., andMazmanian, S.K. (2010). Host-

bacterial symbiosis in health and disease. Adv. Immunol. 107, 243–274.

Clapp, M., Aurora, N., Herrera, L., Bhatia, M., Wilen, E., and Wakefield, S.

(2017). Gut microbiota’s effect on mental health: The gut-brain axis. Clin.

Pract. 7, 987.

Cohen, E., Bieschke, J., Perciavalle, R.M., Kelly, J.W., and Dillin, A. (2006).

Opposing activities protect against age-onset proteotoxicity. Science 313,

1604–1610.

Colenutt, C., and Cutting, S.M. (2014). Use of Bacillus subtilis PXN21 spores

for suppression of Clostridium difficile infection symptoms in a murine model.

FEMS Microbiol. Lett. 358, 154–161.

Croft, D., Mundo, A.F., Haw, R., Milacic, M., Weiser, J., Wu, G., Caudy, M., Ga-

rapati, P., Gillespie, M., Kamdar, M.R., et al. (2014). The Reactome pathway

knowledgebase. Nucleic Acids Res. 42, D472–D477.

Deng, X., Yin, X., Allan, R., Lu, D.D., Maurer, C.W., Haimovitz-Friedman, A.,

Fuks, Z., Shaham, S., and Kolesnick, R. (2008). Ceramide biogenesis is

required for radiation-induced apoptosis in the germ line of C. elegans. Sci-

ence 322, 110–115.

Dobin, A., Davis, C.A., Schlesinger, F., Drenkow, J., Zaleski, C., Jha, S., Batut,

P., Chaisson,M., andGingeras, T.R. (2013). STAR: ultrafast universal RNA-seq

aligner. Bioinformatics 29, 15–21.

Donato, V., Ayala, F.R., Cogliati, S., Bauman, C., Costa, J.G., Leñini, C., and

Grau, R. (2017). Bacillus subtilis biofilm extends Caenorhabditis elegans

longevity through downregulation of the insulin-like signalling pathway. Nat.

Commun. 8, 14332.

Fahn, S. (2015). The medical treatment of Parkinson disease from James Par-

kinson to George Cotzias. Mov. Disord. 30, 4–18.

Ferrazza, R., Cogo, S., Melrose, H., Bubacco, L., Greggio, E., Guella, G., Civ-

iero, L., and Plotegher, N. (2016). LRRK2 deficiency impacts ceramide meta-

bolism in brain. Biochem. Biophys. Res. Commun. 478, 1141–1146.

Foo, J.N., Liany, H., Bei, J.X., Yu, X.Q., Liu, J., Au, W.L., Prakash, K.M., Tan,

L.C., and Tan, E.K. (2013). Rare lysosomal enzyme gene SMPD1 variant

(p.R591C) associates with Parkinson’s disease. Neurobiol Aging 34,

2890.e13-5.

Fung, T.C., Olson, C.A., and Hsiao, E.Y. (2017). Interactions between the mi-

crobiota, immune and nervous systems in health and disease. Nat. Neurosci.

20, 145–155.

Galvagnion, C. (2017). The Role of Lipids Interacting with a-Synuclein in the

Pathogenesis of Parkinson’s Disease. J. Parkinsons Dis. 7, 433–450.

Gan-Or, Z., Ozelius, L.J., Bar-Shira, A., Saunders-Pullman, R., Mirelman, A.,

Kornreich, R., Gana-Weisz, M., Raymond, D., Rozenkrantz, L., Deik, A.,

et al. (2013). The p.L302P mutation in the lysosomal enzyme gene SMPD1 is

a risk factor for Parkinson disease. Neurology 80, 1606–1610.

Gan-Or, Z., Orr-Urtreger, A., Alcalay, R.N., Bressman, S., Giladi, N., and

Rouleau, G.A. (2015). The emerging role of SMPD1 mutations in Parkinson’s

disease: Implications for future studies. Parkinsonism Relat. Disord. 21,

1294–1295.

Garsin, D.A., Villanueva, J.M., Begun, J., Kim, D.H., Sifri, C.D., Calderwood,

S.B., Ruvkun, G., and Ausubel, F.M. (2003). Long-lived C. elegans daf-2 mu-

tants are resistant to bacterial pathogens. Science 300, 1921.

Gems, D., Sutton, A.J., Sundermeyer, M.L., Albert, P.S., King, K.V., Edgley,

M.L., Larsen, P.L., and Riddle, D.L. (1998). Two pleiotropic classes of daf-2

mutation affect larval arrest, adult behavior, reproduction and longevity in Cae-

norhabditis elegans. Genetics 150, 129–155.

Greer, E.L., Dowlatshahi, D., Banko, M.R., Villen, J., Hoang, K., Blanchard, D.,

Gygi, S.P., and Brunet, A. (2007). An AMPK-FOXOpathwaymediates longevity

induced by a novel method of dietary restriction in C. elegans. Curr. Biol. 17,

1646–1656.

Griffioen, K.J., Rothman, S.M., Ladenheim, B., Wan, R., Vranis, N., Hutchison,

E., Okun, E., Cadet, J.L., andMattson,M.P. (2013). Dietary energy intakemod-

ifies brainstem autonomic dysfunction caused by mutant a-synuclein. Neuro-

biol. Aging 34, 928–935.

Guedes, A., Ludovico, P., and Sampaio-Marques, B. (2017). Caloric restriction

alleviates alpha-synuclein toxicity in aged yeast cells by controlling the oppo-

site roles of Tor1 and Sir2 on autophagy. Mech. Ageing Dev. 161, 270–276.

Gusarov, I., Gautier, L., Smolentseva, O., Shamovsky, I., Eremina, S., Mironov,

A., and Nudler, E. (2013). Bacterial nitric oxide extends the lifespan of C. ele-

gans. Cell 152, 818–830.

Hajdu-Cronin, Y.M., Chen, W.J., and Sternberg, P.W. (2004). The L-type cyclin

CYL-1 and the heat-shock-factor HSF-1 are required for heat-shock-induced

protein expression in Caenorhabditis elegans. Genetics 168, 1937–1949.

Hamamichi, S., Rivas, R.N., Knight, A.L., Cao, S., Caldwell, K.A., and Caldwell,

G.A. (2008). Hypothesis-based RNAi screening identifies neuroprotective

genes in a Parkinson’s disease model. Proc. Natl. Acad. Sci. USA 105,

728–733.

Han, S.K., Lee, D., Lee, H., Kim, D., Son, H.G., Yang, J.S., Lee, S.V., and Kim,

S. (2016). OASIS 2: online application for survival analysis 2 with features for

the analysis of maximal lifespan and healthspan in aging research. Oncotarget

7, 56147–56152.

Henry, A.G., Aghamohammadzadeh, S., Samaroo, H., Chen, Y., Mou, K., Nee-

dle, E., and Hirst, W.D. (2015). Pathogenic LRRK2 mutations, through

378 Cell Reports 30, 367–380, January 14, 2020

increased kinase activity, produce enlarged lysosomes with reduced degrada-

tive capacity and increase ATP13A2 expression. Hum. Mol. Genet. 24, 6013–

6028.

Hobley, L., Ostrowski, A., Rao, F.V., Bromley, K.M., Porter, M., Prescott, A.R.,

MacPhee, C.E., van Aalten, D.M., and Stanley-Wall, N.R. (2013). BslA is a self-

assembling bacterial hydrophobin that coats the Bacillus subtilis biofilm. Proc.

Natl. Acad. Sci. USA 110, 13600–13605.

Howe, E., Holton, K., Nair, S., Schlauch, D., Sinha, R., and Quackenbush, J.

(2010). MeV: MultiExperiment Viewer. In Biomedical Informatics for Cancer

Research (Springer), pp. 267–277.

Hsu, A.L., Murphy, C.T., and Kenyon, C. (2003). Regulation of aging and age-

related disease by DAF-16 and heat-shock factor. Science 300, 1142–1145.

Jiang, J.C., Kirchman, P.A., Zagulski, M., Hunt, J., and Jazwinski, S.M. (1998).

Homologs of the yeast longevity gene LAG1 in Caenorhabditis elegans and hu-

man. Genome Res. 8, 1259–1272.

Kaminski Schierle, G.S., Bertoncini, C.W., Chan, F.T.S., van der Goot, A.T.,

Schwedler, S., Skepper, J., Schlachter, S., van Ham, T., Esposito, A., Kumita,

J.R., et al. (2011). A FRET sensor for non-invasive imaging of amyloid formation

in vivo. ChemPhysChem 12, 673–680.

Kautu, B.B., Carrasquilla, A., Hicks, M.L., Caldwell, K.A., and Caldwell, G.A.

(2013). Valproic acid ameliorates C. elegans dopaminergic neurodegeneration

with implications for ERK-MAPK signaling. Neurosci. Lett. 541, 116–119.

Kenyon, C., Chang, J., Gensch, E., Rudner, A., and Tabtiang, R. (1993). A C.

elegans mutant that lives twice as long as wild type. Nature 366, 461–464.

Kim, Y., and Sun, H. (2012). ASM-3 acid sphingomyelinase functions as a pos-

itive regulator of the DAF-2/AGE-1 signaling pathway and serves as a novel

anti-aging target. PLoS ONE 7, e45890.

Kimura, K.D., Tissenbaum, H.A., Liu, Y., and Ruvkun, G. (1997). daf-2, an insu-

lin receptor-like gene that regulates longevity and diapause in Caenorhabditis

elegans. Science 277, 942–946.

Knight, A.L., Yan, X., Hamamichi, S., Ajjuri, R.R., Mazzulli, J.R., Zhang, M.W.,

Daigle, J.G., Zhang, S., Borom, A.R., Roberts, L.R., et al. (2014). The glycolytic

enzyme, GPI, is a functionally conserved modifier of dopaminergic neurode-

generation in Parkinson’s models. Cell Metab. 20, 145–157.

Kobayashi, K., and Iwano, M. (2012). BslA(YuaB) forms a hydrophobic layer on

the surface of Bacillus subtilis biofilms. Mol. Microbiol. 85, 51–66.

