IPMK Mediates Activation of ULK Signaling and Transcriptional Regulation of Autophagy Linked to Liver Inflammation and Regeneration

SUMMARY Autophagy plays a broad role in health and disease. Here, we show that inositol polyphosphate multikinase (IPMK) is a prominent physiological determinant of autophagy and is critical for liver inflammation and regeneration. Deletion of IPMK diminishes autophagy in cell lines and mouse liver. Regulation of autophagy by IPMK does not require catalytic activity. Two signaling axes, IPMK-AMPK-Sirt-1 and IPMK-AMPK-ULK1, appear to mediate the influence of IPMK on autophagy. IPMK enhances autophagy-related transcription by stimulating AMPK-depen-dent Sirt-1 activation, which mediates the deacetylation of histone 4 lysine 16. Furthermore, direct binding of IPMK to ULK and AMPK forms a ternary complex that facilitates AMPK-dependent ULK phosphorylation. Deletion of IPMK in cell lines and intact mice virtually abolishes lipophagy, promotes liver damage as well as inflammation, and impairs hepatocyte regeneration. Thus, targeting IPMK may afford therapeutic benefits in disabilities that depend on autophagy and lipophagy—specifically, in liver inflammation and regeneration.


INTRODUCTION
Autophagy occurs at a basal rate in most cells, eliminating protein aggregates and damaged organelles to maintain cytoplasmic homeostasis. Autophagy can also lead to cell death (Guha et al., 2016) and plays a role in neurodegenerative diseases as well as malignant transformation (Kaur and Debnath, 2015;Galluzzi et al., 2016). Diverse families of genes regulating the autophagic process have been delineated, but how autophagy affects their signaling remains unclear.
Inositol polyphosphates are major signaling molecules generated by a family of inositol phosphate kinases that successively phosphorylate the inositol ring, leading to the formation of inositol hexakisphosphate (IP6) as well as lesser phosphorylated derivatives. IP6, in turn, is phosphorylated to generate inositol pyrophosphates; specifically, one or two isomers of IP7 and IP8 (Maag et al., 2011). Inositol polyphosphate multikinase (IPMK) physiologically generates IP4 and IP5 (Maag et al., 2011). In anon-catalytic fashion, IPMK influences diverse cellular processes, functioning as a co-activator for p53, CREB, p300 (CBP), and serum response factor (SRF) and regulating immediate-early gene transcription (Kim et al., 2011aXu et al., 2013). As one of its kinase-independent activities, IPMK stabilizes the mTORC1 complex (Kim et al., 2011a). IPMK is also a physiological phosphatidylinositol 3-kinase (PI3K), with activity that leads to Akt phosphorylation (Maag et al., 2011). Deletion of IPMK is embryonic lethal in mice, indicating the importance of this enzyme in biology (Maag et al., 2011).
Interactions between IPMK and autophagy have been reported. In yeast, deletion of IPMK leads to virtual abolition of autophagy as well as mitophagy (Taylor et al., 2012). IPMK appears to regulate autophagy genes as well as their link to ULK kinase. Thus, deletion of IPMK markedly reduces transcription of autophagy-associated genes and decreases activation of ULK as well as downstream autophagy signaling. In the present study, we delineate mechanisms whereby IPMK mediates diverse components of autophagy, for which IPMK appears to be a major physiological determinant.
We validated the confocal data using transmission electron microscopy (TEM) (Klionsky et al., 2016). Under basal conditions, IPMK KO MEFs experienced an almost 70% loss of double-membrane autophagic vesicles. Glucose starvation markedly enhanced the numbers of autophagosomes in WT MEFs, which were greatly diminished in KO preparations ( Figure 1B).
Phosphatidylethanol-conjugated ATG8/LC3 is a widely used biochemical marker of autophagy (Sharifi et al., 2016). LC3-I is non-lipidated, whereas LC3-II is the lipidated form. Levels of LC3-II are employed as markers of autophagosome formation and accumulation (Klionsky et al., 2016). To evaluate basal autophagic flux, cells were treated with Baf A1. Deletion of IPMK virtually abolished LC3-II levels, implying a major role of IPMK in determining basal levels of autophagy ( Figure 1C). We also stimulated autophagy through glucose starvation for 8 h. Deletion of IPMK markedly suppressed LC3 lipidation in glucose-starved MEFs ( Figure 1D). We employed 24 h of food deprivation, a process that induced robust LC3-II expression in the livers of F/F mice (IPMK F/F). In contrast, IPMKdeleted KO mice (IPMK F/F-AlbCre) failed to express LC3-II ( Figures 1E and 1F).
To ensure that the findings with glucose starvation can be generalized to other autophagic stimuli, we evaluated H 2 O 2 treatment, which is well-known to elicit autophagy (He et al., 2017). IPMK deletion abolished LC3-II enhancement associated with H 2 O 2 treatment ( Figure S1D).
