The Drosophila melanogaster Neprilysin Nepl15 is involved in lipid and carbohydrate storage

The prototypical M13 peptidase, human Neprilysin, functions as a transmembrane “ectoenzyme” that cleaves neuropeptides that regulate e.g. glucose metabolism, and has been linked to type 2 diabetes. The M13 family has undergone a remarkable, and conserved, expansion in the Drosophila genus. Here, we describe the function of Drosophila melanogaster Neprilysin-like 15 (Nepl15). Nepl15 is likely to be a secreted protein, rather than a transmembrane protein. Nepl15 has changes in critical catalytic residues that are conserved across the Drosophila genus and likely renders the Nepl15 protein catalytically inactive. Nevertheless, a knockout of the Nepl15 gene reveals a reduction in triglyceride and glycogen storage, with the effects likely occurring during the larval feeding period. Conversely, flies overexpressing Nepl15 store more triglycerides and glycogen. Protein modeling suggests that Nepl15 is able to bind and sequester peptide targets of catalytically active Drosophila M13 family members, peptides that are conserved in humans and Drosophila, potentially providing a novel mechanism for regulating the activity of neuropeptides in the context of lipid and carbohydrate homeostasis.

Neprilysin is the founding member of the M13 family of zinc metalloendopeptidases that typically function as "ecto-enzymes". M13 family members typically contain a transmembrane domain and a C-terminal extracellular catalytic domain that cleaves secreted peptides [1][2][3] . Mammalian Neprilysin has a number of potential peptide targets involved in regulating neuronal function, appetite, metabolism, energy homeostasis and inflammation (e.g. tachykinins such as Substance P, galanin, cholecystokinin and neuropeptide Y) [4][5][6][7][8] . Recently, mammalian Neprilysin has been studied as a potential therapeutic target for treating type 2 diabetes 9 .
Drosophila melanogaster (D. melanogaster) behave remarkably similarly to mammals in terms of nutrient metabolism and energy homeostasis and have become a powerful model organism to study these processes [10][11][12] . Like mammals, D. melanogaster have organs for digestion and nutrient absorption (the midgut). They have a circulatory system (the hemolymph) that conveys lipids and other nutrients from one organ to another. The fat body stores carbohydates as glycogen and lipids as triacylglycerides (TAG), and thus combines functions of mammalian liver and adipose tissue. Drosophila use conserved pathways to control metabolism and energy homeostasis. For example, Insulin Like Peptides (ILPs) are released from Insulin Producing Cells (IPCs) in the brain, and signal via a conserved Insulin Receptor pathway that promotes nutrient uptake by cells. Adipokinetic Hormone (AKH), a Drosophila glucagon counterpart, is released from Corpora Cadiaca (CC) cells and signals via a conserved G Coupled Protein Receptor (GCPR) pathway that promotes nutrient mobilization. In addition, a number of peptide targets of mammalian Neprilysins that are known to be involved in feeding behavior, appetite regulation, metabolism and energy homeostasis, have D. melanogaster counterparts that are linked to similar processes 10,[13][14][15] .
Here we show that the D. melanogaster M13 family member, Nepl15, is likely to be secreted and to be catalytically inactive. Nevertheless, we demonstrate that Nepl15 has a role in lipid and carbohydrate storage. Nepl15 is strongly expressed in the fat body, and is differentially expressed between males and females. Knock-out mutations in Nepl15 result in reduced levels of triglycerides and glycogen in adult males but not in females. Further, Nepl15 mutants show an increase in larval developmental time that is consistent with a reduced ability to store nutrients needed to complete metamorphosis. Conversely, over-expression of Nepl15 in certain tissues results in increases in triglyceride and glycogen storage in males. Our results are consistent with a role for Nepl15 in regulating neuropeptide cleavage in the context of nutrient metabolism. www.nature.com/scientificreports/ acids form the S1 and S2 subsites. There does not appear to be a DmNepl15 Asn counterpart, and Ala appears to be replaced with Pro in the Drosophila