A glucose-supplemented diet enhances gut barrier integrity in Drosophila

ABSTRACT Dietary intervention has received considerable attention as an approach to extend lifespan and improve aging. However, questions remain regarding optimal dietary regimes and underlying mechanisms of lifespan extension. Here, we asked how an increase of glucose in a chemically defined diet extends the lifespan of adult Drosophila melanogaster. We showed that glucose-dependent lifespan extension is not a result of diminished caloric intake, or changes to systemic insulin activity, two commonly studied mechanisms of lifespan extension. Instead, we found that flies raised on glucose-supplemented food increased the expression of cell-adhesion genes, delaying age-dependent loss of intestinal barrier integrity. Furthermore, we showed that chemical disruption of the gut barrier negated the lifespan extension associated with glucose treatment, suggesting that glucose-supplemented food prolongs adult viability by enhancing the intestinal barrier. We believe our data contribute to understanding intestinal homeostasis, and may assist efforts to develop preventative measures that limit effects of aging on health.


INTRODUCTION
As nutrition has established impacts on health, optimizing feeding regimes to promote healthy aging has received considerable attention (Kalache et al., 2019). Nutritional deficiencies increase risk of developing a number of age-related chronic diseases, but we have limited understanding of dietary interventions that counter age-dependent deterioration of tissue and organ function (Shlisky et al., 2017). Model organisms, including Drosophila melanogaster, are excellent tools to study interactions between nutrition and organ function with age (Fontana and Partridge, 2015;Lee et al., 2015;Piper and Partridge, 2017). Flies are a genetically tractable system that uses evolutionarily conserved pathways such as the insulin and TOR responses to control nutrient sensing, acquisition, and use. Importantly, researchers can grow flies on chemically defined holidic media that allow investigators to quantify effects of macronutrients on health and lifespan (Piper et al., 2014). Nutritional geometry work emphasized the importance of relative macronutrient levels for Drosophila fitness, and revealed that low protein to carbohydrate ratios extend longevity, with maximal benefits at approximately a 1 to 16 protein to carbohydrate ratio (Lee et al., 2008;Simpson and Raubenheimer, 2009;Solon-Biet et al., 2015). Notably, low protein to carbohydrate ratios also extend longevity in mice (Solon-Biet et al., 2014), suggesting a conserved effect of protein to carbohydrate ratios on animal lifespan.
Work with flies has provided mechanistic insights into the deleterious consequences of excess carbohydrates (Graham and Pick, 2017). For example, flies raised on a high-sucrose diet (1.0M compared to 0.15M controls) have increased weight, alongside elevated triglyceride stores, and insulin resistance (Musselman et al., 2011). High-sucrose treatment (1.0 M compared to 0.15 M controls) reduces Drosophila lifespan (Na et al., 2013), even with transient exposure (1.2 M compared to 0.15 M controls) in young adults (Dobson et al., 2017). However, supplementation with water rescues survival during highsucrose treatment, while weight and insulin activity remain affected, suggesting that dehydration rather than increased sugar levels reduces lifespan (van Dam et al., 2020). Conversely, decreased sucrose supplementation (0-5 mM compared to 50 mM controls) reduces median lifespan in female flies raised on a holidic diet, while higher levels of sucrose (75-100 mM) have no effect (Wu et al., 2020). On a synthetic diet, higher sucrose (5.3% compared to 1.3%) extends median lifespan of Oregon-R females (Reis, 2016). Recently, we found that addition of 0.56M glucose to a holidic medium that contains 0.05M sucrose extends Drosophila lifespan through an unknown mechanism (Galenza et al., 2016). Here, we used a combination of genomic, cellular, and metabolic assays to suggest possible roles for the intestinal epithelial barrier in glucose-dependent extension of longevity.
Biology Open • Accepted manuscript

