The septate junction protein Snakeskin is critical for epithelial barrier function and tissue homeostasis in the Malpighian tubules of adult Drosophila

Transporting epithelia provide a protective physical barrier while directing appropriate transport of ions, solutes and water. In invertebrates, epithelial integrity is dependent on formation, and maintenance, of ‘tight’ septate junctions (SJs). We demonstrated that Drosophila Malpighian (renal) tubules undergo an age-dependent decline in secretory transport capacity, which correlates with mislocalisation of SJ proteins and coincident progressive degeneration in cellular morphology and tissue homeostasis. By restrictively impairing, in adult tubules, the cell adhesion protein Snakeskin, which is essential for smooth SJ formation, we observed progressive changes in cellular and tissue morphology that phenocopied these effects, including mislocalisation of junctional proteins with concomitant loss of cell polarity and barrier function. Resulting in significant accelerated decline in tubule secretory capacity and organismal viability. Our investigations highlight the tubule’s essential role in maintenance of organismal health, while providing measurable markers of compromised epithelial barrier and tissue function that manifest in advanced morbidity and death. Model for epithelial dysfunction arising from failure of smooth septate junctional complexes as a consequence of impaired Snakeskin expression.


INTRODUCTION
In multicellular animals, ageing presents as a progressive decline in tissue homeostasis and organ function, leading to increasing probability of disease and death (Rera et al., 2013b;Rose et al., 2012). Regulation of tissue homeostasis is thus critical to organismal lifespan; yet the molecular and cellular activities responsible for mediating age-dependent changes in tissue function, and how this in turn impacts organismal longevity, remain largely unexplored.
Maintenance of a healthy intestine has recently emerged as a critical determinant of lifespan across taxa, with stereotypic hallmarks of intestinal ageing including augmented stem cell behaviour, blocked terminal differentiation, activation of inflammatory pathways, reduced nutrient uptake, loss-of-barrier integrity and dysbiosis Clark et al., 2015;Davies et al., 2012;Hu and Jasper, 2017;McGee et al., 2011;Regan et al., 2016;Rera et al., 2013a;2012;Resnik-Docampo et al., 2017;Salazar et al., 2018). Notably, a causal link between these phenotypes and age-related remodelling of cell-to-cell junctions has been established in proliferative tissues such as the intestine (Clark et al., 2015;Izumi et al., 2019;Resnik-Docampo et al., 2017;Salazar et al., 2018), indicating that dysregulation of junctional proteins in self-renewing tissues may be a principal driver of ageing. However, whether age-related changes in cell-to-cell junctions also occur in tissues defined by low or no cell-turnover, like the nervous system, heart and kidneys, and whether such alterations significantly contribute to tissue degeneration and/or age-onset remains unresolved.
In insects, the renal (Malpighian) tubules (MTs) constitute the functional analogue of the vertebrate kidneys (Cohen et al., 2020;Dow et al., 2018) and are, as in vertebrates, considered to be a non-proliferative tissue (Skaer, 1993).
The MTs are the principle organs responsible for maintaining water and ion homeostasis, yet serve additional roles in xenobiotic detoxification (Yang et al., 2007) and immunity (Davies et al., 2012;Verma and Tapadia, 2014). In the fruit fly Drosophila melanogaster, the MTs are comprised mainly of two physiologically distinct secretory cell-types; the principal (PC) cell and intercalated or 'stellate' (SC) cell (Beyenbach et al., 2010;Dow, 2012;Sözen et al., 1997). PCs are the sites of active cation transport, energised by V-ATPases localized apically to a prominent brush border (Halberg et al., 2016), whereas the smaller SCs control channel-mediated Cland water fluxes (Beyenbach et al., 2010;Cabrero et al., 2020;Dow, 2012). Primary urine production in Drosophila is achieved through the integrated actions of PCs and SCs, driving transepithelial transport of ions and water. A third population of 'tiny' cells, found only in the ureter and lower tubule, have been proposed to be renal stem cells (Martínez-Corrales et al., 2019;Singh and Hou, 2009;2008;Singh et al., 2011;Takashima et al., 2013;Wang and Spradling, 2020), yet to what extent these potentially proliferative cells contribute to tissue repair/homeostasis is unknown.
Here, we demonstrate for the first time that the adult MT undergo an age-dependent decline in secretory transport capacity, which correlates with mislocalisation of septate junction proteins and coincident progressive degeneration in cellular morphology and overall tissue homeostasis. We further show these effects are phenocopied by acute loss of Ssk expression, in either PC or SC sub-populations, of adult MTs. Ssk impairment in either cell population resulting in mislocalisation of junctional components, again manifesting in overt degeneration in cellular and tissue morphology. Critically, this acute failure of junctional integrity leads to an accelerated reduction in secretory capacity and concomitant loss of systemic fluid homeostasis, which ultimately results in a significant reduction in organismal lifespan. Furthermore, cell-specific manipulations of Ssk expression led to a pronounced increase in SC clustering, a block in SC maturation and a loss in apicobasal polarity, indicating a key role for Ssk in maintaining MT function and stability. Finally, knocking down Ssk expression also led to a striking proliferation of tiny (renal stem) cell as well as a dramatic increase in tubule tracheation, suggesting that MTs can autonomously respond to tissue damage and that Ssk acts as a novel regulator of tissue homeostasis in the tubule.
Taken together, our work demonstrates a crucial link between cell-cell junction integrity, epithelial transport competence and tubule homeostasis in a classically non-proliferative tissue, which provide novel insights into the mechanisms underlying tissue degeneration and age-onset.

