METHOD DETAILS

The full-length cDNAs of Atg1 (LD18893), CPSF5 (SD03330), CPSF6 (LD25239), CstF64 (RE27227), CPSF160 (LD38533), Cbc (LD15072) and Pcf11 (MIP05908), were cloned into the Drosophila Gateway vector pAWG. GFP was cloned into pAWM as a control. GFP-CPSF6^ΔRS (amino acids 1–400) which contains the RRM domain was created by PCR amplification and verified by DNA sequencing. Flag-CDK8 and HA-DOA were constructed following PCR from cDNA clones (RE13344 and RE04477) and cloned into pcDNA3.1-Flag and pWALIUM10-moe, respectively. Using PCR mutagenesis, we generated CDK8-KD (kinase-dead) and DOA-KD mutants by replacing the conserved ATP-binding Lysine 52 and Lysine 193 in the kinase domain with Alanine, respectively. The Atg1-PB-S279A mutant was generated by replacing Serine 297 with Alanine. CPSF6^14A was generated by custom gene synthesis (IDT) and replacing Serines 459, 468, 525, 527, 531, 541, 543, 549, 571, 573, 588, 596, 622, and Threonine 404 with Alanines and cloned into Drosophila Gateway vector pAWG. For the generation of the GST fusion protein, DNA sequences corresponding to amino acids 1–240 and 221–652 of CPSF6 and full length of CPSF5 were PCR-amplified and subcloned into the pGEX-2T vector. GFP-human CPSF6 was obtained from GeneCopoeia (EX-Mm18988-M29). UAS-Flag-Fip200 was a gift of Dr. Jun Hee Lee (Kim et al., 2013)

Second larvae were collected 72–96 hrs after egg laying and then cultured in fresh fly media supplemented with yeast paste (Fed), or in vials containing 20% sucrose (Starved) for 4hrs or 16 hrs.

Total RNA was extracted from larval fat body using TRIzol® reagent (Invitrogen, 15596-018). 4 μg of total RNA was used for cDNA synthesis and 3′RACE by using SMARTer RACE 5′/3′ Kit (Clontech, 634858). Gene-specific forward primers and a universal reverse primer were used for the first-round and nested PCR amplification as indicated. The PCR products were cloned using TOPO® TA Cloning vector (Invitrogen, K4500-01) and sequenced. Gene specific primers are listed in Table S5.

The firefly luciferase reporter plasmids containing long or short 3′ UTRs of Atg1 and Atg8a were constructed using the primers listed in Table S5. Experiments were performed in 96-well plates excluding the outer wells. Cells were transfected with firefly luciferase reporter plasmids and Renilla luciferase reporter plasmid for transfection control. After 72 hrs, luciferase activities were measured using DualGlo (Promega, E2980).

Drosophila wild-type, TSC1KO and TSC2KO S2R+ cells (Housden et al., 2015) were cultured in Schneider’s medium supplemented with 10% fetal bovine serum (FBS) at 25°C. For Rapamycin (LC Laboratories, R-5000), MG132 (Calbiochem, 474791), or Actinomycin D (Calbiochem, 114666) treatment, S2R+ cells were treated with 20nM Rapamycin or 20 μM MG132 for 24hrs, or 10μg/ml Actinomycin D for the indicated time points.

For RNAi experiments, PCR templates for dsRNA against PKA (DRSC03399 and DRSC31381) CDK8 (DRSC28684 and DRSC41558) and DOA (DRSC16650 and DRSC40713) were prepared using the MEGAscript® T7 Transcription Kit (Invitrogen, {"type":"entrez-protein","attrs":{"text":"AMB13345","term_id":"982891170","term_text":"AMB13345"}}AMB13345). DsRNA against the bacterial β-galactosidase gene (lacZ) was used as a control. S2R+ cells were dispensed into assay plates containing dsRNAs at a standard concentration for the ‘bathing’ method (https://fgr.hms.harvard.edu/drsc-cell-rnai).

HEK293T and MCF7 cells were cultured at 37°C in DME M (Invitrogen, 10-017-CV) medium supplemented with 10% FBS (complete medium). For nutrient starvation, cells were starved for 2 hrs in serum-free Earle’s balanced salt solution (EBSS) medium (Sigma, 14155063) plus 100 nM Bafilomycin A1 (Sigma, B1793), 50 μM Senexin A (Tocris, 4875), or 50 μM TG003 (Sigma, T5575) where indicated. The production and infection of lentivirus carrying CPSF6 shRNA clones were performed as described previously (Tang et al., 2013; Tang et al., 2011). The target sequences of these clones are CPSF6 shRNA #1: (Clone ID: TRCN0000237833) 5′ GTTGTAACTCCATGCAATAAA 3′ and CPSF6 shRNA #2: (Clone ID: TRCN0000244314) 5′ GGTGATTATGGGAGTGCTATT 3′. Luciferase shRNA was used as a control.