Kolde, R. (2015). pheatmap: Pretty Heatmaps. R package version 1.0.8.

https://CRAN.R-project.org/package=pheatmap.

Koo, B.M., Kritikos, G., Farelli, J.D., Todor, H., Tong, K., Kimsey, H., Wapinski,

I., Galardini, M., Cabal, A., Peters, J.M., et al. (2017). Construction and Analysis

of Two Genome-Scale Deletion Libraries for Bacillus subtilis. Cell Syst. 4, 291–

305.e297.

Kuwahara, T., Koyama, A., Koyama, S., Yoshina, S., Ren, C.H., Kato, T., Mi-

tani, S., and Iwatsubo, T. (2008). A systematic RNAi screen reveals involve-

ment of endocytic pathway in neuronal dysfunction in alpha-synuclein trans-

genic C. elegans. Hum. Mol. Genet. 17, 2997–3009.

Laaberki, M.H., and Dworkin, J. (2008). Role of spore coat proteins in the resis-

tance of Bacillus subtilis spores to Caenorhabditis elegans predation.

J. Bacteriol. 190, 6197–6203.

Lakowski, B., and Hekimi, S. (1998). The genetics of caloric restriction in Cae-

norhabditis elegans. Proc. Natl. Acad. Sci. USA 95, 13091–13096.

Landré, V., Revi, B., Mir, M.G., Verma, C., Hupp, T.R., Gilbert, N., and Ball, K.L.

(2017). Regulation of transcriptional activators by DNA-binding domain ubiqui-

tination. Cell Death Differ. 24, 903–916.

Lane, D.J. (1991). 16S/23S rRNA sequencing. In Nucleic Acid Techniques in

Bacterial Systematic, E. Stackebrandt and M. Goodfellow, eds. (John Wiley

and Sons), pp. 115–175.

Levine, B., and Kroemer, G. (2008). Autophagy in the pathogenesis of disease.

Cell 132, 27–42.

Li, W., Wu, X., Hu, X., Wang, T., Liang, S., Duan, Y., Jin, F., and Qin, B. (2017).

Structural changes of gut microbiota in Parkinson’s disease and its correlation

with clinical features. Sci. China Life Sci. 60, 1223–1233.

Liao, Y., Smyth, G.K., and Shi, W. (2014). featureCounts: an efficient general

purpose program for assigning sequence reads to genomic features. Bioinfor-

matics 30, 923–930.

Lin, K., Dorman, J.B., Rodan, A., and Kenyon, C. (1997). daf-16: An HNF-3/

forkhead family member that can function to double the life-span of Caeno-

rhabditis elegans. Science 278, 1319–1322.

Lin, G., Lee, P.T., Chen, K., Mao, D., Tan, K.L., Zuo, Z., Lin, W.W., Wang, L.,

and Bellen, H.J. (2018). Phospholipase PLA2G6, a Parkinsonism-Associated

Gene, Affects Vps26 and Vps35, Retromer Function, and Ceramide Levels,

Similar to alpha-Synuclein Gain. Cell Metab. 28, 605–618.e606.

Lin, G., Wang, L., Marcogliese, P.C., and Bellen, H.J. (2019). Sphingolipids in

the Pathogenesis of Parkinson’s Disease and Parkinsonism. Trends Endocri-

nol. Metab. 30, 106–117.

Lloyd-Price, J., Abu-Ali, G., and Huttenhower, C. (2016). The healthy human

microbiome. Genome Med. 8, 51.

Lun, A.T., Chen, Y., and Smyth, G.K. (2016). It’s DE-licious: A Recipe for Differ-

ential Expression Analyses of RNA-seq Experiments Using Quasi-Likelihood

Methods in edgeR. Methods Mol. Biol. 1418, 391–416.

Luo, W., and Brouwer, C. (2013). Pathview: an R/Bioconductor package for

pathway-based data integration and visualization. Bioinformatics 29, 1830–

1831.

Martin, M. (2011). Cutadapt removes adapter sequences from high-

throughput sequencing reads. EMBnet.journal 17, 10–12.

McKay, J.P., Raizen, D.M., Gottschalk, A., Schafer, W.R., and Avery, L. (2004).

eat-2 and eat-18 are required for nicotinic neurotransmission in the Caeno-

rhabditis elegans pharynx. Genetics 166, 161–169.

Minato, T., Maeda, T., Fujisawa, Y., Tsuji, H., Nomoto, K., Ohno, K., and Hir-

ayama, M. (2017). Progression of Parkinson’s disease is associated with gut

dysbiosis: Two-year follow-up study. PLoS ONE 12, e0187307.

Miyake, Y., Kozutsumi, Y., Nakamura, S., Fujita, T., and Kawasaki, T. (1995).

Serine palmitoyltransferase is the primary target of a sphingosine-like immu-

nosuppressant, ISP-1/myriocin. Biochem. Biophys. Res. Commun. 211,

396–403.

Mok, D.Z., Sternberg, P.W., and Inoue, T. (2015). Morphologically defined sub-

stages of C. elegans vulval development in the fourth larval stage. BMC Dev.

Biol. 15, 26.

Morley, J.F., Brignull, H.R., Weyers, J.J., and Morimoto, R.I. (2002). The

threshold for polyglutamine-expansion protein aggregation and cellular

toxicity is dynamic and influenced by aging in Caenorhabditis elegans. Proc.

Natl. Acad. Sci. USA 99, 10417–10422.

Nicholson, W.L., and Setlow, P. (1990). Sporulation, germination, and

outgrowth. In Molecular Biological Methods for Bacillus (John Wiley and

Sons), pp. 391–450.

Ogg, S., Paradis, S., Gottlieb, S., Patterson, G.I., Lee, L., Tissenbaum, H.A.,

and Ruvkun, G. (1997). The Fork head transcription factor DAF-16 transduces

insulin-likemetabolic and longevity signals in C. elegans. Nature 389, 994–999.

Ostrowski, A., Mehert, A., Prescott, A., Kiley, T.B., and Stanley-Wall, N.R.

(2011). YuaB functions synergistically with the exopolysaccharide and TasA

amyloid fibers to allow biofilm formation by Bacillus subtilis. J. Bacteriol.

193, 4821–4831.

Panowski, S.H., Wolff, S., Aguilaniu, H., Durieux, J., and Dillin, A. (2007). PHA-

4/Foxa mediates diet-restriction-induced longevity of C. elegans. Nature 447,

550–555.

Plotegher, N., Bubacco, L., Greggio, E., and Civiero, L. (2019). Ceramides in

Parkinson’s Disease: From Recent Evidence to New Hypotheses. Front. Neu-

rosci. 13, 330.

Poewe, W., Seppi, K., Tanner, C.M., Halliday, G.M., Brundin, P., Volkmann, J.,

Schrag, A.E., and Lang, A.E. (2017). Parkinson disease. Nat. Rev. Dis. Primers

3, 17013.

Pringsheim, T., Jette, N., Frolkis, A., and Steeves, T.D. (2014). The prevalence

of Parkinson’s disease: a systematic review and meta-analysis. Mov. Disord.

29, 1583–1590.

Cell Reports 30, 367–380, January 14, 2020 379

Pujols, J., Peña-Dı́az, S., Lázaro, D.F., Peccati, F., Pinheiro, F., González, D.,

Carija, A., Navarro, S., Conde-Giménez, M., Garcı́a, J., et al. (2018). Small

molecule inhibits a-synuclein aggregation, disrupts amyloid fibrils, and pre-

vents degeneration of dopaminergic neurons. Proc. Natl. Acad. Sci. USA

115, 10481–10486.

Qiao, L., Hamamichi, S., Caldwell, K.A., Caldwell, G.A., Yacoubian, T.A., Wil-

son, S., Xie, Z.L., Speake, L.D., Parks, R., Crabtree, D., et al. (2008). Lysosomal

enzyme cathepsin D protects against alpha-synuclein aggregation and

toxicity. Mol. Brain 1, 17.

Rietdijk, C.D., Perez-Pardo, P., Garssen, J., van Wezel, R.J., and Kraneveld,

A.D. (2017). Exploring Braak’s Hypothesis of Parkinson’s Disease. Front. Neu-

rol. 8, 37.

Ritchie, M.E., Phipson, B., Wu, D., Hu, Y., Law, C.W., Shi, W., and Smyth, G.K.

(2015). limma powers differential expression analyses for RNA-sequencing

and microarray studies. Nucleic Acids Res. 43, e47.

Robinson, M.D., McCarthy, D.J., and Smyth, G.K. (2010). edgeR: a Bio-

conductor package for differential expression analysis of digital gene expres-

sion data. Bioinformatics 26, 139–140.

Romero, D., Aguilar, C., Losick, R., and Kolter, R. (2010). Amyloid fibers pro-

vide structural integrity to Bacillus subtilis biofilms. Proc. Natl. Acad. Sci.

USA 107, 2230–2234.

Roodveldt, C., Bertoncini, C.W., Andersson, A., van der Goot, A.T., Hsu, S.T.,

Fernández-Montesinos, R., de Jong, J., van Ham, T.J., Nollen, E.A., Pozo, D.,

et al. (2009). Chaperone proteostasis in Parkinson’s disease: stabilization of

the Hsp70/alpha-synuclein complex by Hip. EMBO J. 28, 3758–3770.

Ross, C.A., and Poirier, M.A. (2004). Protein aggregation and neurodegenera-

tive disease. Nat. Med. 10, S10–S17.

Ruan, Q., Harrington, A.J., Caldwell, K.A., Caldwell, G.A., and Standaert, D.G.

(2010). VPS41, a protein involved in lysosomal trafficking, is protective in Cae-

norhabditis elegans and mammalian cellular models of Parkinson’s disease.

Neurobiol. Dis. 37, 330–338.

Sampson, T.R., Debelius, J.W., Thron, T., Janssen, S., Shastri, G.G., Ilhan,

Z.E., Challis, C., Schretter, C.E., Rocha, S., Gradinaru, V., et al. (2016). Gut Mi-

crobiota Regulate Motor Deficits and Neuroinflammation in a Model of Parkin-

son’s Disease. Cell 167, 1469–1480.e1412.

