A nucleostemin family GTPase, NS3, acts in serotonergic neurons to regulate insulin signaling and control body size

Abstract

Growth and body size are regulated by the CNS, integrating the genetic developmental program with assessments of an animal's current energy state and environmental conditions. CNS decisions are transmitted to all cells of the animal by insulin/insulin-like signals. The molecular biology of the CNS growth control system has remained, for the most part, elusive. Here we identify NS3, a Drosophila nucleostemin family GTPase, as a powerful regulator of body size. ns3 mutants reach <60% of normal size and have fewer and smaller cells, but exhibit normal body proportions. NS3 does not act cell-autonomously, but instead acts at a distance to control growth. Rescue experiments were performed by expressing wild-type ns3 in many different cells of ns3 mutants. Restoring NS3 to only 106 serotonergic neurons rescued global growth defects. These neurons are closely apposed with those of insulin-producing neurons, suggesting possible communication between the two neuronal systems. In the brains of ns3 mutants, excess serotonin and insulin accumulate, while peripheral insulin pathway activation is low. Peripheral insulin pathway activation rescues the growth defects of ns3 mutants. The findings suggest that NS3 acts in serotonergic neurons to regulate insulin signaling and thus exert global growth control.

Figure 1.

Growth phenotype of ns3 mutant. (A) Domain structure of NS3. (B) ns3 genomic locus (top) and exon/intron organization showing P-element insertion (bottom). (C) Mutation of ns3 retards larval growth. Body length was compared between ns3 heterozygous (het) and mutant (mut) larvae. (*) P = 0.0031, (**) P = 0.0011, (***) P < 0.0001 for ns3 het versus mut (n = 10 per condition per time point). (D) Wing imaginal discs are smaller in ns3 mutant larvae. (E,F) ns3 mutant animals exhibit a profound developmental delay and reduced viability that can be rescued by expression of NS3-YFP using a Tubulin-Gal4 driver (tub). (*) P < 0.0001 for ns3 mutant (n = 146) compared with control (n = 388) or ns3 mutant + NS3-YFP (tub) (n = 221). (G) ns3 mutants are dramatically reduced in size and the size defect can be rescued by ubiquitous expression of an ns3-yfp transgene. (*) P < 0.0001 for ns3 mutant (n = 6 groups) compared with control (n = 3 groups) or ns3 mutant + NS3-YFP (tub) (n = 6 groups). Bar: D, 100 μm. Data in C, F, and G are represented as mean ± SEM.

Figure 2.

Cell size and number are reduced in ns3 mutant wings. (A) Average cell size was determined in the compartment between veins vII and vIII of wild-type or mutant adult wings from male flies. (Right) Images (80 × 80 μm) from wild-type (wt) and ns3 mutant (ns3 mut) wings. (Left) Quantification of cell area in wild-type and ns3 mutant wings. (B) Average cell number was determined by dividing the total wing area of wild-type and ns3 mutant wings by the calculated cell area from A. (Right) Images of wild-type and ns3 mutant wings. (Left) Quantification of cell number in wild-type and ns3 mutant wings. (*) P < 0.0001 for wild type (n = 3 groups) compared with ns3 mutant (n = 6 groups). Data are represented as mean ± SEM.

Figure 3.

NS3 regulates growth in a non-cell-autonomous manner. (A) Production of NS3-YFP in larval wing imaginal discs and adult wings of otherwise ns3 mutant animals using a 71B-Gal4 driver (71B) did not rescue the reduced wing size of the ns3 mutants. (B) Production of NS3-YFP in the posterior compartment of larval wing imaginal discs and adult wings of ns3 mutants using an Engrailed-Gal4 driver (en) did not lead to an increase in cell size in the posterior of the wing. (C) FLP-FRT-mediated recombination was used to generate ns3^−/− cell clones and wild-type twin-spot clones in otherwise ns3 heterozygous animals. The ns3 mutant and wild-type clones were indistinguishable in size, indicating that absence of NS3 from patches of wing cells does not change their size. ns3^−/− cell clones are marked by two copies of GFP, while wild-type twin-spot clones lack GFP. Surrounding heterozygous tissue has one copy of GFP. Data in A and B are represented as mean ± SEM. In A, n = 3 groups for wild type, 4 for ns3 mutant, and 6 for ns3 mut NS3-YFP (71B). In B, n = 3 groups for wild type, 6 for ns3 mutant, and 4 for ns3 mut NS3-YFP (71B).

Figure 4.

NS3 acts in serotonergic neurons to control developmental timing. Shown are growth curves of control animals (blue lines), ns3 mutant animals (harboring, but not expressing, a UAS-NS3-YFP transgene) (red lines), or ns3 mutant animals expressing a UAS-NS3-YFP transgene under the control of the indicated tissue-specific Gal4 driver (green lines). The percent of adult animals emerging is represented on the Y-axis and the number of days after egg lay (AEL) is represented on the X-axis. Only expression of NS3-YFP in all neurons (C, ELAV-Gal4) or in serotonergic neurons (E, DDC-Gal4) results in a rescue of the developmental delay exhibited by ns3 mutant animals.

Figure 5.

