Nuclear translocation ability of Lipin differentially affects gene expression and survival in fed and fasting Drosophila.

Lipins are eukaryotic proteins with functions in lipid synthesis and the homeostatic control of energy balance. They execute these functions by acting as phosphatidate phosphatase enzymes in the cytoplasm and by changing gene expression after translocation into the cell nucleus, in particular under fasting conditions. Here, we asked whether nuclear translocation and the enzymatic activity of Drosophila Lipin serve essential functions and how gene expression changes, under both fed and fasting conditions, when nuclear translocation is impaired. To address these questions, we created a Lipin null mutant, a mutant expressing Lipin lacking a nuclear localization signal (LipinΔNLS), and a mutant expressing enzymatically dead Lipin. Our data support the conclusion that the enzymatic but not nuclear gene regulatory activity of Lipin is essential for survival. Notably, adult LipinΔNLS flies were not only viable but also exhibited improved life expectancy. In contrast, they were highly susceptible to starvation. Both the improved life expectancy in the fed state and the decreased survival in the fasting state correlated with changes in metabolic gene expression. Moreover, increased life expectancy of fed flies was associated with a decreased metabolic rate. Interestingly, in addition to metabolic genes, genes involved in feeding behavior and the immune response were misregulated in LipinΔNLS flies. Altogether, our data suggest that the nuclear activity of Lipin influences the genomic response to nutrient availability with effects on life expectancy and starvation resistance. Thus, nutritional or therapeutic approaches that aim at lowering nuclear translocation of lipins in humans may be worth exploring.


INTRODUCTION
Organisms adjust their metabolism to nutrient availability and, when nutrients become scarce, rely on energy stores in the form of glycogen and triacylglycerols (TAG).
Adjustments require regulation of the balance between the synthesis and breakdown of storage molecules and are associated with changes in feeding behavior (1,2). A linchpin in this regulatory system is provided by lipins, a group of proteins with dual functions in lipid synthesis and the genomic control of energy metabolism. Lipins, which are conserved among all eukaryotes, can convert phosphatidic acid into diacylglycerol as phosphatidate phosphatases (PAP) (3)(4)(5). This activity is required for TAG synthesis and the synthesis of membrane phospholipids (6). In addition, lipins can translocate into the cell nucleus to participate in gene regulation (7)(8)(9)(10)(11). Nuclear translocation of both mammalian lipin 1 and the single Lipin ortholog of Drosophila occurs under fasting conditions and is controlled by the nutrient-sensitive target of rapamycin complex 1 (TORC1) pathway (12,13).
While the mechanism by which lipins participate in gene regulation has been explored to some extent, little is known about the genomic changes that depend on the nuclear translocation of lipins. ChIP experiments indicate that the yeast lipin ortholog Smp2 and mammalian lipin 1 associate with the promoter regions of specific genes, suggesting a function as transcriptional co-regulators (14,15). However, at least some of the genomic effects of mouse lipin 1 depend on its PAP activity and the effect of the protein on nuclear abundance of the transcription factor SREBP1 (13). Genes responding to lipins at UZH Hauptbibliothek / Zentralbibliothek Zuerich, on October 27, 2020 www.jlr.org Downloaded from 5 have been identified through an overexpression approach in mice (15), in relation to glucose feeding in C. elegans (16), and through a candidate gene approach in yeast (14).
Similar to mice lacking lipin 1 (4), Drosophila larvae that express reduced amounts of Lipin exhibit a severe underdevelopment of the fat tissue and reduced TAG stores.
However, Drosophila Lipin is also broadly expressed outside the fat body, the insect equivalent of vertebrate adipose tissue (17). While the lack of energy reserves that Lipin mutant without the NLS allowed us to address how interference with nuclear translocation affects gene expression under both fed and fasting conditions. Our data show that Lipin and its PAP activity both inside and outside the fat body are essential for survival, whereas the NLS appears non-essential. Interference with nuclear translocation of the protein under both fed or fasting conditions leads to substantial changes in the expression of genes involved in energy homeostasis and feeding behavior, and of genes involved in the immune response in the fasting state. Notably, as long as sufficient nutrients are available, nuclear functions of Lipin are not only dispensable, but interference with nuclear entry of Lipin is even beneficial for survival.
This observation suggests that a better understanding of nutrients that influence nuclear at UZH Hauptbibliothek / Zentralbibliothek Zuerich, on October 27, 2020 www.jlr.org Downloaded from 6 entry of lipins may shed light on unhealthy, life-shortening dietary conditions. Ultimately, nuclear translocation may prove to be a property of lipins that can be utilized as a target for specific therapeutic interventions. In summary, our data substantiate the critical role that Lipin has in metabolic adaptation to starvation and they identify gene-regulatory functions in the fed state that have an effect on health and life expectancy.