Sánchez-Blanco, A., Rodrı́guez-Matellán, A., González-Paramás, A., Gonzá-

lez-Manzano, S., Kim, S.K., and Mollinedo, F. (2016). Dietary and microbiome

factors determine longevity in Caenorhabditis elegans. Aging (Albany N.Y.) 8,

1513–1539.

Savitt, D., and Jankovic, J. (2019). Targeting a-Synuclein in Parkinson’s Dis-

ease: Progress Towards the Development of Disease-Modifying Therapeutics.

Drugs 79, 797–810.

Scheperjans, F. (2016). Gut microbiota, 1013 new pieces in the Parkinson’s

disease puzzle. Curr. Opin. Neurol. 29, 773–780.

Scheperjans, F., Aho, V., Pereira, P.A., Koskinen, K., Paulin, L., Pekkonen, E.,

Haapaniemi, E., Kaakkola, S., Eerola-Rautio, J., Pohja, M., et al. (2015). Gut

microbiota are related to Parkinson’s disease and clinical phenotype. Mov.

Disord. 30, 350–358.

Schindelin, J., Arganda-Carreras, I., Frise, E., Kaynig, V., Longair, M., Pietzsch,

T., Preibisch, S., Rueden, C., Saalfeld, S., Schmid, B., et al. (2012). Fiji: an

open-source platform for biological-image analysis. Nat. Methods 9, 676–682.

Schneider, C.A., Rasband, W.S., and Eliceiri, K.W. (2012). NIH Image to Im-

ageJ: 25 years of image analysis. Nat. Methods 9, 671–675.

Smith, J.L., Goldberg, J.M., and Grossman, A.D. (2014). Complete Genome

Sequences of Bacillus subtilis subsp. subtilis Laboratory Strains JH642

(AG174) and AG1839. Genome Announc. 2, e00663-14.

Smolentseva, O., Gusarov, I., Gautier, L., Shamovsky, I., DeFrancesco, A.S.,

Losick, R., and Nudler, E. (2017). Mechanism of biofilm-mediated stress resis-

tance and lifespan extension in C. elegans. Sci. Rep. 7, 7137.

Spillantini, M.G., Crowther, R.A., Jakes, R., Hasegawa, M., and Goedert, M.

(1998). alpha-Synuclein in filamentous inclusions of Lewy bodies from Parkin-

son’s disease and dementia with lewy bodies. Proc. Natl. Acad. Sci. USA 95,

6469–6473.

Stefanis, L. (2012). a-Synuclein in Parkinson’s disease. Cold Spring Harb. Per-

spect. Med. 2, a009399.

Steinkraus, K.A., Smith, E.D., Davis, C., Carr, D., Pendergrass, W.R., Sutphin,

G.L., Kennedy, B.K., and Kaeberlein, M. (2008). Dietary restriction suppresses

proteotoxicity and enhances longevity by an hsf-1-dependent mechanism in

Caenorhabditis elegans. Aging Cell 7, 394–404.

Stiernagle, T. (2006). The C. elegans Research Community. In Maintenance of

C. elegans, WormBook. https://doi.org/10.1895/wormbook.1.101.1.

Subramanian, A., Tamayo, P., Mootha, V.K., Mukherjee, S., Ebert, B.L., Gil-

lette, M.A., Paulovich, A., Pomeroy, S.L., Golub, T.R., Lander, E.S., and Me-

sirov, J.P. (2005). Gene set enrichment analysis: a knowledge-based approach

for interpreting genome-wide expression profiles. Proc. Natl. Acad. Sci. USA

102, 15545–15550.

Thompson, O., Edgley, M., Strasbourger, P., Flibotte, S., Ewing, B., Adair, R.,

Au, V., Chaudhry, I., Fernando, L., Hutter, H., et al. (2013). The million mutation

project: a new approach to genetics in Caenorhabditis elegans. Genome Res.

23, 1749–1762.

Untergasser, A., Nijveen, H., Rao, X., Bisseling, T., Geurts, R., and Leunissen,

J.A. (2007). Primer3Plus, an enhanced web interface to Primer3. Nucleic Acids

Res. 35, W71-4.

van Gestel, J., Weissing, F.J., Kuipers, O.P., and Kovács, A.T. (2014). Density

of founder cells affects spatial pattern formation and cooperation in Bacillus

subtilis biofilms. ISME J. 8, 2069–2079.

van Ham, T.J., Thijssen, K.L., Breitling, R., Hofstra, R.M., Plasterk, R.H., and

Nollen, E.A. (2008). C. elegans model identifies genetic modifiers of alpha-syn-

uclein inclusion formation during aging. PLoS Genet. 4, e1000027.

Verhamme, D.T., Kiley, T.B., and Stanley-Wall, N.R. (2007). DegU co-ordinates

multicellular behaviour exhibited by Bacillus subtilis. Mol. Microbiol. 65,

554–568.

Watson, E., MacNeil, L.T., Ritter, A.D., Yilmaz, L.S., Rosebrock, A.P., Caudy,

A.A., andWalhout, A.J.M. (2014). Interspecies Systems Biology Uncovers Me-

tabolites Affecting C. elegans Gene Expression and Life History Traits. Cell

156, 1336–1337.

Winner, B., Jappelli, R., Maji, S.K., Desplats, P.A., Boyer, L., Aigner, S., Hetzer,

C., Loher, T., Vilar, M., Campioni, S., et al. (2011). In vivo demonstration that

alpha-synuclein oligomers are toxic. Proc. Natl. Acad. Sci. USA 108, 4194–

4199.

Wu, D., Lim, E., Vaillant, F., Asselin-Labat, M.L., Visvader, J.E., and Smyth,

G.K. (2010). ROAST: rotation gene set tests for complex microarray experi-

ments. Bioinformatics 26, 2176–2182.

Ylönen, S., Siitonen, A., Nalls, M.A., Ylikotila, P., Autere, J., Eerola-Rautio, J.,

Gibbs, R., Hiltunen, M., Tienari, P.J., Soininen, H., et al. (2017). Genetic risk

factors in Finnish patients with Parkinson’s disease. Parkinsonism Relat. Dis-

ord. 45, 39–43.

York, K., Kenney, T.J., Satola, S., Moran, C.P., Jr., Poth, H., and Youngman, P.

(1992). Spo0A controls the sigma A-dependent activation of Bacillus subtilis

sporulation-specific transcription unit spoIIE. J. Bacteriol. 174, 2648–2658.

Zeigler, D.R., Prágai, Z., Rodriguez, S., Chevreux, B., Muffler, A., Albert, T.,

Bai, R., Wyss, M., and Perkins, J.B. (2008). The origins of 168, W23, and other

Bacillus subtilis legacy strains. J. Bacteriol. 190, 6983–6995.

Zhang, S., Glukhova, S.A., Caldwell, K.A., and Caldwell, G.A. (2017). NCEH-1

modulates cholesterol metabolism and protects against a-synuclein toxicity in

a C. elegans model of Parkinson’s disease. Hum. Mol. Genet. 26, 3823–3836.

380 Cell Reports 30, 367–380, January 14, 2020

STAR+METHODS

KEY RESOURCES TABLE

REAGENT or RESOURCE SOURCE IDENTIFIER

Antibodies

Mouse monoclonal antibody anti-a-synuclein BD Biosciences Cat# 610786; RRID: AB_398107

Mouse monoclonal anti-b-actin Sigma-Aldrich Cat# A5441; RRID: AB_476744

Mouse monoclonal anti-b-Tubulin Sigma-Aldrich Cat# T4026; RRID: AB_477577

Rabbit Polyclonal Anti-Mouse Agilent Cat# P0260; RRID: AB_2636929

Bacterial and Virus Strains

E. coli: OP50 CGC RRID:WB-STRAIN:OP50

B. subtilis: probiotic PXN21 Bio-Kult by ADM-Protexin N/A

B. subtilis: NCIB 3610 Marburg, undomesticated BGSC BGSCID: 3A1

B. subtilis: NRS2097 NCIB 3610 DbslA::cmlR Nicola Stanley-Wall (Ostrowski et al.,2011) N/A

B. subtilis: NRS2415 NCIB 3610 DtasA::spcR Nicola Stanley-Wall (Ostrowski et al.,2011) N/A

B. subtilis: NRS2450 NCIB 3610 Deps(A-O)::tetR Nicola Stanley-Wall (Ostrowski et al.,2011) N/A

B. subtilis: NRS2543 NCIB 3610 Deps(A-O)::tetR

DtasA::spcR DbslA::cmlR

Nicola Stanley-Wall (Ostrowski et al.,2011) N/A

B. subtilis: NCIB 3610 amyE Phyper-spank-

mKate2::spcR

Ákos T. Kovács (van Gestel et al., 2014) N/A

B. subtilis: NRS5852 NCIB 3610 amyE::Phyper-

spank-mKate2::spcR

Nicola Stanley-Wall N/A

B. subtilis: NRS6296 NCIB 3610 DnosA::kanR Nicola Stanley-Wall N/A

B. subtilis: NRS6297 NCIB 3610, DphrC::kanR Nicola Stanley-Wall N/A

B. subtilis:168 trpC2; DspoIIE::kanR Addgene-BGSC (Koo et al.,2017) Cat# 1000000115

BGSCID: BKK00640

B. subtilis:168 trpC2; DnosA::kanR Addgene-BGSC (Koo et al.,2017) BGSCID: BKK07630

B. subtilis: JH642 BGSC BGSCID: 1A96

Bacteriophage: Bacillus phage SPP1 Anne Moir (University of Sheffield) N/A

Chemicals, Peptides, and Recombinant Proteins

Agar Formedium Cat# AGA02

Select Agar Thermo Fisher Scientific Cat#30391023

Bacto peptone BD Biosciences Cat# 211677

Sodium Hypochlorite solution (4.00-4.99%) Honeywell Cat# 239305

Ca(NO3)2 Sigma-Aldrich Cat# C1396

CaCl2 Sigma-Aldrich Cat# 449709

Coomassie Brilliant Blue G-250 Thermo Fisher Scientific Cat# 20279

Schaeffer and Fulton Spore Stain Solution A Sigma-Aldrich Cat# 90903

Schaeffer and Fulton Spore Stain Solution B Sigma-Aldrich Cat# 39955

Dichloromethane (DCM) Thermo Fisher Scientific Cat# 402152

Nutrient broth No 3 Sigma-Aldrich Cat# 70149

Dried Milk Powder Marvel N/A

Dithiothreitol (DTT) GE Healthcare Cat# 17-1318-01

Ethanol ACROS Organics Cat# AC615095000

Ethyl acetate Sigma-Aldrich Cat# 319902

EDTA Sigma-Aldrich Cat# 798681

FeCl3 Sigma-Aldrich Cat# 451649

FeSO4 Sigma-Aldrich Cat# 450278

L-glutamic acid monosodium salt monohydrate Sigma-Aldrich Cat# 49621

(Continued on next page)