Glucose-supplemented holidic food promotes maintenance of energy stores with age.
In a longitudinal study of relationships between nutrition, age, and metabolism, we found that, regardless of genetic background, glucose-supplemented (100 g/L) holidic food extends the lifespan of adult Drosophila compared to unmodified holidic food, particularly in males (Galenza et al., 2016). As prolonged consumption of sugar-rich food is typically associated with diminished health and lifespan outcomes, we asked how addition of glucose extends longevity. Before addressing this question, we first tested a range of glucose concentrations to identify the optimal amount required for increased longevity.
Specifically, we measured longevity of flies raised on holidic food that we supplemented with 0 to 200 g/L glucose. We found that addition of 50 g/L glucose had the greatest effect, leading to a 27% increase in median lifespan compared to unmodified food ( Fig. 1). Thus, for the remainder of this study we determined the effects of holidic food (HF), and 50 g/L glucose-supplemented holidic food (GSF) on health and longevity.
We then quantified the impact of added glucose on metabolism by comparing weight and macronutrient content in wild-type flies raised on HF or GSF for 20 or 40 days. For each measurement, we performed a 2-way ANOVA to analyze the contributions of age and diet to any detected changes. We found no differences in weights between diets at either at day 20 or 40, and weight increased significantly with age on both diets ( Fig. 2A). Age did not appear to impact protein levels, and protein levels were not affected by GSF-treatment at either age, though they trended lower in flies raised on GSF compared to HF (Fig. 2B). In contrast, we found that diet significantly affected glucose level, as 40-day-old flies raised on GSF had higher glucose than those raised on HF (Fig. 2C). Age and interactions between age and diet also affected changes in triglycerides (Fig. 2D). For both diets, triglyceride levels declined with age, although the decrease was more pronounced in flies raised on HF than GSF, and GSF-fed flies had significantly more triglyceride by day 40. The effects of GSF-treatment on macronutrients reported here largely align with our previous observations using a higher sample size (n=5), where we observed similar effects on weight and glucose levels, though subtle differences exist for protein and triglycerides (Davoodi et al., 2019). In our previous study, we observed that protein levels were reduced significantly with both GSF-treatment and age, and age-dependent triglyceride decreases were more pronounced in flies raised on HF. Despite these differences, our data collectively show that GSF-treatment enhances maintenance of energy stores in older flies.

