The symbiont Lactobacillus plantarum causes intestinal pathogenesis in adult Drosophila.

Under specific circumstances, symbiotic bacteria can cause fulminant, and potentially life-threatening disease in animal hosts. Despite the relevance of symbiont containment for prolonged health, we know very little about the basis for symbiont-dependent disease. We used the Drosophila model to examine the impacts of individual symbiotic species on host longevity. In this study, we found that mono-association of adult Drosophila with Lactobacillus plantarum, a widely reported fly symbiont, and member of the probiotic Lactobacillus genus, curtails adult longevity and causes extensive intestinal pathology within the host. Intestinal disease includes impaired epithelial renewal, a loss of intestinal stem cells, and a gradual erosion of epithelial integrity. Our study uncovers a previously unknown pathogenic aspect of Lactobacillus plantarum association with Drosophila and establishes a simple model for the mechanistic exploration of the biology of symbiont-mediated disease.


INTRODUCTION
An unpredictable landscape of environmental inputs overlaps with the biochemical outputs of prokaryotic and eukaryotic genomes to shape the intestinal microbiome.Extrinsic factors such as oxygen and nutrient supply influence the biogeography of microbiome distribution, while bacteriostatic products such as antimicrobial peptides (AMP) and reactive oxygen species (ROS) prevent bacterial dissemination throughout the host.Combined, these interactions lead to colonization of the metazoan gut by a shifting community of microbes.The metabolic output from this community influences critical events in their host, with profound implications for processes as diverse as growth, immunity, and behavior (Hacquard et al., 2015, Hooper et al., 2012, Hsiao et al., 2013, Kamada et al., 2013a, Kamada et al., 2013b, Wong et al., 2014).This entente cordiale between a host and their prokaryotic symbionts is essential for the long-term health of an animal, as disrupted host-symbiont relationships often lead to the onset of debilitating and potentially deadly host diseases (Blumberg and Powrie, 2012, Carding et al., 2015, Kamada et al., 2013a).
Drosophila melanogaster is a particularly useful tool to study host-microbe relationships.As is often the case, flies lend themselves to sophisticated manipulations that allow investigators to control the expression of almost any gene in specific intestinal cell types (Lemaitre and Miguel-Aliaga, 2013).Of equal importance, there are extensive genetic, developmental, and biochemical similarities between key aspects of fly and mammalian gut biology (Buchon et al., 2013, Jiang and Edgar, 2012, Ma et al., 2015).Thus, discoveries in Drosophila have tremendous value for the illumination of foundational aspects of host-microbe interactions.From a physiological perspective, the posterior midgut of Drosophila is functionally similar to the small intestine of mammals.In both cases, the epithelium contains secretory enteroendocrine cells and absorptive enterocytes that arise from a transitory progenitor cell -the enteroblast of Drosophila, and the transitory amplifying cells of mammals (Lemaitre and Miguel-Aliaga, 2013).
Both types of progenitor are maintained by asymmetric divisions of basally located Intestinal Stem Cells (ISC) (Jiang andEdgar, 2012, Takashima andHartenstein, 2012).In addition, flies and mammals rely on evolutionarily conserved innate immune defenses, ROS generation, and homeostatic intestinal proliferation to prevent bacterial dissemination throughout the host (Buchon et al., 2013).
Although the microbiome of Drosophila is several order of magnitudes less complex than that found in mammals (Broderick and Lemaitre, 2012), populations of Lactobacillus species are common to fly midguts and animal small intestines.Studies of the Drosophila commensal, Lactobacillus plantarum (Lp), uncovered a remarkable degree of interaction between the two species.Lp contributes to larval growth (Storelli et al., 2011), uptake of dietary protein (Erkosar et al., 2015), and management of malnutrition in the host (Schwarzer et al., 2016).Furthermore, Lp activates host defenses such as generation of ROS (Jones et al., 2013), protection from damaging agents (Jones et al., 2015), and defenses against pathogenic microbes (Blum et al., 2013).Remarkably, many host responses to Lactobacilli are conserved across vast evolutionary distances, as Lactobacillus strains coordinate nutrient acquisition (Schwarzer et al., 2016), ROS generation (Jones et al., 2013), and gut defenses in mouse models (Jones et al., 2015).These observations position the fly as a valuable model to examine developmental and homeostatic contributions of Lactobacilli to animal health (Matos and Leulier, 2014).
Our interest in Lp arose from an earlier study where we showed that elimination of the microbiome from adult flies extended their lifespan relative to conventionally-reared (CR) flies (Petkau et al., 2014).Based on this observation, we hypothesized that one or more species of commensal bacteria shortens the lifespan of adult Drosophila.In this study, we tested common Drosophila commensals for effects on adult lifespan.Of all species tested, we found that Lp recapitulated the microbiome-mediated truncation of adult lifespan.Upon closer examination, we observed a number of severe intestinal pathologies in gnotobiotic flies associated with Lp .that included extensive damage to the midgut epithelium, loss of epithelial barrier integrity, and death of midgut progenitors.Remarkably, we found that imd mutants were insensitive to the negative effects of mono-association with Lp, suggesting that Lp diminishes adult longevity in an imd-dependent manner.To our knowledge, ours are the first studies to identify a pathobiont of adult flies, and suggest a link between commensal Lp and innate immune responses in the control of host lifespan.