Mutagenesis
CRISPR/Cas9 mutagenesis was used to create an in-frame deletion of the 18-bp DNA The Lipin KO mutant was created by CRISPR/Cas9 mutagenesis using a guide RNA targeting the region immediately downstream of the Lipin start codon. A protospacer of the sequence 5'-CAGACCAAAGATGAATAGCC-3' was cloned into pCFD3 (Addgene) and the resulting guide RNA expression vector injected into embryos expressing Cas9 in the germline (Bestgene Inc.; yw;;nos-Cas9(III-attP2)). Frame-shift mutations resulting from erroneous non-homologous end joining were identified by screening flies carrying mutagenized chromosomes by PCR and DNA sequencing.
A C-to-G nucleotide exchange leading to an amino acid exchange (D812E) in the catalytic motif of Lipin (Lipin D812E ) was introduced by ends-in gene targeting (19). A 6-kb fragment of the Lipin gene was amplified from Bac-Clone #RP98-9N11 (BACPAC Resources Center) using PCR primers 5'-GCTGCGGCCGCGTTGCTATGGCTGTGGCCAC-3' and 5'-GACTGGGTACCCACCAGCGCCGTCTCCAGCTC-3' and cloned into the KpnI and NotI sites of pBluescriptKSII. The C-to-G nucleotide exchange was then introduced using primers of the sequence 5'-GGTGGTGATCTCGGAGATTGACGGCACCATCA-3' and 5'- the I-Scel homing meganuclease was introduced by PCR using primers 5'-CATCGAACCAGGTATTACCCAGTTATCCCTAGGCGGTCGAACTCCTCGTCCGAGG GTGGT-3' and 5'-GCTGCGGCCGCGTTGCTATGGCTGTGGCCAC-3'. The resulting product was cut at NotI and SexAI sites introduced through the primers at the 5'-and 3'ends, respectively, and used to replace the corresponding fragment in the original mutagenized pBluescript construct. Finally, the modified Lipin DNA was excised from the vector and ligated into the NotI and KpnI sites of targeting vector P[TV2] to create a donor construct for injection into Drosophila embryos (Bestgene Inc.). A mutant stock was established following the published protocol for ends-in targeting (19).

Rescue experiments
To determine if Lipin null mutants can be efficiently rescued by fat body or ubiquitous expression of wild-type Lipin, we carried out crosses to create animals of the genotypes Lipin KO /Df(2R)Exel7095; r4-GAL4/UAS-LipinWT and Lipin KO /Df(2R)Exel7095; tub-GAL4/UAS-LipinWT. Survival of these animals was compared to sibling control animals from the same vials carrying one copy of the Lipin wild-type allele plus GAL4 driver and UAS-responder or one copy of the Lipin wild-type allele plus GAL4 driver or UASresponder. Pupae were collected daily and genotyped. Further development was monitored daily. Eclosing adults were isolated and crossed to flies of the opposite sex to determine fertility. To determine whether r4-GAL4 expression was indeed specific to the fat body, as reported (18), we crossed r4-GAL4 flies with flies carrying UAS-2xeGFP.
Fluorescent microscopy confirmed strong expression in the larval and adult fat bodies that was maintained until at least five weeks after eclosion. In addition, GFP expression at UZH Hauptbibliothek / Zentralbibliothek Zuerich, on October 27, 2020 www.jlr.org Downloaded from was observed in the adult hindgut and Malpighian tubules. Thus, tissue specificity of the r4-GAL4 driver is not quite as strict as assumed in previous studies.

Lipin immunostaining, starvation, and longevity experiments
Larvae of the genotypes Lipin DNLS /Df(2R)Exel7095 and Lipin/Df(2R)Exel7095 were reared on standard food. Feeding third instar larvae were removed from the food, transferred to 1% agar starvation plates and kept in an incubator at 25°C for about 16 hours. Fat bodies dissected from the starved larvae and fed control larvae were stained with an affinity-purified antibody directed against Lipin and DAPI to visualize DNA as previously described (12).
To examine starvation resistance of Lipin DNLS /Df(2R)Exel7095 and control flies carrying the wild-type Lipin allele of the injection stock over Df(2R)Exel7095, newly eclosed flies were collected over 24 hours and kept for three days on standard food. Flies were then separated by sex and groups of 25 flies transferred into starvation vials with watersaturated Fly Plugs at the bottom (Genesee Scientific). Survival was monitored daily and dead flies removed. Flies were regularly transferred to fresh vials with water-saturated plugs to avoid feeding on mold or, in the case of the females, eggs deposited onto the plugs. Experiments were carried out in biological triplicates.
To compare the longevity of Lipin DNLS /Df(2R)Exel7095 flies to the longevity of control flies, we proceeded as for the starvation experiments, but kept flies on standard food at UZH Hauptbibliothek / Zentralbibliothek Zuerich, on October 27, 2020 www.jlr.org Downloaded from instead of starving them. Flies were transferred to fresh food every 3-4 days and survival was monitored daily. Experiments were carried out in biological triplicates.

