HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 276, Issue 36, 33336-33344, September 7, 2001

This Article Abstract Full Text (PDF) All Versions of this Article: 276/36/33336    most recent M105193200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Baulande, S. Articles by Pairault, J. Search for Related Content PubMed PubMed Citation Articles by Baulande, S. Articles by Pairault, J. Social Bookmarking                      What's this?

Adiponutrin, a Transmembrane Protein Corresponding to a Novel Dietary- and Obesity-linked mRNA Specifically Expressed in the Adipose Lineage^*

Sylvain Baulande, Françoise Lasnier, Marguerite Lucas, and Jacques Pairault^§

From the Université Pierre et Marie Curie, UMR Physiologie et Physiopathologie, CNRS, 15 rue de l'Ecole de Médecine, 75270 Paris, Cedex 06, France

Received for publication, June 6, 2001, and in revised form, June 27, 2001

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

We have used a mRNA differential display technique to identify new genes involved in the reprogramming of gene expression during the adipose conversion of mouse 3T3 preadipocyte cell lines. We report here on the identification and cloning of a novel adipose-specific cDNA encoding a predicted membrane protein of 413 amino acids. The level of the corresponding 3.2-kilobase mRNA is tremendously increased during 3T3-L1 and 3T3-F442A differentiation into adipocytes. A single, very abundant 3.2-kilobase transcript is also found in inguinal and epididymal white adipose tissues and in interscapular brown adipose tissue but not in any other tissues examined. Its expression in adipose tissue is under tight nutritional regulation. The level of this novel 3.2-kilobase transcript becomes virtually nondetectable during fasting but is dramatically increased when fasted mice are refed a high carbohydrate diet. Based on its adipose specificity and dietary regulation, the novel gene product has been designated adiponutrin. The expression of adiponutrin mRNA is also 50-fold elevated in genetically obese fa/fa rats, indicating a link between adiponutrin and obesity. Western blot and confocal imagery analyses with epitope-tagged protein transiently expressed in 3T3-L1 adipocytes, and COS cells show that adiponutrin strictly localizes to membranes and is absent from the cytosol. Sequence analysis reveals homologies with several other members of related eukaryotic proteins, including a human paralog, which has been recently described in vesicular transport mechanisms. This leads us to suggest that adiponutrin could be involved in vesicular targeting and protein transport restricted to the adipocyte function.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Adipocytes, the main cellular components of adipose tissue, serve an important function in the energy economy of vertebrate organisms through a unique and quite specialized complement of enzymes and regulatory proteins needed to carry out de novo lipogenesis, lipolysis, and several other physiological functions including secretion of paracrine and endocrine factors (for reviews see Refs. 1 and 2). Studies of adipogenesis became experimentally feasible with the availability of immortal cell lines that differentiate in vitro into white adipocytes. The most compelling evidence that 3T3 preadipocyte cell lines represent faithful models of preadipocyte differentiation in vivo comes from transplantation studies into athymic mice (3). Whether in vivo or in vitro, the development of these preadipocyte cell lines led to functionally, morphologically, and biochemically identical cells to tissue adipocytes as a result of the execution of an elaborate program of differential gene expression. For differentiation to occur, a large number of genes has to be activated or repressed in a selective and coordinated manner. Thus, in vitro the differentiation of preadipocytes to adipocytes involves a series of events including morphological changes, triacylglycerol accumulation, and alterations in the levels of over 300 proteins (4). Conversely, the time course of differentiation is reflected by the appearance of numerous early and late mRNA markers. During the past few decades, considerable progress has been made in identifying molecular details of adipocyte differentiation with the identification of a great number of related genes. They include genes coding for enzymes and structural proteins (5), hormone receptors (6, 7), secreted proteins, and the regulatory network of trans-acting factors for preadipocyte differentiation (8).

So far, the molecular characterization of these genes has been arisen from the screening of cDNA libraries of differentiated adipocytes by differential hybridization, subtractive methods, and/or various shotgun techniques and more recently by microarray analysis (9, 10). In our laboratory, we have been studying for several years the differentiation of 3T3 adipoblasts into mature adipocytes by using the mRNA differential display approach developed by Welsh et al. (11). By comparing PCR amplification of mRNA isolated either from differentiation-defective (3T3-C2 clones) or differentiation-competent (3T3-L1 or 3T3-F442A clones) cell types at various stages, we isolated a great number of cDNA fragments whose corresponding transcripts are several dozen-fold increased or decreased during adipose conversion. Nucleotide sequences were systematically analyzed, and our interest was focused on cDNA species revealing no homology with sequences listed in GenBank^TM/EMBL data bases. We further explored these unknown mRNA sequences for additional characteristics, namely tissue specificity and physiological regulation. Indeed, whether in vivo or in vitro, adipocyte development requires a complex process of terminal differentiation that is controlled by the interplay of intracellular factors and various signals, namely hormonal and nutrimental. So the scope of the present study was to identify among the set of these unknown cDNA fragments those corresponding both to adipose-restricted and diet-controlled mRNA species. First, we used them as probes in Northern analysis of tissue mRNAs from various origin. Second, we further screened our panel of cDNA against adipose tissue RNA isolated from rodents submitted to starvation-refeeding regimens. Among the set of cDNA fragments, one of them appeared particularly interesting, and we decided to pursue further its characterization.