Cell Reports 30, 367–380.e1–e7, January 14, 2020 e1

Continued

REAGENT or RESOURCE SOURCE IDENTIFIER

Glycerol R 99.5% Thermo Fisher Scientific Cat# BP229-1

HEPES sodium salt Sigma-Aldrich Cat# H7006

Hydrogen peroxide solution Sigma-Aldrich Cat# 216763

Kanamycin Sulfate Corning Cat# 61-176-RG

KCl Sigma-Aldrich Cat# P9541

KH2PO4 ACROS Organics Cat# AC424200025

K2HPO4 ACROS Organics Cat# AC424190025

L-Arginine Alfa Aesar Cat# A15738

Levamisole Hydrochloride MP Biomedicals Cat# 155228

Luminol 97% Sigma-Aldrich Cat# 123072

Luria-Bertani (LB) broth Sigma-Aldrich Cat# L3022

Lysozyme Thermo Fisher Scientific Cat# 89833

MgCl2 Sigma-Aldrich Cat# M8266

MnCl2 Sigma-Aldrich Cat# 244589

MOPS Sigma-Aldrich Cat# M1254

Na2HPO4 ACROS Organics Cat#AC204851000

NaCl Thermo Fisher Scientific Cat# BP358-1

NaF Sigma-Aldrich Cat# S7920

NaOH Thermo Fisher Scientific Cat# S612-3

NativePAGE 4-16% Bis-Tris-gels Thermo Fisher Scientific Cat# BN1002BOX

NativePAGE 20x Running buffer Thermo Fisher Scientific Cat# BN2001

Nitrocellulose membrane 0.2 mm Biorad Cat# 1620112

MAHMA NONOate (NO donor) Sigma-Aldrich Cat# M1555

NuPAGE 4-12% Bis-Tris-gels Invitrogen Cat# NP0322PK2

PageRuler Protein Ladder, 10 to 250 kDa Thermo Fisher Scientific Cat# 26620

PBS Sigma-Aldrich Cat# P4417

P-Coumaric acid Sigma-Aldrich Cat# C9008

Penicillin Streptomycin GIBCO Cat# 15070063

PFA Sigma-Aldrich Cat#

Protease Inhibitor Mix GE Healthcare Cat# 80-6501-23

Proteinase K BioVision Cat# 9211-5

Thiamine hydrochloride Sigma-Aldrich Cat# T4625

Triton X-100 MP Biomedicals Cat# 807423

TWEEN! 20 Sigma-Aldrich Cat# P1379

ZnCl2 Sigma-Aldrich Cat# 229997

Critical Commercial Assays

OneTaq! 2X Master Mix with Standard Buffer New England Biolabs (NEB) Cat# M0482S

GoTaq! G2 Green Master Mix Promega Cat#M782A

QIAquick PCR Purification Kit QIAGEN Cat#28104

Qubit RNA BR kit Thermo Fisher Scientific Cat# Q10210

Quick Start Bradford Protein Assay Kit Bio-Rad Cat# 5000201

Quick-RNA Microprep Kit ZYMO Research Cat# R1050

Q5 High-Fidelity 2X Master Mix NEB Cat# M0492S

SuperScript IV First-Strand Synthesis System Thermo Fisher Scientific Cat# 18091050

TruSeq Stranded mRNA kit Illumina Cat# 20020594

LightCycler! 480 SYBR Green I master Roche Cat# 04707516001

Zymoclean Gel DNA Recovery Kit ZYMO Research Cat# D4001

Deposited Data

C. elegans RNA-Seq reads (fastq files) This study ArrayExpress: E-MTAB-8164

(Continued on next page)

e2 Cell Reports 30, 367–380.e1–e7, January 14, 2020

LEAD CONTACT AND MATERIALS AVAILABILITY

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Maria Doitsidou ([email protected]).

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Nematode and bacterial strains All bacterial and nematode strains used and generated in this study can be found in the Key Resources Table.

Continued

REAGENT or RESOURCE SOURCE IDENTIFIER

Experimental Models: Organisms/Strains

C. elegans: N2 Bristol CGC CGCRRID:WB-STRAIN:N2

C. elegans: NL5901 pkIs2386[Punc-

54::a-synuclein::YFP + unc-119(+)]

Ellen Nollen (van Ham et al., 2008) RRID:WB-STRAIN:NL5901

C. elegans: MDH586 daf-2(e1370) III; pkIs2386 This study N/A

C. elegans: MDH585 daf-16(mu86) I; pkIs2386 This study N/A

C. elegans: MDH587 hsf-1(sy441) I; pkIs2386 This study N/A

C. elegans: MDH657 daf-2(e1370) III; daf-16(mu86) I;

pkIs2386

This study N/A

C. elegans: MDH611 eat-2(ad465) II; pkIs2386 This study N/A

C. elegans: MDH711 lagr-1(gk331) I, pkIs2386 This study N/A

C. elegans: MDH724 asm-3(ok1744) IV; pkIs2386 This study N/A

C. elegans: MDH725 sptl-3(ok1927) II; pkIs2386 This study N/A

Oligonucleotides

For information regarding oligonucleotide sequences

used in this study please refer to Table S9

This study Table S9

Software and Algorithms

Cutadapt version cutadapt-1.9.dev2 Martin, 2011 https://cutadapt.readthedocs.io/

en/stable/

EdgeR package version 3.16.5 Lun et al., 2016; Robinson et al., 2010 https://bioconductor.org/packages/

release/bioc/html/edgeR.html

FeatureCounts package version 1.5.3 Liao et al., 2014 http://subread.sourceforge.net

Fiji Schindelin et al., 2012 http://fiji.sc/

GraphPad Prism GraphPad Software, La Jolla

California USA

https://www.graphpad.com

ImageJ Schneider et al., 2012 https://imagej.nih.gov/ij/

Limma package version 3.30.13 of Bioconductor Ritchie et al., 2015 https://bioconductor.org/packages/

release/bioc/html/limma.html

Multiple Experiment Viewer version 4.9.0_r2731 Howe et al., 2010 http://mev.tm4.org/

OASIS 2 Han et al., 2016; Oncotarget 11269 https://sbi.postech.ac.kr/oasis2/

Pathview package version 1.24.0 of Bioconductor. Luo and Brouwer, 2013 https://bioconductor.org/packages/

release/bioc/html/pathview.html

Pheatmap package (R package version 1.0.8.) Kolde, 2015 https://rdrr.io/cran/pheatmap/

Photoshop CC Adobe Systems Inc. https://www.adobe.com/Photoshop

Primer3Plus Untergasser et al., 2007 http://bioinfo.ut.ee/primer3/

R-3.5.1 R Core Team https://www.r-project.org/

SnapGene SnapGene software from GSL Biotech https://www.snapgene.com/

STAR version 2.5.2b Dobin et al., 2013 http://code.google.com/p/rna-star/

Illustrator CC Adobe Systems Inc. https://www.adobe.com/products/

illustrator

WormLab tracking platform MBF Bioscience, Williston, VT USA https://www.mbfbioscience.com/

wormlab

Cell Reports 30, 367–380.e1–e7, January 14, 2020 e3

C. elegans NL5901 pkIs2386[Punc-54::a-synuclein::YFP + unc-119(+)] was kindly provided by Ellen Nollen. C. elegans N2 and all the mutant strains used for the generation of the different pkIs2386 derived strains were obtained from the Caenorhabditis Genetics Center (CGC) (https://cgc.umn.edu) and the Million Mutant collection (Thompson et al., 2013).

Strains obtained from the CGC are: CB1370 daf-2(e1370), CF1038 daf-16(mu86), PS3551 hsf-1(sy441), DA465 eat-2(ad465), VC747 lagr-1(gk331), RB1579 sptl-3(ok1927), RB1487 asm-3(ok1744).

The molecular identity of these alleles is as follows: daf-2(e1370) (Gems et al., 1998) is a missense reference allele, daf-16(mu86) (Lin et al., 1997) and lagr-1(gk331) (Deng et al., 2008) are deletion, loss of function alleles. eat-2(ad456) (Lakowski and Hekimi, 1998) and hsf1(sy441) (Hajdu-Cronin et al., 2004) are nonsense alleles. asm-3(ok1744) is a complex substitution allele removing most of the last 7 exons, sptl-3(ok1927) is a deletion removing exons 5 to 9.

The following strains were generated in this study: MDH586 daf-2(e1370) III; pkIs2386, MDH585 daf-16(mu86) I; pkIs2386, MDH587 hsf-1(sy441) I; pkIs2386, MDH657 daf-2(e1370) III; daf-16(mu86) I; pkIs2386, MDH614 daf-2(gk390525) III; pkIs2386, MDH611 eat-2(ad465) II; pkIs2386, MDH711 lagr-1(gk331) I, pkIs2386, MDH725 sptl-3(ok1927) II; pkIs2386, MDH724 asm- 3(ok1744) IV; pkIs2386.