Triglyceride assays and lipid staining
TAG was measured using a colorimetric assay (20). Briefly, flies were collected as for the starvation assays and samples were obtained by pooling and homogenizing 10 male  Fat body from wandering third instar larvae was stained with Bodipy 493/503 (Invitrogen/Molecular Probes) to visualize fat droplets as described previously (12). After data normalization, differential expression analyses were performed using the R/Bioconductor software package limma (21). Reads were aligned to the Drosophila melanogaster assembly BDGP6. 22   Overrepresentation among differentially expressed genes of gene ontology (GO) terms was analyzed using the Bioconductor clusterProfiler package (22). If not indicated otherwise, tissue-specific gene expression data presented in Tables 1, 2, S1, S2, and S3 were derived from the Drosophila gene expression atlases FlyAtlas 1 and 2 (23,24).

Metabolic rate measurements
Metabolic rate was measured as volume CO2 (VCO2; µL/min) produced per fly using open-flow respirometry. A minimum of seven independent fly cohorts were measured for each genotype and sex. Flies were kept on standard food for one to two weeks after eclosure and then separated by sex into groups of seven flies that were transferred to fresh standard food. Flies were allowed to recover from the CO2 anesthesia for 24 hours prior to transfer to respirometry chambers (25). Respirometry chambers were made by modifying 5.0 mL plastic Luer lock syringe barrels with plastic mesh and glass wool to prevent animal escape. The resulting headspace of each chamber was 1.25 mL. A total of eight chambers (one baseline and seven animal chambers) were measured during each respirometry run. All measurements were performed at the same time of the day at 25°C in the dark. During measurements, CO2 and water free air was pushed through each chamber at a flow rate of 50 mL/min. A multiplexer (Sable Systems, Las Vegas, at UZH Hauptbibliothek / Zentralbibliothek Zuerich, on October 27, 2020 www.jlr.org Downloaded from NV) was used to divert the excurrent air from each chamber to a LI-COR CO2 analyzer (LI-COR, Lincoln, NE) for measuring CO2 (ppm) production. During each respirometry run, each chamber of seven flies was measured for 7.5 minutes per hour; baseline was measured for 3.75 min at the top and bottom of every hour. Each set of flies were measured for a minimum of six hours. Expedata (Sable Systems) was used to control the multiplexer and record data every five seconds. Expedata was also used to run baseline corrections, calculate VCO2 per fly (µL/min), and filter data through the use of a nadir function (26). The nadir function selected the lowest continuous values of VCO2 for a span of 50 sec per chamber per hour. The resulting filtered data was analyzed in R using the car package. CO2 production by the flies was stable for at least four hours of measurement. The values for the first hour of measurements were used for statistical analysis, as they were the lowest recorded. CO2 production was compared between strains and sexes using a two-way ANOVA. In addition to the significant difference detected between strains (p=0.021; Fig. 5), metabolic rates were also significantly different between sexes (p=0.023).

Lipin is an essential gene with vital functions outside the fat body
We used CRISPR/Cas9 mutagenesis to create a bona fide null allele of Lipin, Lipin KO .
Lipin KO carries a frame-shift mutation immediately after the start codon, introducing an early stop codon and resulting in a nucleotide sequence encoding a 29-amino acid random peptide (Fig. 1A). Consistent with the prediction that Lipin KO is a null allele, animals homozygous for Lipin KO , or carrying the allele over the deficiency Df(2R)Exel7095 that uncovers the Lipin locus, are not viable as indicated by the absence of homozygous larvae, pupae or adults. Having this mutant available, we asked whether Lipin is required for viability in tissues other than the fat body. Such a requirement is suggested by the broad expression pattern of Lipin, which is not only expressed in the fat body but also in many other tissues, including the gut, the Malpighian tubules, the brain, and the endocrine ring gland (17). We used the r4-GAL4 driver to express wild-type Lipin in the larval fat body in a Lipin KO  The longest-lived male died after 104 days, and the longest-lived female after 66 days.
In striking contrast, none of the flies that had been rescued by fat body expression, and hadn't already died at eclosion, lived for more than 14 days. Although flies rescued by ubiquitous expression appeared much healthier and longer lived, flies rescued by either ubiquitous or fat body-expression were infertile. Together, these data strongly suggest that Lipin has vital functions outside the fat body and that proper Lipin expression is particularly critical for fertility.

