MRT, Functioning with NURF Complex, Regulates Lipid Droplet Size Yan Yao, Xia Li, Wei Wang, Zhonghua Liu, Jianming Chen, Mei Ding, Xun Huang  Cell Reports  Volume 24, Issue 11, Pages 2972-2984 (September 2018) DOI: 10.1016/j.celrep.2018.08.026 Copyright © 2018 The Author(s) Terms and Conditions

Cell Reports 2018 24, 2972-2984DOI: (10.1016/j.celrep.2018.08.026) Copyright © 2018 The Author(s) Terms and Conditions

Figure 1 Overexpressing mrt Increases the Size of LDs (A) BODIPY staining (green) of larval fat bodies. Nuclei were stained by DAPI (blue). Scale bar represents 25 μm. (B) Quantification of the size of LDs in (A) (n = 30 for each genotype). Data were analyzed by one-way ANOVA with a post hoc Tukey’s multiple-comparison test. Each point represents a single LD. (C) The genomic structure of mrt and the domain structure of MRT. The RNA isoform A (RA), RB, and RC of mrt have alternative 5′ exons. White boxes: UTRs of mrt isoforms, black boxes: mrt coding regions, yellow box: region encoding the Myb/SANT-like domain. The transposon insertion positions of the UAS lines mrtF931 and mrtEY10964, the double-stranded RNA (dsRNA) target regions of mrtGD10947 and mrtkk112537, and the deleted region in the CRISPR/Cas9 allele mrtA-3, which results in a frameshift, are indicated. (D) Relative mrt expression levels in the fat bodies were determined by qRT-PCR in two technical repeats (>15 flies for each genotype). (E and F) The relative levels of neutral lipids (E) and phospholipids (F) (normalized to total polar lipids) in the larval fat bodies of different backgrounds were determined in biological triplicates (each replicate contains 15 flies). Data were analyzed with the Kruskal-Wallis test. Error bars represent means ± SDs. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ns: non-significant. See also Figures S1 and S2. Cell Reports 2018 24, 2972-2984DOI: (10.1016/j.celrep.2018.08.026) Copyright © 2018 The Author(s) Terms and Conditions

Figure 2 mrt Mutants Have Small LDs (A, D, and G) The LD of mrt RNAi knockdown (A), fed mrt mutant (D), and starved mrt mutant (G) visualized by BODIPY staining of larval fat bodies. Scale bar represents 25 μm. (B, E, and H) Quantification of the size of LDs (n = 30 for each genotype). (B) was analyzed by one-way ANOVA with a post hoc Tukey’s multiple-comparison test; (E) was analyzed by one-way ANOVA with a post hoc Dunnett’s comparison test; (H) was analyzed by Student’s t test. (C) Relative mrt expression levels in the fat bodies were determined by qRT-PCR in four technical repeats (>30 flies for each genotype). (F) Glucose levels and trehalose levels of w1118 and mrtA-3 were determined in at least four biological replicates (each replicate contains five flies). The difference was analyzed by Student’s t test. (I) Relative total glyceride levels of w1118 control and mrtA-3 third-instar larvae after 24 hr of starvation were determined in seven biological replicates (each replicate contains five flies). Total glyceride levels were normalized to total protein levels. Data were analyzed by Student’s t test. Error bars represent means ± SDs. ∗p < 0.05; ∗∗∗p < 0.001; ns: non-significant. Cell Reports 2018 24, 2972-2984DOI: (10.1016/j.celrep.2018.08.026) Copyright © 2018 The Author(s) Terms and Conditions

Figure 3 The Function of MRT Depends on PZG and NURF (A) Immunofluorescent staining of FLAG-MRT (red) in third-instar larval fat bodies. Nuclei were stained by DAPI (blue). Scale bar represents 25 μm. (B and F) Analyzing the phenotype of pzg or NURF complex subunits (B) and their genetic interactions with mrt (F) by BODIPY staining of larval fat bodies. Scale bar represents 25 μm. (C and G) Quantification of the size of LDs in (B) and (F) (n = 30 for each genotype), respectively. Each point represents a single LD. Data were analyzed by one-way ANOVA with a post hoc Dunnett’s comparison test (C) and post hoc Tukey’s multiple-comparison test (G). (D) Eyes of flies from the PEV reporter line wm4 on wild-type and mrtA-3 heterozygous backgrounds. Scale bar represents 100 μm. (E) Spectrophotometric quantification of eye pigment in flies from (D) were determined in at least five biological replicates (each replicate contains 10 fly heads). Data were analyzed by Student’s t test. Error bars represent means ± SDs. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ns: non-significant. See also Figures S3 and S4 and Table S1. Cell Reports 2018 24, 2972-2984DOI: (10.1016/j.celrep.2018.08.026) Copyright © 2018 The Author(s) Terms and Conditions