Nepl15 orthologs. Moreover, the aromatic amino acid in the motif that lines the S1 subsite is replaced by a His (Fig. 2C,C' ,C"; Figure S1), though the S1 subsite appears to be less important for specificity in M13 family members compared to other subsites 3,17,[38][39][40][41][42][43][44][45] . However, other aspects of the Drosophila Nepl15 sequences and the DmNepl15 homology model structure suggest that it could interact with substrates, albeit perhaps through a different mechanism than typical M13 Figure 2. DmNepl15 lacks critical catalytic residues. Homology models of DmNepl15 (gray) based on models of crystal structures of HsNeprilysin in complex with (left, magenta) LBQ657, the active metabolite of sacubitril (5jmy) 21 ; and (center, green) with the inhibitor phosphoramidon (1dmt) 16 ; (right, blue) homology model of DmNepl15 based on a model of a crystal structure of HsECE-1 in complex with the inhibitor phosphoramidon (3dwb) 17 . Yellow, white and gray dotted lines denote interactions between Zn(II) and its coordinators, hydrogen bonds, and other interactions described in the text, respectively. (A-A") Two of the three residues involved in Zn(II) coordination in HsNeprilysin (His583, His587 and Glu646) and in HsECE-1 (His607, His611, and Glu667) are replaced by Arg in DmNepl15. Asp residues involved in Asp-His-Zinc triads in HsNeprilysin (Asp590 and Asp650) and in HsECE-1 (Asp614 and Asp671) [16][17][18]27,115 are present at these positions in DmNepl15. The base that deprotonates water in HsNeprilysin (Glu584) and in HsECE-1 (Glu608) is replaced by Gln in DmNepl15. (B-B") In HsNeprilysin, His711 is thought to help stabilize the transition state during catalysis; Asp709 interacts with His711 and is crucial for catalysis 16,21,[28][29][30][31] . The corresponding residues are Asp730 and His732 in HsECE-1 17,27 . There appears to be no counterpart to HsNeprilysin His711/732 in the Drosophila Nepl15 orthologs. (C-C") In the HsNeprilysin and HsECE-1 crystal structures the N-A-Ar-Ar motif includes the Asn542/Asn562 side chains and backbone CO groups that hydrogen bond with backbone elements of the LBQ657 and phosphoramidon inhibitors. There does not appear to be a DmNepl15 counterpart to Asn542/Asn562. Ala543/Ala563 in HsNeprilysin appears to be replaced with Pro in the Drosophila Nepl15 orthologs, and Phe544/Tyr565 is replaced by His. HsNeprilysin Arg717 and its HsECE-1 Arg738 counterpart hydrogen bond with backbone CO groups on the LBQ657 and phosphoramidon inhibitors, participate in salt bridges with Asp650/Asp671 and are important for catalytic activity 16,17,21,24,46,47 . Drosophila Nepl15 counterparts of these residues are Arg654 and Asp586. These drawings were made using the DeepView-Swiss-PdbViewer (version 4.1.1) 114 16,17,21,24,46,47 is present in the Drosophila Nepl15s ( Fig. 2C; Figure S1). The hydrophobic residues of the S1′ subsite interact preferentially with bulky hydrophobic residues at the P1′ site are generally the most important drivers of specificity in M13 family members 1,3,16,17,48 . The corresponding S1′ site residues are also hydrophobic in the Drosophila Nepl15s ( Fig. 3A; Figure S1).
The S2′ subsites in M13 family members vary considerably in specificity and appear to be somewhat flexible 3 . For instance, two Arg residues (Arg 102 and Arg 110) present in the HsNeprilysin S2′ subsite are thought to enable this enzyme to function as a dipeptidylcarboxypeptidase in addition to its endopeptidase activity 16,21,34,47-51 . In the Drosophila Nepl15 orthologs Arg or Lys are found in the corresponding position to Arg102, but Arg110 is replaced with Gln (Fig. 3B, Figure S1). Taken together, these results suggest that the Drosophila Nepl15 orthologs are likely able to bind to peptides containing a hydrophobic amino acid in the P1′ position. The contributions of the S1 and S2′ subsites to specificity are less certain.
Based on published estimates of divergence dates among these species 52 , for the most part these similarities and differences in Drosophila Nepl15 orthologs compared to other M13 family members have been conserved over ~ 35 million years of evolution 53,54 . Thus, despite the probable lack of catalytic activity, the differences are likely to be important for Drosophila Nepl15′s unique function.