The present paper describes the molecular cloning and sequencing of the corresponding full-length mRNA. The new gene whose function is unknown so far is termed adiponutrin. In vitro the level of adiponutrin mRNA is induced at a very early stage during the onset of adipose differentiation in 3T3-L1 cells. An adiponutrin cDNA fragment hybridizes to a unique 3.2-kilobase mRNA species abundantly and specifically expressed in adipose tissue both of brown and white origin. Adiponutrin transcripts are coding for a 413-amino acid deduced protein. In vivo adiponutrin mRNA abundance is submitted to a dramatic down-regulation after starvation of animals while refeeding promptly restores mRNA levels. Furthermore, adiponutrin mRNA is dramatically overexpressed (more than 50-fold) in adipose tissue of genetically obese fa/fa rats as compared with their congenic thin littermates. Given the quite specific adipose tissue expression, its overabundance in a rodent model of genetic obesity, and its pattern of nutrient modulation, adiponutrin could represent an important clue in the adaptative function of adipose tissue for the regulation of energy homeostasis.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Cell Lines and Cell Culture-- Murine 3T3-C2 fibroblasts (12) and 3T3-L1 (13) and 3T3-F442A (14) preadipocytes were grown in 4.5 g/liter D-glucose DMEM supplemented with 10% fetal calf serum (FCS) at 37 °C in an atmosphere of air/CO₂ (90:10, v/v). Differentiation of 3T3-F442A cells was achieved by the addition of 1 µg/ml insulin at confluence. For 3T3-L1 cells, induction of differentiation was initiated according to Rubin et al. (15) by the addition of 1-methyl-3-isobutylxanthine (0.1 mM), dexamethasone (0.25 µM) and 1 µg/ml insulin to 24-h postconfluent cell cultures (day 0). After a 48-h treatment period, induction medium was removed, and cells were then refed with DMEM supplemented with 10% FCS and insulin (1 µg/ml). Eight days later, following confluence (day 8), more than 90% of 3T3-L1 or 3T3-F442A showed mature adipocyte morphological features. When grown in regular medium containing 10% FCS and insulin (1 µg/ml), 3T3-C2 cells were unable to undergo adipose conversion and kept a fibroblastic shape at all stages of culture.

COS cells were cultivated in DMEM supplemented with 10% FCS. Cell stocks were cultivated in these conditions until nearly confluent and subcultured at 1:20 dilution or plated for experiments.

RNA Arbitrarily Primed PCR (RAP-PCR) and cDNA Cloning and Sequencing-- Cellular total RNA was prepared according to the procedure described by Cathala et al. (16), and poly(A)⁺ RNA was isolated using oligo(dT)-cellulose columns. RAP-PCR was achieved according to the protocol of Welsh et al. (11) using sets of RAP-PCR primers (Stratagene). Briefly, 100 ng of poly(A)⁺ RNA were used in a reverse transcription reaction using the arbitrary 18-mer oligonucleotide primer A (primer A sequence, 5'-AATCTAGAGCTCCCTCCA-3') driven by 25 units of Moloney murine leukemia virus reverse transcriptase in presence of 40 units of RNase block ribonuclease inhibitor and 25 pmol of each dNTP. Then PCR was performed using 10 µl of a 1:10 cDNA dilution in a 50-µl final volume containing 2.5 pmol of each dNTP, 10 µCi of [-³²P]dCTP (ICN Pharmaceuticals), 50 pmol of primer A, and 1 unit of Taq DNA polymerase. All components for reverse transcription and PCR were provided by the RAP-PCR kit (Stratagene). PCR was conducted as follows: first, one cycle of denaturation step at 94 °C for 1 min, low stringency annealing at 36 °C for 5 min, and extension at 72 °C for 5 min and then 40 cycles of amplification (one cycle consisted of denaturation at 94 °C for 1 min, high stringency annealing at 50 °C for 2 min, and elongation at 72 °C for 2 min). Finally, a last elongation step at 72 °C for 10 min was performed. 5-µl aliquots of PCR mixtures were loaded on a 4% acrylamide, 7 M urea sequencing gel prepared in 1× TBE buffer. PCR products were visualized by autoradiography of the dried sequencing gel. Candidate differentially amplified PCR fragments were excised from the gel, and DNA strips were overlaid in 50 µl of TE buffer (1 mM EDTA, 10 mM Tris-HCl, pH 7.5) for a 1-h incubation at 60 °C followed by an overnight incubation at room temperature. Eluted DNA fragments were reamplified by PCR using the same arbitrary primer and subsequently cloned into a T/A cloning vector (pGEM-T Easy Vector System; Promega). The sequence of a 491-bp cDNA fragment, temporarily termed a5u6, was performed according to the dideoxysequencing procedure using a T7 DNA Polymerase Sequencing Kit (Sequenase version 2.0; U.S. Biochemical Corp.).

Rapid amplification of cDNA ends (RACE)-PCR strategies, as described by Frohman (17) were used to obtain the 5'- and 3'-flanking regions of a5u6 cDNA. For the 3'-RACE protocol, 200 ng of mature 3T3-L1 adipocyte poly(A)⁺ RNA were reverse transcribed using 10 units of avian myeloblastosis virus reverse transcriptase (Finzyme) in the presence of an oligo(dT)₁₅ adaptor primer B. The first PCR round was achieved using a first adaptor primer (AP1) and an a5u6 3'-end-specific primer (primer P1). A second PCR round was led using a second adaptor (AP2) and primer P1. A second 3'-chain extension was performed by using another distal a5u6 3'-end specific primer (primer P2). To obtain 5'-end cDNA clones, reverse transcription of 200 ng of poly(A)⁺ RNA was achieved in the presence of an a5u6 5'-end-specific primer (primer P3). After removal of the primer excess through Microcon 100 spin filters (Amicon) the 3'-end first strand cDNA was elongated by homopolymeric A tailing with 10 units of terminal deoxynucleotidyl transferase (Roche Molecular Biochemicals) in the presence of 4 nmol of dATP. The second strand cDNA was synthesized in the presence of the oligo(dT)₁₅ adaptor primer B and 2.5 units of Taq polymerase (Life Technologies, Inc.) with a 2-min annealing time at 50 °C followed by a 40-min elongation step at 72 °C. Finally, two rounds of PCR were achieved, the first one using the first adaptor primer AP1 and a5u6 5'-end-specific primer P3 and the last one using the second adaptor primer AP2 and the same a5u6 5'-end-specific primer P3. Sequences of primers used above were as follows: 5'-CCAGTGAGCAGAGTGACGAGGACTCGAGCTCAAGCT₁₅-3' for the oligo(dT)₁₅ adaptor primer B and 5'-CCAGTGAGCAGAGTGACG-3' and 5'-GAGGACTCGAGCTCAAGC-3' for the AP1 and AP2 adaptor primer, respectively. Sequences of specific primers for a5u6 were as follows: P1, 5'-ACCAACCTCAGCCTCCGC-3'; P2, 5'-GTATGATCAGTTAAGTGG-3'; and P3, 5'-GGAGCCCGTCTCTGATGC-3'. PCR products were visualized on a 1% agarose gel, and cDNA fragments were cloned and sequenced as described above. Final DNA inserts sequencing was carried out by Genome Express (Montreuil, France) using the ABI PRISM Dye Terminator Cycle Sequencing Core kit (PerkinElmer Life Sciences) and a model 377 sequencer (Applied Biosystems).