Lipin's PAP motif, but not its NLS, is required for the development and viability of adult flies
Next, we asked whether Lipin's PAP activity and/or nuclear functions of the protein are required for development and survival. To generate a mutant that lacks Lipin's PAP activity, we used ends-in gene targeting to introduce a single amino acid exchange into Lipin's catalytic DIDGT motif at position 812, changing it to EIDGT (19) (Fig. 1A). The Dto-E exchange in the motif leads to a complete loss of PAP activity (15). Animals homozygous for the Lipin D812E allele developed into first instar larvae, but none of these larvae reached the second larval instar. The animals could be rescued from the lethality at UZH Hauptbibliothek / Zentralbibliothek Zuerich, on October 27, 2020 www.jlr.org Downloaded from by expressing wild-type Lipin from a transgene in the Lipin D812E /Lipin D812E background, confirming that the observed lethality was indeed caused by the lack of PAP activity normally provided by Lipin. Most homozygous larvae (83%; n=75) were able to survive beyond the 24 hours that it normally takes to reach the next stage, but all of them died within 60 hours as first instar larvae. This clearly demonstrates that Lipin's PA phosphatase activity is required for development beyond the embryonic and early larval stages. It seems likely that it is also required for embryogenesis, because Lipin is a maternally expressed gene and Lipin protein can be detected in the oocyte cytoplasm (17). This suggests that maternal Lipin rescues homozygous mutants through embryogenesis.
To create a mutant that is defective in carrying out nuclear functions of Lipin, we used CRISPR/Cas9 mutagenesis to introduce an 18-bp deletion into Lipin that removes the protein's NLS (Fig. 1A). We opted to remove the NLS rather than introducing point mutations into the transcriptional co-regulator motif, because the latter had been shown to also interfere with enzymatic activity of the protein (15). We were successful in isolating a Lipin DNLS allele and found that, in contrast to the Lipin KO and Lipin D812E mutants, Lipin DNLS mutants were viable. To avoid homozygosity for other loci on the mutagenized chromosome, we used animals carrying one copy of Lipin DNLS and control animals carrying one copy of the wild-type Lipin allele of the CRISPR injection stock for all phenotypic characterizations reported here. We accomplished hemizygosity by using a previously characterized deficiency chromosome, Df(2R)Exel7095 (17). To confirm that the mutant protein is excluded from the nucleus, we examined intracellular at UZH Hauptbibliothek / Zentralbibliothek Zuerich, on October 27, 2020 www.jlr.org Downloaded from distribution of Lipin DNLS and wild-type Lipin under conditions that promote nuclear translocation of Lipin. Nuclear translocation of Lipin is observed when signaling through the nutrient-sensitive TORC1 signaling pathway is attenuated (12). To accomplish this, we subjected pre-wandering third instar mutant and wild-type control larvae to starvation conditions. As expected, starvation caused nuclear translocation of Lipin in fat body of control larvae. In contrast, we did not observe nuclear translocation of Lipin DNLS under these conditions (Fig. 2). These data imply that Lipin DNLS mutants are deficient in carrying out functions that Lipin normally has in the cell nucleus, including generegulatory functions.

Lipin DNLS flies have an improved life expectancy
Lipin DNLS mutant flies were not only viable, but they also did not show any obvious phenotypic deviations from wild-type flies. Although the absence of Lipin's NLS did not appear to have major detrimental effects on survival and fecundity, we decided to examine these animals more closely by asking whether they had the same life expectancy as wild-type control flies. To determine whether this was the case, we compared the survival of Lipin DNLS /Df(2R)Exel7095 flies to that of Lipin WT /Df(2R)Exel7095 control flies (Fig. 3). Males and female flies were kept separately on standard food and deaths were monitored regularly until all flies had died.
The median lifespan of both male and female Lipin DNLS flies was significantly increased, an effect that was most pronounced in the females with an increase of 24 days, which amounts to an impressive 34% jump in life expectancy. Females also showed a significant increase in maximum lifespan by 14 days, whereas maximum lifespan in males was not significantly changed (Fig. 3). These data suggest that nuclear functions of Lipin are not only not required but may even be detrimental when flies are raised on a standard diet.

Lipin DNLS flies have increased susceptibility to starvation
In stark contrast to fed flies, Lipin DNLS flies subjected to starvation showed significantly decreased survival (Fig. 3). Lipin DNLS /Df(2R)Exel7095 and Lipin WT /Df(2R)Exel7095 control flies were separated by sex and transferred to vials containing a source of water but no food. Survival was monitored daily and dead animals were removed. Both Lipin DNLS males and females showed significantly reduced maximum and median lifespans. Median lifespan of males was reduced from 9 to 3 days, and for females from 7 to 6 days. Maximum lifespan for both males and females was reduced from 10 to 6 days. Almost 40% of control females were still alive on day 7 of starvation after all Lipin DNLS females had died. These data suggest that Lipin's ability to enter the nucleus is of critical importance for physiological adaptation to nutrient deprivation.

Fat stores are not reduced in Lipin DNLS flies
Reduced survival of Lipin DNLS flies under starvation conditions could be due to reduced fat stores of these animals. Therefore, we measured triglyceride levels of mutant and control flies (Fig. 4A). TAG levels proved to be not reduced in the mutants. On the at UZH Hauptbibliothek / Zentralbibliothek Zuerich, on October 27, 2020 www.jlr.org Downloaded from 20 contrary, they appeared slightly elevated, although differences were not statistically significant. Staining with the lipid dye Bodipy confirmed that the fat body of Lipin DNLS larvae contained normally sized fat droplets, while indicating slightly elevated rather than decreased levels of neutral lipids. This strongly suggests that nuclear function of Lipin, but not its function in TAG synthesis, is responsible for adaptation to starvation conditions.