We extended our findings to a different cell type. Using small hairpin RNA (shRNA), we stably knocked down IPMK in 786-0 renal cancer cells ( Figure S1E). Knockdown of IPMK with shRNA clone 3 in 786-0 cells significantly reduced enhancement of LC3-II levels by glucose starvation ( Figure S1F). IPMK possesses distinct inositol phosphate kinase activity and PI3K activity ( Figure S1G; Maag et al., 2011). To ascertain the importance of IPMK's catalytic activity in regulating autophagy, we overexpressed IPMK WT or IPMK-KSA (IPMK K129A/S235A), which is kinase-dead (devoid of inositol triphosphate [IP 3 ] and phosphatidylinositol 3-phosphate [PIP3] kinase activity) and verified their enzymatic activity through inositol profiling ( Figure  S1H). We attempted to reverse the decreased autophagy associated with IPMK deletion by rescuing IPMK KO MEFs with WT or kinase-dead IPMK mutants ( Figure 1G). Ki-nasedead IPMK mutants rescued the loss of LC3-II in IPMK KO cells as effectively as IPMK WT ( Figure 1G). Thus, the catalytic activity of IPMK is not required for its enhancement of autophagy.
Removal of exclusive autophagic substrates (not proteasomal substrates) provides an independent way to analyze autophagy. Neomycinphosphotransferase II (NeoR) is an exclusive autophagic substrate (Nimmerjahn et al., 2003;Chauhan et al., 2013;Yang et al., 2011). As established earlier, NeoR-GFP degradation is completely inhibited by autophagic inhibitors like 3-methyladenine (3-MA) but does not respond to inhibitors of proteasomal degradation. Inhibition of autophagy leads to accumulation of NeoR-GFP, resulting in enhanced GFP fluorescence (Nimmerjahn et al., 2003;Chauhan et al., 2013;Yang et al., 2011). We transfected WT and KO MEFs with NeoR-GFP plasmids, and 24 h after transfection we analyzed sequestration of NeoR-GFP using confocal imaging and western blotting. Under basal conditions, WT MEFs displayed uniform cytoplasmic and nuclear fluorescence. However, in KO cells, brightly fluorescent protein aggregates were evident in nuclear proximal regions, with a greatly enhanced mean fluorescence intensity ( Figure 1H). Western blots showed stronger NeoR-GFP bands in KO than in WT cells ( Figure 1I), implying defects in basal autophagy.

Deletion of IPMK Profoundly Suppresses Transcription of Autophagy-Related Genes by Deactivating Sirtuin 1
We showed previously that IPMK can function as a transcriptional co-activator and control transcription of immediate-early genes (Xu et al., 2013). IPMK has also been found to control the transcriptional activity of HDAC (Watson et al., 2012;Bosch and Saiardi, 2012). Arg82, the yeast homolog of IPMK, controls transcription of a set of genes important for arginine metabolism (Bosch and Saiardi, 2012).
To analyze IPMK's role as a transcriptional regulator, we performed qPCR of 6 autophagyrelated genes. We selected LC3B and GABARAPL1, which facilitate elongation and closure of autophagic vesicles (Joachim et al., 2015;Slobodkin and Elazar, 2013); BNIP3 and BNIP3L, which help initiate macro-autophagy and selective forms of autophagy, such as mitophagy (Zhang and Ney, 2009;Quinsay et al., 2010); ATG12, a ubiquitin-like protein involved in autophagic vesicle formation (Fader and Colombo, 2009); and P62 (sqstm1), an adaptor protein that recruits cargo to autophagic vesicles (Kaur and Debnath, 2015). Deletion of IPMK in MEFs markedly impaired mRNA expression of these genes in untreated preparations and under glucose starvation ( Figure 2A). Furthermore, we performed western blotting of BNIP3L, ATG12-ATG5, and GABARAPL1 to confirm the qPCR data. The protein levels of the above genes were induced by glucose starvation and virtually abolished in IPMK KO MEFs ( Figure 2B).
We extended our findings to 786-0 cells, in which deletion of IPMK markedly reduced the protein levels of BNIP3L, ATG12, and GABARAPL1 ( Figure S2A). In vivo, western blot analysis of mouse liver samples after 24 h of food starvation showed significant increases in BNIP3L, ATG12-ATG5, and GABRAPL1 in IPMK F/F mice; they were markedly diminished in IPMK F/F-AlbCre mice (KO) ( Figure 2C). We also examined MEFs treated with H 2 O 2 . Within 1 h of H 2 O 2 exposure, we observed a substantial increase in the mRNA levels of LC3B, BNIP3, BNIP3L, p62, GABARAPL1, and ATG12; they were markedly decreased in IPMK KO MEFs ( Figure S2B).