Plasmid Construction and Expression of FLAG-tagged Adiponutrin Construct-- A FLAG-tagged protein bearing only the open reading frame of adiponutrin was constructed by subcloning a reverse transcription-PCR product of adiponutrin into the C-terminal p3×FLAG-CMV-14 expression vector (Sigma). Briefly, forward and reverse primers were synthesized that corresponded to regions just 5' and 3' of the open reading frame of adiponutrin: forward (5'-TATCATGAATTCCACCACCATGTATGACCCAGAGC-3') and reverse (5'-TATCATTGAATTCAAACCAGCAGAGTTGCC-3'). These primers were used for reverse transcription-PCR for 20 cycles at 60 °C using 3T3-L1 adipocyte poly(A)⁺ mRNA as the template. The product of PCR amplification was digested with EcoRI and inserted into the corresponding site of pGEMT vector. After transformation into Escherichia coli DH5, the plasmid was digested with EcoRI, and the adiponutrin fragment was inserted in frame into EcoRI-digested p3×FLAG-CMV-14 vector (Sigma). The sequence of the resultant plasmid was confirmed by restriction mapping and sequencing.

Transient Transfection of Plasmid DNA into COS Cells and 3T3-L1 Adipocytes-- Cells were harvested after trypsinization and resuspended in PBS, and 500-µl cell suspension aliquots (6 × 10⁶ cells/assay) were electroporated at a voltage of 200 V and a capacitance of 975 microfarads (Gene-pulser apparatus; Bio-Rad) in the presence of 20 µg of CsCl-purified plasmid DNA. Cells were then seeded in 60-mm dishes (for protein analysis) or on coverslips in 35-mm culture dishes (for immunofluorescence analysis) and cultivated in DMEM containing 10% FCS for an additional 30-36-h period for maximal FLAG-tagged adiponutrin expression. Cells transfected with empty vector or mock-transfected cells were used as controls.

Cell Fractionation and Immunobloting Analysis of Epitope-tagged Adiponutrin-- Transiently transfected COS cells (see above) were grown for 36 h, washed with ice-cold PBS, resuspended in homogenization buffer (0.25 M sucrose, 1 mM EDTA, 10 mM Hepes, pH 7.4, supplemented with a mixture of protease inhibitors), and sonicated. The homogenate was centrifuged at 1,500 × g for 5 min. The low speed supernatant was then centrifuged for 1 h at 150,000 × g at 4 °C to obtain the high speed particulate and cytosolic fractions, respectively. The conditioned medium was collected from cell cultures that were overlaid for 24 h with fresh serum-free medium. It was filtered through a 0.45-µm pore size filter (Costar) and concentrated 30-fold by centrifugation in a 30-kDa cut-off spin column (Microcon). The pelleted membranes were resuspended in Laemmli sample buffer, and equal amounts of protein from each fraction were separated in an SDS-polyacryamide gel (18). Proteins were transferred to polyvinylidene difluoride membranes (Millipore Corp.). Membranes were blocked with TBS (20 mM Tris, 500 mM NaCl, pH 7.5) containing 5% nonfat milk, incubated with polyclonal anti-FLAG rabbit antibody (1:50,000; Sigma; catalog no. F7425). Bound antibody was detected with the use of anti-rabbit IgG conjugated with horseradish peroxidase (1:20,000; Amersham Pharmacia Biotech) and enhanced chemiluminescence reaction (PerkinElmer Life Sciences).

Immunofluorescence and Confocal Analysis-- Subconfluent cells were grown on coverslips for 36 h in DMEM supplemented with 10% FCS. Cells were washed three times with PBS and fixed for 30 min in paraformaldehyde (2% w/v) in sodium phosphate buffer, pH 7.5. Cells were washed again three times in PBS and permeabilized for 30 min in PBS containing 0.1% (w/v) saponin. After washing with PBS, cells were treated for 10 min with 2% paraformaldehyde and for 10 min with 100 mM glycine in PBS. Cells were then washed four times with PBS and kept at 4 °C in PBS containing 0.2% gelatin until immunolabeling. Cells were incubated for 1 h with primary antibody, washed twice with PBS and twice with PBS/gelatin, and stained for 2 h with secondary antibody. After washing with PBS, the coverslips were mounted on glass slides with Vectashield (Vector Laboratories). Primary antibodies were as follows: polyclonal anti-FLAG rabbit antibody (1:60) purchased from Sigma and polyclonal anti--enolase rabbit antibody (1:400) given by Dr. A. Keller (19). Secondary antibody was a Texas Red-conjugated sheep anti-rabbit IgG (1:200; Jackson Laboratories). All antibodies were diluted in PBS containing 0.2% gelatin. Confocal laser-scanning microscopy was performed using a Zeiss LSM510 microscope (Zeiss, Jena).

