Imp-L2, a putative homolog of vertebrate IGF-binding protein 7, counteracts insulin signaling in Drosophila and is essential for starvation resistance

Abstract

Background: Insulin and insulin-like growth factors (IGFs) signal through a highly conserved pathway and control growth and metabolism in both vertebrates and invertebrates. In mammals, insulin-like growth factor binding proteins (IGFBPs) bind IGFs with high affinity and modulate their mitogenic, anti-apoptotic and metabolic actions, but no functional homologs have been identified in invertebrates so far.

Results: Here, we show that the secreted Imaginal morphogenesis protein-Late 2 (Imp-L2) binds Drosophila insulin-like peptide 2 (Dilp2) and inhibits growth non-autonomously. Whereas over-expressing Imp-L2 strongly reduces size, loss of Imp-L2 function results in an increased body size. Imp-L2 is both necessary and sufficient to compensate Dilp2-induced hyperinsulinemia in vivo. Under starvation conditions, Imp-L2 is essential for proper dampening of insulin signaling and larval survival.

Conclusion: Imp-L2, the first functionally characterized insulin-binding protein in invertebrates, serves as a nutritionally controlled suppressor of insulin-mediated growth in Drosophila. Given that Imp-L2 and the human tumor suppressor IGFBP-7 show sequence homology in their carboxy-terminal immunoglobulin-like domains, we suggest that their common precursor was an ancestral insulin-binding protein.

Figure 1

Imp-L2 overexpression suppresses dInR-induced growth. (a-e) Scanning electron micrographs of compound eyes. All flies (females) carry the GMR-Gal4 and UAS-dInR^wttransgenes. The dInr-dependent big eye phenotype (a) is suppressed by EP5.66 (b). UAS-Imp-L2 (c) and the stronger UAS-s.Imp-L2 (d) also suppress, but EP5.66 driving the mutant Imp-L2^MG2 allele can no longer suppress the dInR overexpression phenotype (e). (f) Genomic organization of the Imp-L2 locus. The mutant alleles and P-element insertions used in this study are indicated. MG2 marks the point mutation in the EMS allele Imp-L2^MG2 that generates a premature stop codon. (g) Alignment of Imp-L2, its orthologs in invertebrates and the putative human ortholog IGFBP-7. Black and gray boxes indicate amino acid identity and similarity, respectively. The triangle marks the premature stop codon in Imp-L2^MG2. Asterisks mark the cysteines forming the two disulfide bridges. The gray bars indicate the Ig domains. Dm, Drosophila melanogaster Imp-L2; Ag, Anopheles gambiae CP2953; Sf, Spodoptera frugiperda IBP; Ce, Caenorhabditis elegans zig-4; Hs, Homo sapiens IGFBP-7.

Figure 2

Imp-L2 controls body and organ size. (a) Tangential section through an adult eye containing an Imp-L2 overexpression clone marked by the lack of red pigment. Within the clone, the size of the ommatidia is reduced. Wild-type ommatidia close to the clone are also smaller (compare black circled areas). (b) Eye-specific overexpression of UAS-s.Imp-L2 reduces male body weight (-38.3%, P = 7 × 10^-42). (c) Overexpression of UAS-Imp-L2 by ppl-Gal4 results in a 56.1% weight reduction in male flies, whereas ppl-Gal4 driven expression of UAS-s.Imp-L2 results in lethality (†). P = 3 × 10^-47. (d) Loss of Imp-L2 function increases body size in males (top) and females (bottom). (e) Analyses of male and female weights. Wing area, cell number and cell size were assessed in female adult wings. GR indicates Imp-L2 genomic rescue construct. P-values are indicated by numbers as follows: 1, 2 × 10^-33; 2, 8 × 10^-18; 3, 9 × 10^-16; 4, 6 × 10^-7; 5, 3 × 10^-46; 6, 1 × 10^-24; 7, 8 × 10^-31; 8, 2 × 10^-7; 9, 4 × 10^-4; 10, 4 × 10^-7; 11, 3 × 10^-7; 12, 1 × 10^-4. Genotypes: 'control' y, w/w; 'Imp-L2^-/-' y, w; Imp-L2^Def42/Imp-L2^Def20; 'Imp-L2^+/-' for the weight analysis (e) y, w; Imp-L2 ^Def20/+; 'Imp-L2^+/-' for the wing analysis (e) y, w; Imp-L2 ^Def42/+; 'Imp-L2^-/-, GR' y, w; Imp-L2 ^Def42/Imp-L2 ^Def20, GR-57; 'Imp-L2^+/-, GR' y, w; Imp-L2 ^Def20, GR-57/+. P-values were determined using unpaired Student's t-test against the control except in (4) where the weight of IMP-L2^+/- was compared to IMP-L2^{+/-, GR}. n = 40 for the weight analysis in (b,c,e); n = 12 for the wing analysis in (e). Error bars represent s.d.