D. melanogaster Nepl15 transcript expression is enriched in the fat body.
To further elucidate the function of the Nepl15 gene, we first checked the Nepl15 mRNA expression profile based on genome-wide microarray 55 and RNAseq data 56 ( Figure S7). In both microarray data and RNAseq data the larval fat body had the highest levels of Nepl15 expression among the different larval tissues tested. Both microarray and RNAseq data also showed that the adult fat body has among the highest levels of expression of Nepl15 compared to other adult tissues. We also noted from the RNAseq data that Nepl15 expression levels are more than four-fold higher  21 ; and (center, green) with the inhibitor phosphoramidon (1dmt) 16 ; (right, blue) homology model of DmNepl15 based on a model of a crystal structure of HsECE-1 in complex with the inhibitor phosphoramidon (3dwb) 17 . Yellow, white and gray dotted lines denote interactions between Zn(II) and its coordinators, hydrogen bonds, and other interactions described in the text, respectively (A-A") Hydrophobic residues line the S1′ subsite in DmNepl15 as they do in HsNeprilysin and in HsECE-1. (B-B") Interactions between the N-A-Ar-Ar motif, the S2′ subsite and the inhibitors reveal flexibility in binding to peptide substrates, as well as differences between the S2′ subsites of HsNeprilysin, HsECE-1 and DmNepl15. In particular, HsNeprilysin Val541, Arg110 and Arg102 are involved in positioning the terminal -COOH group that enables the dipeptidylcarboxypeptidase actvitiy of which HsNeprilysin is capable. Whereas Val563 appears to be the HsECE-1 counterpart to HsNeprilysin Val541, there is no DmNepl15 counterpart to this residue. An Arg is present in the HsECE-1 and DmNepl15 counterparts of HsNeprilysin Arg102. However, the HsNeprilysin Arg110 counterparts are Trp153 in HsECE-1 and Gln88 in DmNepl15. These drawings were made using the DeepView-Swiss-PdbViewer (version 4.1.1) 114 (http://www.expas y.org/spdbv /).  Figure S7). The 1.6-fold difference in fat body expression levels between males and females suggests it is one of the primary contributors to the difference in Nepl15 levels in adult males versus females.
To corroborate the Nepl15 transcript expression data available from the genome-wide studies described above, we performed RT-qPCR to quantify the relative abundance of Nepl15 transcripts in the third instar larval central nervous system, gut and fat body, and in the head, thorax and abdomen of w 1118 , 2-3 days old, adult female and male flies (Fig. 4). Consistent with the genome-wide data described above, Nepl15 transcript levels in the larval fat body were significantly higher compared to levels in larval brain and gut. Nepl15 transcript levels in the adult female head and thorax, as well as the adult male head, thorax and abdomen, were similar to each other and to levels in the larval fat body. Strikingly, however, Nepl15 transcript levels in the abdomens of adult female flies were significantly lower than those in abdomens of adult male flies and were comparable to the much lower levels of Nepl15 transcripts observed in the larval gut and brain.
The difference in Nepl15 transcript levels in the adult male versus female abdomen are interesting, but are not easily compared to the genome-wide studies, which looked at individual organs. Furthermore, there is substantial variability in Nepl15 transcript levels among biological replicates. This variability is likely to result from variability in expression rather than variability due to experimental technique, because we were careful to follow MIQE guidelines that stipulate that standard deviations between technical replicates be ≤ 0.3, and that standard deviations between biological replicates be ≤ 0.5 for the RpL32 control (see "Methods"). Nevertheless, the higher Nepl15 transcript levels in the larval and adult fat body suggest a role for Nepl15 expression in the fat body.
Nepl15 knockout (w 1118 ; Nepl15 ko ) flies have reduced lipids. To study the function of Nepl15, we generated knockout mutants (w 1118 ; Nepl15 ko ) by ends-out homologous recombination 57 . We have confirmed the knockouts by performing genomic PCR followed by sequencing as well as by RT-qPCR (see Materials and Methods). Homozygous knockout flies are viable and fertile. Following our initial observation that Nepl15 is highly expressed in the larval and adult fat body compared to other tissues, we assessed whether Nepl15 knockout mutants have defects in fat body function.
One major role of the fat body is storing lipids in the form of triacylglycerides (TAG) in lipid droplets. Therefore, we stained the larval ( Fig. 5A-C) and adult fat bodies ( Fig. 5D-F) from wild type and Nepl15 ko flies with the red fluorescent dye Nile Red to observe lipid droplets. Lipid droplets in the larval and adult fat body of the Nepl15 ko flies appeared smaller than those in the wild type fat body and the % area of Nile Red staining was significantly less in Nepl15 ko larvae and adult fat bodies compared to wild type. The images shown in Fig. 5A-F are representative of the images used for the quantification, and we did not observe obvious differences between different regions of the fat body. We also stained the larval fat body with periodic acid-Schiff (PAS) ( Figure S8A,B) to detect polysaccharides, including glycogen 58 . Under bright-field illumination, the stain results in accumulation of a dark pink pigment in areas with high polysaccharide concentrations. Nepl15 ko larval fat body showed less of this dark pigment, suggesting reduced concentrations of glycogen.