Northern Blot Analysis-- Total RNA from 3T3-L1 cells was extracted according to Cathala et al. (16), while tissue RNA was purified by using a RNA-PLUS kit (Quantum Appligene). For analysis (20) cell RNA (10 µg) or tissue RNA (20 µg) was loaded on 1.2% agarose electrophoresis gel containing 2.2 M formaldehyde and transferred to Nylon membranes (Biodyne B; PALL Gelman Laboratory). Blots were stained using methylene blue to estimate quantitative loading variations. Hybridization was carried out as described by Church and Gilbert (21). The 491-bp RAP-PCR cDNA fragment was used as a probe after labeling by random priming (Prime-a-Gene Labeling System; Promega). Final washing was performed in 0.2× SSC, 0.1% SDS for 15 min at 60 °C for mouse samples, nonstringent conditions (1× SSC, 0.1% SDS for 15 min at 50 °C) were used for rat samples. Membranes were exposed to X-AR5 films (Eastman Kodak Co.) and autoradiographs were scanned using a computing videodensitometer (Vilbert Lourmat).

Animals and Dietary Studies-- Care of rodents was according to institutional guidelines. Male SWISS mice (7-8 weeks old) were housed (six mice/cage) and maintained with a 12-h light/12-h dark photoperiod. Animals were fed with a commercial diet, and water was given ad libitum. In fasting experiments, food was removed at the beginning of the light period and reintroduced 19 h and 30 min later. Animals were killed at the indicated time, epididymal and interscapular fat pads (white adipose tissue and brown adipose tissue, respectively) were removed, and blood was collected for glucose and insulin monitoring. Glucose concentration was measured by the glucose oxidase procedure, and insulin concentration was measured by a radioimmunoassay (INSULIN-CT in vitro test; CIS Bio International). For lean and obese comparative studies, 5-week-old female Zucker fa/fa rats (generous gift of Dr. R. Bazin, INSERM U465, Paris) were fed ad libitum and killed after a dark period.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Characterization of Adiponutrin, a Novel mRNA Differentially Expressed during Adipose Conversion-- To identify novel genes that are differentially expressed during adipocyte conversion, we used the RAP-PCR protocol (see "Experimental Procedures"). Poly(A)⁺ RNA was prepared from undifferentiated and fully differentiated 3T3-L1 cells. As a control for the differential display, poly(A)⁺ RNA was also extracted from 3T3-C2 cells, a fibroblastic cell line that is unable to turn into adipocytes. After reverse transcription of these RNAs, PCR amplification using an arbitrary primer, and comparative gel analysis of [-³²P]dCTP-labeled PCR products (see "Experimental Procedures"), we focused our interest on a partial cDNA fragment, termed a5u6, only amplified from the mature adipocyte RNA pool. The profile of the differential display experiment is represented in Fig. 1A. This cDNA insert was prepared by PCR reamplification, cloned into an A/T vector, and sequenced by the dideoxy chain termination method. Northern blot analysis using the 491-bp a5u6 DNA fragment as a probe was carried out to confirm the pattern of differential expression (Fig. 1B). Total RNA extracted from 3T3-L1 and 3T3-F442A preadipose cells in growing or resting state did not exhibit any hybridization signal with the 491-bp a5u6 probe (Fig. 1B). Similarly, no detectable signal was observed in 3T3-C2 fibroblastic cells kept in a resting state for 8 days following confluence (Fig. 1B). In contrast, 3T3-L1 and 3T3-F442A fully differentiated adipocytes contained large amounts of a 3.2-kilobase mRNA species. This newly identified transcript was called "adiponutrin" due to the related properties that will be explained below. RNA loading variations were estimated using an unknown housekeeping mRNA, arbitrarily termed a4d4, as a normalizing probe. During preliminary experiments, the abundance of this short mRNA species (about 500 bp in length) appeared as invariant according to growth and differentiation. Moreover, its expression was strongly correlated with methylene blue-stained 18 S ribosomal RNA (Fig. 1B) so that it was a very useful tool to normalize RNA loading and estimate mRNA integrity.

View larger version (45K): [in this window] [in a new window]   Fig. 1.   Identification of a novel adipocyte-specific mRNA. A, pattern of mRNA differential display. Reverse transcriptions were performed on poly(A)+ RNA isolated from 3T3-C2 fibroblastic cells (F), 3T3-L1 preadipocytes (P), and 3T3-L1 adipocytes (A). RAP-PCR was conducted in the presence of [32P]dCTP as described under "Experimental Procedures." 32P-Labeled PCR products were separated on denaturing polyacrylamide gel and visualized by autoradiography. B, Northern blot analysis performed with 20 µg of total RNA from 3T3-C2 (undifferentiating cells) and 3T3-L1 and 3T3-F442A (differentiating cells) at different stages. The 491-bp a5u6 cDNA fragment was used as a probe. RNA loading was normalized by methylene blue staining of 18 S ribosomal RNA and hybridization to a4d4 probe.

These data suggested that adiponutrin was a genuine, differentially regulated mRNA species.

Adipose Tissue-specific Expression of Adiponutrin in Mice-- Northern analysis with clone a5u6 as a probe using various tissue RNAs from mice indicated that adiponutrin mRNA was satisfying the criterion of adipose tissue specificity. As shown in Fig. 2, a single abundant mRNA species was readily detected in total RNA from mouse epididymal (white adipose tissue) and interscapular (brown adipose tissue) fat pads. The size of adiponutrin mRNA in adipose tissue was similar to that found in 3T3- adipocytes. Adiponutrin mRNA was virtually undetected in brain, heart, liver, intestine, lung, muscle, spleen, kidney, and testis. In this set of tissues, no signal was detectable even after a very long time of film exposure (4 days). These results were consistent with the increased expression of adiponutrin observed during adipocyte differentiation in established cell lines. The highly restricted adipose tissue expression of this new mRNA species in vivo led us to pursue further its characterization.