Mono-association with Lp diminishes adult lifespan.
We and others found that germ-free (GF) adult Drosophila outlive CR flies (Clark et al., 2015, Petkau et al., 2014).To identify the species responsible for diminished adult longevity, we inoculated GF, wild-type flies with one of three common Drosophila commensals.Specifically, we generated gnotobiotic flies that we mono-associated with lab-specific isolates of Acetobacter pasteuranius (Ap), Lactobacillus brevis (Lb), and Lactobacillus plantarum (Lp).In each case, we worked with GF adults that we derived from CR larvae.This allowed us to examine microbial effects on adult health without compromising larval development.We then monitored longevity of the respective gnotobiotic flies to determine species-specific impacts on adult lifespan.As expected, GF flies significantly outlived CR flies (Fig. 1A, B).Among the mono-associated populations, we found that Lb and Ap had little to no effect on fly longevity (Fig. 1A, B).In both cases, the mono-associated populations had lifespans that closely mirrored GF flies.In contrast, mono-association with Lp significantly decreased adult lifespan (χ 2 = 32.30,p<0.0001) compared to CR flies, with a substantially diminished median survival of 20 days compared to 29 days for CR flies.These findings indicate that mono-association with Lp shortens the lifespan of adult Drosophila.

Abundance of Lp impacts adult fly lifespan
As CR flies significantly outlived Lp mono-associated flies (Fig. 1B), we hypothesized that additional commensal species partially dampen the detrimental effects of Lp on CR populations.To test this hypothesis, we compared the longevity of CR flies to gnotobiotic flies that we mono-associated with Lp, or gnotobiotic flies that we poly-associated with defined commensal species.For these experiments, we fed GF flies a culture of live Lp alone, or Lp mixed with equal amounts of Ap and Lb.In this assay, we noted a substantial increase in the numbers of bacteria that associated with adult guts over time (Fig. 1C).These findings overlap with earlier data that intestinal colonization by commensal bacteria increases as flies age (Blum et al., 2013, Broderick et al., 2014).However, we noticed that intestinal colonization by Lp was several orders of magnitude higher in mono-associated flies than in poly-associated flies (Fig. 1D), suggesting that partner commensals limit intestinal colonization by Lp.As Ap colonizes the gut to similar extents as Lp (Fig. 1C), we asked if Ap attenuates the pathogenic effects of Lp.To determine the impact of Ap on Lp pathogenesis, we measured the lifespan of gnotobiotic adult Drosophila that we co-associated with differing rations of Ap and Lp.Here, we observed a clear relationship between Ap:Lp input ratios and adult lifespan -the greater the ratio of Ap to Lp, the longer the lifespan of co-associated flies (Fig. 1E, F).Together, these data argue that common fly commensals such as Ap limit intestinal colonization by Lp and mitigate against the pathogenesis of Lp.