Figure 3

Imp-L2 binds Dilp2 and counteracts its activity. (a,b) Antibody staining of larval brains with an Imp-L2 antibody (green). (a) Specific neurons of both brain hemispheres, the subesophageal ganglion region (gray arrow) and the corpora cardiaca (white arrow) express Imp-L2 protein. The corpora allata are innervated by Imp-L2 expressing axons. White arrowheads mark the Dilp-producing m-NSCs. (b) In larvae carrying a dilp2-lacZ.nls transgene, co-staining with β-galactosidase and Imp-L2 antibodies reveals that the seven dilp-expressing m-NSCs also produce low levels of Imp-L2. (c) The size increase of arm-Gal4, UAS-dilp2 flies is dominantly enhanced by reducing Imp-L2 levels. In an Imp-L2^-/- background, dilp2 overexpression results in lethality, which can be rescued by a copy of the Imp-L2 genomic rescue construct (GR). (d) Overexpression of dilp2 as well as of Imp-L2 at high levels by ppl-Gal4 causes lethality, whereas concomitant overexpression of dilp2 and Imp-L2 yields flies of wild-type size. The lacZ transgene was introduced to rule out a dosage effect of the UAS/Gal4-system. (e) Imp-L2 binds Dilp2. In-vitro-translated, ³⁵S-labeled wild-type (ImpL2-IVT, about 32 kDa) or mutant (ImpL2^MG2-IVT, about 30 kDa) Imp-L2 (lane 1) was incubated with cell lysates of either non-transfected (lane 3) or stably transfected S2 cells expressing Flag-Dilp2 (lane 2). Imp-L2 could only be pulled down in the presence of Dilp2. The Imp-L2^MG2 mutation abolished Dilp2 binding. Genotypes in (c): 'Imp-L2^+/-' Imp-L2^Def42/+; 'Imp-L2^-/-' Imp-L2^Def42/Imp-L2^Def20; 'Imp-L2^-/-, GR' Imp-L2^Def42/Imp-L2^Def20, GR-57; 'control' (black bar) arm-Gal4, UAS-GFP. P-values were determined using unpaired Student's t-test (n = 40, except for bars 1–3 in (c): bar 1, n = 31; bars 2 and 3, n = 17). Error bars represent s.d.

Figure 4

Imp-L2 is necessary for blocking dInR signaling under starvation. (a,b) tGPH fluorescence (green, showing PIP₃ levels and thus indicating IIS activity) in the fat body of feeding third instar larvae under different nutritional conditions. Nuclear staining (Hoechst) is shown in blue in the right panels. (a) Under normal conditions ('yeast'), IIS activity is high in wild-type feeding third instar larvae. Upon starvation, only little PIP₃ localizes to the membranes of fat body cells. (b) In Imp-L2 mutants, IIS activity is higher than in control larvae and only slightly reduced after 4 h PBS starvation. (c) Survival of Imp-L2^Def42/Imp-L2^Def20early third instar larvae is severely compromised under starvation conditions. One copy of the genomic rescue construct (GR) suffices to restore viability. Heterozygous larvae were Imp-L2^Def42/+, control larvae y, w/w. Larvae (40) were subjected for 24 h to 20% glucose, 1% glucose or PBS. The experiment was repeated twice. (d) In starved larvae (y, w), Imp-L2 protein expression (green) is induced in fat body cells after 24 h PBS starvation. Imp-L2 is localized to vesicle-like structures but not detectable under normal nutritional conditions. Genotypes: (a,d) y, w; (b) y, w; Imp-L2^Def42/Imp-L2^Def20.