Antibodies used for the study were: anti-Atg8 (LSBio, LS-B4021), anti-GFP (Molecular Probes, A6455), anti-phospho-Ser (Santa Cruz, sc-81514), anti-phospho-Threnonine (Cell Signaling, 9381), anti-Phospho-PKA substrate (Cell Signaling, 9624), anti-Flag (Sigma, F3165), anti-HA (Covance/BioLegend, MMS-101P), anti-CDK8 (Santa Cruz, sc-13155 ), anti-LC3B (Cell Signaling, 2775), anti-human CPSF6 (Santa Cruz, sc-100692), anti-ubiquitin FK2 (Enzo life Science, BML-PW8810-0100), and anti-Tubulin (Sigma, T5168). Anti-Atg1 was a gift of Dr. Jun Hee Lee (Kim et al., 2013). Anti-CLK2 was produced by cloning via PCR a 1.3 kb fragment encoding the catalytic domain of human CLK2 from a full-length cDNA in frame into the BamHI site in the pMAL-C2 vector (New England Biolabs). It was expressed in E. coli as a fusion protein with the Maltose Binding Protein. Following purification on an amylose column, the recombinant protein was injected into two rats for antibody production.

Eye imaginal discs from wandering third instar larvae were dissected, fixed with 4% paraformaldehyde, and mounted. GFP-marked flip clones in the larval fat body were generated through heat shock-independent induction as previously described (Tang et al., 2013). For lipid droplet staining, larvae were dissected in PBS and fixed in 4% paraformaldehyde for 30 min at room temperature. Larval fat body was then washed with PBS, incubated for 15 min in 2μg/ml Nile red/PBS, and then mounted. S2R+ cells were fixed with 4% paraformaldehyde and MCF7 cells were fixed with ice-cold 100% methanol. Cells were then permeabilized with 0.1% triton and processed for immunostaining. DAPI (1 μg/ml) was used to stain nuclei. Samples were examined using a confocal laser scanning microscope (LSM780; Carl Zeiss Inc.) equipped with a 63× Plan-Apochromat (NA1.4) objective lens.

Larvae were collected 72–96 hrs after egg laying, cultured in fresh fly media supplemented with yeast paste (Fed) and 10 mg/ml CQ (Fed+CQ), or in vials containing 20% sucrose (Starved) and 10 mg/ml CQ (Starved+CQ) for 4hrs or 16 hrs prior to dissection. After dissection, the samples were then boiled in SDS sample buffer, run on a 4–20% polyacrylamide gel (Bio-Rad, 4561096), and transferred to an Immobilon-P polyvinylidene fluoride (PVDF) membrane (Millipore). The membrane was blocked by 5%BAS in TBST (TBS with 0.1% Tween-20) in room temperature for 1 hr and then probed with primary antibody in 1X TBST with 5% BSA overnight, followed by HRP-conjugated secondary antibody, and signal was detected by enhanced chemiluminescence (ECL; Amersham, RPN2209; Pierce, 34095).

Total RNA was extracted from larval fat body using TRIzol® reagent (Invitrogen 15596-018). We synthesized first strand cDNA with 1 μg of total RNA using iScriptTM Reverse Transcription Supermix (BIO-RAD, 1708896) and then performed RT-PCR using GoTaq Green Master mix (Promega, M7122) or quantitative PCR with CFX96 Real-Time System (BIO-RAD) using iQTM SYBR Green Supermix (BIO-RAD, 1708880). All expression values were normalized to RpL32 (also known as rp49). All assays were performed in triplicate. The primer sequences used for PCR are listed in Table S5.

MCF7 cells or six larvae from each group were homogenized in PBS supplemented with 0.1% Triton X-100 and Proteinase inhibitor cocktail (Pierce, 78440), heated at 70C for 5 min, and the supernatant collected after centrifugation at 14,000 rpm for 10 min. Supernatants were then subjected to TAG and Protein measurement using a Serum Triglyceride Determination kit (Sigma, T2449) and a BCA protein assay (Pierce, 23227) following the manufacturer’s protocol. For ATP assay, larvae or cells were lysed in CellTiter-Glo buffers and lysates were subjected to ATP measurement using the CellTiter-Glo luminescent cell viability assay kit (Promega, G7573). For the Drosophila studies, ATP, triglyceride, and protein values were normalized to larval weight.