Adult midguts mount a limited defensive response against Lp
At present, we know very little about adult intestinal responses to Lp.As our data point to detrimental effects of Lp mono-association for adult flies, we examined critical aspects of intestinal responses to microbial challenges in flies mono-associated with Lp.Drosophila adults typically respond to commensal or pathogenic bacteria with production of AMP via the IMD pathway, generation of ROS, and activation of homeostatic proliferation in ISCs (Buchon et al., 2014).We first examined the production of ROS and AMP in the guts of CR, GF and Lp monoassociated flies.For these assays, we quantified the intestinal expression of dual oxidase (DUOX), a source of gut ROS, and diptericin (dpt), a reporter of IMD pathway activity.Consistent with previous reports that commensals activate antibacterial responses in the adult gut (Buchon et al., 2009), we observed elevated expression of DUOX and dpt in CR guts compared to GF guts (Fig. S1A, B).In contrast, we found that Lp mono-associated flies expressed DUOX and dpt at lower levels than CR flies (S1A, B).These data suggest that Lp mono-associated flies have a diminished antibacterial responses relative to CR flies.
We then asked if intestinal colonization by Lp stimulates a proliferative response to maintain tissue homeostasis.For this assay, we examined the EGF growth factor pathway, an integral component of proliferative responses to intestinal bacteria (Buchon et al., 2010).Specifically, we quantified expression of the EGF ligand spitz (spi) and the spi-activating endopeptidase rhomboid (rho).As expected, we detected higher levels of growth factor signaling in CR flies than in GF flies (Fig. S1C, D).In contrast, we found that spi and rho were expressed at lower levels in the guts of Lp-associated flies than in CR flies (Fig. S1C, D).In fact, we found that spi was expressed at significantly lower levels in the midguts of Lp monoassociated flies relative to GF flies after 15 days (Fig. S1C).Combined, these data suggest that the adult midgut response to Lp is phenotypically distinct to that observed in CR flies.

Lp gnotobiotic flies lack intestinal progenitor cell markers.
As our data suggest that adult midguts do not engage the EGF pathway after association with Lp, we examined the morphology of adult posterior midguts that we associated exclusively with Lp.To determine the influence of Lp on midgut morphology, we visualized the posterior midguts of CR, GF, Lb mono-associated and Lp mono-associated flies at days 2 and 15-post association.We used anti-Armadillo and anti-Prospero immunofluorescence stains in esg[ts]>GFP flies to visualize cell borders, enteroendocrince cells, and intestinal progenitor cells, respectively.We did not observe differences between the different treatment groups at the early time point (Fig. S2).In each case, the midgut contained evenly spaced nuclei, regular GFP positive progenitors, and neatly organized cell boundaries.As expected, older CR midguts showed signs of tissue dysplasia and disorganization compared to the midguts of GF flies (Fig. 2B, C).Interestingly, we found that flies mono-associated with the non-pathogenic commensal Lb were phenotypically indistinguishable from CR flies (Fig 2B , D).In both cases, posterior midguts contained unevenly spaced nuclei, irregular cell borders, and widespread GFP fluorescence (Fig. 2B, D), hallmarks of the dysplasia typically observed in older, CR flies (Biteau et al., 2008).In contrast, Lp mono-associated flies appeared to maintain a remarkably wellorganized midgut epithelium as they aged.Despite, the physiological burden of Lp monoassociation, we observed regularly spaced nuclei, defined cell borders, and an even distribution of enteroendocrine cells at day 15 (Fig. 2E).However, in contrast to the apparently pristine organization of the epithelium, we noticed a striking absence of GFP positive progenitors throughout the midgut of flies that we mono-associated with Lp for 15 days (Fig. 2E).These data suggest an unexpected impact of Lp on the maintenance of progenitor cells in adult flies.