Biology Open • Accepted manuscript
As GSF elevated total glucose content, we asked if GSF also impacted circulating glucose and trehalose, the primary blood sugar in insects. We found that diet had a mild effect on total circulating sugars, in older flies (Fig. 2E). Focussing on component circulating sugars, this difference is likely attributable to increased free glucose (Fig. 2F), with no detectable effects on trehalose (Fig. 2G). Combined, our data suggest that increased GSF-treatment contributes to the maintenance of energy-rich triglycerides and sugars, particularly as flies age.
Glucose-supplemented food increases calorie intake.
As our flies are fed ad libitum, we do not know if GSF-dependent effects on macronutrients are the indirect result of changes in feeding. We consider this an important question to address, as calorie intake and feeding frequency have been associated with lifespan changes in several experimental organisms (Fontana and Partridge, 2015).
To measure feeding frequency, we used the flyPAD (Itskov et al., 2014) to count individual sips; bursts, which are clusters of sips; and bouts, which are clusters of bursts, in flies raised on HF or GSF. For this assay, we raised flies on their respective foods for 20 days, then starved them for 2 hours prior to feeding in a flyPAD arena for 1 hour. We saw no difference in sips (Fig. 3A), bursts (Fig. 3B), or bouts ( Fig. 3C), between flies raised on HF or GSF, suggesting that GSF does not significantly alter feeding behavior over short periods. However, it is worth considering that food contact may not correlate with consumption.
To address this and determine if GSF impacts feeding behavior over longer timeframes, we used the capillary feeding (CAFE) assay (Ja et al., 2007), to calculate food consumption across three days. In the CAFE assay, flies are fed through capillary tubes that allow us to quantify liquid food consumption. We raised flies on HF or GSF for 20 days before transfer to the CAFE setup for a 3-day period, where flies were fed a liquid version of their respective food. We found that flies raised on HF consumed a greater volume than those raised on GSF, about a 1.2-fold daily increase (Fig. 3D). Accounting for macronutrient composition, this translates to a 2.3-fold increase in calorie intake for GSF-treated flies compared to HFtreated (Fig. 3E). The increased calorie intake is a result of elevated carbohydrate consumption, as flies raised on GSF consumed approximately 3.2-fold more calories from carbohydrates per day than their counterparts raised on HF (Fig. 3F). Conversely, amino acids provided approximately 20% fewer calories to flies raised on GSF than on HF (Fig. 3G). We note that both the flyPAD and CAFE assays are performed in different conditions than those used for the lifespan studies, so it is worth consideration that these data may not reflect the standard lab conditions used in this study. Nonetheless, our data show that flies Biology Open • Accepted manuscript raised on GSF consume significantly more calories in the form of carbohydrate, and fewer in the form of protein than flies raised on HF.
To test if the lifespan extension observed for flies raised on GSF is simply a consequence of feeding adults a higher calorie food, we measured lifespans of flies raised on modified holidic food isocaloric to GSF, where extra energy was provided either from lard, or casein. As expected, flies raised on GSF lived significantly longer than counterparts on HF (Fig. 3H). In contrast, casein-supplemented holidic food had no detectable effects on lifespan, whereas lard-supplemented holidic food shortened lifespan (Fig. 3H), and significantly increased the risk of early death (Fig. 3I). As increased levels of protein often decrease lifespan (Lee et al., 2008;Simpson and Raubenheimer, 2009), it was unexpected that the supplementation with casein had no effect. It is possible that the increased calories from casein offsets the effect of increased protein to carbohydrate ratio on lifespan, although further studies are required to test this hypothesis. However, as lard supplementation decreased lifespan, and protein supplementation had no effect, we conclude that simply adding extra calories to HF is not sufficient to extend longevity, indicating that GSF extends lifespan through a more specific mechanism.
Glucose-supplemented food extends lifespan independent of insulin activity.
As we observed increased total and circulating glucose in flies that we raised on GSF, we wondered what effects GSF has on the insulin pathway, a known modifier of longevity (Clancy et al., 2001;Tatar et al., 2003).
To answer this question, we quantified transcription of the insulin-like peptides (Ilp) ilp2, ilp3, and ilp5, in flies raised on HF or GSF for 20 or 40 days. Expression of ilp2 and ilp5 was lower in 40-day old flies raised on GSF compared to flies raised on HF (Fig. 4A,C), while expression of ilp3 was unaffected (Fig. 4B).
In flies, ilp gene expression is complex, and does not necessarily reflect amounts of peptide in storage, or circulation (Park et al., 2014). Therefore, we used an ELISA to quantify total, and circulating amounts of FLAG and HA epitope-tagged Ilp2 (Ilp2-FH) in flies raised on GSF or HF. In this line, Ilp2-FH expression is controlled by the ilp2 promoter, and accurately reports Ilp2 peptide levels (Park et al., 2014). We observed significantly lower total amounts of Ilp2-FH in GSF-treated flies compared to age-matched HFtreated controls (Fig. 4D). However, we did not detect food-specific effects on levels of circulating Ilp2-FH (Fig. 4E).
To determine if GSF-dependent shifts in insulin peptide expression translate into effects on insulin activity, we measured starvation resistance and oral glucose tolerance in flies raised on GSF and HF. In Biology Open • Accepted manuscript flies, insulin impairs starvation resistance, and improves glucose tolerance. Thus, we expect that any effects of GSF on insulin signaling will have measurable impacts on starvation resistance or glucose tolerance. For starvation assays, we raised flies on HF, or GSF, for 15 or 30 days, and followed survival after switching to nutrient-deficient medium. For both ages, we did not detect food-dependent effects on starvation resistance (Fig. 4F,G). Interestingly, increased triglyceride levels, as observed in GSF-fed flies, typically corresponds with increased starvation resistance (Heier and Kühnlein, 2018), though we did not observe that here.
For the oral glucose tolerance test (OGTT) we raised flies on HF or GSF for 20 or 40 days, followed by a 16h fast, prior to a 2h ad libitum feed on a 10% glucose medium, followed by re-fasting. We quantified total glucose in flies following the initial fast (0h), after feeding on 10% glucose (2h), and twice during the re-fast period (4h, 6h). In insulin-sensitive flies, glucose levels rise during feeding, and drop during the fast, due to insulin-dependent stimulation of glucose uptake. We found that flies raised on either food processed glucose with equal efficiency at all time points in both ages ( Fig. 4H,I), arguing that GSF does not significantly impair insulin sensitivity as flies age.
Finally, we measured the lifespans of HF and GSF-treated ilp2-3, 5 mutant flies. ilp2-3, 5 mutants are deficient for systemic insulin signaling, and normally outlive wild-type controls. Thus, if GSF extends lifespan by suppressing systemic insulin activity, we expect that ilp2-3, 5 mutants will not benefit from lifelong culture on GSF. As expected, w 1118 controls raised on GSF outlived those raised on HF, though not to the same extent as is Figure 1, likely reflecting inherent variability in lifespan assays. Contrary to our hypothesis, ilp2-3, 5 mutants raised on GSF significantly outlived ilp2-3, 5 mutants raised on HF (Fig. 4J, K), a phenotype we replicated in three independent assays (Fig. 4K). Thus, although GSF has effects on the expression of two insulin peptide genes, we did not detect GSF-dependent effects on insulin activity, or survival of insulin-deficient flies. As we did not observe a sign of functional insulin defects, we believe our data are most consistent with a hypothesis that that GSF extends life through insulin-independent means.