To confirm the Nile Red stainings, we performed thin layer chromatography (TLC) to determine TAG levels in lysates generated from whole adult wild type and two independent Nepl15 ko isolates. Levels of TAG were were analyzed using a one-way ANOVA and were assigned to statistical groups using Tukey's multiple comparison test. Groups sharing at least one letter are not significantly different; groups not sharing any letter are significantly different (P < 0.05). Values in F were analyzed using a two-tailed student's t test and the P value is shown. Error bars represent the standard error of mean (SEM). www.nature.com/scientificreports/ significantly reduced in w 1118 ; Nepl15 ko flies compared to w 1118 flies (Fig. 5G). Taken together, our results suggest that Nepl15 is required for TAG and glycogen storage in the fat body during both larval and adult stages. One possibility behind the reduced levels of stored nutrients is impaired capacity of food intake by Nepl15 ko mutant flies compared to wild type. To examine any changes in feeding behavior, we first monitored the amount of liquid food ingested by these adult flies using the capillary feeding (CAFE) assay 59 . Interestingly, food intake was significantly increased in the adult Nepl15 ko flies compared to the adult wild type flies (Fig. 5H). To substantiate this finding, we also measured food ingestion via radioisotope labeling 60 and observed a similar outcome, though the results were not statistically significant when normalized to fly weight (Fig. 5I). These results suggest that reduced storage of nutrients in the Nepl15 ko flies does not result from reductions in food intake.
Adult Nepl15 ko males but not females have reduced glycerolipids and glycogen. To avoid the possibility that the reductions in stored TAG and glycogen observed in the w 1118 ; Nepl15 ko flies arose from background effects or differences in population densities, we back-crossed w 1118 ; Nepl15 ko flies to our w 1118 ; + (hereafter w 1118 ) lab strain for 6 generations. In addition, we raised flies at similar densities (see Materials and Methods). Intrigued by the differences in expression of Nepl15 in males versus females, we also began to separate males and females during experiments. Using these methods, we repeated the food ingestion via radioisotope labeling experiments, with similar results: in two trials, w 1118 ; Nepl15 ko males and females consumed approximately the same amount of food compared to their w 1118 counterparts (results of one trial shown in Figure S8C). We also weighed w 1118 and w 1118 ; Nepl15 ko adult males and females. As expected, female D. melanogaster of both genotypes are larger than males of both genotypes, and therefore weigh more. In multiple trials and at different ages, there were no consistent differences in weight between w 1118 and w 1118 ; Nepl15 ko males or between w 1118 and w 1118 ; Nepl15 ko females. Two representative trials conducted independently in the Curtiss and Ja labs, respectively are shown in Figure S8D,E. For the experiments in Figures S8C & E that were performed in the Ja lab, flies were raised for several generations in the Ja lab after being shipped from the Curtiss lab, prior to conducting the experiments. These data provide strong evidence that there are no differences between w 1118 and w 1118 ; Nepl15 ko adult in terms of feeding or ability to reach and maintain a normal body weight.
To confirm the differences between w 1118 and w 1118 ; Nepl15 ko flies in lipid and carbohydrate storage described above, we used more quantitative enzymatic assays coupled with colorimetric analysis for measuring lipid, carbohydrate and protein levels in age-matched, whole adult fly lysates 61 (see also Materials and Methods). Total glycerolipid as well as glycogen concentrations were significantly reduced in whole adult w 1118 ; Nepl15 ko male flies compared to controls (Fig. 6A,B). Accordingly, adult w 1118 ; Nepl15 ko male flies succumbed significantly earlier to starvation than did w 1118 control males (Fig. 6C). Only minor reductions in total glycerolipids were observed for whole adult w 1118 ; Nepl15 ko female flies compared to controls, and they had significantly elevated glycogen levels (Fig. 6D,E). In spite of these differences, w 1118 ; Nepl15 ko females succumb significantly earlier to starvation compared to controls (Fig. 6F), similar to males. The Ja lab independently performed similar starvation assays, with similar results ( Figure S8F). Together with the staining and TLC data (Fig. 5), these results suggest that Nepl15 is required for normal TAG and glycogen storage in both males and females.
We were intrigued by the differences between male and female w 1118 ; Nepl15 ko flies in terms of nutrient storage, particularly of glycogen, with males having significantly less glycogen and females having significantly more. To determine whether the metabolic rates for either male and female w 1118 ; Nepl15 ko flies differed from controls by measuring the amount of CO 2 produced per fly over the course of a few hours. Accordingly, metabolic rate was slightly but significantly reduced in mutant females compared to control females, but did not differ in mutant males compared to control males (2 trials, similar results, representative data from 1 trial are shown in Fig. 6G,H).
We hypothesize that the differences in lipid and carbohydrate storage (elevated in Nepl15 mutant females compared to males) as well as metabolic rate (reduced in Nepl15 mutant females compared to males) reflects the fact that females but not males are obliged to supply resources to developing eggs. The increased levels of carbohydrates in females may be present in the eggs and not metabolically available for the females to use under starvation conditions, and the females may have to reduce their metabolic rates due to this lack of resources.