Impaired intestinal proliferation in Lp mono-associated flies.
The apparent loss of progenitor cells in Lp mono-associated flies prompted us to examine the proliferative capacity of ISCs in midguts associated with Lp.For this study, we used the MARCM clonal marking method to assess stem cell proliferation and subsequent differentiation in adult flies.MARCM allows investigators to identify mitotic cells and their progeny as GFP positive groups (Wu and Luo, 2006).As expected, GF flies had a limited number of small mitotic clones, whereas CR flies had more clones that contained greater numbers of cells (Fig. 2F, G).In contrast, we observed a near total absence of mitotic clones in the midguts of flies that we mono-associated with Lp (Fig. 2H).The few clones that observed invariably held fewer cells than age-matched clones in CR flies (Fig 2K).In addition, cells in clones from Lp mon-associated flies were morphologically distinct from those observed in CR flies (Fig. 2I, J).These results, in conjunction with our findings regarding impaired growth factor signaling and loss of progenitor cells, suggest that adult Drosophila fail to activate proliferative homeostatic repair mechanisms in response to the growing burden of Lp.

Lp disrupts posterior midgut ultrastructure.
To fully examine the effects of Lp on adult midgut ultrastructure, we used transmission electron microscopy to visualize the posterior midguts of age-matched CR and Lp monoassociated flies.Inspection of CR midguts revealed the anticipated sheath of visceral muscle that surrounds basal progenitor cells (Fig. 3A, B).Progenitors, in turn, support the generation of columnar epithelial cells that with an apical surface that faces the intestinal lumen (Fig. 3B, C).
Upon examination of midguts associated with Lp, we were struck by substantial alterations to the intestinal epithelium.In CR flies, the epithelium consisted of continuous, regularly shaped cells with clearly visible nuclei and a flat apical surface (Fig. 3A-C).In contrast, the epithelium of Lp mono-associated flies contained an undulating population of cells (Fig. 3D, E) with large vacuoles (Fig. 3E, F) and poorly-discernible nuclei (Fig. 3D-F).We also noticed severe alterations to the morphology of presumptive progenitor cells.In place of the small, densely staining progenitors intimately associated with the visceral muscle of CR flies (Fig. 3B), monoassociation with Lp led to the appearance of misshapen cells that did not associate properly with the muscle, had large, lightly staining nuclei, and contained numerous cytosolic vacuoles (Fig. 3D, E).Combined, these findings show that colonization of the adult midgut with Lp causes a pronounced intestinal pathology, characterized by thinning of the epithelium, formation of large cytosolic vacuoles, and a loss of progenitor cells.The extent of intestinal damage led us to determine the effects of Lp mono-association on midgut barrier integrity.For this experiment, we used the smurf assay, a convenient measure of intestinal barrier function (Rera et al., 2012).
We found that mono-association of adult flies with Lp invariably caused a loss of gut barrier function immediately prior to adult death (Fig. 3G).In short, our data show that association of the adult midgut with Lp leads to a number of intestinal pathologies, loss of epithelial barrier function, and a premature death of the host.

Imd is essential for Lp-mediated pathogenesis.
In agreement with a recent study that showed partial activation of IMD responses by Lp in larvae (Erkosar et al., 2015), we noted a modest induction of dpt in Lp flies at day 15-post inoculation (Fig. S1).This prompted us to examine interactions between imd, Lp and host viability.To address this question, we determined the viability of CR, GF and Lp monoassociated imd mutants flies (Imd EY08573 ).Similar to wild-type flies, we found that GF imd mutant flies outlived CR adults (Fig. 4A, B).However, in contrast to CR wild-type flies, we did not observe a negative impact of Lp on the viability of imd mutants.Whereas Lp shortened the lifespan of wild-type flies, we found that imd mutants mono-associated with Lp significantly outlived CR imd mutants (Fig. 4A, B).Indeed, Lp only moderately shortened the lifespans of imd mutants relative to GF imd mutants (Fig. 4B).Together, these data show that IMD is required for the negative effects of Lp on adult longevity.