Genes involved in energy homeostasis, feeding behavior and the immune response are mis-regulated in fed and food-deprived Lipin DNLS flies
Together, our results suggested that nuclear functions of Lipin are less important when nutrients are plentiful, but essential under conditions of nutrient deprivation. Since lipins can act as transcriptional co-regulators (14,15), it seemed likely that nuclear Lipin brings about changes in gene expression that promote adaptation to nutrient deprivation. To test this hypothesis, we carried out RNA-seq analyses with Lipin DNLS and wild-type control flies that had been kept on a standard diet or starved for 24 hours. RNA was extracted from male and female flies separately and mRNA used for library preparation.
Consistent with the observed differences in life expectancy, we found that many genes Taken together, the data suggest that fatty acid metabolism is shifted from b-oxidation to lipogenesis in Lipin DNLS flies. Thus, when nutrients including fats are plentiful, it seems to be an important function of wild-type Lipin to suppress the de novo synthesis of fatty acids and to promote the use of fatty acids for energy production. At the same time, Lipin appears to impede the use of stored carbohydrates for energy production. In both males and females, one of the most strongly up-regulated genes in Lipin DNLS was target of brain insulin (tobi), which encodes an a-glucosidase involved in glycogenolysis (Table   1). Tobi is specifically activated when nutritional sugar is low (33).
Another group of genes that were identified as enriched in the sex-independent analysis consisted of genes encoding serine-type proteases (Supplemental Table S1). Half of these proteases that respond to Lipin are specifically expressed in the digestive system. These data suggest that Lipin, which shows both nuclear and cytoplasmic expression in the digestive system (17) Table S1). The latter suggests a role of Lipin in male fertility, a conclusion that is supported by our observation that rescued Lipin KO at UZH Hauptbibliothek / Zentralbibliothek Zuerich, on October 27, 2020 www.jlr.org Downloaded from mutants remain infertile. Interestingly, the expression of certain cytochrome P450 genes is also changed in females, among them Cyp4g1, which encodes an oenocyte-specific w-hydroxylase that regulates triacylglycerol composition (30).
Finally, again specifically in males, genes encoding transmembrane receptors showed enrichment (Tables 1 and S1). The gene encoding one of these receptors, methuselahlike 8 (mthl8), also showed strong up-regulation in females. Interestingly, mthl8 is also up-regulated when the transcription factor Cabut is reduced, which is involved in nutrient sensing and rapidly induced by sugar feeding (34). In addition, three gustatory receptors showed altered expression in Lipin DNLS males. Gustatory receptor GR64f is expressed in sugar-sensing neurons of the proboscis and functions as a co-receptor for the detection of sugars in the food (35). Sugar-sensing neurons also mediate fatty acid taste, specifically through activation of the phospholipase C pathway (36). Therefore, it is interesting to note that phospholipase C encoded by CG14945 is altered in females.
Together, these data suggest an involvement of Lipin in the control of responses to nutritional sugars and, possibly, fatty acids. A role in the control of feeding behavior is further supported by altered expression of several other genes, tiwaz in males, which encodes an adult feeding regulator that negatively regulates meal size (37), the takeout (to) feeding regulator in both males and females (38,39), and the intestinal feeding regulator hodor in females (40) ( Table 1).
The diminished starvation resistance of Lipin DNLS flies suggested that genes that normally respond to starvation do not do this, or to a lesser degree, in the mutant. As  Tables S6 and S7). Not surprisingly, gene ontology analysis revealed that metabolic genes, including genes involved in lipid and carbohydrate metabolism, were enriched among these genes. To test the prediction that the starvation response of some of the genes depended on the presence of wild-type Lipin, we filtered for genes that showed an at least 50% reduction of the starvation response in the Lipin DNLS mutant.
This led to the identification of 141 genes in females and 96 genes in males that showed a blunted response to starvation. Many of these genes were immune response genes or involved in energy metabolism (Tables 2 and S2).
Interestingly, Lipin DNLS flies showed a blunted response of genes encoding enzymes that produce key intermediates of fatty acid metabolism, acyl-CoAs and malonyl-CoA, and that are normally up-regulated during starvation. Three acyl-CoA synthetases (ACS) involved in fatty acid activation, which normally show increased expression during starvation in females, showed no or reduced increases in the Lipin DNLS mutant. One of them, pudgy (pdgy), localizes to mitochondria and fatty acid b-oxidation is reduced in pdgy mutants, suggesting that acyl-CoAs produced by pdgy are primarily used for energy production (41). In another study, pdgy mutants were shown to have decreased starvation resistance (42). Thus, at least some of the reduced starvation resistance of Lipin DNLS flies may be due to the impaired activation of pdgy and the other ACS. In females, starvation also activated the gene encoding acetyl-CoA carboxylase (ACC), which catalyzes the rate limiting step in fatty acid synthesis. In males, ACC was significantly increased as well, but below the 1.5-fold threshold applied here. ACC is the at UZH Hauptbibliothek / Zentralbibliothek Zuerich, on October 27, 2020 www.jlr.org Downloaded from single gene encoding acetyl-CoA carboxylase in Drosophila. Knockdown of ACC in the fat body reduces TAG storage, which is consistent with its role in producing malonyl-CoA for the de novo synthesis of fatty acids (43). Malonyl-CoA is also known to block the carnitine shuttle that transports fatty acids into mitochondria for b-oxidation.