Several bacterial strains were used in this study, E. coli OP50 was obtained from the CGC. The B. subtilis PXN21 strain (Colenutt and Cutting, 2014) was isolated from Bio-Kult Advanced Multi-Strain Formulation dietary supplement (https://www.bio-kult.com, ADM-Protexin) and genotyped using universal 16S rRNA primers (Lane, 1991) (See Table S9 for primers). The wild-type undomes- ticated B. subtilis NCIB 3610 and the laboratory 168 and JH64102 strains were obtained from the Bacillus Genetic Stock Center (BGSC) (http://www.bgsc.org). 168-based deletion strain DspoIIE was obtained from Addgene (www.addgene.org) as part of the B. subtilis Single Gene Deletion Library with Kanamycin resistance (Koo et al., 2017). The NCIB 3610 deficient derivatives strains DtasA::cml, DbslA::spc, Deps(A-O)::tet, Dnos::kan and DphrC::kan and the triple DbslA::spc; eps(A-O)::tet; tasA::::cml, were ob- tained from the Nicola Stanley-Wall lab. SPP1 phage transductions were used to introduce DNA into B. subtilis NCIB 3610 strains from 168 derivatives (Verhamme et al., 2007). Drug resistance cassettes are indicated as follows: cml, chloramphenicol resistance; kan, kanamycin resistance; erm, erythromycin resistance; tet, tetracycline resistance and spc, spectinomycin resistance.

METHOD DETAILS

C. elegans growth conditions Nematodes were handled according to standard practices (Brenner, 1974; Stiernagle, 2006). Worm strains were grown on NGM plates for experiments with mixed spores and vegetative cells, NGM plus 0.5 mM of arginine for experiments with vegetative cells (to avoid sporulation), or NGM without peptone for experiments with spores only (to avoid germination). All strains were grown at 20"C unless otherwise indicated. Worms were synchronized by the alkaline hypochlorite method (Stiernagle, 2006) and left nutating overnight to hatch in M9 supplemented with kanamycin 50 mg/ mL (Sigma) and 1x antibiotic-antimycotic (Thermo Fisher Scientific). For the continuous feeding regime, synchronized L1 worms were plated, grown until day 1 adults, and then transferred to new plates every two days. For the food switch experiments, worms were grown on E. coli OP50 until L4 stage, then shifted to a new diet and transferred to new plates every two days thereafter.

Bacterial growth conditions Bacterial cultures were grown until an OD600 of 1 in Luria-Bertani (LB) media at 37"C with agitation (220 rpm). 330 mL of a 2x concen- trated culture were seeded on 55mm unvented NGMplates. Seeded plates were left to dry and grow for 3 days at room temperature for experiments with mixed spores and vegetative cells, or overnight for experiments with vegetative cells only. To obtain spore-pure bacterial cultures, PXN21 B. subtilis bacteria were grown in Schaeffer’s sporulation medium (SSM) as previously described (Donato et al., 2017) (containing per liter: 8 g of Difco Bacto-nutrient broth, 10mL of 10%w/v KCl, 10mL of 1.2%w/vMgSO4$7H2O,!1.50mL of 1MNaOH up to pH 7.6, 1.0mL of 1MCa(NO3)2, 1.0mL of 0.010MMnCl2, and 1.0mL of 1mMFeSO4). Briefly, bacteria were grown in SSM medium at 37 "C for 48 h. The culture was heat-treated for 20 min at 80 "C to kill vegetative cells and then spun down. To obtain pure spores, the heat-treated pelleted cells was treated three times with lysozyme (25 mg/mL; for 30 min at 37 "C), washed each time with cold deionised water and centrifuged until the culture consisted of only phase-bright spores. The efficiency of the purification was tested by the Schaeffer Fulton staining method.

Characterization of biofilm formation by B. subtilis strains B. subtilis biofilms were grown on MSgg medium (5 mM potassium phosphate and 100 mM MOPS at pH 7.0 supplemented with 2 mM MgCl2, 700 mM CaCl2, 50 mM MnCl2, 50 mM FeCl3, 1 mM ZnCl2, 2 mM thiamine, 0.5% glycerol, 0.5% glutamate) (Branda et al., 2001) solidified with 1.5% select agar (Invitrogen) at 30"C at the indicated time points. To set up a biofilm, a 3 mL aliquot of LB medium was inoculated with an individual colony taken from an overnight plate and grown at 37"C to an OD600 of 1. Then 5 mL of the culture was placed onto an MSgg plate which was incubated at 30"C for morphology and hydrophobicity studies. Images of colony biofilms were recorded using a Leica MZ16FA stereomicroscope. Biofilm hydrophobicity was determined by placing a 5 ml droplet of water on the upper surface of biofilms that had been grown for 48 hours at 30"C (Hobley et al., 2013). The water droplet was allowed to equilibrate for 5 minutes prior to imaging using a ThetaLite TL100 optical tensiometer (Biolin Scientific).

e4 Cell Reports 30, 367–380.e1–e7, January 14, 2020

Experiments with killed bacteria For dietary restriction (DR) with killed E. coliOP50 experiments, bacterial cultures were grown as previously described until an OD600

of 1. DRwas induced by seeding 200 mL of a 1x concentrated culture on 55mm unvented NGM plate and the same amount but twice concentrated (2x final) was used for normal growth conditions. E. coli culture was completely spread in the plates and left to dry and grow for 24 h or 48 h. Bacteria were killed by UV irradiation (Watson et al., 2014) (254 nm, 5 J/cm2), using a UV crosslinker (CL 508, Cleaver Scientific). For the experiments with killed B. subtilis PXN21, bacterial cultures were grown as previously described until an OD600 of 1 to have

only vegetative cells present. 200 mL of a 2x concentrated culture were completely spread on 55 cmunventedNGM+0.5mMarginine plate and left to dry and grow for only 24h. Bacteria were killed by a combination of UV irradiation (254 nm, 5 J/cm2) and antibiotics treatment (200 mg/mL kanamycin and 1mg/mL carbenicillin) for 3h before transferring the worms onto them (Smolentseva et al., 2017). In all the conditions, the efficiency of the killing protocol was tested by sampling and streaking the killed bacteria in LB agar plates and incubated overnight at 37"C.

Experiments with bacterial extracts B. subtilis PXN21was inoculated from a fresh colony into 1 L of LB and left to grow at 37"C and 220 rpm for 48 h. Under this condition, the cultures of this strain are very saturated and reach a final OD600 of 4-5. Bacteria were pelleted at 14000 rpm for 30 min at 4"C and the supernatant was separated from the pellet. The supernatant was consecutively filtered twice with 0.45 mm and 0.22 mm vacuum cellulose acetate filters to completely get rid of the cells. The pellet was washed twice with 250 mL of cold water and then once with 60% cold ethanol, centrifuged each time at 5000 rpm and 4"C and resuspended with vortex. The final clean pellet was resuspended in 100 mL of PBS and the bacteria were killed with 3 flash freeze-thaw cycles with liquid nitrogen/water bath at 60"C, followed by 1 h of incubationwith lysozyme (25 mg/mL) on ice. Cells were finally disrupted by sonication using 5 cycles of 30 s at 20Hz (MSESoniprep 150), with incubation on ice for 30 s between cycles to avoid overheating. For both the supernatant and the cell lysate, 3 sequential organic extractions with 1:2 ratio of supernatant to diclorometane (DCM) and 1:10 of cell lysate and DCM were performed, respec- tively. The organic phases were dried separately to fully remove the DCM using an EZ-2 Elite personal evaporator (GeneVac), in the very low BP mode. The evaporator was maintained at 40"C throughout the process and 15 mL glass tubes rinsed clean with DCM were used for concentrating the organic phase. The final dry extracts were resuspended in 1mL of ethyl acetate with vortex and kept it at #80"C afterward. Since we started with 1L of material, both extracts were considered to be 1000x concentrated. Appropriate dilutions from the concentrated stock were prepared in ethyl acetate, mixed in a glass falcon tube with 100 ml of water and spread on the top of E. coli seeded 35 mm plates. Ethyl acetate alone was added to E. coli as a vehicle-only control.

Nitric Oxide (NO) experiments Freshly prepared NGM agar plates were placed open in a tissue culture hood for 30 min to dry and facilitate rapid absorption. Next, 50 mL of 2x OD600 = 1 bacterial culture was spread atop the plate, and then a freshly prepared solution of 200 mMNO donor MAHMA NONOate (Sigma) in water was applied to NGM agar plates to achieve a final concentration of 2 mM, and 4 mM. Immediately after- ward, !70 synchronized L1s worms were quickly transferred to the plate. This protocol was shown to be efficient to extend C. elegans lifespan, even though MAHMA NONOate has a very short half-life at

pH 6 (!1 min) (Gusarov et al., 2013). For control experiments, the NO donor was substituted with an equal amount of distilled water. For measurements of the effect of NO on a-synuclein aggregation, worms weremoved to freshly prepared NO plates every day start- ing from L1 and scored at day 1 and 3 of adulthood.

Quantification of aggregation NL5901 pkIs2386[Punc-54::a-synuclein::YFP + unc-119(+)] worms were anaesthetized using 50 mM Levamisole (Sigma) and high magnification (40x objective) z stack images of the head region were obtained by using a Zeiss Axio imager 2 microscope. Fluores- cent spots bigger than 1 mm2, present in the region between the tip of head and the end of the pharyngeal bulb, were quantifiedmanu- ally, assisted by the Fiji analyze particle function applied to maximum intensity projections of the z stacks. To do so, background subtraction (rolling ball radius of 10 pixels) and adjustment of the threshold (automatic) were applied to the images before the analysis of the number of particles. The total area of the aggregates was extracted from the particle analysis with Fiji and themean aggregates size per diet was simply calculated considering the total number of aggregates in the corresponding area. 72 hours after plating of the L1s was counted as day 1 adult. At least 25 worms were quantified per time point per condition. Each experiment was performed in triplicate, unless stated otherwise.