View larger version (29K): [in this window] [in a new window]   Fig. 2.   Tissue expression of adiponutrin. 20 µg of total RNA from different mouse tissues was analyzed by Northern blot using a5u6 as a probe. Methylene blue staining was used to normalize for RNA loading.

Molecular Cloning of the Entire Coding Sequence of Adiponutrin-- We decided to isolate and determine the complete nucleotide sequence of the corresponding full-length transcript and the amino acid sequence of its encoded protein. To obtain the sequence information of the 5' and 3' regions flanking the original a5u6 491-bp fragment, we used the RACE-PCR protocol (See "Experimental Procedures"). The cDNA strand that contained the major open reading frame was assumed to be the (+)-strand, a hypothesis supported by the existence of a predicted human gene (GenBank^TM data base) encoding for a conceptual translation protein with a high degree of homology. Initial 3'-RACE extensions (first one, 1018 bp; second one, 823 bp) and 5' extension (272 bp) confirmed that our assumed orientation of clone a5u6 was correct. The 5'-RACE reaction was repeated several times, and different PCR products were sequenced to eliminate PCR errors and validate the completion of 5' extension. By combining the sequence of the original a5u6 clone (491 bp) together with the ones obtained by 3'- and 5'-RACE reactions and assembling with a partial mouse cDNA sequence (403 bp) from the dbEST data base, we thereby obtained a cDNA sequence of 3007 bp (Fig. 3A). A major open reading frame of 1239 bp was identified, resulting in a predicted protein of 413 amino acids. The 5'-most ATG found within the coding region of the adiponutrin cDNA is likely to serve as the start site for translation initiation, although the sequence flanking this ATG does not conform very well with Kozak's rules (22).

View larger version (75K): [in this window] [in a new window]   Fig. 3.   Nucleotide and deduced amino acid sequences of adiponutrin (GenBankTM accession number AY037763). A, full-length sequence of adiponutrin cDNA (3007 bp). The amino acid sequence is derived from the longest open reading frame (413 amino acids; one-letter code in boldface letters). The underlined nucleotides are the first upstream in-frame stop codon and the two putative polyadenylation signals, respectively. The asterisk indicates the stop codon. B, conserved domains between mouse adiponutrin and related proteins of other species and predicted architecture. Two highly homologue motifs are represented (motifs 1 and 2). Eight related proteins were analyzed by the ClustalW alignment program (33). Homologous proteins are identified by their accession number. The amino acid sequence deduced from AK025665 cDNA corresponds to the human ortholog of adiponutrin. The mouse ortholog of human TTS-2 corresponds to BAB22387. A schematic illustration of the general structure of the family is given. Amino acids from position 1 to 280 of mouse adiponutrin sequence (gray-shaded) correspond to the highly conserved region between the related molecules, which includes motif 1 and 2 signatures (arrows). The variable region between the proteins is represented by the white-shaded portion corresponding to C-terminal residues. Four putative transmembrane spans (TM, boxed in black) have been predicted in all members of the family using the TMpred program (23).

Data Base Searches and Sequence Analysis of Adiponutrin-- The multiple genome sequencing projects and data bases of ESTs and proteins have now made it possible to search for adiponutrin-like members in several different organisms spanning the plant and animal kingdoms. A BLAST search of the GenBank^TM data base has showed that adiponutrin cDNA sequence exhibits a noticeable identity (~75% in aligned exons) with an uncharacterized gene in the locus 22q13 of the human chromosome 22, suggesting that this locus could contain the human ortholog of this gene. Homology searches have identified related sequences in a range of metazoans including vertebrate and invertebrate genomes. But neither homologous sequences have been identified in prokaryotic genomes or in fungi from the completed S. cerevisiae genome. The N terminus part of mouse adiponutrin shows a high degree of amino acid identity with protein from several species, including human, Caenorhabditis elegans, Drosophila melanogaster, and Arabidopsis thaliana. In humans, four homologous proteins bear a strong degree of similarity: GS2 protein (accession number P41247), transport secretion protein-2 (TTS-2; CAC01132), a putative GS2-like protein (CAB09789), and the probable human orthologous protein forming a cluster of two genes in the locus 22q13. Several other proteins exhibit a high degree of homology: two from the D. melanogaster genome, CG5295 and CG5560 gene products (AAF49704 and AAF48418, respectively for accession numbers) and three others from the C. elegans genome (C05D11.7, B0524.2, and D1054.1 proteins (Q11186, AAF39747, and T20303, respectively for accession numbers)). All of these proteins are deduced by conceptual translation from genome sequencing projects without any further functional information, except for TTS-2, which has been cloned and described as an vesicular transport-secretion protein of the cell surface receptor ICAM-3. In mice, two hypothetical translated protein are described (BAB22387 and BAB29543 as accession numbers). Otherwise, several short open reading frames encoding weak adiponutrin homologs are issued from sequencing projects in the GenBank^TM/EMBL/DBJ data bases. These open reading frames have shown several obvious reading frame jumps and dislocations of base pairs, probably due to sequencing mistakes. These findings provide an indication that there are multiple paralogs each deriving from separate but related genes. For instance, the sequence identity/similarity may be higher for particular paralogs compared between different mammals (~90% for human TTS-2 versus mouse adiponutrin) than for the ones of distinct paralogs of the same species (~60% for mouse proteins).