Dampening of the carnitine shuttle is also suggested by Lipin-dependent downregulation of two genes in males encoding enzymes required for carnitine synthesis, trimethyllysine dioxygenase and g-butyrobetaine dioxygenase. This suggests that Lipin contributes to a down-regulation of the carnitine shuttle during long-term starvation in both males and females.
Another lipid metabolic gene that was down-regulated in response to starvation, but less so in Lipin DNLS males, was FASN1. FASN1 is a broadly expressed fatty acid synthase that is strongly expressed in the fat body and required for the accumulation of normal TAG stores (24,27). In addition to FASN1, the starvation response of several lipases was altered in Lipin DNLS flies. Consistent with a shift from fatty acid synthesis to lipolysis, a fat body-expressed lipase (CG7367) was up-regulated in females. In males, the fat body-expressed lipase 4 was down-regulated, but lipase 4 is also strongly expressed in the digestive system, suggesting that overall down-regulation of this lipase primarily reflects the reduced requirement for digestion of nutritional fats. Conspicuously, many other genes that showed a reduced response to starvation in Lipin DNLS are predominantly or exclusively expressed in the digestive system, especially the midgut (Tables 2 and S2). This suggests that Lipin regulates genes in the gut that are involved in the uptake and processing of nutrients. Among these genes is p38c, which encodes a at UZH Hauptbibliothek / Zentralbibliothek Zuerich, on October 27, 2020 www.jlr.org Downloaded from MAP kinase functioning in intestinal lipid homeostasis. Flies deficient in p38c accumulate neutral lipids in the gut, suggesting that the kinase controls lipid processing by the tissue (44). Another example is Nplp2, a gene that is important for dietary lipid extraction and effective lipid storage in the fat body (45).
Interestingly, our data show that the up-regulation under starvation conditions of lipase 3 (Lip3), another lipase that is expressed in the digestive system, is considerably inhibited in Lipin DNLS flies ( Table 2). In the fed state, Lip3 is negatively regulated by Lipin (Table   1). Lip3 is regulated by the fatty acid-activated nuclear receptor HNF4, which upregulates Lip3 under starvation conditions (46,47). Thus, the reduced up-regulation of Lip3 in Lipin DNLS flies suggests that HNF4 and Lipin cooperate in Lip3 activation. In mammals, lipin 1 directly interacts with HNF4a to regulate lipid metabolic genes (15).
The Drosophila HNF4 ortholog has an important role during starvation when it activates genes that act in fatty acid b-oxidation and lipolysis, such as Lip3 (47). We therefore asked whether the two proteins share additional potential target genes. HNF4 expression is not changed in the Lipin DNLS mutant, neither under fed nor under starved conditions. Thus, potential responses of HNF4 target genes to Lipin should be independent of HNF4. Genes regulated by HNF4 in Drosophila have been identified by microarray studies using third instar larvae (47). When we compare these data with our RNA-seq data, we find that, in the fed state, HNF4 and Lipin share no target genes in females and only four genes in males. However, in the starved state, this number increases for males and females together to 22 genes, corresponding to 7% of the genes that respond to HNF4 under starvation conditions (Supplementary Table S12).  Table S12).
In summary, our data reveal complex changes in the starvation response of genes involved in lipid metabolism that depend on Lipin. Overall, these changes seem to promote lipolysis while lowering fatty acid synthesis, as one would expect based on the energy demands of starved flies.
In addition to genes involved in lipid metabolism, a similar number of genes involved in carbohydrate and mitochondrial energy metabolism showed a reduced response to starvation in the Lipin DNLS mutant. Among these are genes promoting glycogen breakdown (Gbs-76, CG9485) and sugar transport (CG9657, CG17930) that are downregulated during starvation in males. CG9485, which encodes an ortholog of glycogen debranching enzyme AGL, is also down-regulated when insulin signaling is reduced in the larval fat body of Drosophila (48). Another gene associated with glycogen storage, Acer, is reduced in starved females. Acer mutants of Drosophila have reduced glycogen stores. It is, therefore, somewhat surprising that these mutants exhibit slightly increased starvation resistance (49). Thus, reduced down-regulation of Acer in the Lipin DNLS mutant may contribute to the starvation sensitivity of the mutant. The up-regulation of several at UZH Hauptbibliothek / Zentralbibliothek Zuerich, on October 27, 2020 www.jlr.org Downloaded from genes by starvation that promote energy production by glycolysis and oxidative phosphorylation, including 6-phosphofructo-2-kinase, Phosphoglucose isomerase and klumpfuss (50), was impaired in Lipin DNLS (Table 2). Thus, combined, the data obtained for fed and fasting conditions suggest that it is the basic function of nuclear Lipin to stimulate catabolic processes and energy production.
Consistent with the observed effects of Lipin on genes controlling feeding behavior in the fed state, our data suggest that Lipin also contributes to changes in feeding behavior induced by starvation (Table 2). FIT, a satiety hormone produced by the fat body, which is essential for feeding control in females and strongly down-regulated during starvation (51), shows reduced down-regulation in the Lipin DNLS mutant. Likewise, reduced expression of Root is blunted in the mutant, a gene that is required for the normal function of sugar-sensing neurons (52). Thus, our data support a role of Lipin in controlling feeding behavior in both fed and starved flies.
Interestingly, for some genes, the starvation response was enhanced in the Lipin DNLS mutant, indicating that Lipin normally limits the starvation response of these genes.
Enriched among these genes in males were, again, genes involved in lipid metabolism (Tables 2 and S3 Tables S2 and S3).