Figure 4 MRT Physically Interacts with PZG and NURF55 (A) Co-immunoprecipitation of FLAG-MRT with Myc-PZG by Myc antibody from a lysate of co-transfected Drosophila S2 cells. Blot was representative of triplicate experiments. IP, immunoprecipitate. (B) Co-immunoprecipitation of FLAG-MRT with Myc-NURF55 by Myc antibody from lysates of co-transfected Drosophila S2 cells. Blot was representative of four replicate experiments. (C) Coomassie blue staining of purified FLAG-MRT, Myc-PZG, and Myc-NURF55 from transfected Drosophila S2 cells. (D) Purified FLAG-MRT co-immunoprecipitates with purified Myc-NURF55 and Myc-PZG. Blot was representative of duplicate experiments. (E) Schematic diagrams of full-length or truncated MRT. FL, full length; NT, N-terminal alone; ΔN, N-terminal deletion; SD, Myb/SANT-like domain alone; ΔC, C-terminal deletion; CT, C-terminal alone. (F) ΔN, SD, and ΔC truncations of MRT co-immunoprecipitate with FLAG-PZG from lysates of co-transfected Drosophila S2 cells. Blot was representative of four replicate experiments. (G) ΔN, SD, and ΔC truncations of MRT co-immunoprecipitate with FLAG-NURF55 from lysates of co-transfected Drosophila S2 cells. Blot was representative of four replicate experiments. Cell Reports 2018 24, 2972-2984DOI: (10.1016/j.celrep.2018.08.026) Copyright © 2018 The Author(s) Terms and Conditions

Figure 5 mrt Regulates the Size of LDs through Transcriptional Regulation of plin1 (A) Relative mRNA levels of LD size- and lipid storage-regulating genes in ppl-GAL4 control and ppl>mrt larval fat bodies were determined by qRT-PCR in two technical repeats (>15 flies for each genotype). (B) Nile red (red) staining of larval fat bodies. Nuclei were stained by DAPI (blue). Scale bar represents 25 μm. (C) Quantification of the size of LDs in (B) (n = 30 for each genotype). Data were analyzed by one-way ANOVA with a post hoc Tukey’s multiple-comparison test. Each point represents a single LD. (D) Relative mRNA levels of mrt and Cct1 in larval fat bodies of different genotypes were determined by qRT-PCR in two technical repeats (>15 flies for each genotype). (E) BODIPY staining of larval fat bodies. Scale bar represents 25 μm. (F) Quantification of the size of LDs in (E) (n = 30 for each genotype). Data were analyzed by one-way ANOVA with a post hoc Tukey’s multiple-comparison test. Each point represents a single LD. (G) Results showing the occupancy of FLAG-MRT on the promoter of plin1 (P1–P8) was determined by ChIP-qPCR in two technical repeats (>150 flies for each genotype). rp49 and the 3′-UTR of plin1 (P9) served as negative controls. (H) Relative mRNA levels of plin1 in larval fat bodies of different genotypes were determined by qRT-PCR in at least two technical repeats (>15 flies for each genotype). (I) The occupancy of Myc-NURF55 and PZG-GFP on the P4 region of the plin1 promoter was determined by ChIP-qPCR in two technical repeats (>150 flies for each genotype). The 3′-UTR of plin1 (P9) served as a negative control. (J) The abundance of H3K4me3, H3K9ac, and H3K27ac on the P4 region of the plin1 promoter was determined by ChIP-qPCR in two technical repeats (>1,000 flies for each genotype). Error bars represent means ± SDs. ∗∗∗p < 0.001; ns: non-significant. See also Figures S5 and S6 and Table S2. Cell Reports 2018 24, 2972-2984DOI: (10.1016/j.celrep.2018.08.026) Copyright © 2018 The Author(s) Terms and Conditions

Figure 6 MRT and PLIN1 Affect LD Coalescence In Vitro (A and B) Time-lapse images showing the coalescence of LDs isolated from larval fat bodies of plin138 mutants (A), mrt-overexpressing larvae (B), and their controls. Red dotted circles show the LDs that undergo fusion in the first hour; green dotted circles show the LDs that undergo fusion in the second hour. Scale bar represents 20 μm. (C) The coalescence frequencies of LD pairs of different genotypes. (D and E) LD size comparison of coalesced (D) and non-coalesced (E) pairs from plin138 mutants. Lines connect the two LDs in a pair. (F and G) LD size comparison of coalesced (F) and non-coalesced (G) pairs from mrt-overexpressing larvae. Lines connect the two LDs in a pair. (H) Diameter ratios of paired LDs from different genotypes (w1118, n = 41; plin138, n = 34; ppl-GAL4/+, n = 50; ppl>mrt, n = 54). Data were analyzed by one-way ANOVA with a post hoc Tukey’s multiple-comparison test. Each point represents the diameter ratio of one LD pair. (I and J) Diameter ratios of coalesced (plin138, n = 32; ppl>mrt, n = 32) and non-coalesced paired LDs (plin138, n = 2; ppl>mrt, n = 22) from plin138 mutants (I) or mrt-overexpressing (J) larvae. Each point represents the diameter ratio of one LD pair. (K) Schematic model showing how mrt regulates LD size. The transcriptional levels of LD size regulatory genes, including plin1, are affected by the chromatin state. The MRT-PZG-NURF axis mediates the “closed” state, while some unknown factor(s) may mediate the “open” chromatin state. Error bars represent means ± SDs. ns: non-significant. Cell Reports 2018 24, 2972-2984DOI: (10.1016/j.celrep.2018.08.026) Copyright © 2018 The Author(s) Terms and Conditions