DNA was transfected into S2R+ or HEK293T cells in a 10cm plate with Effectene transfection reagent (Qiagen, 301427) following the manufacturer’s protocol. After 3 days of incubation, cells were lysed with lysis buffer (Pierce 89901 and 87788) with 2X protease and phosphatase inhibitor cocktail (Pierce, 78440) or RNasin Plus RNase inhibitor (Promega, N2611). Lysate was incubated with Chromotek-GFP-Trap (Bulldog Biotechnology, gta-20), anti-HA agarose (Sigma, A2095) or anti-Flag agarose (Sigma, A2220) for 1–2hrs at 4°C to precipitate the protein complexes. Beads were washed 3–4 times with 1 ml lysis buffer. Phosphorylated protein-RNA complexes were subjected to label-free quantitative mass spectrometry with in-gel digestion using Chymotrypsin or detected by Western blotting or qPCR. The primer sequences used for RNA-Immunoprecipitation are listed in Table S5.

Recombinant human proteins used for the study were: His-tagged CDK8/CycC (Thermo Fisher Scientific, PV4402), CLK2-cat (amino acids 137–499) (Thermo Fisher Scientific, PV4201), GST-tagged CPSF6-N (NovoPro Bioscience, 506834), and GST-tagged CPSF6 (Abnova, H00011052-P01). To generate recombinant Drosophila proteins, various segments of CPSF6 cDNA were cloned into pGEX-2T and recombinant CPSF6 proteins were purified from bacteria and eluted from Glutathione Sepharose (Clontech, 635607) as suggested by the manufacturer. S2R+ cells expressing HA-DOA, GFP-ATG1-PB, or ATG1-PB-S297A or HEK 293T cells expressing FLAG-CDK8 were lysed in lysis buffer (Pierce, 87788) with protease and phosphatase inhibitor cocktail (Pierce, 78440) and the lysates were immunoprecipitated. Kinase reactions were carried out in a kinase reaction buffer containing the immune complex, recombinant proteins, and [r-32P] ATP as described previously (Tang et al., 2011).

Total RNA was extracted from larval fat body using TRIzol® reagent (Invitrogen, 15596-018). After assessing RNA quality with an Agilent Bioanalyzer, libraries constructed with an Illumina TruSeq Stranded Total RNA Library Prep Kit with Ribo-Zero Gold were sequenced using an Illumina HiSeq 4000 at the Columbia Genome Center (http://systemsbiology.columbia.edu/genome-center). We multiplexed samples in each lane, which yields a targeted number of paired-end 100-bp reads for each sample. The raw data files of sequencing reads were processed at Harvard Chan Bioinformatics Core with the bcbio-nextgen pipeline, version 1.0.0a0-4708de9 (http://bcbio-nextgen.readthedocs.io/). For quality control purposes, the reads were aligned to Drosophila genome version BDGP6 with STAR version 2.5.3 (https://github.com/alexdobin/STAR) and the alignments were evaluated based on the mappability to transcripts and the complexity of the transcriptome as well as the quality among other custom metrics. The expression of transcripts was quantified using Sailfish 0.10.1 (https://github.com/kingsfordgroup/sailfish) taking 30 bootstrap samples for each sample. Gene level expression was calculated by collapsing the transcript-level quantification with tximport (https://github.com/mikelove/tximport). Differentially expressed genes were analyzed with DESeq2 version 1.14.1 (https://bioconductor.statistik.tu-dortmund.de/packages/3.4/bioc/html/DESeq2.html) while differentially expressed transcripts were analyzed with sleuth version 0.28.1 (https://github.com/pachterlab/sleuth), leveraging the bootstrap samples to account for technical variability assigning reads to individual transcripts. The final hits of differentially expressed genes as well as transcripts were selected based on both fold changes comparing to the control larval fat body as well as adjusted p value calculated by DESeq or sleuth. TPM values were used to calculate the fold changes. Hits were selected if 2 or more-fold changes were consistently observed among the replicates and the adjusted P value is less than 0.05. The genes or transcripts that do not meet the 2-fold change cutoff with any of the replicates but with adjusted p-value of 0.01 or less are included as low confidence hits in Table S2 but were not selected for enrichment analysis. Gene set enrichment analysis was performed using an in-house java program based on hyper-geometric distribution. For Drosophila genes, gene sets were assembled using gene ontology annotation, pathway annotation from GLAD, and protein complex annotation from COMPLEAT (Hu et al., 2017). Human pathway annotation of Reactome and KEGG were mapped to Drosophila gene sets using DIOPT (Hu et al., 2017). The heatmap of the selected gene sets were obtained using TM4 software suite (http://mev.tm4.org/). We selected the starvation responsive genes that are CPSF5 dependent and also part of selected gene sets (energy metabolism, ribosome, lipid metabolism, glycolysis and autophagy) and built a protein-protein interaction network based on integrated network from BioGRID, InAct, MINT, DIP, DroID, DPiM and FlyBase. Solid edges are protein-protein interactions identified in Drosophila while the dotted lines are for interologs, protein-protein interactions derived from data in other species. Network visualization was done using Cytoscape vs 3.1.0 (http://www.cytoscape.org/) (https://github.com/cytoscape).