Glucose-supplemented food increases expression of intestine-associated cell-cell junction genes.
To determine how GSF extends longevity, we used RNA sequencing (RNA-Seq) to compare transcription in whole flies raised on GSF or HF for 20 days. When we looked at differential gene expression, we found 488 upregulated genes and 555 downregulated genes in GSF-fed flies compared to HF-fed controls (Fig. 5A). Gene ontology analysis of downregulated processes showed that GSF primarily Biology Open • Accepted manuscript leads to diminished expression of genes required for metabolism, and energy use (Fig. 5B). In particular, we noticed significant decreases in expression of genes involved in gluconeogenesis and lipid catabolism ( Fig. 5B), likely a result of the increased availability of glucose as an energy source, and consistent with our observation that flies raised on GSF have elevated triglycerides relative to HF-treated counterparts ( Fig. 2D).
In contrast to the dominance of metabolic terms among downregulated gene ontologies, we found that GSF enhanced expression of genes involved in a number of distinct cellular processes, including immunity, cell adhesion, and cell mobility (Fig. 5A,B). In fact, many of the genes with the highest GSFdependent changes in gene expression encode antimicrobial peptides such as attacins and diptericins To confirm this, we compared transcription of representative cell-cell junction genes in whole flies, dissected heads as a control tissue, and dissected intestines. For all genes, we noted enriched expression in the intestinal tract relative to whole flies, or dissected heads (Fig. S2), raising the possibility that GSF impacts organization of the gut epithelial barrier.