Cell-specific Ssk depletion impairs systemic osmoregulation and reduced organismal lifespan
Failure to form sSj's properly in early development, either ubiquitously or restricted to the MTs, is lethal (Beyenbach et al., 2020;Jonusaite et al., 2020;Izumi et al., 2019;Yanagihashi et al., 2012). We therefore employed temperature-sensitive tubulin GAL80 (McGuire et al., 2004), in conjunction with either a SC-specific, c724GAL4 (Sözen et al., 1997), or PC-specific, UroGAL4 (Terhzaz et al., 2010), GAL4 driver acting on UAS-Ssk RNAi  to restrictively knock down Ssk expression in adult tubules (Figure S1A) (full genotypes see Methods). At the permissive temperature (18ºC), GAL4 expression is repressed and c724 ts >Ssk RNAi and Uro ts >Ssk RNAi (respectively designated as SC SskRNAi and PC SskRNAi ) flies develop into viable, fertile adults ( Figures 1A and S1A). However, when transferred to the restrictive temperature (29ºC) at the pre-pupal stage GAL4-drives expression of Ssk RNAi in a tissue and developmentally restricted manner ( Figure S1A-B). These SC SskRNAi and PC SskRNAi experimental animals successfully eclosed as adults. However, over time they progressively developed a 'bloated' phenotype due to increased water content ( Figure 1A), a clear indicator of compromised MT epithelial function which resulted in significantly reduced viability ( Figure 1B).

Cell-specific Ssk depletion compromises cellular morphology
We then set out to determine what specific cellular deficits occur when Ssk is depleted in adult MTs that might contribute to this compromised epithelial function. When Ssk is knocked down in either SCs or PCs of adult tubules, junctional complex organisation -as realised by staining for Discs large (Dlg), a junctional protein required for structure, cell polarity, and proliferation control in epithelia (Bilder et al., 2000;Khoury and Bilder, 2020;Woods et al., 1996) is overtly compromised compared to the more regular distribution exhibited in controls ( Figure 2A As it has been shown that dysregulation of Dlg results in disruption of micro-filament maintained cytoarchitecture (Izumi et al., 2019;Woods et al., 1996), we examined MTs in which Ssk had been knocked down and found Factin micro-filaments were clearly absent in SC SskRNAi SCs, while filaments associated with bicellular boundaries appeared reduced ( Figure 3B). This loss of internal cytoarchitecture is almost certainly a contributing factor in the failure of SC SskRNAi SCs to develop mature stellar morphology. This loss of cytoarchitecture was also apparent in PC SskRNAi MTs ( Figure S2C).
Dysregulation of Dlg, and the associated loss of cytoarchitecure, also results in failure of apicobasal polarity (Laprise and Tepass, 2011). This is consistent with the observed loss of apical bias in GFP expression in experimental PC SskRNAi MTs as compared with controls ( Figure S2A-B). To further test this idea we employed an antibody to the Na + /K + ATPase alpha-subunit -a transporter known to localize to the basolateral membrane (Patrick et al., 2006) -as an indicator of loss of apicobasal polarity. These data showed an overall reduction of Na + /K + ATPase expression in SC SskRNAi MTs as compared to controls, with a marked decrease in basal localisation in SCs indicative of a loss of apicobasal polarity ( Figure S3). This loss is perhaps not surprising considering the functional requirement for proper formation of junctional complexes in determining overall cellular polarity (Bonello et al., 2019;Nelson, 2003), but importantly these phenotypes appear progressive, becoming more overt as experimental adults age.
In SCs, kinin-modulated Clshunt conductance occurs specifically through the chloride channel ClC-a localizing to the basolateral membrane (Cabrero et al., 2014(Cabrero et al., , 2020. We therefore employed anti-ClC-a to assay polarity in SC populations in experimental adult MTs and found ClC-a expression absent in the cuboidal SC population in SC SskRNAi MTs ( Figure S4B) but apparently unaffected in SCs in PC SskRNAi MTs ( Figure S4C). To test the physiological significance of these cellular defects, we exposed animals to conditions known to induce osmotic stress and assayed for organismal survival.
These results revealed that SC SskRNAi adults demonstrated a significant reduction in survival when allowed access to water only (non-desiccating starvation) ( Figure S5A), but intriguingly this compromised viability is absent when exposed to high salt loading ( Figure S5B). These data are consistent with the observed defects in SC function and compromised hormone-induced change in Cland water fluxes in SC SskRNAi animals (Cabrero et al, 2014;Denholm et al., 2013;Feingold et al, 2019), which result in a reduced capacity to respond and adapt to hypoosmotic challenges and regulate systemic fluid balance. Taken together, our data demonstrates a necessary role for Ssk in maintaining the SC transport machinery and by extension tubule transport competency and organismal fluid balance.