The decreased mRNA expression of LC3B, BNIP3, BNIP3L, p62, GABARAPL1, and ATG12 was rescued by both the WT and kinase-dead forms of IPMK ( Figure 2D). Thus, the regulation of transcription of autophagy-related genes by IPMK is independent of its kinase activity.
Collectively, the data above detail a signaling cascade whereby IPMK helps to activate AMPK, and binding of IPMK and AMPK to Sirt 1 facilitates dissociation of Sirt-1 from DBC1, stimulating Sirt-1 activation. Activated Sirt-1 further enhances deacetylation of H4K16 and transcription of autophagy-related genes ( Figure S2F).
Because IPMK deletion abolished the activating phosphorylation events on ULK, one might anticipate decreases in ULK-dependent phosphorylation with IPMK deficit. Accordingly, we monitored phosphorylation of the ULK substrate ATG 13 (Egan et al., 2015;Orsi et al., 2012; Figure 3A). IPMK deletion abolished phosphorylation of ATG 13 under glucose starvation.
We rescued the lost ULK phosphorylation of IPMK-deleted cells by overexpressing IPMK WT and the kinase-dead form, which restored these phosphorylation events ( Figure 3C).
We showed that regulation of ULK phosphorylation by IPMK occurs in intact animals ( Figure S3B). Food deprivation markedly augmented ULK-S-555 phosphorylation, which was substantially reduced in the livers of IPMK-deleted mice.
IPMK can regulate AMPK phosphorylation under nutrient starvation ( Figure S2C), which might mediate the influence of IPMK on ULK. However, H 2 O 2 can directly induce AMPK phosphorylation by oxidative modification of the AMPKα subunit (Zmijewski et al., 2010). Consistent with this model, in IPMK-deleted MEFs, H 2 O 2 -stimulated levels of phos-pho-AMPK were comparable with the WT ( Figure 3D). Interestingly, ULK phosphorylation at the AMPK site after H 2 O 2 treatment was still significantly reduced in IPMK KO MEFs ( Figure 3E). Thus, the loss of ULK phosphorylation in IPMK KOs is not just secondary to any alteration in AMPK phosphorylation.

IPMK Regulates ULK Phosphorylation by Direct Binding Interactions
Because IPMK's regulation of autophagy does not require its kinase activity, we studied direct binding of IPMK to ULK. IPMK bound ULK regardless of whether the pull-down employed ULK or IPMK ( Figure 4A). We also showed that endogenous ULK binds IPMK ( Figure 4B). The absence of satisfactory antibodies to IPMK precluded evaluation of endogenous IPMK binding interactions. Utilizing in vitro systems, we did demonstrate direct binding of ULK and IPMK ( Figures S4A and S4B). To facilitate manipulation of the IPMK and ULK system, we mapped sites on IPMK responsible for binding ULK ( Figures  4C and 4D). Fragment 3, comprising amino acids 182-252, appeared to be a candidate dominant-negative structure because it substantially inhibited IPMK and ULK binding ( Figure 4E). Acting as a dominant-negative fragment, fragment 3 reduced the influence of glucose starvation on LC3-II in HEK293 cells ( Figures 4F-4H). In contrast, fragment 1, comprising amino acids 1-92, failed to influence LC3-II levels or ULK phosphorylation ( Figure S4C). These findings indicate that binding of IPMK to ULK mediates ULK phosphorylation and autophagy ( Figure 4I).

Direct Binding of IPMK to AMPK Is Required for IPMK's Influence on Autophagy
In intact cells, overexpressed IPMK and AMPK bound to each other ( Figure 5A). To determine whether binding was direct, we monitored the interactions of the purified IPMK and AMPK proteins ( Figures S5A and S5B). We observed substantial direct binding of IPMK and AMPK. We mapped binding sites on IPMK, establishing that the binding is primarily associated with fragment 2, comprising amino acids 92-182 ( Figure 5B). Fragment 2 may offer promise as a dominant-negative fragment because it abolished IPMK and AMPK binding ( Figure 5C). We employed fragment 2 as a dominant-negative fragment to explore the importance of IPMK in regulating ULK. Overexpressing fragment 2 greatly reduced ULK-S-555 phosphorylation as well as LC3 lipidation ( Figures 5D-5F). Although fragment 5 bound AMPK, it failed to serve as a dominant-negative fragment ( Figure S5C). Intriguingly, dominantnegative fragment 2 suppressed LC3b gene expression at mRNA levels ( Figure S5D). These findings indicate that binding of IPMK to AMPK mediates ULK phosphorylation and autophagy ( Figure 5G).