Circulating carbohydrates are unaltered in adult w 1118 ; Nepl15 ko flies. The nonreducing glucose dimer trehalose comprises the vast majority of circulating carbohydrates in insects including D. melanogaster. Trehalose is synthesized exclusively in the fat body from where it is released into the hemolymph 10,58,62-65 . Although trehalose constitutes the majority of circulating carbohydrates, hemolymph glucose levels appear to be more responsive to environmental and genetic alterations to metabolism 65 . Hemolymph glucose concentrations were slightly elevated in 2-4 day old and 8-10 day old w 1118 ; Nepl15 ko adult male flies relative to the w 1118 control, but the difference was not statistically significant ( Figure S9A,B). Additionally, we found no significant differences in whole-body trehalose concentrations in w 1118 ; Nepl15 ko adult male flies relative to w 1118 flies, though whole-body glucose levels were significantly elevated ( Figure S9C). Taken together, these results suggest that loss of Nepl15 has little to no effect on circulating carbohydrates.

Overexpressing Nepl15 results in elevated lipid and carbohydrate storage. To test whether
Nepl15 is capable of driving increased lipid and carbohydrate storage when overexpressed, we used the Gal4-UAS system to drive expression of either UAS-GFP or UAS-Nepl15 (1) in all tissues using the armadillo-Gal4 (arm-Gal4) driver 66 , (2) in the fat body using the Cg-Gal4 driver 67 , and (3) in the midgut using the mex1-Gal4 driver 68 . We used two independent UAS-GFP lines (one on the second chromosome and one on the third chromosome), as well as two independent UAS-Nepl15 lines (UAS-Nepl15 37 on the second chromosome, and UAS-Nepl15 1 on the third). As for the loss-of-function experiments we back-crossed all strains 6 generations into the www.nature.com/scientificreports/ lab w 1118 strain, and carefully controlled population densities, as described in the "Methods". We analyzed the concentrations of different nutrients in whole-body lysates as described above. Both total glycerolipids and glycogen were higher in flies expressing Nepl15 in all tissues using arm-Gal4 (Fig. 7A,B). In contrast, we did not observe consistent results with the tissue-specific drivers Cg-Gal4 (Fig. 7C,D) or mex1-Gal4 (Fig. 7E,F), with glycogen but not glycerolipids elevated only in Cg > Nepl15 37 flies, and glycerolipids but not glycogen elevated in mex1 > Nepl15 flies. Nevertheless, these results suggest that overexpression of Nepl15 is sufficient to promote increased nutrient storage.
Given that Nepl15 is apparently expressed at highest levels in the fat body, we were concerned by the fact that over-expressing Nepl15 in the fat body using the Cg-Gal4 driver did not consistently result in elevated glycerolipids or glycogen. We therefore used r4-Gal4 to drive expression of either GFP or Nepl15 in the fat body from the late embryonic to adult stage 69 . Interestingly, whereas most of the r4 > GFP embryos laid developed to pupariation (63 ± 5.2%), most of the r4 > Nepl15 embryos laid did not develop to pupariation (7.5 ± 1.5%). These results suggest that maintaining appropriate levels of Nepl15 in the fat body is critical for proper development. Values were analyzed using a one-way ANOVA and were assigned to statistical groups using Tukey's multiple comparison test. Groups sharing at least one letter are not significantly different; groups not sharing any letter are significantly different (P < 0.05). Error bars represent the standard error of mean (SEM). n = 3 biological replicates with 3 technical replicates for all panels. shown to increase the amount of whole-body total glycerolipids in wild type adult flies [70][71][72] . To further investigate the role of Nepl15 function in lipid storage, we cultured 5 days old w 1118 and w 1118 ; Nepl15 ko adult male flies for 5 days on either normal food (NF) or on a diet containing 10% w/v coconut oil added to NF (10% HFD) (see Materials and Methods). We then used the colorimetric assays described above to determine the concentrations of free glycerol and total glycerolipids. Exposure to a HFD elevated the whole-body glycerolipid concentrations in both w 1118 and w 1118 ; Nepl15 ko adult males compared to flies cultured on NF, though the differences were not statistically significant (Fig. 8A). In contrast, the whole-body glycerolipid concentrations were significantly less in the HFD-fed w 1118 ; Nepl15 ko males as compared to either the w 1118 adult male flies fed on a HFD or on NF. Nevertheless, the change in glycerolipid concentrations resulting from a HFD vs. NF were not significantly different between w 1118 and w 1118 ; Nepl15 ko adult males. These results suggest that Nepl15 is important during the larval stage for nutrient uptake, that adult w 1118 ; Nepl15 ko flies start out with a deficit of stored nutrients, and that Nepl15 has a less important role in nutrient storage during the adult stage.
Glycogen, TAG, trehalose and total protein levels increase 62,73 as Drosophila larvae take up nutrients to reach a critical weight during its third instar stage, at which they have stored enough nutrients and grown to a sufficient size to complete metamorphosis. Larvae can reach the critical weight faster if nutrients are readily available, whereas when nutrients are in short supply the larval period is prolonged [74][75][76] .