Multiple sequence alignments have revealed the highly conserved domain structure of the adiponutrin family. All family members share a common core domain starting at the N terminus and ending at position 280 of the adiponutrin sequence. Using the program TMpred (23), this region encompasses three predicted transmembrane spans of similar length and conserved amino acid sequence (Fig. 3B; blocks TM1-TM3). In this regard, the motif 1 sequence G(C/A)SAG(A/S)L located in transmembrane span TM2 appears as a hallmark of the family. A fourth transmembrane span (TM4) is also predicted by the program TMpred; it is located in the variable C terminus part of the proteins, but sequences preceding and following TM4 are divergent in the length and nature of amino acids. In contrast, loops between the transmembrane spans TM1-TM3 are well conserved in length, while the amino acid sequence linking TM3 to TM4 (Fig. 3B, motif 2, sequence ¹⁸¹TITVSFPXGEXDICP¹⁹⁵) is especially conserved in all adiponutrin homologs. A Prosite pattern search of functional motifs in the deduced amino acid sequence of adiponutrin did not reveal any interesting features.

Subcellular Distribution and Localization of Adiponutrin Protein-- In order to validate the predicted topology of adiponutrin as an integral membrane component, an epitope-tagged adiponutrin construct was introduced into COS cells by transient transfection. Then, cell homogenates were submitted to fractionation by ultracentrifugation and the fusion protein was analyzed by immunoblotting of membrane and cytosolic fractions. The C-terminal FLAG-tagged adiponutrin was readily detected in crude homogenates as a protein migrating with an apparent molecular mass of 48 kDa (Fig. 4). This is the size expected from the amino acid sequence of the chimera. This also indicates that the protein is not likely to be processed by posttranslational modifications such as glycosylation. Furthermore adiponutrin-FLAG protein was seen to be present in the membrane fraction and absent in the cytosolic one (Fig. 4). As a control, we used an antibody raised against the glycolytic enzyme -enolase. We detected this soluble protein almost exclusively in the cytosolic fraction with a faint band in the microsomal fraction. Even after a longer time exposure, no adiponutrin-FLAG signal was apparent in the 50-fold concentrated culture medium. This finding indicates that the protein is not secreted and is well in line with the absence of N terminus hydrophobic signal peptide sequence.

View larger version (41K): [in this window] [in a new window]   Fig. 4.   Subcellular distribution of adiponutrin-FLAG fusion protein. Conditioned medium, cell lysate, and subcellular fractions of transiently transfected COS cells with p3×FLAG-CMV-14 expression vector were prepared, and proteins were analyzed by Western blotting (see "Experimental Procedures"). Cell proteins (20 µg of protein) and conditioned medium (50 µl of 50-fold concentrated medium) were separated on 10% SDS-polyacryamide gel electrophoresis and immunoblotted with anti-FLAG antibody (upper panel) and anti--enolase antibody (lower panel).

The intracellular localization of adiponutrin-FLAG was investigated by confocal scanning microscopy of transiently transfected 3T3-L1 adipocytes and COS cells (Fig. 5). In both types of cells, the pattern of adiponutrin appears more compact near the periphery of the plasma membrane by forming patch-like structures. Overall, the staining has a granular appearance and often consists of numerous punctate structures throughout the cytoplasm. Although we are aware of the difficulties associated with drawing precise conclusions about the localization of an endogenous protein from data obtained with an overexpressed one, the association of adiponutrin with endoplasmic reticulum membranes remains plausible. As a control, the staining pattern of -enolase appears different, uniformly dispersed in the cell cytosol, in agreement with the known localization of this glycolytic enzyme (data not shown).

View larger version (74K): [in this window] [in a new window]   Fig. 5.   Immunolocalization of adiponutrin-FLAG fusion protein in COS cells (A) and 3T3-L1 (B) by confocal microscopy. Cells were transfected with the epitope-tagged adiponutrin expression vector, seeded for 48 h, and then fixed and were processed with anti-FLAG primary antibody. After staining with Texas Red-conjugated sheep anti-rabbit IgG, preparations were analyzed by confocal microscopy (see "Experimental Procedures"). A punctate staining of overexpressed adiponutrin is seen both for COS cells and 3T3-L1 adipocytes. The staining appears brighter close to the cell membrane but never around lipid droplets of adipocytes. Confocal images are shown at the middle level of the cell.

Regulation of Adiponutrin Expression in 3T3 Adipocytes-- In order to examine the time course of adiponutrin mRNA expression during the adipose conversion of 3T3-L1 and 3T3-F442A preadipocytes, total RNA was extracted at various time points of cell cultures and analyzed by Northern blot. Adiponutrin mRNA expression was very well correlated with cell differentiation and changes in cell morphology (round shape and apparition of cytoplasmic lipid droplets) (Fig. 6). Adiponutrin levels were found to increase dramatically during early differentiation, as soon as day 2 following induction of adipose conversion. A maximal level of adiponutrin mRNA was reached at day 8, so that the message appeared as a prominent mRNA species in mature adipocytes (Fig. 6). The expression profile of adiponutrin gene paralleled those of fatty acid synthase (FAS) and ADD1/SREBP1c. Thus, our results demonstrated a simultaneous in vitro emergence of adiponutrin mRNA expression and adipocyte phenotype.

View larger version (31K): [in this window] [in a new window]   Fig. 6.   Adiponutrin mRNA induction during adipocyte differentiation of 3T3-L1 and 3T3-F442A cell lines. Cells were grown in DMEM with 10% FCS. Confluent preadipocytes (day 0) were exposed to a differentiation mixture as described under "Experimental Procedures." 10 µg of RNA prepared from cells at the indicated time points, day 1, day 0 (confluence), day 1, day 2, day 3, day 4, day 5, and day 8, were subjected to Northern blot analysis for adiponutrin, FAS, and ADD1 mRNAs. As a control for RNA loading, the a4d4 hybridization signal is shown.