IPMK Is Essential for AMPK and ULK Interactions
IPMK is required for binding of ULK and AMPK because their binding was abolished in IPMK KO MEFs with or without glucose starvation ( Figure 6A). This action is selective because IPMK deletion did not influence binding of ULK to FIP200 ( Figure 6B), ATG101 ( Figure 6C), or ATG 13 ( Figure 6D). We extended this finding to H 2 O 2 treatment, which acted the same as glucose starvation ( Figures S6A-S6D). We buttressed these conclusions in experiments employing in vitro ULK phosphorylation by AMPK. Addition of recombinant human IPMK (hIPMK) (100 ng, 500 ng, and 1 μg) augmented AMPK-dependent ULK phosphorylation ( Figure S6E). ULK phosphorylation reached saturation at 500 ng of IPMK. Thus, IPMK is essential for AMPK and ULK interactions and AMPK-dependent ULK phosphorylation ( Figure 6E).

IPMK Is Required for Lipophagy and Regulates Liver Inflammation and Hepatocyte Regeneration
Abundant data implicate IPMK in regulation of autophagy. One form of macroautophagy, called lipophagy, has been shown to contribute to hydrolysis of triacylglycerol stored in cytoplasmic lipid droplets. Accordingly, we evaluated a potential role of IPMK in regulating lipophagy. IPMK-deleted MEFs displayed a doubling of lipid droplets both in regular medium and with oleate treatment, indicating substantial diminution of lipophagy ( Figure  7A). Starvation induces hepatic autophagy and increases delivery of free fatty acids (FFAs) from adipose tissue lipolysis to the liver. Electron microscopy and oil red O staining revealed that both under untreated conditions and overnight starvation, the numbers of lipid droplets were substantially increased in F/F-AlbCre (liver-specific IPMK KO) mice compared with F/F (WT) mice ( Figure 7B; Figure S7A), indicating impaired lipophagy in IPMK-deleted livers. The effect of IPMK and AMPK signaling on lipid droplet formation was analyzed by overexpressing fragment 2 (dominantnegative for IPMK and AMPK binding). The number of lipid droplets increased significantly in fragment 2 ( Figure S7D).
We wondered whether IPMK deficiency affected overall liver function, which we assessed by monitoring the serum levels of alanine-leucine transaminase (ALT), which were unchanged in IPMK KOs and F/F-AlbCre mice ( Figure S7B). We examined liver morphology by H&E staining, which was not altered in F/F-AlbCre mice. However, we observed a mild increase in inflammatory cell number in F/F-AlbCre livers as well as a modest enhancement of apoptotic cells ( Figure 7C).
We evaluated the response of IPMK-deleted livers (F/F-AlbCre) to cytotoxic insults utilizing carbon tetrachloride (Ccl4) ( Figure 7D). Damage was increased substantially in IPMK KO livers ( Figure 7E). The damage associated with Ccl4 was especially notable, with increased numbers of inflammatory cells and apoptotic cell profiles as well as serum ALT levels ( Figure S7C). These observations indicate that IPMK is cytoprotective. We also monitored liver regeneration following Ccl4 administration ( Figures 7E and 7F). We measured Ki67, an index of replicative cell activity, as well as 5-ethynyl-2'-deoxyuridine (EDU) incorporation into regenerating cells. Cell regeneration appeared to decrease by about 50% in IPMK KO livers.

DISCUSSION
The present study establishes IPMK as a principal determinant of autophagy and of cytoprotection in the liver. Deletion of IPMK abolishes autophagy, monitored in multiple ways, indicating that IPMK is a physiological regulator of autophagy ( Figure 1). Catalytic activity of IPMK is not required for this. Mechanistically, IPMK regulates autophagy in two different ways. (1) IPMK influences transcription of autophagy-related genes by regulating H4K16 deacetylation. (2) IPMK mediates AMPK-dependent ULK phosphorylation. We also showed that deletion of IPMK impairs lipophagy in cell lines and intact liver ( Figures 7A  and 7B). Most importantly, deletion of IPMK augments liver inflammation and impedes hepatocyte regeneration (Figures 7D and 7E).
How does IPMK influence the autophagic process? IPMK regulates autophagy through AMPK. AMPK initiates autophagy by regulating the transcription of autophagic genes and activating ULK. Nutrient deprivation promotes Sirt-1-mediated deacetylation of H4K16, followed by transcription of autophagy-related genes (Fullgrabe et al., 2013). AMPK enhances dissociation of Sirt-1 from its inhibitor DBC1 and stimulates transcription of autophagy-related genes . We found that deletion of IPMK markedly diminishes AMPK activation during glucose starvation and impairs AMPK-mediated H4K16 deacetylation (Figure 2). On the other hand, AMPK phosphorylates ULK at serines 555, 317, and 777 and activates autophagy by recruiting the beclin1 complex and activating the class III PI3K VPS34 (Lamb et al., 2013). AMPK-dependent ULK phosphorylation is abolished with deletion of IPMK ( Figure 3). It could be anticipated that IPMK influences ULK phosphorylation by activating AMPK. However, with H 2 O 2 treatment, IPMK-deleted MEFs have increased levels of phospho-AMPK, comparable with the WT (Figure 3F), although ULK phosphorylation at the AMPK site is significantly diminished ( Figure 3G). Protein-protein interaction studies confirmed that IPMK acts as a scaffold protein, linking AMPK with ULK (Figures 4, 5, and 6).