If Nepl15 is important during larval stages for nutrient storage, then w 1118 ; Nepl15 ko larvae should take longer to reach metamorphosis. In multiple separate trials, the median time to pupariation was significantly longer for the w 1118 ; Nepl15 ko flies compared to w 1118 flies (one representative example is shown in Fig. 8B) with a ~ 24-h delay in the time it took for w 1118 ; Nepl15 ko larvae to initiate pupariation (median time: 8 days) compared to w 1118 larvae (median time: 7 days). Once the pupal stage was reached, there was no further delay in time to eclosion w 1118 (median time: 11 days) and w 1118 ; Nepl15 ko flies (median time: 11-12 days). These results support the idea that w 1118 ; Nepl15 ko larvae require more time to reach critical weight, likely because of their reduced ability to store nutrients.
To further test this idea, we used the TARGET system 77 in combination with tubulin-Gal4 78 to overexpress Nepl15 in all cells (1) throughout development (embryonic to adult stages cultured at 30 • C), (2) during just the embryonic and larval stages (embryonic and larval stages cultured at 30 • C, then shifted to 18 • C for pupal through adult stages), (3) during just the pupal and adult stages (embryonic and larval stages cultured at 18 • C, then shifted to 30 • C for pupal through adult stages), or (4) no overexpression during any stage (embryonic to adult stages cultured at 18 • C).
When we assayed time to pupariation (Fig. 8C), we found that the time to pupariation for flies overexpressing Nepl15 during the embryonic and larval stages (30 • C and 30 • C --> 18 • C) was approximately a day shorter (median time to pupariation = 5 days) compared to controls expressing GFP (median time to pupariation = 6 days). In addition, the time to pupariation for flies overexpressing Nepl15 only during pupal and adult stages (18 • C --> 30 • C) was statistically significantly shorter, though the median time to pupariation was the same (7 days). Time to pupariation for flies never overexpressing Nepl15 (18 • C) did not differ from control flies expressing GFP. However, when we assayed for sensitivity to starvation, we found no significant differences for any of our treatments ( Figure S10). These data do not support the hypothesis that Nepl15 is more important for nutrient storage during the larval period than the adult period, but they are consistent with the idea that Nepl15 has a role in regulating nutrient storage during the larval period.
Nepl15 does not affect thor, Akh or AkhR expression. In theory, Nepl15 ko flies could show reductions in stored lipids and carbohydrates due to reduced activity in processes that lead to nutrient storage or to hyperactivity in processes that lead to nutrient mobilization. In flies and other insects, the insulin receptor (InR) signaling pathway promotes nutrient storage, and the adipokinetic hormone (Akh) signaling pathway promotes nutrient mobilization 79 .
Output of the InR can be monitored by transcription of thor, which encodes 4E Binding Protein (4EBP), a direct transcriptional target of dFOXO, a transcription factor that is negatively regulated by InR signaling 80,81 . Furthermore, transcription of Akh and the Akh Receptor (Akhr) are affected by nutrient availability as well as by activity of Insulin Producing Cells (IPCs), Akh Producing Cells (APCs) and activity of cells that express the neuropeptide Allatostatin A [82][83][84][85][86] . To determine whether Nepl15 ko mutations affect transcription of thor, Akh or AkhR, we performed quantitative qPCR on w 1118 and w 1118 ; Nepl15 ko whole adult males. We observed a modest but non-significant increase in thor transcript levels, suggesting a decrease in InR signaling in the absence of Nepl15, but no changes in Akh or AkhR transcript levels (Fig. 9A).
A HFD is known to alter expression of transcripts from hundreds of Drosophila genes 87 . To determine whether a HFD affects Nepl15 transcription, we performed qPCR on w 1118 adult males subjected to a HFD versus NF. We detected no difference in Nepl15 transcription in flies raised on a HFD versus NF (Fig. 9B).

Discussion
We have identified a role for the D. melanogaster family member Nepl15 in lipid and carbohydrate storage. Nepl15 transcripts are enriched in the fat body, particularly in larvae, and Nepl15 ko adults show reductions in glycerolipids and glycogen, as well as increased susceptibility to starvation, while overexpressing Nepl15 results in increases in glycerolipids and glycogen. The fact that Nepl15 ko adults can accumulate glycerolipids on a high fat diet at the same apparent rate as w 1118 flies, coupled with the increase in developmental time for Nepl15 ko larvae compared to w 1118 larvae, suggests that Nepl15 functions during larval development to promote nutrient storage. www.nature.com/scientificreports/ There appear to be differences between adult males and adult females, both in Nepl15 expression levels, and in the effects of the Nepl15 ko mutation on lipid and carbohydrate storage. Recent studies have found intriguing differences in triglyceride storage and metabolism between males and females, as well as in expression of genes involved in these processes [88][89][90][91] . Consistent with these studies, our results suggest that Nepl15 has a sex-specific role in regulating nutrient storage.