Glucose-supplemented food improves intestinal barrier integrity.
Intestinal barrier integrity deteriorates with age and a weakened barrier is associated with reduced lifespan (Rera et al., 2012). As we observed increased expression of cell-cell junction genes in GSF-treated flies, we asked what effects GSF has on barrier integrity.
The fly gut epithelial barrier is maintained by septate junctions, which are analogous to mammalian tight junctions. Coracle (Cora), a Drosophila protein 4.1 homolog, is an essential component of septate junctions. As flies age, Cora and other septate junction proteins partially lose their cell junction localization and accumulate in the cytosol, leading to breaches in the barrier, paracellular leak of lumenal material into interstitial tissue, and ultimately, death (Rera et al., 2012;Resnik-Docampo et al., 2017). To Biology Open • Accepted manuscript determine effects of GSF on the intestinal barrier, we used immunofluorescence to examine the cellular distribution of Cora in intestines of 40-day old flies raised on HF or GSF compared to 5-day old flies raised on HF. The intestines of 5-day old flies raised on HF contained orderly arrangements of large, polyploid nuclei of absorptive enterocytes, and smaller, evenly spaced nuclei of progenitor cells or secretory enteroendocrine cells (Fig. 6A, Hoechst). At this young age, septate junctions are easily identified as fine margins of Cora staining (Fig. 6A, Coracle). In 40-day old flies raised on HF, we noted classic hallmarks of age-dependent epithelial degeneration. Specifically, we detected unevenly distributed, large enterocyte nuclei, interspersed by irregular populations of smaller nuclei from progenitor/enteroendocrine cells ( Fig. 6A, Hoechst). In addition, we detected cytosolic accumulations of Cora (Fig. 6A, asterisk), including enrichment in punctae (Fig. 6A, arrowhead). In contrast, age-matched intestines of flies raised on GSF looked more similar to younger flies raised on HF, with regularly spaced nuclei (Fig. 6A, Hoechst), while Cora distribution appeared more localized to junctions than in HF-fed samples (Fig. 6A, Coracle). 3D reconstruction of 40-day old intestines highlighted the difference in Cora localization between flies raised on HF or GSF (Fig. 6B). In flies raised on GSF, Cora retained a reticulated pattern associated with points of cell-cell contact at septate junctions. In contrast, we detected uneven, diffuse Cora distribution in intestines from age-matched flies raised on HF.
To quantify food-dependent impacts on subcellular distribution of Cora, we determined the junction to cytosol ratio of Cora in midguts of flies raised on HF or GSF for 40 days. Here, we detected significantly higher junction to cytosol ratios of Cora in 40-day old GSF-treated flies than in age-matched HF-treated flies ( Fig. 6C), supporting the hypothesis that GSF sustains Cora association with septate junctions as flies age.
To test if GSF functionally improves barrier integrity in aged flies, we performed a smurf assay, in which a non-permeable dye, that only crosses the epithelium upon loss of barrier integrity, is added to the food. Flies raised on GSF smurfed significantly later than those on HF (Fig. 6D), confirming enhanced barrier integrity in GSF-treated flies. Finally, we asked if disrupting the epithelial barrier reverts the lifespan benefits associated with GSF. For this experiment, we raised flies on GSF or HF for 20 days, at which point we transferred them to HF or GSF that we supplemented with 5% dextran sodium sulfate (DSS), a detergent that disrupts the gut barrier (Fig. 6E). By increasing intestinal permeability with DSS, we found that flies raised on GSF completely lost their survival advantage (Fig. 6F), perishing at the same time as flies raised on HF, suggesting that GSF-dependent lifespan requires an intact intestinal epithelial barrier. Biology Open • Accepted manuscript

DISCUSSION
Aging and age-related diseases pose a growing global challenge. Dietary interventions offer a promising approach to improve aging, but questions remain regarding optimal regimes. Here, we asked how glucose-supplemented food (GSF) extends adult male Drosophila longevity. Our data suggest that GSF-dependent lifespan extension is not an effect of lower caloric intake, or systemic insulin activity, two frequently-studied regimes of lifespan extension. Instead, we found that flies raised on GSF increased expression of cell junction genes, and had an extended duration of gut barrier function. Furthermore, our work showed that chemical disruption of the intestinal barrier removed the lifespan extension associated with GSF-treatment. Thus, our data are consistent with a hypothesis that GSF prolongs adult viability by maintaining intestinal barrier integrity, although detailed mechanistic studies are required to test this hypothesis.
While we found that glucose supplementation extends lifespan in flies, in contrast, studies in other organisms have found differing effects from glucose. In particular, recent studies using the Caenorhabditis elegans model have found mixed outcomes on lifespan from glucose supplementation.
Interestingly, high glucose (2% or 111 mM) treatment in young worms (1-3 days old) reduces lifespan, but beginning glucose treatment after worms are at a post-reproductive age (7 days old) extends lifespan While our study suggests that supplementation of glucose to holidic food extends lifespan through enhanced intestinal barrier integrity, it is important to acknowledge that limitations in our study prevent us from establishing a causal mechanism. Disruption of the intestinal barrier through DSS-treatment removed the survival advantage of GSF-fed flies, supporting a role for barrier integrity in GSF-mediated lifespan extension. However, DSS treatment was also detrimental to flies raised on either diet compared to untreated controls. Thus, further studies of how GSF affects the intestinal barrier with interventions that are less harmful will help to provide this mechanistic insight. As we mainly compared two diets in this study, we cannot definitively state whether increased glucose intake, or if an alternative difference between the two diets, such as the protein to carbohydrate ratio or restriction of protein intake, leads Biology Open • Accepted manuscript to improved barrier integrity and lifespan extension. A thorough, comprehensive study in line with the Geometric Framework for Nutrition will be required to determine if glucose-supplemented food extends lifespan because of its lower protein to carbohydrate ratio relative to unmodified holidic food or reduced protein intake (Lee et al., 2008). Furthermore, while we examined the effect of glucose supplementation, it may be possible that other carbohydrates, such as sucrose or fructose, impact lifespan through a similar mechanism and future studies will be required to investigate this.
In this study, we performed experiments on virgin male flies, though we previously found that 100 g/L glucose-supplemented food also improved lifespan in virgin females ( This study shows that moderate levels of glucose may extend Drosophila lifespan through improved intestinal barrier integrity. In humans, the intestinal barrier deteriorates with age, as well as in chronic diseases such as inflammatory bowel disease. With population aging becoming a growing global concern, further investigation of how dietary components can help maintain intestinal barrier integrity will be essential. We believe that these findings contribute to our understanding of intestinal health and may help efforts to develop preventative measures to limit the effects of aging and disease.