Cell-specific depletion of Ssk results in absence of septa and impairment of junctional, cellular and tissue organisation
We next set out to determine the cellular mechanisms through which knock down of Ssk in specific sub-population of cells in the adult MTs could then affect overall tissue morphology and homeostatic capabilities. While both SC SskRNAi and PC SskRNAi MTs appear hyperplastic ( Figures 2B and S2B), there is only a small, though significant, increase in overall population of cells in PC SskRNAi MTs ( Figure 4A). In contrast there is a slight decrease in the SC population in anterior SC SskRNA MTs ( Figure 4A-B), presumably due to progressive loss of SCs through apoptosis.
In SC SskRNAi SCs, not only is mature stellar morphology affected, but so too is spatial distribution; with a significant increase in the %SC population exhibiting clustering of 2 or more cells ( Figure 4C). Tubules are traditionally regarded as developmentally 'static', that is the SC population has intercalated and been positioned within the tubule primordia by mid-embryogenesis and are thought merely to 'mature' physiologically (developing their characteristic stellar or bar morphology during final stages of pupariation) (Denholm, 2013;Denholm et al., 2013;2003;Dow, 2012). This dysregulation in positioning of the SC population is indicative that the processes that determine, and maintain, the MTs cellular architecture may extend past embryogenesis in the fly's development.
Both SC SskRNAi and PC SskRNAi MTs also have an increased crosssectional diameter ( Figure 4D), a phenotype iterating pathological changes observed during intestinal barrier dysfunction (Izumi et al., 2019;Rera et al., 2013a;Salazar et al., 2018) and when Mesh or Tsp2A are impaired in MTs (Beyanbach et al., 2020;Jonusaite et al., 2020). Complementing these observations, while TEM scans show the septal junction span is significantly reduced ( Figure 4E), the junctional spans are in fact highly irregular (as evidenced by the frequency-distribution in septal junction measurements; Figure 4F). Importantly there is a complete absence of septa ( Figure 4H and H'), a consequence of which is loss of para-cellular barrier function (Baumgartner et al., 1996;Genova and Fehon, 2003;Lamb et al., 1998).
Absence of septa does not itself result in loss of septal-gap structural integrity, as demonstrated by septal mutants such as sinuous and coracle (Lamb et al., 1998;Wu et al., 2004). Rather, structural integrity is a function of the adherens junctions (AJ) (Tepass and Hartenstein, 1994). Dlg localisation however has been shown to be regulated by, and in turn modulate, AJ formation (Bilder et al., 2000;2003;Bonello, et al., 2019;Harris and Peifer, 2004) Taken together these results demonstrate that cell-specific depletion of Ssk not only compromises individual cellular junctional complexes but affects tubule morphology and organisation as a whole, compromising barrier integrity and homeostatic capabilities and allowing opportunistic pathogenic invasion, all of which contributes to significantly reduced organismal viability. That our findings show impairment of junctional complexes in a restricted population of cells affects overall tissue function must also speak to the idea that these cellcell 'tight' junctions are required to direct appropriate non-cell autonomous communication(s).