Since in vitro adipose differentiation is elicited or maintained by various extracellular signals, we decided to further investigate hormone or nutrient control of adiponutrin mRNA expression in 3T3 adipocytes. Mature 3T3-L1 adipocytes were differentiated for 8 days in regular medium (DMEM supplemented with 10% FCS), and then cells were challenged to a serum-free, glucose-free medium supplemented with pyruvate, lactate, and bovine serum albumin (Fig. 7). In these conditions, adiponutrin mRNA levels were found to decrease dramatically within a 24-h period. After that, various nutriments or hormones were added to the culture medium for an extra 6-h period. The addition of D-glucose elicited a 7-fold increase of adiponutrin mRNA levels. Insulin had a slight and discrete effect whether it was added alone or together with glucose. Furthermore, the increase of adiponutrin gene expression obtained with glucose treatment could be counteracted by the addition of forskolin or isoproterenol and MIX, agents known to mimic catecholamine effects by raising intracellular cAMP levels.

View larger version (29K): [in this window] [in a new window]   Fig. 7.   In Vitro regulation of adiponutrin mRNA expression in 3T3-L1 adipocytes submitted to glucose deprivation and complementation experiments. As established in cell culture section, 3T3-L1 adipocytes were obtained from cell cultures induced for 2 days by dexamethasone-isobutylmethylxanthine-insulin treatment and further differentiated for 6 days in regular medium containing glucose (4.5 g/liter), insulin (170 nM), and 10% FCS (regular adipocytes, day 8). A, Northern analysis of adiponutrin (10 µg of RNA/lane) in adipocytes (Regular) or adipocytes submitted to a 16-h period of serum, insulin, and glucose deprivation (Deprived). B, deprived adipocytes (control t = 0) were submitted to an additional 6-h period of incubation without glucose (no addition) or supplemented with 170 nM insulin (INS). In all other lanes, adipocytes were incubated in the presence of 5 g/liter glucose (GLU) alone or associated with various effectors including 170 nM insulin (INS) or 100 µM isobutylmethylxanthine (MIX), 10 µM isoproterenol (ISO), or 10 µM forskolin (FSK). 10 µg of total RNA was analyzed by Northern blotting. Relative signal intensities were estimated by videodensitometric scanning. The hybridization signals obtained with a4d4 probe were used for normalization.

In Vivo Changes of Adiponutrin mRNA Levels in Nutritional States-- Due to the glucose-dependent gene expression observed in vitro, possible effects of changes in nutritional states on adiponutrin expression were investigated in mice submitted to fasting-refeeding regimens. The level of adiponutrin mRNA was high in white and brown adipose tissue from fed animals, but it dropped to a barely detectable level after 19 h of fast (Fig. 8). Refeeding animals for 8 h promptly reversed the drop and rescued the initial expression level observed in control animals. As positive controls, FAS and ADD1/SREBP1c mRNA exhibited a similar pattern of expression. Conversely, -enolase mRNA, the encoded product of which is a housekeeping glycolytic enzyme, remained unchanged in these various nutritional states both in white and brown adipose tissue. This demonstrates that adiponutrin levels are responsive to acute metabolic changes in vivo. All of these findings are reminiscent of its functional relevance in lipogenesis and/or adipogenesis in vivo.

View larger version (78K): [in this window] [in a new window]   Fig. 8.   Adiponutrin expression in mouse adipose tissue is dependent upon changes in nutritional status. 8-week-old Swiss mice were divided into three experimental groups and were either allowed free access to food (Fed) or subjected to 19.5 h of fast (Fast) or 12 h of fast followed by 7.5 h of refeeding (Fast/Refed). The animals were sacrificed at the end of 19.5 h, and 10 µg of total RNA extracted from epididymal fat pads was used for Northern blot analysis for adiponutrin, ADD1, and FAS mRNAs. Mean values of insulinemia and glycemia corresponding to each nutritional status are indicated at the top. -Enolase mRNA, probed with the 3'-untranslated region of the murine mRNA (accession number ACX52379) was used as invariant control. 18 S ribosomal RNA from ethydium bromide-stained gels is shown.

Alterations of Adiponutrin Gene Expression in a Rat Obesity Model-- Dysregulated expression of a gene in states of obesity may provide valuable clues on its functional relevance in metabolic disease states. Accordingly, we investigated the Zucker (fa/fa) obese rat for modulation of gene expression. 5-week-old female lean and homozygous obese rats were fed ad libitum and sacrificed in the morning at the end of the feeding period. Total RNA was extracted from inguinal and interscapular adipose tissues and analyzed by Northern blot. We observed a 30-50-fold elevation of the adiponutrin mRNA level in obese animals relative to their congenic lean controls, both in white and brown fat (Fig. 9). This observation was representative, since it was based on six animals per group.

View larger version (34K): [in this window] [in a new window]   Fig. 9.   Dysregulation of adiponutrin gene expression in obese Zucker rat (fa/fa). Northern analysis of adiponutrin (10 µg RNA/lane) in white inguinal (white adipose tissue (WAT)) and brown scapular (brown adipose tissue (BAT)) fat tissues from obese Zucker (fa/fa) rats and lean congenic controls are shown. Blots were hybridized with the 491-bp a5u6 fragment as a probe and washed in no stringent conditions. 18 S ribosomal RNA from ethydium bromide-stained gels was used to correct for RNA loading.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The characterization of novel genes that are highly expressed during adipose differentiation in vitro has been so far a privileged way to assess the mechanisms controlling adipogenesis and the maintenance of adipocytic terminal phenotype. Both in vitro and in vivo adipogenesis is profoundly influenced by a variety of hormones and nutritional signals. In the present study, we have focused our attention on a novel and very abundant mRNA species that is adipose-specific and responds to changes in animal nutrition. Corresponding to a single size message in adipocytes, we have cloned and sequenced a full-length cDNA of 3007 bp encoding an integral membrane protein of 413 amino acids. The novel gene product, called adiponutrin, possesses all of the hallmark features of an important marker of adipocyte development and homeostasis with respect to tissue specificity, nutritional status, and physiopathologic states in animals.