NS3 acts in serotonergic neurons to control organismal growth. Average weights of control animals (blue bars), ns3 mutant animals (harboring, but not expressing, a UAS-NS3-YFP transgene) (red bars), or ns3 mutant animals expressing a UAS-NS3-YFP transgene under the control of the indicated tissue-specific Gal4 driver (green bars) are shown. Only expression of NS3-YFP in all neurons (C, ELAV-Gal4) or in serotonergic neurons (E, DDC-Gal4) results in rescue of the reduced weight exhibited by ns3 mutant animals. Data are represented as mean ± SEM. (*) P-values: (C) ns3 mut (n = 4 groups) versus ns3 mut + NS3-YFP (ELAV) (n = 5 groups), P = 0.0007. (E) ns3 mut (n = 3 groups) versus ns3 mut + NS3-YFP (DDC) (n = 3 groups), P = 0.0002.

Figure 6.

Projections from serotonergic neurons and IPCs are closely apposed. (A) Serotonin levels are elevated in ns3 mutants. Levels of serotonin (5-HT) were measured by ELISA and are represented as mean ± SEM per head, normalized to body weight. (*) P < 0.05 (n = 4 groups each). (B) Low-magnification 3D projection of a confocal Z-stack of a whole-mount wild-type Drosophila brain from a wandering third instar larva expressing GFP in the IPCs under control of a dilp2 promoter (green) and immunostained for 5-HT (red). (C,D) Higher-magnification images of brain lobes (C) and ventral ganglion (D), acquired as in A. Arrowheads in C denote regions where the neuropil containing serotonergic processes is in close proximity to the IPCs. Arrow in C denotes apposition of processes from the IPCs and serotonergic neurons in the commissure between the brain lobes. Arrowheads in D denote regions of apposition between serotonergic and IPC processes along the length of the ventral ganglion. Individual channels are shown in grayscale images at left. Ninety-degree rotation about the Y-axis of the Z-stacks are shown at right. Note that the serotonergic and IPC projections are in the same plane. Bars: B, 40 μm; C,D, 20 μm.

Figure 7.

NS3 acts in serotonergic neurons as an upstream regulator of insulin signaling. (A–G) DILP2 levels are elevated in ns3 mutant brains. Brains from third instar larvae were immunostained using an antibody to DILP2 and imaged by confocal microscopy using identical laser power and scan settings. High-magnification images are shown in A–C, and lower-magnification images are shown in D–F, and in these images, DILP2 is shown in green and DNA counterstain is shown in blue. Fluorescence intensity in the IPCs is quantified in G. (*) P < 0.0001 for ns3 mutant (n = 35) versus wild type (n = 8) or for ns3 mutant versus ns3 mutant + NS3-YFP (DDC) (n = 30). (H) Insulin pathway activation is impaired and small ribosomal subunit protein levels are reduced in ns3 mutant larvae. Immunoblot for indicated proteins from equal amounts of ns3 heterozygous (ns3 het) or mutant (ns3 mut) larval extract. Data are from one experiment that is representative of three independent experiments. (I) Akt activation and ribosomal protein S6 (RPS6) levels are not altered in ns3 mutant cell clones, but Akt activation is reduced in TOR^ΔP mutant clones. For ns3 mutant clones, wing imaginal disc tissue in which mitotic clones have been generated is marked as mutant by two copies of GFP (green in merge) or wild type by lack of GFP. Surrounding heterozygous tissue has one copy of GFP. For TOR^ΔP mutant clones, wing imaginal disc tissue in which mitotic clones have been generated is marked as mutant by lack of GFP or wild type by two copies of GFP (green in merge). Surrounding heterozygous tissue has one copy of GFP. Tissue was immunostained for phospho-S505-Akt1 (top and bottom panels, red in merge) or RPS6 (middle panel, red in merge). (J–L) Expression of constitutively-active Akt1 T342D, S505D (Akt1) in the eye using a GMR-promoter fusion rescues the reduction in eye size (J), ommatidia size (K), and ommatidia number (L) observed in ns3 mutant flies. Images in J show scanning electron micrographs of whole eyes (eye size quantified at right). Images in K show scanning electron micrographs of ommatidia (ommatidia size quantified at right and ommatidia number is quantified in L). (*) P-values: (J) wild type versus ns3 mut, P = 0.001; ns3 mut versus wild type + Akt1, P < 0.0001; ns3 mut versus ns3 mut + Akt1, P < 0.0001; wild type + Akt1 versus ns3 mut + Akt1, P = 0.77. (K) Wild type versus ns3 mut, P = 0.0023; ns3 mut versus wild type + Akt1, P = 0.0009; ns3 mut versus ns3 mut + Akt1, P = 0.0018; wild type + Akt1 versus ns3 mut + Akt1, P = 0.43. (L) Wild type versus ns3 mut, P = 0.0048; ns3 mut versus wild type + Akt1, P = 0.0034; ns3 mut versus ns3 mut + Akt1, P = 0.011; wild type + Akt1 versus ns3 mut + Akt1 P = 0.55. Bars: A–F, 25 μm; J, 100 μm; K, 15 μm. Data are represented as mean ± SEM. n = 3 per condition.