Locomotion analysis Thrashing assays were performed as described before (Pujols et al., 2018), with somemodifications.C. elegansNL5901wormswere synchronized as described above and were cultured at 20"C on E. coli OP50 strain until they reached L4 developmental stage. They were then either remained on OP50 diet or transferred to B. subtilis PXN21 (regular 3 days seeded protocol). Thrashing was assayed on days 1, 3, 5, 7 and 10 of adulthood. 5 animals were placed in a 40 ml drop of M9 buffer on an unseeded plate. Movies of thrashing worms were recorded for 3 min using the WormLab tracking platform (MBF Biosciences) at 7.5 frames/second. Waves per minute were obtained by analyzing the last minute of each video (allowing the animals to recover for 120 s after picking them into the drop).

Cell Reports 30, 367–380.e1–e7, January 14, 2020 e5

Average frequencies were determined every 0.4 s. Experiments were performed in duplicate. 10 videos per condition and 5 animals per video were analyzed (a total of 100 worms per condition).

Lifespan assays Lifespan assays were performed at 20"C as previously described with modifications (Greer et al., 2007). Briefly, 200-250 synchro- nized L1s were placed on to corresponding food conditions and were, starting at d1Ad, transferred every 2 days onto fresh food and assessed for survival. Worms that failed to respond to the transfer process and repeated gentle prodding were declared dead and removed. Individuals that were missing or needed to be removed due to internal hatching were marked as censored. Experiments were performed in triplicate.

Quantification of life-traits To determine developmental growth rates, 40-65 worms were mounted on a 3%w/v agarose pad in a drop of 50 mM levamisole and their developmental stages were assessed under compoundmicroscope (DIC, 40Xmagnification) at exactly 48 h, 60 h, and 72 h after the synchronized L1s were placed on food. Individuals were staged as early L4, late L4, or adult by using the 9 stages of vulva devel- opment as reference points as described before (Mok et al., 2015). Stages L4.1 to L4.4 were considered as early L4, L4.5 to L4.9 considered as late L4 and young adult category was based on a fully formed vulva. For size measurements, worms were photo- graphed at 10x and the images analyzed using Fiji (ImageJ). A segmented line was drawn along the center line of each the worm, quantified with the measure function, and then calibrated based on the scale bar. To assess the egg-laying rate and brood size, L4 worms were singled on to 10 separate plates per condition and transferred every 24 hours on to fresh plates until day 5 adult. The numbers of progeny resulting from each day of egg-laying were counted 2 days later.

Immunoblot analysis Day 1 adult worms (!4000) were rinsedwithM9 + 0.01%Triton X-100, washed 3 times to remove bacteria, pelleted and resuspended in 400 mL of HEPES-based detergent buffer (50 mM HEPES pH 8, 0.2% v/v Triton x 100, 150 mM NaCl, 10 mM NaF, 5 mM DTT) + 1x Protease Inhibitor Mix (GE Healthcare 80-6501-23). Worms were centrifuged at 14000 rpm for 1 min, flash freeze-thawed 5 times with liquid nitrogen/water bath at 80"C and kept at #80"C. Worm pellets were disrupted mechanically using a TissueLyser II (QIAGEN) for 4 cycles of 40 s at 30 Hz, with 200ul of 0.7mm zirconia beads (Biospec). The lysates were centrifuged at 14000 rpm for 1 min and total amount of protein was quantified by Bradford assay (Bio-Rad). NuPAGE (4%–12%) Bis-Tris-gels (Invitrogen) were used to analyze a-synuclein (from 3 mg total protein) and b-actin (from 20 mg total protein) under denaturing conditions as previously described (Landré et al., 2017). Following transfer to nitrocellulose (PALL), membranes intended for a-synuclein analysis were fixed for 10 min using 4% PFA and washed 3 times with PBS containing 0.1% v/v tween-20 prior to blocking. The immunoblots were probed using anti-a-synuclein monoclonal antibody (BD Biosciences) 1:2000 and anti-b-actin (Sigma) 1:500 with appropriate HRP-labeled secondary antibodies (DAKO) at 1:2000. Bound antibodies where detected using ECL.

For the blots corresponding to the time course experiments, 50 worms (in duplicates) from day 1, day 3, day 5, day 7 and day 10 adults in the different diets were manually picked into 50 ml of M9 + 0.01% Triton X-100, washed 3 times to remove bacteria and re- suspended in 4xLDS sample buffer supplemented with 10mMdithiothreitol. Worms were flash freeze on dry ice, sonicated at 4"C for 10 cycles of 40 s at intensity II (Bioruptor! Plus) and boiled at 95"C for 10 min. The lysates were centrifuged at 14000 rpm for 1 min and around 2.5 ul of samples from day 1 adults to day 10 adults were loaded in NuPAGE (4%–12%) Bis-Tris-gels (Invitrogen) and transfer to nitrocellulose (PALL). Membranes were probed with 1:6000 of anti-b-tubulin monoclonal antibody (Sigma) to adjust the volumes manually. Tubulin was specifically selected for these blots because of the higher sensitivity versus actin for samples with low protein content. A second blot was performed by using the previously adjusted volumes for the samples and probed with both 1:6000 of the anti-b-tubulin antibody and 1:2000 of the anti-a-synuclein antibody.

Native protein analysis was carried out using NativePAGE 4%–16% Bis-Tris-gels (Invitrogen) loaded with 30 mg total protein in 1x NativePAGE sample buffer (Invitrogen) containing 0.5% w/v Coomassie Brilliant Blue G-250 per lane. The gel was run at 150V for 2 hours in 1x NativePAGE running buffer (minus G-250; Invitrogen). Proteins where subsequently transferred to nitrocellulose and a-synuclein was detected as described above.

Nematode RNA Sequencing 2000 young adult worms (approximately 50-55h after plating of the L1s) grown on E. coli OP50, B. subtilis PXN21 or a 1:1 mix of E. coli: B. subtilis, were collected and washed three times with M9 + 0.01% Triton X-100 buffer. The pellet was resuspended in 400 ml of RNA Lysis buffer (Quick-RNA Microprep Kit, Zymo Research) and worms were mechanically disrupted as previously described and kept at #80"C. Total RNA was extracted from the samples according to the manufacturer’s instructions. Three inde- pendent biological replicates were used for each experimental condition.

RNA samples were sent to Edinburgh Genomics for QC check and sequencing. Briefly, quality check of the samples was performed using Qubit with the broad range RNA kit (Thermo Fisher Scientific) and Tapestation 4200 with the RNA Screentape for eukaryotic RNA analysis (Agilent). Libraries were prepared from 5 mg of total RNA using the TruSeq Stranded mRNA kit (Illumina), and then validated. Samples were pooled to create 9 multiplexed DNA libraries, which were paired-end sequenced on an Illumina HiSeq 4000 platform. At least 290M + 290M 75 nt PE reads were obtained (one lane).

e6 Cell Reports 30, 367–380.e1–e7, January 14, 2020

Sequence reads were trimmed using Cutadapt (version cutadapt-1.9.dev2; Martin, 2011) for quality at the 30 end using a quality threshold of 30 and for adaptor sequences of the TruSeq stranded mRNA kit (AGATCGGAAGAGC), with a minimum length of 50. After trimming, reads were aligned against the C. elegans (WBcel235_ens8) genome from Ensembl release 84 with STAR) (version 2.5.2b; Dobin et al., 2013) with default parameters, except for specifying paired-end reads and the option ‘‘–outSAMtype BAM Unsorted.’’ Count tables for the different feature levels were obtained from bam files using the featureCounts (Liao et al., 2014) pack- age version 1.5.3 with custom R scripts. Strandness was set to ‘reverse’ and a minimum alignment quality of 10 was specified. Gene names and other fields were derived from input annotation and added to the count/expression matrices. Count tables at the gene level presented a good correlation overall between replicates and samples. Differential gene expression analysis was conducted for 17,708 genes whose expression was above a minimum ‘‘counts per

million’’ (CPM) threshold level (CPM 0.1) in at least three samples. Differential gene expression was estimated using the edgeR pack- age (Lun et al., 2016; Robinson et al., 2010) version 3.16.5 and resulting p values were adjusted using a false discovery rate (FDR) criterion. Geneswith p values lower than 0.05, FDR values lower than 0.05 and a log2 fold change > 0.58were considered to be differ- entially expressed. Heatmaps were generated using R with the pheatmap package (Kolde, 2015) (R package version 1.0.8.) or MeV (Multiple Experiment Viewer) version 4.9.0_r2731 (Howe et al., 2010). Differential gene set analysis was carried out with the ROAST method (Wu et al., 2010) from the Limma package (version 3.30.13;

Ritchie et al., 2015) of Bioconductor, using the samemodels and contrasts as used in differential expression. The following gene sets were used: Gene Ontology Biological Process, Molecular Function and Cellular Component downloaded from Ensembl version 91 (Ashburner et al., 2000; Subramanian et al., 2005); Reactome Pathways (Croft et al., 2014) (Reactome, downloaded January 2018). ROAST was executed using 9,999 rotations (randomizations). Each gene set was annotated with those genes individually differential (in the same direction as indicated for the gene set) to an unadjusted p value of 0.05. KEGG pathways were visualized utilizing the pathview package (Luo and Brouwer, 2013) version 1.24.0 of Bioconductor.

Reverse transcription and quantitative real-time PCR (qRT-PCR) For quantitative real time PCR, 3 mg of total RNA and poly-T(V) (20 nt long) and random hexamers (6 nt long) were used for cDNA synthesis using SuperScript III Reverse Transcriptase (Thermo Fisher Scientific) and following manufacturer’s instructions. Quanti- tative PCR analyses were performed with 1:50 dilutions of the cDNAs using LightCycler 480 SYBR Green I Master mix and a Light- Cycler 480 II (Roche), following manufacturer’s instructions. Expression of all the genes was normalized to the geometric mean of cdc-42 and F25B5.5 references genes. Expression of these genes was not variable between E. coli and B. subtilis in the RNaseq results. Data was analyzed by using the standard curve method and normalized to E. coli or E. coli arginine samples, respectively. For primer sequences see Table S9, all the primers were ordered from Sigma.