Metabolic rate is reduced in Lipin DNLS flies
The RNA-seq analyses indicated that interference with Lipin's ability to translocate into the cell nucleus has broad effects on the expression of genes involved in energy homeostasis. In particular, they suggested a shift to reduced b-oxidation in fed Lipin DNLS flies, which predicts that energy production by oxidative phosphorylation is reduced in these flies. To test this hypothesis, we examined whether Lipin DNLS and control flies exhibited differences in their metabolic rates. Metabolic rate in male and female flies was measured as CO2 production by open-flow respirometry. We found that the metabolic rates were indeed significantly reduced by 8% in male and by 9% in female flies (Fig. 5). These results confirm our prediction that energy production is throttled in Lipin DNLS flies and are consistent with our observation that Lipin DNLS flies were longer lived and, thus, seemed healthier than control flies.

DISCUSSION
We created three different Lipin mutants, Lipin KO , Lipin D812E , and Lipin DNLS , to distinguish between requirements for different activities of the Lipin protein during development and different metabolic states. We had previously shown that animals homozygous for a hypomorphic allele of Lipin exhibit delayed development and late-larval and pupal lethality (17). Our data with the newly generated Lipin KO and Lipin D812E mutants now shows that zygotically expressed Lipin and its enzymatic activity are absolutely required for larval development beyond the L1 stage. This result is consistent with our previous observation that expression of Lipin D812E from a transgene does not rescue Lipin mutant phenotypes (12). Rescue experiments show that ubiquitous, but not fat body-restricted expression, of wild-type Lipin can fully rescue the Lipin null mutant (except for fertility), indicating that Lipin has essential functions outside the fat body. Flies rescued by ubiquitous expression appeared healthy and had a life expectancy that was comparable to that of wild-type flies, whereas most of the flies rescued by fat body expression died shortly after eclosion. Interestingly, however, all rescued flies remained sterile, indicating a requirement of Lipin for reproduction that is sensitive to the timing and/or amount of Lipin expression in one or more tissues. In contrast to Lipin's PAP activity, interference with the ability of the protein to translocate into the cell nucleus does not impair survival under normal feeding conditions. Animals that entirely rely on Lipin without an NLS are not only viable, but they also appear healthier and live longer than wild-type flies. This finding is consistent with our previous observation that Lipin DNLS  Our RNA-seq analysis provides a first insight into the genomic response to a loss of Lipin function in Drosophila. Previous studies in yeast, mice, and C. elegans identified candidate genes that respond to loss of lipin (14), were based on overexpression of the protein (15), or addressed lipin-dependent changes in response to sugar-rich diet (16).
Overexpression of lipin 1 in mouse liver led to the changed expression of almost 4,000 genes (15). Among the up-regulated genes were the genes encoding nuclear receptors PPARa and HNF4a, which were also shown to physically interact with lipin 1 (15). Thus, at least some of the response to lipin 1 is likely mediated by these two transcription factors and lipin 1 in a cooperative fashion. PPAR's, which are known as key regulators of lipid and energy metabolism in mammals, have not been identified in Drosophila, whereas an HNF4 ortholog has. Drosophila HNF4 acts predominantly during starvation, activating genes involved in lipolysis and fatty acid b-oxidation. Since target genes of Drosophila HNF4 are similar to target genes of PPARa, it has been proposed that HNF4 may, at least in part, functionally substitute for PPARa in Drosophila (47). Comparison of our RNA-seq data with microarray data obtained for an HNF4 null mutant did not indicate a substantial overlap between potential target genes of Lipin and HNF4. Our RNA-seq data provide a first glimpse into the organism-wide genomic response to Lipin under feeding and fasting conditions. A large number of genes were differentially expressed in the fed state in both males and females between wild-type and Lipin DNLS flies (total n=413). This does not come as a surprise, because Lipin shows partial nuclear localization in the fed state in some tissues, such as the gut and the brain, and preferential nuclear localization in at least one tissue, the Malpighian tubules (17). Downloaded from find substantial differences in metabolic gene expression between the sexes (54)(55)(56).
Especially in insects, with their yolk-rich eggs and extensive resource allocation for reproduction, sex-specific differences in metabolic gene expression do not come as a surprise. Despite these differences, the biological processes affected, which are summarized in figure 6, are similar in both males and females. The majority of the affected genes have functions in energy metabolism, in particular lipid and carbohydrate metabolism.
A prediction resulting from the RNA-seq data was that energy production is reduced in fed Lipin DNLS flies. We confirmed this by showing that the flies indeed exhibit a lowered metabolic rate. Reductions in reactive oxygen species associated with low metabolic rates are being discussed as determining factors of aging and longevity (57,58). Thus, the reduced metabolic rate may contribute to the observed increase in life expectancy of Lipin DNLS flies. However, although it does not refute this possibility, it should be noted that Drosophila strains selected for longevity exhibit normal metabolic rates (59,60).