Analysis of the Nepl15 protein sequence, together with molecular modeling, suggest that Nepl15 is secreted. One other D. melanogaster M13 family member, Nep2, has been experimentally shown to be secreted 92 , and there is a secreted version of a vertebrate Neprilysin called Membrane Metallopeptidase Like 1 (MMEL1) that is generated by alternative splicing [93][94][95] . In addition, HsNeprilysin is itself found in soluble form in blood, urine and cerebrospinal fluid (reviewed in 96 ). This analysis has also shown that critical catalytic residues differ in Nepl15 in ways that suggest that it lacks peptidase catalytic activity. One mammalian M13 family member, PHEX, has been experimentally shown to bind substrates and prevent their cleavage by other peptidases, a role that does not require catalytic residues 97,98 . It is possible that Nepl15 similarly binds to substrates to prevent their cleavage. The fact that these differences in the catalytic site are conserved across the evolution of the Drosophila genus, suggests that these differences are essential for this protein's function.
Existing evidence indicates that invertebrate M13 metalloendopeptidases, like most of their vertebrate counterparts, prefer to cleave N-terminal to bulky hydrophobic residues (reviewed in 1,3 , as well as cleavage patterns of predicted D. melanogaster M13 target peptides 99 and references therein). The fact that Drosophila genus Nepl15 homologs appear to share a hydrophobic S1′ subsite bolsters the idea that Nepl15 could bind and sequester substrates of catalytic M13 family members, preventing hydrolysis by catalytically active peptidases, and leaving them available to continue performing their functions.
What substrates might be targets of Drosophila Nepl15? The list includes Neuropeptide F (NPF) and short NeuroPeptide F (sNPF), counterparts to the mammalian orexigenic NeuroPeptide Y (NPY), the D. melanogaster tachykinins DTK and Leucokinin 100-106 , Allatostatin (D. melanogaster galanin) 4,86 , and Drosulfakinins (D. melanogaster cholecystokinins) 8,107-109 . All of these D. melanogaster peptides have known or predicted Neprilysin cleavage sites 99 . Many of them affect activity of InR and AkhR pathways, and vice versa 10,[13][14][15]110 . Identification of which of these or other neuropeptides might have their activitie(s) regulated by Nepl15, as well as the mechanisms by which Nepl15 affects nutrient storage, awaits further experimentation.
The work described here provides evidence for a novel mechanism by which neuropeptide activity is regulated by M13 family peptidases in the context of lipid and carbohydrate metabolism. Further study is likely to inform thinking about how M13 family members such as Neprilysin function in diabetes, and how their activities may be controlled to treat diabetes. www.nature.com/scientificreports/ Population densities were controlled by collecting embryos 0-12 h after egg laying (AEL) laid by 2-to 5-dayold control and experimental flies. 100 embryos of each strain were seeded into vials containing 10 mL of food and allowed to develop at 25 °C. Time to pupariation was determined by counting the number of pre-pupae/ pupae formed within a given 12-h time period. Time to eclosion was similarly determined by counting the number of adults that emerged within a given 12-h time period. Newly emerged adult male and female flies were collected every 12 h and kept separately in groups of 20 in vials containing regular food at 25 °C for 3-5 days. Age-matched 3 to 5 day-old flies were subsequently used for RT-qPCR, Bradford, triglyceride and glycogen assays, and starvation assays. Molecular modeling. Alignments of the predicted Nepl15 sequence to other M13 family members were generated using Clustal Omega. Molecular modeling of Nepl15 was performed using SwissModel and the models were analyzed using MolProbity. The Swiss-PdbViewer was used to further analyze the models and to generate the images of the Nepl15 structure.
Quantitative RT-PCR. RNA was extracted from tissues using TRI reagent (Sigma-T9424) and cDNA synthesis was performed using iScript Reverse Transcriptase Supermix for RT-qPCR (Bio-Rad-1708841). Primers (see list below) either had sequences based on previous publications or were designed using the Integrated DNA Technology qPCR primer designing tool and were checked for their efficiency and any off-target amplification.