Drosophila husbandry
Virgin male w 1118 flies were used for all experiments unless otherwise specified. Other fly lines used were Df(3L)Ilp2-3,Ilp5 3 and ilp2 1 gd2HF (Park et al., 2014). Flies were maintained at 25˚C on a 12-hour light: 12-hour dark cycle and flipped to fresh food every two-three days. Flies in this study were allowed to develop on BDSC cornmeal food (https://bdsc.indiana.edu/information/recipes/bloomfood.html).
Upon emergence, adults were transferred to their respective holidic food. The holidic food (HF) was Biology Open • Accepted manuscript prepared following the published protocol and recipe using the original amino acid solution (Oaa) at 100mM biologically available nitrogen (Table S1) (Piper et al., 2014). Variants to this diet included supplementation with either 50 g/L glucose (GSF), 50 g/L casein, or 22.2 g/L lard. For starvation assays, flies were maintained on 1% agar vials.

Lifespan analysis
Virgin male flies were used for all lifespan studies. Lifespan studies were performed with 30 flies/vial.
Flies were maintained at 25˚C on a 12-hour light: 12-hour dark cycle in a humidified incubator. Flies were flipped to fresh food every 2-3 days. Deaths were recorded daily.
To measure circulating sugars, each assay was performed with three biological replicates consisting of hemolymph drawn from 15-20 flies per replicate (Tennessen et al., 2014). Flies were carefully pierced in the thorax with a 26G needle and placed in a filter collection tube. Tubes were centrifuged at 9000g for 5 min at 4˚C yielding at least 1 µL of hemolymph. 1 µL of hemolymph was diluted 1:100 in trehalase buffer (5 mM Tris pH 6.6, 137 mM NaCl, 2.7 mM KCl), and placed in a 70˚C water bath for 5 min. Each sample was split into two 50 µL aliquots, one to measure glucose and one to measure trehalose.
Trehalase was prepared by diluting 3 µL porcine trehalase (1 UN) in 1 mL trehalase buffer. 50 µL of this trehalase solution was added to one aliquot of each sample while 50 µL trehalase buffer was added to the other, then samples were incubated at 37˚C for 24 hours. 30 µL of samples and standards were added to a 96-well plate and glucose was measured using the Glucose Oxidase (GO) Assay kits (Sigma, GAGO20).
Total circulating sugars was measured from the trehalase-treated sample, free glucose was measured from the untreated sample, and trehalose was calculated as the difference between treated and untreated samples.