In contrast to the fed state, the ability of Lipin to translocate into the cell nucleus clearly Lipin DNLS males suggests that these flies break down fat stores more rapidly than control flies ( Table 2). At the same time, the data suggest that they also exhaust glycogen faster. Down-regulation of a gene encoding glycogen debranching enzyme is attenuated in Lipin DNLS males and of Gbs-76A, which encodes a putative protein phosphatase 1 (PP-1) inhibitory subunit. PP-1 plays a critical role in glycogenolysis by inhibiting glycogen phosphorylase (61). Interestingly, the conserved NLIP domain of Lipin contains an HVRF motif which constitutes a binding site for the catalytic subunit of PP-1 (62). This raises the possibility that Gbs-76A contributes to the regulation of Lipin phosphorylation. Female Lipin DNLS flies less efficiently down-regulate Acer, a gene that is involved in glycogen metabolism as well (49) and they less efficiently up-regulate glycolytic enzymes. This suggests that, similar to the males, females cannot make efficient use of energy reserves, although lipolytic enzymes are not as much affected in females as they are in males. Together with the presence of an additional energy source provided by degenerating oocytes, the latter may explain why starvation resistance is not as severely affected in females as it is in males (Fig. 3). The reduced activity predicted by the data of several acyl-CoA synthetases in starved Lipin DNLS males and females, and of cytosolic glycerol-3-phosphate dehydrogenase in males, is consistent with the interpretation that Lipin DNLS flies deplete energy reserves more quickly than control flies.
The transcriptional up-regulation of ACC during starvation, which is dampened in Lipin DNLS females, was somewhat unexpected. ACC produces malonyl-CoA from acetyl-CoA for the de novo synthesis of fatty acids. Malonyl-CoA also has the property of blocking the carnitine shuttle that transports fatty acids into mitochondria for b-oxidation.
Thus, up-regulation of ACC has the potential of counteracting the shift from lipogenesis at UZH Hauptbibliothek / Zentralbibliothek Zuerich, on October 27, 2020 www.jlr.org Downloaded from to b-oxidation that occurs during starvation. However, it is important to note that the activity of ACC is tightly regulated at the posttranslational level, in particular by AMPK (63). Moreover, malonyl-CoA produced by ACC is not only used for fatty acid synthesis, but also for protein malonylation (64). Interestingly, activity of the nutrient-sensitive TOR kinase can be reduced by malonylation, which may contribute to the inhibition of TORC1 during starvation (65). The possibility of increased inhibition of the carnitine shuttle after 24 hours of starvation, which must be considered as long-term starvation for flies, is not only suggested by up-regulation of ACC, but also by decreased expression of two enzymes acting in the carnitine synthesis pathway, trimethyllysine dioxygenase and gbutyrobetaine dioxygenase. It is possible that after this extended period of starvation flies start to conserve energy stores for long-term survival.
An interesting and novel observation was that Lipin is involved in the control of immune response genes. Starvation-induced changes in the expression of a number of immune response genes depend on Lipin (Tables S2 and S3). Recent evidence suggests that interactions between Lipin and the immune response are dynamic. Lipin transcript levels decrease by 50% when an immune response is stimulated by expression of a constitutively active Toll receptor in the fat body. This decrease is associated with a reduction in triglyceride storage (66). Together, these observations further strengthen previously observed links between lipid metabolism and the immune response in The improved life expectancy of normally fed flies that results from interference with nuclear translocation of Lipin is of particular interest in light of the recent finding that high levels of glucose induce nuclear translocation of C. elegans Lipin 1 and that decreased levels of Lipin 1 increase glucose toxicity (16). Mis-regulation in Lipin DNLS flies of the aglucosidase Tobi as well as Root and the Mthl8 receptor, which is regulated by the sugar-sensitive Cabut transcription factor, suggest a similar role of Drosophila Lipin in the metabolic response to glucose. These observations support the idea that nuclear translocation of lipins is sensitive to specific nutrients, and that lipins mediate or block genomic effects of these nutrients. It will be interesting to further investigate how nuclear translocation of Drosophila Lipin is fine-tuned in this manner and how this translates into healthy or unhealthy metabolic outcomes.

DATE AVAILABILITY STATEMENT
All data described in the manuscript are contained in the manuscript and the supplementary data files.   Table 1S) GO: Oxidoreductase Activity (see Table 1S)  Proteases (see Table S2) Immune Response Genes (see Table S2) Feeding third instar larvae of the indicated genotypes were removed from the food and subjected to starvation conditions. Fat body dissected from starved larvae and fed control larvae was stained with an antibody against Lipin (L) and DAPI to visualize DNA.