Cell-specific Ssk depletion results in proliferation of tiny cells and hyperplasia of trachea supplying the MTs
We assessed the efficacy of the Ssk RNAi transgene via qPCR analysis and immunolocalisation using an antibody specific to Ssk. We were able to confirm a tissue-specific knockdown of Ssk of ~35% in 5 day old adult SC SskRNAi MTs as compared with controls (  Figure S6C). Application of an antibody specific to Mesh, Ssk's obligate partner , demonstrated that distribution of Mesh at the junctional complexes appeared reduced and disorganised in SC SskRNAi MTs, but did not manifest the same increased expression in hyperplastic trachea as observed with Ssk ( Figure 5D).
In both SC SskRNAi and PC SskRNAi MTs we also observed an increased number of nuclei associated with 'tiny cells' in the proximal area of the tubule and ureter (Figures S7). These tiny cells are described as renal and nephritic stem cells (Martinez-Corrales et al, 2019;Singh and Hou, 2008;2009;Singh et al., 2011;Wang and Spradling, 2020) and are observed to increase in number in response to stress or as MTs degenerate over time. In keeping with these observations, some of these cells co-express the proliferative cell marker Delta It is intriguing that the observed proliferation of (potential) stem cells and hyperplasia of associated trachea is coincident with mislocalisation of the cell growth regulator Dlg (Bonello et al., 2019;Elsum et al., 2012;Humbert et al., 2008;2003;Khoury et al., 2020). Whether these effects are causally related or merely coincidental, reflecting overall failure in tissue homeostatic functions due to dysregulation of the junctional complexes, warrants further investigation.

Ssk impairment accelerates age-related changes in tubule septate junction integrity and secretory capacity
To assess potential age-related changes in tubule physiology, we examined junctional integrity by analysing Dlg localization as well as renal secretory capacity, at progressive time-points during ageing. We observed a significant decrease in Dlg intensity associated with cell-cell junctions in older as compared with younger animals ( Figure 6A-B), as well as a progressive decline in secretory activity, as evidenced by significant reductions in both basal and stimulated rates of tubule secretion ( Figure 7A). These data suggest that junctional stability declines with age and that there is a measurable 'natural' deterioration in transport capacity of the tubule that correlates with physiological ageing ( Figure 7A-B).
We also investigated whether compromising junctional integrity, through cell-specific depletion of Ssk, could phenocopy these age-related changes in tubule function. Knocking down Ssk expression, in either SC SskRNAi or PC SskRNAi MTs, resulted in an accelerated and progressive decline in tubule secretory capacity relative to age-matched controls ( Figures 7A and S8). This decline is evident as early as 15 days post-eclosion, with this decrement in function mirrored by an associated degeneration in tissue morphology; with overt signs of apoptosis, mislocalised cytoplasmic Dlg accretions and a dramatic increase in 'tiny' cells ( Figures 6A-B, 7A-B). Compellingly, when comparing secretory rates between 45 day old control and 15 day old experimental animals, while the experimental animals' basal secretory rate were marginally better, there was no significant difference in stimulated rates of tubule secretion between control and experimental animals ( Figure S8). That is, the 15 day old experimentally aged tubule recapitulated the reduced secretory capacity of 45 day old controls that was a result of 'natural' physiological ageing. That compromised junctional integrity resulted in an accelerated loss in tissue homeostasis is supported by the observed significant increase in Dlg mislocalisation in SC SskRNAi MT cellular junctions as compared with controls of the same age ( Figures 6A-B).
Our investigations demonstrate, for the first time, a measurable agedependent decline in secretory transport capacity in adult tubules, that correlates with mislocalisation of SJ proteins and coincident progressive degeneration in cellular and tissue morphology. Critically, by cell-specific depletion of Ssk, we were able to phenocopy the failure of junctional complexes, leading to an accelerated reduction in secretory capacity. Providing insight into the mechanism by which failure in barrier integrity can advance loss of systemic fluid homeostasis, morbidity and, ultimately, death.

DISCUSSION
Transporting epithelia must be able to proliferate appropriately in response to growth directives (organogenesis, wound response etc.), while acting to ensure cellular compartmentalization to ensure key systems are protected from physiological and xenobiotic stresses and provide a physical barrier against invasive pathogens, all the while allowing appropriate transport of ions, solutes and water. Pivotal to these diverse functional demands is the formation, and maintenance, of 'tight' cell-cell junctional complexes. The profound consequences associated with a failure in these processes have been demonstrated in intestinal epithelia dysfunction, resulting in physiological decline contributing to onset of age-related disease and, ultimately, death Clark et al., 2015;Davies et al., 2012;Hu and Jasper, 2017;McGee et al., 2011;Regan et al., 2016;Rera et al., 2013a;2012;Resnik-Docampo et al., 2017;Salazar et al., 2018).