QUANTIFICATION AND STATISTICAL ANALYSIS

All assays were performed in triplicate, unless stated otherwise. Graphs and statistical analysis were performed using Graphpad Prism 7. Data shown are presented as mean ± SEM. Statistical significance was calculated by unpaired t test, one-way or two- way ANOVA and Bonferroni multiple comparisons post hoc test, with p < 0.05 considered statistically significant. Statistical signif- icance levels are denoted as follows: ****p < 0.0001; ***p < 0.001; **p < 0.01 and *p < 0.05. All statistical tests were two tailed where applicable. For the expression levels by qRT-PCR, 3 independent biological samples with technical triplicates were analyzed by using two-way

ANOVA with Bonferroni’s post hoc test of logNRQ. ImageJ and Fiji were used to quantify the number and size of the aggregates as described above. For the lifespan assay, Kaplan-Meier survival curves were generated using the statistical analysis software Graphpad Prism 7.

Comparisons were made using the Log-rank (Mantel-Cox) test in both Graphpad Prism 7 and the online survival analysis tool OASIS 2 (Han et al., 2016). For the immunoblots, quantification of the signals was performed using the gel analysis function of ImageJ and a-synuclein inten-

sity was normalized against b-actin/tubulin signals. Locomotor analysis was performed using WormLab tracking platform and software (MBF Biosciences).

DATA AND CODE AVAILABILITY

The accession number for the C. elegans RNA sequencing data reported in this paper is ArrayExpress: E-MTAB-8164. The authors declare that all data supporting the findings of this study are available within the article and its Supplementary Infor-

mation files or upon request. Code is also available either online where indicated or upon request.

Cell Reports 30, 367–380.e1–e7, January 14, 2020 e7

Cell Reports, Volume 30

Supplemental Information

Probiotic Bacillus subtilis

Protects against a-Synuclein

Aggregation in C. elegans

María Eugenia Goya, Feng Xue, Cristina Sampedro-Torres-Quevedo, So!a Arnaouteli, Lourdes Riquelme-Dominguez, Andrés Romanowski, Jack Brydon, Kathryn L. Ball, Nicola R. Stanley-Wall, and Maria Doitsidou

Figure S1 | B. subtilis PNX21 strain prevents and reverses α-synuclein aggregation in the C. elegans model NL5901 (Punc-54::α-syn::YFP). Related to Figure 1. (A) Normalised protein levels of α-syn vs

β-actin signals shown in Fig. 1D. (B) Quantification of α-syn aggregates larger than 1 μm2 per animal in

the head region for the food-switching experiment (food switch at L4, see Fig. 1G). Most of the

reversibility effect is reached 24 h after shifting to a B. subtilis PXN21 diet. ****P<0.0001 vs E. coli; a

vs b, ****P<0.0001; n=25 worms per time point per condition. (C) Quantification of the total area of all

aggregates larger than 1 μm2 and the mean aggregates size per animal in the head region shown in Fig. 1G. (D) Expression levels by qRT-PCR of unc-54 and α-syn transcripts from day 1 adult worms on E.

coli or 24 h after switching to B. subtilis PXN21 mixed lawns, normalized to the E. coli diet. Expression

level of each gene in E. coli was taken as 1. ***P<0.0004; n=3 per condition, with 3 technical replicates

each (n represents a population of ~4000 worms). (E) SDS/PAGE from α-syn transgenic and wild type

day 1 adult worms before and after the food-switching (food switch at L4, see Fig. 1G). Arrow with *

indicates α-syn monomeric isoform. (F) Normalised protein levels of α-syn vs β-actin signals shown in E. (G, H) Quantification of α-syn aggregates in worms before and after food-switching at day 1 adult, from

E. coli to B. subtilis PXN21. Aggregates progressively clear at 24 h, 48 h and 72 h after shifting to a B.

subtilis PXN21 diet (H). ****P<0.0001 vs E. coli (G); *P=0.015, ****P<0.0001 (H); n=25 worms per

time point per condition. Data shown are mean ± SEM from one representative experiment out of three

with similar results, unless stated otherwise. ns: no significant differences.

Figure S2 | Biofilm formation and active metabolites contribute to the B. subtilis effect. Related to Figure 3. (A) Biofilm formation ability of B. subtilis PXN21 with respect to other well characterized B.

subtilis NCIB 3160 strains. In the upper panels, representative B. subtilis strains in solid MSgg biofilm-

inducing medium are shown. PXN21 is able to form highly complex and elaborate biofilm structures

similar to the undomesticated NCIB 3610. These structures are severely disrupted in the isogenic NCIB

3610 deletion strains ΔtasA, ΔbslA, and Δesp(A-O). Hydrophobicity tests (an important assessment of

biofilm function), performed by adding a drop of water on top of the biofilm structure, are shown for each

strain on the lower panels of each image and quantified on the table (bottom left). Biofilms with contact

angles above 90° are consider hydrophobic. Scale bars=0.5 cm. Mean ± SEM, n=25 biofilm/strain from 5

independent experiments are shown. (B) Average α-syn aggregation of worms fed with E. coli or

switched from E. coli to B. subtilis wild isolate NCBI 3610 or its matrix biofilm-deficient derivatives,

ΔtasA, ΔbslA, Δeps(A-O) and the triple mutant (see Fig. 3A). ****P<0.0001, **P=0.0049; n≥25 worms

per time point per condition. (C) Average α-syn aggregation in worms fed with either E. coli, switched

from E. coli to B. subtilis NCBI 3610, or the NO and CSF deficient derivatives ΔnosA and ΔphrC,

respectively (see Fig. 3B). ****P<0.0001, **P<0.01 and ***P=0.0001; n≥25 worms per time point per

condition. Data shown are mean ± SEM from one representative experiment out of three with similar

results, unless stated otherwise. ns: no significant differences.

Figure S3 | B. subtilis spores and vegetative cells both protect against α-synuclein aggregation. Related to Figure 4. (A) Phenotypes of B. subtilis PXN21 tested in this work. Upper panels show

Schaeffer Fulton staining of B. subtilis in different growing conditions. Scale bars=50 µm. In regular

NGM medium and 3 days after seeding, B. subtilis forms a mix of spores (blue spheres) and vegetative

cells (pink rods, left). In NGM supplemented with arginine (+ arg) and only 24 h after seeding, vegetative

cells are predominant (middle). The spore-only condition is achieved by using a minimal NGM medium

without peptone (- pep) and a short drying time (24 h) to prevent germination (upper right). Day 1 adult

α-syn transgenic worms (fed from L1) with their corresponding diets are shown at the bottom. (B)

Quantification of α-syn aggregates per animal in the head region for the food-switching experiment from

E. coli to different B. subtilis PXN21 diets (see Fig. 4C). Most of the reversibility effect is reached only

24 h after shifting to any of the B. subtilis diets. ****P<0.0001; n=25 worms per time point per condition.

(C) Expression levels by qRT-PCR of unc-54 and α-syn transcripts from day 1 adult worms grown

continuously on E. coli or B. subtilis PXN21 vegetative cells and after the switching to B. subtilis

vegetative cells. Expression level of each gene was normalised to E. coli levels, which was taken as 1.

n=3 per condition, with 3 technical replicates each (n represents a population of ~4000 worms). (D)

SDS/PAGE from α-syn transgenic and wild type day 1 adult worms fed on a vegetative B. subtilis diet

from L1 or L4, compared to E. coli (food switch at L4, expression levels shown in D). Arrow with *

indicates α-syn monomeric isoform. (E) Normalised quantification of protein levels of α-syn vs β-actin

signals shown in D. (F) Average number of α-syn aggregates in worms grown on E. coli, B. subtilis

PNX21, B. subtilis 168, and its derivative ΔspoIIE. As shown in Fig. 4D for pure vegetative cells, in

mixed lawns, wild type and ΔspoIIE B. subtilis 168 strains both protect from α-syn aggregation similarly

to the PXN21 strain. n=50 worms per time point per condition. Data shown are mean ± SEM from one

representative experiment out of three with similar results, unless stated otherwise. ns: no significant

differences.

Figure S4 | B. subtilis reduces α-synuclein aggregation through dietary restriction dependent and independent mechanisms. Related to Figure 5. (A) Undigested spores visible in the gut of day 1 adult

α-syn transgenic worms fed on B. subtilis PXN21 mixed cell-lawns (Nomarski images). (B) Day 1 adult

worms grown on a B. subtilis PNX21 spores-only diet from L1 larval stage show severe signs of DR

(strong developmental delay, very few eggs, and small size) (right) in comparison with animals fed on a

mixed (left) or vegetative exclusive diet (middle). (C) Developmental stage at 60 h of α-syn-expressing

worms grown on E. coli or B. subtilis mixed-cell lawns or vegetative cells (nearly 100% of worms have

reached adult stage by 60 h). n≥100 worms per condition from two biological replicates are shown. (D-G)

Quantification of brood size from L4 stage by day (D, F), or as total number (E, G), of worms fed with

the different diets. (D, F) *P=0.0113, **P=0.0019, ****P<0.0001. (E, G) ***P=0.0009, ****P=0.0001.

n≥34 worms per condition from two biological replicates are shown. (H) Average number of α-syn

aggregates in wild type or eat-2(ad456) worms grown on E. coli or vegetative B. subtilis lawns. Note that

eat-2 mutant worms grown on E. coli supplemented with arginine no longer show reduced levels of

aggregation in day 1 adults compared to wild type. ****P<0.0001; n=50 worms per time point per

condition from two biological replicates are shown. (I) pha-4 expression levels quantified by qRT-PCR

and normalized to E. coli levels of worms grown on the diet conditions shown in H. eat-2 or wild type

worms grown on arginine supplemented plates show no pha-4 upregulation in either bacterial diet

compared to E. coli. The pha-4 expression level in worms fed with E. coli was taken as 1. n=3 per

condition, with 3 technical replicates each (n represents a population of ~4000 worms). (J) Quantification

of α-syn aggregates in day 1 adult worms fed on 2x concentrated, alive or UV-killed E. coli, 48h after

seeding. ****P<0.0001; n=25 worms per condition. (K) Quantification of α-syn aggregates in day 1 adult

worms fed on different concentrations of alive or UV-killed E. coli, 24 h after seeding. ****P<0.0001,

n≥15 worms per condition. Worms grown on 2x concentrated UV-killed E. coli show no further reduction

in aggregation compared to the diluted UV-killed, indicating that the protective effect of the less

concentrated UV-killed E. coli is not due to reduced pathogenicity but due to dietary restriction. (L)

Average α-syn aggregates of worms before and after day 1 adult switching to regular or dietary restriction

inducing conditions. ****P<0.0001 indicates comparison of each diet vs its respective E. coli control; a

vs b, **P=0.0026 for E. coli to UV-killed E. coli vs E. coli; ****P=0.0001 for E. coli to B. subtilis vs E.

coli; n=25 worms per time point per condition. Data shown are mean ± SEM from one representative

experiment out of three with similar results, unless stated otherwise. ns: no significant differences.