Enzyme-linked immunosorbent assay (ELISA)
To measure total and circulating Ilp2 levels, the ilp2 1 gd2HF fly stock and protocols were provided by Dr. Seung K. Kim (Park et al., 2014). Note that a different published protocol for hemolymph extraction was used compared to circulating sugar measurement. Each assay was performed with five biological replicates. To prepare each replicate, the black posterior was removed from 10 males, and the remaining bodies were transferred to 60 µL PBS, followed by a 10 min vortex at maximum speed. Tubes were centrifuged at 1000 g for 1 min, then 50 µL of the supernatant was transferred to a PCR tube as the circulating Ilp2-FH sample. To the tubes with the remaining flies, 500 µL of PBS with 1% Triton X-100 was added, homogenized with a pestle and cordless motor (VWR 47747-370), and followed by a 5 min vortex at maximum speed. These tubes were centrifuged at maximum speed for 5 min, then 50 µL of the supernatant was transferred to a PCR tube, as the total Ilp2-FH sample.
For the ELISA, we used FLAG(GS)HA peptide standards (DYKDDDDKGGGGSYPYDVPDYA amide, 2412 Da: LifeTein LLC). 1 µL of the stock peptide standards (0-10 ng/ml) was added to 50 µL PBS or PBS with 1% Triton X-100. Wells of a Nunc Maxisorp plate (Thermo Scientific 44-2404-21) were coated with 100 µL of anti-FLAG antibody diluted in 0.2M sodium carbonate/bicarbonate buffer (pH 9.4) to 2.5 µg/mL, then the plate was incubated at 4°C overnight. The plate was washed twice with PBS with 0.2% Tween 20, then blocked with 350 µL of 2% bovine serum albumin in PBS at 4°C overnight. Anti-HA-Peroxidase, High Affinity (clone 3F10) (Roche 12013819001, 25 µg/mL) was diluted in PBS with 2% Tween at a 1:500 dilution. 5 µL of the diluted anti-HA-peroxidase was added to the PCR tubes containing 50 µL of either samples or standards, vortexed, and centrifuged briefly. Following blocking, the plate was washed three times with PBS with 0.2% Tween 20. Samples and standards were transferred to wells of the plate, the plate was sealed with adhesive sealer (BIO-RAD, MSB-1001), and then placed in a humid chamber at 4°C overnight. Samples were removed with an aspirator and the plate was washed with PBS with 0.2% Tween 20 six times. 100 µL 1-Step Ultra TMB -ELISA Substrate (Thermo Scientific 34028) was added to each well and incubated at room temperature for 30 mins. The reaction was stopped by adding 100 µL 2M sulfuric acid and absorbance was measured at 450 nm on a Spectramax M5 (Molecular Devices).

Consumption assays
Both the flyPAD and the CAFE assays were used to characterize consumption in this study. The fly Proboscis and Activity Detector (flyPAD) records changes in capacitance to detect physical interaction of an individual fly with their food (Itskov et al., 2014). For the flyPAD assay, flies were starved for 2 hours Biology Open • Accepted manuscript prior to the assay. HF and GSF was prepared as described, but with agarose substituted for the agar.
Prepared food was melted at 95°C and then maintained at 60°C to facilitate pouring. Individual flies were placed in each flyPAD arena using a mouth aspirator at n=32 for each sample. Eating behaviour was recorded for 1 hour.
The Capillary Feeder (CAFE) assay allows quantification of ingested liquid food over an extended period (Ja et al., 2007). For the CAFE assay, flies were maintained in empty vials at 10 flies/vial with 10 vials per sample (n=10) and fed liquid food through capillary tubes. To prepare liquid food for this assay, HF and GSF were prepared as described, but without the addition of agar. Flies were fed the liquid version of their respective diets for a period of 3 days. Food consumption was measured every 24 hours, and fresh food was provided each day.

Oral glucose tolerance test (OGTT)
Glucose tolerance was measured using an OGTT (Palu and Thummel, 2016). Each assay was performed with five biological replicates consisting of five flies per replicate. Flies were starved overnight for 16 hours on 1% agar, switched to vials containing 10% glucose and 1% agar for 2 hours, and then re-starved on vials of 1% agar. Samples were obtained after initial starvation, after 2 hours on 10% glucose, and then at both 2 hours and 4 hours following re-starvation. Samples of 5 flies were weighed and then mashed in 125 µL TET buffer (10mM Tris, 1mM EDTA, 0.1% Triton X-100, pH 7.4). Glucose was measured using the Glucose Oxidase (GO) Assay kits (Sigma, GAGO20).