STAR★METHODS CONTACT FOR REAGENT AND RESOURCE SHARING
Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Solomon H. Snyder MD (ssnyder@jhmi.edu).

EXPERIMENTAL MODEL AND SUBJECT DETAILS
Animals-All protocols were approved by the Johns Hopkins University Animal Care and Use Committee. Mice were housed according to institutional guidelines, in a controlled environment at a temperature of 22°C ± 1°C, under a 12-h dark-light period and provided with standard chow diet and water ad libitum. Male and female IPMK F/F, IPMK F/F-Alb Cre, (between 1 and 3-month-old) were used. Specifically, male mice were used for liver regeneration study, electron microscopy and tissue histology. All mice were maintained in 129SV-C57BL/6 mixed background.
Cell culture-Mouse embryonic fibroblasts (MEF) and human embryonic kidney HEK293T cells (American Type Culture Collection) were maintained in a humid atmosphere of 95% air and 5% CO2 at 37°C in DMEM 786-0 renal cancer cell was marinated in RPMI-1640 supplemented with 10% FBS, L-glutamine (2 mM), and penicillin (100 units/mL)/streptomycin (100 mg/mL). Retroviral transfection and generation of stably transfected cells: Retroviral constructs (5) were transiently transfected into a Platinum-E retroviral packaging cell line for 48 h by using Lipofectamine 2000 transfection reagent. The high-titer viral stocks were produced by passing the supernatant using a 0.45 μm pore size-filter. For infection, MEFs were incubated with the viral supernatant in the presence of polybrene (8 μg/mL) for 48 h. Stably infected MEFs were selected with blasticidin (4 μg/mL) for 1-2 weeks. Selected stable cell lines were always maintained in respective medium containing blasticidin (4 μg/mL).
Generation of brain specific IPMK knock out mice-IPMK F/F mice were generated as described previously (5). To develop brain specific IPMK KO, IPMKfl/fl mice were crossed with FLP delete (Neomycin) and Nestin Cre driver (brain specific cre driver) mice. FRT sites flanking the neomycin resistance gene facilitate its removal by FLP recombinase,and loxP sites facilitate removal of the targeted exon 6 by Cre recombinase. IPMKfl/fl mice were mated with mice expressing FLP recombinase to excise the neomycin resistance gene to generate IPMKflpped/flpped mice (we refer IPMK F/F in the paper). Homozygous IPMK F/F mice were crossed with the Albumin Cre+/− mice, which mediate excision of floxed alleles in liver, mostly in hepatocytes. Genotyping was performed using a transnetyx genotyping facility. All mice were maintained on a 129SV-C57BL/6 mixed background. Animal care and experimentations were approved by the Johns Hopkins University Animal Care and Use Committee. Mice were housed in a 12-h light/12-h dark cycle, at an ambient temperature of 22°C, and fed standard rodent chow.
Generation of brain specific IPMK knock out mice-IPMK F/F mice were generated as described previously (5). To develop brain specific IPMK KO, IPMKfl/fl mice were crossed with FLP delete (Neomycin) and Nestin Cre driver (brain specific cre driver) mice. FRT sites flanking the neomycin resistance gene facilitate its removal by FLP recombinase,and loxP sites facilitate removal of the targeted exon 6 by Cre recombinase. IPMKfl/fl mice were mated with mice expressing FLP recombinase to excise the neomycin resistance gene to generate IPMKflpped/flpped mice. Homozygous IPMKflpped/flpped mice were crossed with the Albumin Cre+/− mice, which mediate excision of floxed alleles in liver, mostly in hepatocytes. Genotyping was performed using a transnetyx genotyping facility. All mice were maintained on a 129SV-C57BL/6 mixed background. Animal care and experimentations were approved by the Johns Hopkins University Animal Care and Use Committee. Mice were housed in a 12-h light/12-h dark cycle, at an ambient temperature of 22°C, and fed standard rodent chow.
Autophagy activation in mouse liver-Mice (male) were starved overnight followed feeding for 6 h to synchronize the food cycle. Then one group of mice was allowed to feed, while a second group of mice (IPMK F/F and IPMK F/F-Albumin Cre) was starved for 24 h. Starvation of food induced autophagy in liver was analyzed by western blot of LC3-II and ULK ser 555. Each group comprised 3 mice.