Figure S5 | B. subtilis protects against α-synuclein aggregation by changing the sphingolipid metabolism in the host. Related to Figure 7. (A) Heatmap showing the hierarchical clustering of

samples by normalised expression values for all samples, generated by pheatmap. The 1:1 mixture of B.

subtilis PXN21: E. coli OP50 shows a gene expression profile closer to E. coli than B. subtilis. (B, C) Normalised log2 fold-change expression levels of 10 random upregulated (B) and downregulated (C)

genes from the RNAseq by qRT-PCR from α-syn-expressing young adult worms (approx. 50-52 h after

hatching). Expression level of each gene in worms fed with E. coli was taken as 1. Dashed lines indicate

the 1.5-fold change threshold (log2 FC 0.58 and -0.58). *P<0.05, **P<0.01 and **P<0.001; n=3 per

condition, with 3 technical replicates each (n represents a population of ~4000 worms). (D) Normalised

expression levels, as quantified by qRT-PCR, for a selection of dietary restriction transcription factors

from young adult worms. Expression level of each gene in worms fed with E. coli was taken as 1. The

bottom table shows the extracted values from the RNAseq analysis. **P=0.034; n=3 per condition, with 3

technical replicates each (n represents a population of ~4000 worms). (E) Summary of the top 50 non-

redundant BP GO terms of B. subtilis:E. coli mix vs E. coli by log10 P value. (F) Visualization of B.

subtilis and B. subtilis: E. coli mix vs E. coli beta scores over the sphingolipid pathway generated by

Pathview. The left and right portions of a gene box represent B. subtilis and B. subtilis:E. coli beta scores,

respectively. Red indicates a positive beta score, blue indicates a negative beta score, and grey marks

genes that are neither positively nor negatively selected. White gene boxes correspond to genes that have

not been found in the C. elegans genome. Data shown are mean ± SEM from one representative

experiment out of three with similar results, unless stated otherwise. ns: no significant differences.

Table S9. | Oligonucleotides used in this study. Related to the STAR Methods.

PCR genotyping of C. elegans mutants Primer name Primer sequence daf16(mu86) Fw1 TCCGTCACAATCTGTCTCTTCA daf16(mu86) Rv1 AAGTGTCGAGTGAAGGGAGC daf16(mu86) Fw2 CGACAAGACAGGCGGTATCC daf16(mu86) Rv2 TATCCTCTTCTTGGCTCCGC daf2(e1370) Fw GAGTCGCTCAAGTTTTGCCAT daf2(e1370) Rv CTCGACGTTCTCAACATCCG hsf-1(sy441) Fw CCGCAACAAGACTATTCGGG hsf-1(sy441) Rw ACAAATCCTCGGCTCCATCA eat-2(ad465) Fw TATGACCCAGTAGAACGGCC eat-2(ad465) Rv TGGAAGTAGTTGGTGGAGGG lagr-1(gk331) Fw AGGTGTCGGAGGTCGATG lagr-1(gk331) Rv AGTACCCGAATCGTTCTGG sptl-3(ok1927) Fw1 AGGTTCTACAACATGGGTGG sptl-3(ok1927) Rv1 GTTCCAGCAGAGTATCGACG sptl-3(ok1927) Fw2 CTTCTCGCATCGATCTGGAG sptl-3(ok1927) Rv2 TTGGGCTAACTCCACACAAC asm-3(ok1744) Fw1 ACGACCCGGAGATTAGTAGG asm-3(ok1744) Fw1 GACTTCGCGCGTAATTGAAG asm-3(ok1744) Fw2 TGTGGACCCCATCAATTGTG asm-3(ok1744) Rv2 GGGACACGGTGTGGTATAAC

PCR genotyping of bacterial mutants Primer name Primer sequence Universal16S-27Fseq-Fw AGAGTTTGATCMTGGCTCAG Universal16S-907Rseq-Rv CCGTCAATTCMTTTRAGTTT UP2-B. subtilis 168 library Fw GAGGGAGGAAAGGCAGGA UP3-B. subtilis 168 library-Rv CGCCGTATCTGTGCTCTC SpoIIE-B. subtilis 168 Fw CGAAGATTTCTTTGGTATTG SpoIIE-B. subtilis 168 Rv AATGTCCGTTGTTCACATTCA tasA-B. subtilis NCIB3610 Fw1 GGAAACAGATACAAAGGACAGC tasA-B. subtilis NCIB3610 Rv1 CATCGAGACGGCCCAGTATATG tasA-B. subtilis NCIB3610 Rv2 TCTTCTGGAGATGTATTGCTGC eps(A-O)-B. subtilis NCIB3610 Fw TCATTAACAGAAAGGGGCGC eps(A-O)-B. subtilis NCIB3610 Rv1 AACACTAGTGCAGGACCAG eps(A-O)-B. subtilis NCIB3610 Rv2 TTCCGAAGCCTTCGCCTC bslA- B. subtilis NCIB3610 Fw GTAAAGCAGAAAACGCCTGG bslA- B. subtilis NCIB3610 Rv1 GTTCCCCCGTTCGTGTTTG bslA- B. subtilis NCIB3610 Rv2 TCTGTGTTGTACGCAAGGC phrC-B. subtilis NCIB3610 Fw TCGAACGATGTTTCCAGTAC phrC-B. subtilis NCIB3610 Rv CTTGATCATTGTGGGAACTGC nosA-B. subtilis NCIB3610 Fw GTTTTCCAGGAAGTTGCAGA nosA-B. subtilis NCIB3610 Rv TGGCGGTTTCAATATGGTGA

Quantitative real time PCR (qRT-PCR) Primer name Primer sequence cdc-42 Fw CTGCTGGACAGGAAGATTACG cdc-42 Rv CTCGGACATTCTCGAATGAAG F25B5.5 Fw GAAAGCAGTGTTCGACAATTCG F25B5.5 Rv CCTATGAGTTGCGCTTTCAATG unc-54 Fw AAGACCCAGAAGAAGCAGGTTG unc-54 Rv GATCGCATCTTTGAGAGGGAGT alfasyn Fw ATGTAGGCTCCAAAACCAAGGA alfasyn Rv TGCTCCTCCAACATTTGTCACT pha-4 Fw CGGCTGTTAATCACAGTCAACC pha-4 Rv GGTAAGGAGACGCGTATTGACC sir-2.1 Fw GTTCGTCTTGCTCATCAAATGC sir-2.1 Rv GCTCGTTGCAAGTCCAGATGTA aak-2 Fw GAGGCGAGTATTGAGAAAATGG aak-2 Rv AGTCTGGGAAAGCTTAGCTCCT acs-17 Fw GTAGAGGCATTGCAAAAGGAGT acs-17 Fw TTCTTTTGAGCTTCAAGGCTTC asp-6 Fw GATTCGGACCATCATGGATTCT asp-6 Rv CGAATCCCATTCTCTTGTTTCC C28H8.5 Fw CCGAATCAAGTGAAGTTTAGGC C28H8.5 Rv TCACGGAACATAATTCGAACAG cpt-3 Fw GTACAATAGCCCGTTCCTTGAC cpt-3 Rv CAATGGAGTTTGTCATGTTTGG ctl1-3 Fw CGACTGGAGATGTTGATCGTTA ctl1-3 Rv AGCACTTTCTCCCAGAACTGAC cyp-34A7 Fw GACGAGGAAACGTTCAAAAATC cyp-34A7 Rv TCAAATACAGCTCAGCTTTTGC F56H6.9 Fw CATGCAGCTATCTCTGGGTAGA F56H6.9 Rv TGCACAATATGTTTTCCTGTCC lagr-1 Fw GACAATCGTGAATGGGACACAA lagr-1 Rv GTTCTTCGGTTTAAGATGGCAC T25B9.2 Fw GCTGGATCTTCAAATGGTTATCCGC T25B9.2 Rv CAATACTAAACACAGCCGCTGTA acdh-2 Fw GGATCAAAATGGGGAATCAGTA acdh-2 Rv GATCTCGATCCACAATGAAACA acs-11 Fw GATATTAAGGCTTGGTGCAAGG acs-11 Rv CACTTTTCCAGTCACTGTCAGC gpd-3 Fw CACACTTTGTCAAGCTCGTCTC gpd-3 Rv TAGGCCTTGGTAGCAATGTAGG R09H3.3 Fw CAACTCTCAGCCATCGTACCAC R09H3.3 Rv GTCGGTTTCCTGTTGGTCTCTC sdz-35 Fw ACGAAACAATCCAACAAAATCC sdz-35 Rv TATCCTCCTCCAACTTTTCCAA sod-3 Fw ACCTTCAAAGGAGCTGATGGAC

sod-3 Rv AGCCTTGAACCGCAATAGTGAT ugt-44 Fw AGATTTGGAAAACCACATGGAC ugt-44 Rv GCTCTTAGAACCTTCGAAACGA ver-4 Fw ACGTTCACACAAAAATCGGATG ver-4 Rv TCAAATACAGCTCAGCTTTTGC W03F11.1 Fw CAACGTAGCTGCGGAGAAG W03F11.1 Rv TTAACCTCATTTGGTGGGTAGG hcf-1 Fw ATTGCTGCAAGAAATGAAAAGG hcf-1 Rv AAACGAGCTCTCTTTTGCTGAC