The Drosophila estrogen-related receptor directs a metabolic switch that supports developmental growth

Abstract

Metabolism must be coordinated with development to provide the appropriate energetic needs for each stage in the life cycle. Little is known, however, about how this temporal control is achieved. Here, we show that the Drosophila ortholog of the estrogen-related receptor (ERR) family of nuclear receptors directs a critical metabolic transition during development. dERR mutants die as larvae with low ATP levels and elevated levels of circulating sugars. The expression of active dERR protein in mid-embryogenesis triggers a coordinate switch in gene expression that drives a metabolic program normally associated with proliferating cells, supporting the dramatic growth that occurs during larval development. This study shows that dERR plays a central role in carbohydrate metabolism, demonstrates that a proliferative metabolic program is used in normal developmental growth, and provides a molecular context to understand the close association between mammalian ERR family members and cancer.

Figure 1. dERR mutants exhibit metabolic defects

(A) A schematic representation of the dERR locus is depicted along with the flanking genes atg18 and CG7979. The dERR¹ lesions are shown, including an 84 bp deletion that removes the exon 2 splice acceptor (Δ) and point mutations in exons 2 and 3 (*). The dERR² deletion removes the entire dERR coding region and portions of the neighboring genes. The arrows at the 3′ end of atg18 indicate that it extends beyond what is depicted. (B–E) w¹¹¹⁸ control and dERR¹/dERR² mutants (dERR^–) were collected as mid-second instar larvae and whole animal homogenates were analyzed for concentrations of (B) ATP, (C) trehalose, (D) glycogen, or (E) TAG. Amounts of ATP, glycogen, and TAG were normalized to soluble protein levels. Mutant animals contain lower levels of (A) ATP (p < 1 × 10⁻⁴) and (E) TAG (p < .001), but have higher concentrations of (C) trehalose (p < 1 × 10⁻¹⁸) and normal levels of (D) glycogen (p = 0.44). n>20 independently collected samples per value with 25 animals per sample. Error bars are ± S.E.

Figure 2. Genes involved in carbohydrate metabolism are down-regulated in dERR mutants

(A) Gene ontology (GO) analysis of the 572 up-regulated and 334 down-regulated genes in dERR¹/dERR² mutant animals relative to w¹¹¹⁸ controls. The top GO categories for each gene set are listed in order of significance along with the number of genes affected in that category, the total number of genes in that category (in parentheses), and the statistical significance of the match. (B) A diagram of glycolysis is depicted that displays the glycolytic genes that are down-regulated in dERR mutants followed by their fold-change in expression from the microarray. (C) Trehalose was measured in extracts from Pfk⁰⁶³³⁹/+ (Control) and Pfk⁰⁶³³⁹/Df(2R)BSC303 (Pfk⁻) mid-second instar larvae, revealing elevated levels in Pfk mutants. (D) Trehalose concentrations were determined for dERR²/+, da-GAL4 controls (grey box), dERR¹/dERR²; da-GAL4 mutants (black box), and UAS-Pgi; dERR¹/dERR², UAS-Pfk, da-GAL4 animals (white box). Trehalose levels are rescued when both transgenes are expressed using the ubiquitous da-GAL4 driver.

Figure 3. dERR mutants display changes in the levels of specific metabolites

GC/MS was used to compare the relative levels of small metabolites in CantonS (blue) and w¹¹¹⁸ (green) controls with dERR¹/dERR² (orange) and dERR¹/Df(3L)Exel6112 (red) mutant second instar larvae. dERR mutants exhibit (A) elevated levels of glucose-6-phosphate, mannose-6-phosphate, and sorbital, along with (B) diminished concentrations of lactate. (C) The relative amounts of citrate, isocitrate, and succinate are similar among the four strains, while α-ketoglutarate, fumarate, and malate levels are decreased in mutant larvae. (D) Methionine levels are normal in mutant animals, while proline concentrations are significantly lower. Glutamine and alanine levels appear to be slightly decreased in mutant strains, and aspartate is the only amino acid that is elevated in dERR mutants. All data are graphically represented as a box plot, with the box representing the lower and upper quartiles, the horizontal line representing the median, and the bars representing the minimum and maximum data points. n = 6 samples collected from independent populations with 25 larvae per sample (See Table S3 for p values). Similar results were observed in two additional independent experiments (Table S3).

Figure 4. dERR is required for the coordinate induction of glycolytic gene expression during mid-embryonic development

(A–B) Total RNA from staged (A) w¹¹¹⁸ control and (B) dERR¹/dERR² mutant embryos, first instar larvae (L1), and second instar larvae (L2) was analyzed by northern blot hybridization to detect the expression of transcripts encoding glycolytic enzymes. (A) Glycolytic genes are coordinately induced during mid-embryogenesis in control animals, but (B) not in dERR mutants. A sample of RNA from 24–28 hr w¹¹¹⁸ first instar larvae (Ctrl) was included on the blot of dERR mutant RNA to facilitate comparisons between control and dERR mutant animals. (C) The temporal expression pattern of a lacZ reporter construct that carries a multimerized dERR binding site from the Pfk locus was analyzed by northern blot hybridization in staged w¹¹¹⁸ control or dERR¹ mutant larvae to detect the expression of lacZ or Pfk mRNA. The reporter is expressed in synchrony with zygotic Pfk expression in control animals and is dependent on dERR function. Hybridization to detect rp49 mRNA is included as a control for loading and transfer.

Figure 5. dERR protein accumulation is temporally regulated during embryonic development

Staged dERR¹ mutant embryos that carry a dERR-GFP rescue transgene were visualized by confocal microscopy to detect GFP fluorescence. (A) No GFP was detected between 6–12 hrs after egg laying (AEL), while (B) dERR-GFP accumulates in the nuclei of several cell types, including the muscle and epidermis, between 12–18 hrs AEL.

Figure 6. dERR functions in peripheral tissues to control carbohydrate metabolism

Mid-second instar larvae of five genotypes were tested for trehalose levels or the expression of specific dERR target genes: dERR¹/+; UAS-dERR controls (white boxes), dERR¹/dERR²; UAS-dERR mutants (black boxes), dERR²/+; GAL4 controls (blue boxes), dERR¹/dERR²; GAL4 mutants (red boxes), or rescued dERR¹/dERR²; UAS-dERR, GAL4 animals (grey boxes). Six GAL4 transgenes were used to drive UAS-dERR expression: da-GAL4 (Ubq, ubiquitous expression), r4-GAL4 (fat body), dmef2-GAL4 (muscle), A58-GAL4 (Epi, epidermis), mex-GAL4 (midgut), and dilp2-GAL4 (IPCs, Insulin-Producing Cells). Total RNA was analyzed by northern blot hybridization to detect UAS-dERR expression, three genes involved in glycolysis (Pgi, Pfk, and ImpL3), and a gene in the pentose phosphate pathway (Pgd). Hybridization to detect rp49 mRNA was used as a control for loading and transfer. The apparent reduced level of mex>dERR relative to A58>dERR is an artifact of the blot hybridization. These levels of expression are comparable. In contrast, the low level of dilp2>dERR expression likely reflects the small number of cells that express the dilp2 driver.