Activation of autophagy in cells-1× 10 5 Cells (MEF and others) were glucose starved for 8 h followed by lysing them for biochemical analysis(30). Oxidative stress was elicited by treatment with H2O2 (500 uM) for 30 min, which induced substantial autophagy in MEFs and other cell lines. Autophagic flux was analyzed using Baf A1 100 nM to inhibit lysosomal degradation of autophagic vesicles (12). Autophagy was analyzed either by western blot of LC3-II, transmission electron microscopy, or confocal microscopy.

Transmission Electron Microscopy (TEM)-MEFs were treated with indicated
concentrations of H 2 0 2 , bafilomycin, or by glucose starvation. The cells were fixed in 2.5% glutaraldehyde, 3 mM MgCl2, in 0.1 M sodium cacodylate buffer, pH 7.2, for one h at room temperature. After buffer rinse, samples were post-fixed in 1% osmium tetroxide in buffer (1 h) on ice in the dark. The cells were stained with 2% aqueous uranyl acetate (0.22 mm filtered, 1 h in the dark), dehydrated in a graded series of ethanol solutions and embedded in Eponate 12 (Ted Pella) resin. Samples were polymerized at 37°C for 2-3 days before moving to 60°C overnight. Thin sections, 60 to 90 nm, were cut with a diamond knife on a Reichert-Jung Ultracut E ultramicrotome and picked up with 2×1 mm copper slot grids. Grids were stained with 2% uranyl acetate in 50% methanol and lead citrate at 4°C and observed with a Hitachi 7600 TEM. Images were captured with an AMT CCD XR50 (2K × 2K) camera. Classification and counting of autophagic vacuoles were done by double-blinded independent observers.
To study liver tissue we starved mice for 24 h followed by surgical excision of the liver. Livers were fixed in 2% (wt/vol) paraformaldehyde (freshly prepared from EM grade prill form), 2% (vol/vol) glutaraldehyde, 3 mM MgCl2, in 0.1 M sodium cacodylate buffer, pH 7.2, overnight. Regions of interest were dissected and samples were washed in 0.1 M sodium cacodylate buffer with 3 mM MgCl2 and 3% (wt/vol) sucrose. Samples were postfixed in reduced 2% (wt/vol) osmium tetroxide, 1.6% (wt/vol) potassium ferrocyanide in buffer (2 h) on ice in the dark. Samples were stained with 2% (wt/vol) aqueous uranyl acetate (0.22 μm filtered, 1 h in the dark), dehydrated in a graded series of ethanol propylene oxide solutions, and embedded in Eponate 12 (Ted Pella) resin. Samples were polymerized at 60°C overnight. Thin sections (60-90 nm) were cut with a diamond knife on a Reichert-Jung Ultracut E ultramicrotome and picked up with 2 × 1 mm copper slot grids. Grids were stained with 2% (wt/vol) uranyl acetate in 50% (vol/vol) methanol and lead citrate at 4°C and observed with a Hitachi 7600 TEM. Images were captured with an AMT CCD XR50 (2K × 2K) camera. qPCR analysis-After extracting the total RNA using Gen Elute Mammalian Total RNA miniprep kits (Sigma), and checking its integrity by electrophoresis, the cDNA was synthesized from 1 mg of purified total RNA using Revert Aid H minus first strand cDNA synthesis kit (Fermentas Life Sciences, Ontario,Canada). Expression of mouse and human IPMK, mouse LC3B, mouse BNIP3 and mouse GAPDH was detected using suitably designed Taqman primers (Invitrogen).Other genes such as mouse BNIP3L, P62,GABARAPL1 and ATG12 were determined using primers given in the table. qPCR of these genes was performed using power Sybr green pcr mater mix from Invitrogen. Expression of the designated enzymes was normalized against glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as the internal reference. The experiments were performed (realtime PCR Systems StepOne plus, Applied Biosystems) in triplicate. Data were quantified for the above genes using the comparative Ct method, as described in the Assays-on-Demand Users Manual (Applied Biosystems).