DISCUSSSION
Conventional narratives partition intestinal microbes into convenient functional groupspathogens that damage their host; benign passengers that exert little to no influence; or beneficial symbionts that equip their hosts with probiotic metabolites.The reality of host-microbe interactions is considerably more complex and requires an integrated appreciation of the environment, the microbiome, and the host genotype.Shifts in any one of these factors can drive dysbiosis among the commensal community and convert symbionts to pathogens (pathobionts) that fuel chronic diseases within their host (Kamada et al., 2013a).Clostridium difficile is a classic example of this paradigm: 2 -5% of the North American population carry C. difficile, often without notable effects.However, simple changes such as antibiotic exposure can eradicate neighboring microbes and permit expanded colonization of the colon by C. difficile, leading to the development of colitis in afflicted individuals (Keller and Kuijper, 2015).In this study, we used Drosophila melanogaster as a tool to identify and phenotypically characterize a novel fly pathobiont, Lactobacillus plantarum.Lactobacilli are widely reported fly and mammal commensals with numerous established benefits for their animal hosts (Walter, 2008).
To our knowledge, our data are the first to show a detrimental impact of a common Drosophila commensal bacteria on the maintenance or proliferation of ISCs.We found that mono-association with Lp impairs ISCs proliferation, damages the intestinal epithelium, causes a breakdown of gut barrier function, and shortens adult lifespan.A previous study showed that Gluconbacter morbifer, causes disease in adult Drosophila if allowed to expand within the host (Ryu et al., 2008).However, G. morbifer is a comparatively rare fly commensal, and disease onset requires impaired immunity within the host.In contrast, this report identify a novel disease phenotype associated with unchecked expansion of a common fly commensal in a wild-type host.We believe these findings represent a valuable model to define the mechanistic basis for host damage by a pathobiont.
As imd mutants appear relatively indifferent to mono-association with Lp, it appears the innate immune IMD response is critical for Lp-dependent pathogenesis in the fly.It is attractive to assume that this phenotype arises from collateral damage through deregulated expression of toxic immune effectors such as AMP or ROS.However, we believe it unlikely that IMDresponsive bactericidal effectors are responsible for the negative effects of Lp on host lifespan, as CR flies outlive Lp mono-associated flies despite increased AMP and DUOX expression.In this context, we consider it important to consider that several transcriptional studies demonstrated that a relatively small fraction of IMD-responsive transcripts are easily categorized as bacteriostatic or immune modulatory (De Gregorio et al., 2002).In fact, it seems that intestinal IMD activity primarily modifies metabolic gene expression (Broderick et al., 2014, Erkosar et al., 2014).As intestinal microbes are known to control nutrition and metabolism in their Drosophila host (Wong et al., 2014), we consider it possible that the Lp-dependent pathologies described in this study reflect an underlying imbalance in IMD-dependent regulation of host metabolism.
Consistent with possible links between Lp, IMD and host metabolism, it is noteworthy that a recent study established a direct link between Lp and the IMD-dependent expression of intestinal peptidases (Erkosar et al., 2015).Our data show that intestinal colonization by Lp is much greater in mono-associated flies than in poly-associated flies.We speculate that the elevated levels of Lp, combined with the absence of additional commensal, alters metabolic responses in the host, leading to impaired intestinal function.This hypothesis includes the possibility that Lp directly affects host diet, as proposed for other Drosophila-associated microbes (Chaston et al., 2014, Yamada et al., 2015).We are currently testing this hypothesis in flies with modified intestinal IMD activity.
Our work was initially inspired by reports from our group and others that GF adults outlive CR flies (Clark et al., 2015, Petkau et al., 2014).However, other studies reported variable impacts of the effects of microbiome removal on adult lifespan (Brummel et al., 2004, Ren et al., 2007).
We believe that the differences between the individual reports reflects the intricate nature of interactions within a host-microbe-environment triad.For example, research groups typically raise their flies on an incompletely defined, oligdic diet that exerts underappreciated influences on the metabolic outputs of intestinal bacteria, and the transcriptional outputs of the host.We believe that a complete evaluation of the relationship between microbes and their hosts requires consideration of environmental inputs such as diet.Likewise, it is important to consider genotypic inputs of the symbiont and the host.For example, the beneficial contributions of Lp to mouse and larval nutrition display strain-specific effects (Schwarzer et al., 2016), suggesting variable effects of individual Lp strains on host phenotypes.In addition, most research groups study lab-raised strains of experimental models, a situation that allows unexplored genetic drift within the host to influence microbiome composition and phenotype.
In summary, this report identifies Lp as a novel pathobiont of adult flies.To our knowledge, ours is the first study to report such a class of microbe in flies.Given the negative effects of pathobionts on animal health, and the experimental accessibility of Drosophila and Lactobacilli, we believe these findings represent a valuable tool for the definition of the mechanisms by which shifts in symbiotic populations lead to pathologies in tolerant hosts.