Fat body staining for glycogen (PAS) and triglycerides (Nile Red). Third instar larval fat bodies
were dissected in ice cold 1 × PBS (10 × PBS:1.37 M NaCl, 27 mM KCl, 100 mM Na 2 HPO 4, 18 mM KH 2 PO 4 , pH 7.4), and fixed in 4% paraformaldehyde in PBS for 20 min at room temperature (RT). For Periodic acid Schiff (PAS) staining tissue was washed in 1% BSA in PBS, incubated in periodic acid solution (Sigma-3951, 1 g/dl) for 5 min at room temperature, washed with 1% BSA in PBS incubated in Schiff 's reagent (Sigma-3952) for 15 min at room temperature and again washed with 1% BSA in PBS. For Nile Red staining tissue was washed in PBS, incubated in PBS containing 0.2 µg/ml Nile Red (Sigma-072485) and 0.5 µg/ml Hoescht 33342 (Sigma-94403) in PBS. Stained tissue was mounted in 80% glycerol (PAS staining) or in 80% plus 5% propyl gallate (Sigma-P3130). Images of PAS-stained fat bodies were taken using bright-field microscopy. Images of Nile-Red-stained fat bodies were obtained on a Leica TCS SC5 Laser Scanning Confocal Microscope. Confocal images are projections of multiple sections obtained using FIJI is Just Image J. The FIJI Analyze Particles function was used to determine the total area of lipid droplets stained by Nile Red, and % area was determined by dividing the total area of lipid droplets by the total area of the fat body.
Thin layer chromatography. 10 adult males were homogenized in 250 µl chloroform:methanol solvent (2:1) on ice and centrifuged to remove debris. 2 µl of the resultant lipid supernatant was pipetted on a silica gel plate (Sigma-60805) and separated using a diethyl ether:hexane solvent system. Lipids were visualized using a heat activated phosphomolybdic acid solution (Sigma-02553).
Feeding assays. Capillary Feeding (CAFE) and food labeling with radioactive tracer assays were performed as described 60 . Colorimetric assays for glycerolipids, glycogen, trehalose and glucose. Colorimetric Assays for these metabolites were performed essentially as described 61 . In all cases three biological and three technical replicates were included.
The Sigma GAGO-20 (G0885) kit was used to determine free glucose concentrations, to determine glycogen concentrations from homogenates treated with amyloglucosidase (Sigma A160220) and to determine trehalose concentrations from homogenates treated with porcine trehalase (Sigma-T8778-1UN). A D-( +)-Glucose solution (Sigma-G3285), a glycogen solution (Ambion-9510) and a trehalose solution (Sigma-90208) were used to generate standard curves. Absorbance was measured at 540 nm using an Eon BioTek plate reader. www.nature.com/scientificreports/ Triglyceride Reagent (Sigma-T2449) was used to determine triglyceride concentrations; a glycerol solution (Sigma-G7793) was used to generate standard curves. Absorbance was measured at 540 nm using an Eon BioTek plate reader.
Bradford Reagent (Sigma-B6916) was used to determine protein concentrations; Bovine Serum Albumin (Sigma-A7906) was used to generate standard curves. Absorbance was measured at 595 nm using the Eon BioTek plate reader.
Briefly, for assays on whole adult flies, 10 age matched adult flies were rinsed with ice cold 1× PBS and homogenized in 200 µl of ice cold 1× PBS, 1× PBST or Trehalase Buffer (5 mM Tris pH 6.6, 137 mM NaCl, 2.7 mM KCl). 20 μl of the homogenized sample was cleared by centrifugation, diluted 2× with ice-cold buffer, and used to measure protein content using a Bradford assay. Remaining homogenates were heat-inactivated (10 min at 70 °C) and cleared by centrifugation. Resulting supernatants were either not diluted (free glucose, trehalose assays, glycerolipids), or were diluted 15× (males, glycogen), 20× (females, glycogen) with ice cold buffer prior to performing the assay.
For assays on hemolymph, 50 adult male flies were punctured in the thorax with a needle and placed in a 0.65 ml microcentrifuge tube with a small hole in the bottom. The 0.65 ml tube was placed in a 1.7 ml microcentrofuge tube and spun for 2 min at 10,000 × g generating approximately 1 µl of transparent hemolymph, which was mixed with 19 µl ice-cold 1XPBS. 3 µl was used to measure free glucose concentrations, and 3 µl was used to determine protein concentration.
Metabolic rate assays. A few flies were placed in a sealed chamber containing soda lime to adsorb CO 2 and a capillary to measure changes in pressure. The rate of change in the height of liquid in the capillary provides a measure of CO 2 production and therefore metabolic rate. The amount of CO 2 produced was normalized by dividing by the total number of flies in the sealed chamber.
TARGET assays. Flies were allowed to lay for 10 h at 25 • C. 100 embryos were transferred to a new vial of complete media and allowed to develop at the permissive temperature for Gal80ts (18 • C) or at the nonpermissive temperature (30 • C) until the first pupae were noted. At this point vials were either retained at the original temperature, were transferred from the permissive to the non-permissive temperature (18 • C --> 30 • C), or were transferred from the non-permissive temperature to the permissive temperature (30 • C --> 18 • C).
Starvation assays. Newly emerged adult flies were incubated separately in vials containing normal food at