TCATCGTCTCCTCCTGAGCA-p62 (mR)
ChIP-Approximately 7×10 6 cells were fixed with 1% formaldehyde (Sigma-Aldrich, Cat#: F8775) at room temperature for 10 min followed by ChIP using ChIP assay Kit from Millipore-Sigma (17-295). Cells were then harvested and lysed in 500 mL of ChIP lysis Buffer (50 mM Tris-HCl pH 8.0, 5 mM EDTA, 150 mM NaCl, 0.5% Triton X-100, 0.5% SDS, 0.5% NP-40, 1 mM sodium butylate) containing protease inhibitor cocktail. The lysates were subjected to sonication to shear DNA to a length of approximately between 150 and 900 bp. The lysate was then diluted in 1.2 mL of ChIP dilution buffer and incubated with control IgG (Cell Signaling Technology, Cat#: 2729S) or primary antibody H4k16ac (Active Motif 39168) 4C overnight. Then the lysate was incubated with Protein A/ Salmon sperm slurry (provided in kit) for 1 h at 4C. The beads were washed sequentially with wash buffer provided in kit. The immunocomplexes were eluted with 75 mL of elution buffer (1% SDS, 0.1 M NaHCO3) twice at 65C for 30 min. After elution, the cross-link was reversed by adding NaCl and incubated together with Proteinase K (Thermo Fisher Scientific, Cat#: EO0491) overnight at 65C. ChIP DNA was purified using ChIP DNA purification kit (Actif motif 58002). The purified DNA was analyzed on a StepOnePlus using power SYBR Green Master Mix. The results are presented as percentage of input. qPCR analyses were done in triplicate. We used LC3B primers:(F) CATGCCTTGGGACACCAGAT, (R) ACCTTCTTCAAGTGCTGTTTGT (45).
Confocal microscopy of autophagy-IPMK WT and KO MEFs were stably transfected with LC3 GFP. IPMK WT and KO cells were transiently transfected with NeoR-GFP and GFP-WIPI2 constructs using lipofectamine 3000. Cells were treated with different conditions and images were captured in a confocal microscope using Zeiss LSM 800. Images were analyzed with Zenlite software. The puncta were counted using Imaris x64 7.7.2 software.The intensity of NeoR-GFP was analyzed using ImageJ.
Immunoprecipitations-pCMV mycIPMK or pCMV myc were cotransfected with pCMV Flag or pCMV ULK1 plasmids into HEK293 cells using polyfect (QIAGEN). Fortyeight hours after transfection, immunoprecipitation of the myc or Flag tagged protein was performed with 500 μg of protein lysates in lysis buffer (150 mM NaCl, 0.5% CHAPS, 0.1% Triton, 0.1% BSA, 1 mM EDTA, protease inhibitors, phosphatase inhibitors) incubated overnight at 4°C EZview myc or Flag beads (Sigma). Beads were pelleted and washed with lysis buffer 3 times, and SDS sample buffer loading dye was added. Immunoprecipitated samples were resolved by polyacrylamide gel electrophoresis. To assess interactions between IPMK and AMPK, HEK293 cells were co-transfected with pCMV mycIPMK or pCMV myc with Flag AMPK followed by immunoprecipitaion of myc using the abovementioned protocol. In the same way, different IPMK gst fragments were co-transfected either with myc ULK1 or myc AMPK to map the IPMK binding site.
To analyze Sirt-1/AMPK and IPMK/Sirt-1 interaction we overexpressed myc Ampk and Flag Sirt in one reaction myc IPMK and Flag Sirt-1 with emty myc as control.We immunoprecipiated myc using EZview myc beads and western blot flag to check the interaction.
In vitro AMPK mediated ULK phosphorylation-In Vitro assay was performed as previously reported (30).Recombinant GST-hUlk1 protein (ThermoFisher) 500 ng was incubated with 10 ng of recombinant purified AMPK complex (Signal Chem) in kinase assay buffer (20 mM HEPES at pH 7.4, 1 mM EGTA, 0.4 mM EDTA, 5 mM MgCl2 and 0.05 mM DTT) supplemented with 0.2 mM AMP and 0.1 mM cold ATP, for 20 min. Recombinant myc-human IPMK was added to the kinase reaction (100 ng, 500 ng and 1 ug). After the reaction, western blot of ULK S 777 was performed to test the importance of IPMK.
Mice Liver toxicity and regeneration study-IPMK F/F and F/F-AlbCre mice were treated acutely with carbon tetrachloride (Ccl4) 2 ul/g. 48 h after treatment serum was collected from blood through cardiac puncture followed by analysis of liver specific serum chemistry for alanine-leucine transaminase (ALT). Liver tissue was fixed in 4% formalin followed by sectioning at 5 μm thickness. Sections were stained with hematoxylin and eosin (H&E),and immunostained with F4/80 (biorad), Ki67 (Abcam) and TUNEL staining followed by light microscopy. To study liver regeneration mice were intraperitoneally injected with 100 μg/g Edu 2h before harvesting liver. Further liver sections were stained with Click-it Edu staining kit (Thermofisher) followed with imaging by confocal microscopy.
Analysis of lipophagy-IPMK WT and KO MEFs were cultured in regular medium (RM) or oleate (OL) 0.25mM for 24h in serum free medium followed by staining lipid droplets with BODIPY 493/503 dye as per manufacturer's protocol (Thermofisher). To study accumulation of lipid droplets in mice liver, mice were either maintained on regular diet or starved overnight, liver tissue was either processed for transmission electron microscopy or tissue samples were cryosectioned for oil red o (ORO) of lipid droplets in liver as per manufacturer's protocol, abcam.