RNA isolation and RT-qPCR
To isolate RNA for both RT-qPCR and RNA-seq, samples of 5 whole flies (or 10 dissected heads, thoraxes, or intestines where specified) were homogenized in 250 µL TRIzol, then incubated at room temperature for 5 min. Samples were centrifuged at 12000 g for 10 min at 4˚C. Clear homogenate was transferred to a 1.5 mL Eppendorf tube, then 50 µL of chloroform was added, shaken vigorously for 15 seconds, and incubated at room temperature for 3 min. Samples were centrifuged at 12000 g for 15 min at 4˚C. The upper aqueous layer was transferred to a 1.5 mL Eppendorf tube, 125 µL isopropanol was added, then left at -20˚C overnight. Samples were centrifuged at 12000 g for 10 min at 4˚C. The RNA pellet was washed with 500 µL 75% ethanol, centrifuged at 7500 g for 5 min at 4˚C, then allowed to air dry. The RNA pellet was dissolved in RNAse free water, then incubated at 37˚C for 30 min with 1 µL DNAse.

Biology Open • Accepted manuscript
For RT-qPCR, the following primers were used in this study: ilp2 ( replicates per sample (n=3), and relative expression values were calculated using delta delta Ct calculations. Expression levels were normalized to rp49.

Immunofluorescence and microscopy
Flies were briefly washed with 95% ethanol then dissected in PBS to isolate intestines. Samples were fixed for 30 min at room temperature in 4% formaldehyde. Samples were quickly washed in PBS + 0.3% Triton-X (PBT), followed by 3x 10 min washes in PBT. Samples were blocked for 1 hour in PBT + 3% bovine serum albumin (BSA) at room temperature, then incubated overnight at 4˚C in PBT + 3% BSA with 1˚ anti-Cora 1:100 (DSHB, C615.16). Samples were washed 3x for 10 min in PBT, then incubated for 1 hour at room temperature with 2˚ Alexa anti-mouse 1:500. Samples were briefly washed with PBT, followed by 3x 10 min washes in PBT. Hoechst DNA stain 1:500 was added to the second 10 min wash. Samples were washed in PBS, then mounted on slides in Fluoromount (Sigma-Aldrich F4680).

Biology Open • Accepted manuscript
Slides were visualized on a spinning disk confocal microscope (Quorum WaveFX; Quorum Technologies Inc). The R4/R5 region of the posterior midgut of each sample was located by identifying the midgut-hindgut transition and moving 1-2 frames anterior from the attachment site of the Malpighian tubules. Images were acquired using Velocity Software (Quorum Technologies). 3D reconstruction was performed with Icy.

Quantification of Coracle
Quantification of localization of coracle in images was performed in FIJI. Three representative cells were selected per 40X image. For each cell, a transverse line was drawn across the bicellular junction into the cell to measure coracle expression. Peak expression was recorded as the junction value and 2.24 µm (10 px) into the cell from this peak level was recorded as the cytosol value. The junction/cytosol ratio was calculated from these two values. This was performed in triplicate for each cell, and the average of these three measurements was recorded as the value for the cell. Sample sizes for flies raised on HF (n = 7 guts, 66 cells) and GSF (n = 8 guts, 84 cells).

Barrier function assays
For the smurf assay, HF and GSF were prepared as described with the addition of 1% erioglaucine disodium salt (Brilliant Blue FCF). Flies were raised on their respective diets and monitored daily for extraintestinal leakage of dye or 'smurfing'. For the dextran sulphate sodium (DSS) challenge, flies were raised on either HF or GSF for 20 days, then transferred to either HF or GSF with 5% DSS added, respectively. Deaths were recorded daily and flies were transferred to fresh food every 2-3 days.

Statistical analysis
Statistical analysis was performed using Graphpad Prism (Version 7.0). Statistical significance was set at p < 0.05. Significance between two samples was determined by Student's T-tests. Significance in experiments with two independent variables were determined by two-way analysis of variance (ANOVA).
For lifespan and survival analysis, significance was determined using log-rank (Mantel-Cox) test. Hazard function was determined with 5-day bins.
Biology Open • Accepted manuscript