Ssk depletion phenocopies natural age-related changes in tubule junctional integrity affecting organismal viability
Our investigations demonstrate that a measurable 'natural' decline in tubule transport capacity, which correlates with physiological ageing, occurs over time.
This decrement in tubule function occurs in conjunction with mislocalisation of cellular junctional components, indicative of a failure in cell-cell junctional integrity, resulting in a progressive decline in organismal homeostatic capabilities and viability. This is supported by Salazar et al., 2018, who & Tepass, 2003;Woods et al., 1996;Wu & Beitel, 2004) as evidenced by the absence of basolateral expression of Na + /K + ATPase. That cellular function and polarity is compromised is further demonstrated by the absence of the SC-specific chloride channel (ClC-a), normally present at the basolateral membrane (Cabrero et al., 2014), in the SCs of SC SskRNAi animals.
The absence of Clc-a in the SCs also means that the Clshunt, necessary to create the osmotic gradient required for water flux (Cabrero et al., 2020;Denholm et al., 2013), is compromised, which again speaks to the decrement in osmoregulatory capacity observed in these tubules.

Impairment of polarity cues and cell growth regulators is a necessary
presage for epithelium-mesenchymal transition in developing cancers (Elsum et al., 2012;Humbert et al., 2008;2003;Royer and Lu, 2011;Wodarz and Näthke, 2007). It is of note that, during development, the SC cell population migrates and intercalates into the tubule primordia, undergoing a programmed mesenchymal-epithelium transition (Denholm et al., 2003). It would be tempting to speculate that in SC SskRNAi MTs, should the mutant SCs not undergo premature apoptosis, this loss of polarity, in conjunction with failure to develop a mature stellar morphology, presentation of an extruded cellular profile and breakdown in cell-to-cell communication would indicate a shift towards development of further tumorigenic characteristics.

Failure of septa and junctional structure results in compromised tubule barrier integrity
The MTs' compromised fluid integrity is due to a loss of septa and therefore para-cellular barrier function. While reduction in viability in experimental animals must, initially at least, be a consequence of the significant progressive decline in osmoregulatory capacity resulting in physiological failures, this effect is compounded by loss of barrier functions, which allows opportunistic invasion of pathogens and has been shown to lead to dysbiosis and activation of immune/inflammatory responses Clark et al., 2015;Izumi and Furuse, 2014;Rera et al., 2013b;Resnick-Docampo et al., 2018;Salazar et al., 2018;Xu et al., 2019). These effects phenocopying those described for age-related loss of intestinal integrity (Rera et al., 2012;Resnik-Docampo et al., 2017;Salazar et al., 2019).
Concomitant with loss of septa was an observed irregularity of the junctional spans, indicating a loss in junctional structural integrity, likely due to impairment of AJs. The AJs also provide cues for appropriate localisation of, and may then be modulated by, Dlg in junctional complexes (Bilder et al., 2000;2003;Bonello et al., 2019;Harris and Peifer, 2004). AJs are also required during the integration of the SCs in the developing tubule, for proper localisation and polarity of the SC population during formation of the junctional complexes (Campbell et al., 2010). Therefore, any compromise of these structures may again contribute to the observed atypical SC distribution, loss of polarity and failure of the junctional complexes. Further, the fact that increased clustering of SCs occurs when Ssk expression is knocked down in pupal and adult stages might indicate that the MTs are not completely developmentally static and are able to respond to environmental cues in order to assure tissue homeostasis throughout the animal's life.