Endogenous immunoprecipitation analysis-To analyze binding of IPMK to endogenous ULK1, gst IPMK was transfected in HEK293 cells. Forty-eight h after transfection, immunoprecipitation of the gst tagged IPMK was performed with 500 μg of protein lysates in lysis buffer (150 mM NaCl, 0.5% CHAPS, 0.1% Triton, 0.1% BSA, 1 mM EDTA, protease inhibitors, phosphatase inhibitors) incubated overnight at 4°C EZview gst beads (Sigma). Beads were pelleted and washed with lysis buffer 3 times, and SDS sample buffer loading dye was added. Immunoprecipitated samples were resolved by polyacrylamide gel electrophoresis followed by western blotting of ULK1 using ULK1 specific antibody (Santacruz Biotechnology). Immunoprecipitation of endogenous ULK1 was further confirmed using siRNA analysis (Santacruz Biotechnology). For DBC1/Sirt-1 interaction studies we immunoprecipitated endogenous Sirt-1 using Sirt-1 specific antibody and western blot with DBC1. We immunoprecipitated endogenous ULK1 using ULK1 antibody and western blot for AMPK, FIP200, ATG101, and ATG 13 from IPMK WT and KO MEF to analyze endogenous interaction of these proteins with ULK.
In vitro binding assay-Recombinant myc IPMK (Origene) was co-incubated with either recombinant gst ULK1 (Signalchem) or His AMPK (Signalchem) in lysis buffer and the complex maintained for 30 min at 4°C. After the addition of myc beads incubation was continued for an additional 15 min and washed 3 times with ice cold lysis buffer. SDS sample buffer was added. Binding was confirmed by western blotting of gst or His. The purity of recombinant proteins was confirmed by resolving single bands in NuPAGE protein gels that were stained with Simply Blue Safe Stain (Invitrogen).
Transfections a. Retroviral transfection and generation of stably transfected cells: Retroviral constructs (5) were transiently transfected into a Platinum-E retroviral packaging cell line for 48 h by using Lipofectamine 2000 transfection reagent. The high-titer viral stocks were produced by passing the supernatant using a 0.45 μm pore size-filter. For infection, MEFs were incubated with the viral supernatant in the presence of polybrene (8 μg/mL) for 48 h. Stably infected MEFs were selected with blasticidin (4 μg/mL) for 1-2 weeks. Selected stable cell lines were always maintained in respective medium containing blasticidin (4 μg/ mL).

Intracellular inositol content profiling:
MEFs were plated at a density of 250,000 cells per 60 mm plate, then labeled with 60 μCi (1 Ci = 37 GBq) [3H]myo-inositol (PerkinElmer) in conventional cell culture media for three days. To extract soluble inositol phosphates, cell pellets were suspended in 300 μL of ice-cold 0.6 M perchloric acid buffer (0.1 mg/mL IP6, 2 mM EDTA) and incubated on ice for 1 min. Ninety μL of 1 M potassium carbonate with 5 mM EDTA were added and incubated on ice for 1 h. Extracts were centrifuged at 12,000 rpm for 15 min. The supernatant was collected and analyzed by HPLC using a Partisphere SAX column (Whatman Inc.). The column was eluted with a gradient generated by mixing Buffer A (1 mM EDTA) and Buffer B (Buffer A plus 1.3 M (NH4)2HPO4, pH 3.8 with H3PO4). The 1 mL fractions were collected and counted using 5 mL of Ultima-Flo AP mixture (PerkinElmer). Lentiviral Transduction-786-0 human renal cancer cell was transduced with shRNA particles (Sigma) to knock down IPMK and selected cells using puromycin (0.5 micro gram/ ml) antibiotic.

QUANTIFICATION AND STATISTICAL ANALYSIS
Error bars in the figures represent standard error of the mean and number of experiments is indicated by n in figure legends. n indicates animals employed for the experiment or times an experiment was performed. Specifics are indicated in the figure legends. Statistical significance (two-tails) was tested with Student's T-Test for two groups or one-way ANOVA for multiple groups with similar size. The differences were considered significant when p < 0.05. All the statistical analysis was performed with Prism 7 program (GraphPad).
(H) HEK293 cells were transfected with FLAG Sirt-1 and myc AMPK or empty vector of myc. Immunoprecipitation of myc was followed by FLAG western blotting. (I) HEK293 cells were transfected with FLAG Sirt-1 and myc IPMK and empty vector of myc. Immunoprecipitation of myc was followed by FLAG western blotting. Data are means ± SD.