Fly husbandry
All experiments were performed at 29°C with virgin female w 1118 flies.Additional genotypes are provided in supplemental material.Flies were raised on standard corn meal medium (Nutri-Fly Bloomington Formulation, Genesse Scientific).Germ-Free flies were generated by raising adult flies on autoclaved standard media supplemented with an antibiotic solution (supplemental material).For longevity studies, flies were raised in vials with 10 flies per vial, and transferred to fresh vials weekly.Colony forming units per fly were determined by independent quantification of 4 replicates with 5 flies/replicate.Flies were sterilized in 50% bleach, 75% ethanol and rinsed in water.Sterilized flies were homogenized in MRS broth (Fluka Analytical) and serial dilutions of homogenate were plated on MRS to identify Lactobacillus species and GYC agar plates to identify Acetobacter species.Lactobacillus species were distinguished by colony morphology.

Generation of gnotobiotic Drosophila
Virgin females were raised on antibiotic supplemented medium for 5 days at 29°C.Flies were starved in sterile empty vials for 2 hours prior to bacterial association.Lab isolate Ap was grown in Mannitol broth at 29°C with shaking 2 days prior to association.Lab isolates Lb and Lp were grown in MRS broth at 29°C 1 day prior to association.For mono-associations, bacterial cultures were re-suspended in 5% sucrose/PBS to an OD 600 of 50.For co-associations, bacterial cultures of Ap and Lp were prepared to an OD 600 of 50 each in 5% sucrose/PBS.The bacterial cultures were then mixed at the selected ratios.For poly-associations, bacterial cultures of Ap, Lb, and Lp were prepared to an OD 600 of 50 in 5% sucrose/PBS, and mixed at a 1:1:1 ratio.For all bacterial associations, 12 flies/vial were fed 1ml bacterial suspension on autoclaved cotton plugs for 16 hours at 29°C.To ensure maintenance of mono-association or GF conditions, respective flies were homogenized in MRS broth and plated on MRS or GYC agar plates periodically (1-1.5 weeks) throughout the study.

Immunofluorescence
Flies were washed with 95% ethanol and dissected in PBS to isolate adult intestines.The complete immunofluorescence protocol is provided in the supplementary material.The primary antibodies used in this study were as follows: mouse anti-Armadillo (1:100; DSHB N2 7A1), mouse anti-Prospero (1:100; DSHB MR1A).Guts were mounted on slides in Fluoromount (Sigma-Aldrich F4680).The posterior midgut was visualized with a spinning disk confocal microscope (Quorum WaveFX, Quorum Technologies Inc.).Images were collected as z-slices and processed with Fiji software to generate a single Z-stacked image.For transmission electron microscopy, posterior midguts were excised and immediately placed into fixative.
Fixative preparation, sectioning, staining and visualization were performed at the Faculty of Medicine and Dentistry Imaging Core at the University of Alberta.

qPCR
Realtime PCR was performed on the dissected guts of adult Drosophila as described previously (Guntermann and Foley, 2011).Sequences of the individual primers is provided in the supplementary materials as methods. .

Figure 2 .
Figure 2. Lp blocks ISC proliferation.(A) Schematic representation of the generation of mono-

Figure 4 .
Figure 4. Lp pathogenesis involves IMD activity.(A) Survival curve of CR, GF, and Lp monoassociated Imd EY08573 virgin flies (100 flies/group).Table show Log-rank (Mantel-Cox) test for data in A. χ 2 and p values are relative to CR flies.(B) Quantification of fly death in CR, GF and Lp gnotobiotic w 1118 and Imd EY08573 flies.