Cell-specific depletion of Ssk in tubule epithelium has global consequences
Impairment of the sSJ junctional complexes, with consequent loss of cell cytoarchitecture and polarity, has profound effects on the individual cells.
However, these effects are not limited to the cells in which Ssk expression has been specifically impaired, but rather affects morphology and function across the tubule, with this degradation of epithelium competence affecting viability.
That impairment of junctional complexes in a specific sub-population of cells may then affect overall tissue function and morphology speaks to the idea that non-cell autonomous communication(s) are required across these 'tight' junctions. This is supported by the observations that mutant gut clonal cells in which Tsp2A expression is impaired induced non-cell autonomous stem cell proliferation (Izumi et al., 2019) and that mesh knock-down in PCs resulted in dysmorphic SC development (Jonusaite et al., 2020).
Realisation of the apparent dysregulation of Dlg, a known tumoursuppressor gene (Elsum et al., 2012;Humbert et al., 2008;2003), in individual junctional complexes may also speak to the effects that occur more universally throughout the mutant MTs. Polarity proteins have been shown to directly modulate signalling pathways controlling tissue growth (Royer and Lu, 2011;Tepass et al., 2001) and certainly one of the most striking global effects observed in both SC SskRNAi and PC SskRNAi MTs is the proliferation of 'tiny' cells at the ureters and the hyperplasia that occurs in the associated trachea supplying the MTs. Whether the trachea hyperplasia is an indirect effect of local tissue hypoxia due to a compensatory upregulation of metabolic function, as seen during fly development (Texada et al., 2019), or is a direct consequence of impairment of appropriate tissue growth regulation due to dysregulation of the junctional complexes warrants further investigation. However, it is clear that the progressive proliferation of the tiny cells, which have been described as potential renal and nephritic stem cells (Martinez-Corrales et al., 2019;Hou, 2008, 2009;Singh et al, 2011;Wang and Spradling, 2020) Altogether, our findings describe an essential role for Snakeskin in appropriate formation of smooth septate junctions, ensuring epithelial integrity and acting, directly or indirectly, as a modulator for tissue growth in Drosophila Malpighian tubules. We demonstrate that a significant and progressive decline in secretory transport capacity in the adult MTs occurs naturally over time as result of, or in conjunction with, a failure of cellular junctional components, with the overall tissue displaying progressive degeneration in cellular morphology.
We demonstrate that cell-specific impairment of Ssk in the adult MTs accelerates the mislocalisation of junctional components and degenerative effects in tubule morphology, with an associated measurable decline in functional capacity, resulting in significantly reduced viability. These investigations highlight the pivotal role that cell-cell junctional integrity plays in assuring epithelial competence, providing insight into how age-related progressive degeneration of junctional components can lead to failure of cellular, tissue and organismal homeostasis resulting in advancing morbidity and, ultimately, death.

Acknowledgements
We would like to thank Prof. Mikio Furuse for the kind supply of fly lines and reagents. We would also like to thank Dr. Selim Terhzaz for invaluable help with

Immunocytochemistry
Unless stated adult flies were reared at 29°C and aged for 5 days prior to dissection in Schneider's Drosophila medium (Gibco, Thermo Fisher Scientific).

Fly Weight Measurements.
To measure wet-body weight, 20 female flies (n=8-12 groups) were anesthetized, transferred to pre-weighed Eppendorf tubes and weighed. For dry-body weight, flies were killed by freezing for 20 min and then dried at 60°C for 48 hr. Dry flies were weighed after reaching room temperature. All weighings were performed on a GR-202 analytical balance (A&D Instruments Ltd., Abingdon, UK).

Survival assays
All survival assays comprised a minimum of 180 flies (n= 30 flies per vial, 3 vials per sex) and were performed in 12:12 hr LD at the restrictive (29 °C) temperature. Starvation assay vials contained 7 ml of 1% LMP-Agarose in water. Sucrose assay vials contained 7 ml of 1% LMP-Agarose in water with addition of 5% sucrose. Osmotic stress assays comprised vials containing either 7 ml of standard fly medium with the addition of 3% NaCl or with the addition of 1% H2O2. The age of the flies ranged between 3 -5 days. Data for all assays were plotted as Kaplan-Meier curves and analysed using the Mantel-Cox (Log-rank) test using Prism v6.0 (GraphPad, CA, USA).

Ramsay fluid secretion assay
Fluid secretion assays using Drosophila Malpighian tubules were performed as described previously (Dow et al., 1994). Malpighian tubules were dissected in ice cold Schneider's medium and transferred to a 9 µl drop of 1:1 of Schneider's medium and Drosophila saline. Baseline secretion was measured every 10 minutes for 30 minutes, after which 1 µl of DromeKinin peptide (DK; 10 -6 M) (Terhzaz et al., 1999) was added to the drop and stimulated secretion was measured every 10 minutes for a further 30 minutes. Percentage change of basal secretion rates were calculated as previously shown (Dow et al., 1994).

Author contributions
AJD, KAH and JATD designed and conceptualized the study with input from SAD. AJD, KAH and LKB performed the experiments. AJD and KAH analyzed the data, wrote the manuscript and produced the figures. All authors reviewed and edited the manuscript.