While belonging to an extended, well conserved family of homologous products, the adiponutrin transcript is only expressed in adipose tissue. Adiponutrin message is virtually absent in 10 other major tissues, including liver. The specificity of adiponutrin expression is corroborated by analysis of data bases of ESTs. Indeed, most similar ESTs originate from the Soares mammary gland NbMMG Mus musculus cDNA bank, a finding attributed to the constant fat contamination of mammary epithelium. Adiponutrin gene expression is more elevated in brown interscapular than in inguinal or epididymal white adipose tissue. Such a situation is common for most of coexpressed genes and probably reflects some differences between metabolic activities of both kinds of tissue. Despite this minor difference, both tissues exhibit similar dramatic changes of adiponutrin mRNA expression in the response to an altered nutritional status of mice. Thus, the levels of adiponutrin mRNA in brown and white adipose tissue are reduced dramatically upon fasting and rescued upon refeeding of animals. This pattern of gene regulation generally occurs in metabolically responsive tissues that synthesize lipids for storage or export. In adipose tissue as well as in the liver, the control of gene expression by a nutritional transition is the attribute of those genes carrying out the lipogenic or lipolytic programs. Thus, in adipose tissue, the increase in lipogenesis after weaning to a commercial diet (high carbohydrate) is accompanied by a large increase in three major lipogenic enzyme activities and levels of corresponding mRNAs (i.e. FAS, acetyl-CoA carboxylase, and ATP-citrate lyase (24, 25)). Nuclear spot 14 protein, the expression of which matches that of lipid synthesis (26), is also induced by premature weaning and disappears in catabolic circumstances such as fasting (27). Hence, leptin, the encoded protein of the ob gene, which is secreted by the adipocyte as part of the sensing program of adipose stores, also behaves as nutritionally regulated (28). Similarly, the regulation of FAS and leptin genes has been shown to be closely related to that of the ADD1/SREBP1c transcriptional factor in fasting-refeeding experiments in rodents (28).

An interesting point to consider is the nature of hormone and nutrient signals that could be involved in the regulation of adiponutrin gene expression. The control of gene expression by feeding has been attributed to the major feeding-associated hormone, insulin. In the present study, insulin alone appears to slightly modulate adiponutrin mRNA content in 3T3-L1 adipocytes. There seems to exist an unusual situation, since we here show that adiponutrin gene expression is clearly up-regulated by glucose in mature 3T3 adipocytes. In fine, the pattern of adiponutrin responsiveness to glucose is similar to regulation of several other lipid-metabolizing enzymes genes such FAS, acetyl-CoA carboxylase, or stearoyl-CoA reductase, which are all induced by glucose in adipose cells or adipose tissue explants (29, 30). However, insulin does appear to slightly enhance the effects of glucose on adiponutrin adipocyte expression. This contrasts with the strong enhancing effect of insulin with glucose observed for FAS and acetyl-CoA carboxylase in adipose tissue but is reminiscent of the glucose-dependent and insulin-independent regulation of stearoyl-CoA reductase in adipocytes (30). Since insulin partially restores stearoyl-CoA reductase mRNA levels in the liver of diabetic mice (31), the final clarification of this point for adiponutrin should be addressed with in vivo experiments.

Concerning the catabolic point of view, our in vitro experiments indicate that isoproterenol and forskolin, both activators of cAMP production, elicit a profound decrease of adiponutrin mRNA levels in 3T3-L1 adipocytes. Thus, the classic hormones of the fasted state (i.e. catecholamines) that function by raising cAMP levels could explain the dramatic drop of adiponutrin expression observed upon fasting in vivo.

Further evidence for a role of adiponutrin in adipose development and maintenance of adipose tissue homeostasis is provided by experiments with genetically obese fa/fa rats. Adiponutrin mRNA levels of adipose tissue of homozygous obese fa/fa rats are 50-fold higher than those of lean animals. This very strong up-regulation of adiponutrin mRNA in fa/fa rats may contribute to some of the pathophysiological features of these animals. Further investigations of adiponutrin gene expression in other animal models of genetic or induced obesity will be required to extend this observation.

Our studies have substantially identified a new family of proteins sharing significant homologies. At this time, the structure and sequence of adiponutrin protein are poorly informative, and its function remains to be elucidated. Adiponutrin belongs to a family of related proteins that is restricted to mammals, insects, and nematodes. The murine genes for the adipose-specific adiponutrin and the ubiquitous human TTS-2 are clearly paralogs. Obviously, it cannot be established whether the homologs found in C. elegans and D. melanogaster are orthologs of adiponutrin, TTS-2, or a common ancestor of the two genes. The Clusters of Orthologous Groups of Proteins data base (available on the World Wide Web at ncbi.nlm.nih.gov/COG (32) represent a phylogenetic classification of protein from 30 completes genomes of bacteria, archaea, and the yeast S. cerevisiae. No homologs appear in any shared or unshared clusters in inferior organisms, suggesting that members of the family are not related to enzymes found in all three divisions of cellular life. The absence of any obvious yeast homolog suggests that adiponutrin-related proteins serve a function that is observed only in multicellular eukaryotes. Furthermore, adiponutrin appears as a functional hallmark of the adipocyte. Its large increase during the early stage of adipogenesis, its exclusive expression in adipocytes, its dramatic repression/induction during fasting/refeeding in mice, and its dysregulated overexpression in obese animals strongly suggest that adiponutrin may pertain to some regulatory aspects of the pathway of lipogenesis or lipolysis among others.

The present data are compatible with a complex and dynamic role in adipocyte, for example in the regulation of triacylglycerol storage and/or lipid vacuole processing or in the modulation of certain regulatory/secretory function of the cell. In this regard, the closest related mammalian protein to adiponutrin is TTS-2. This protein has been functionally annotated as a novel protein implicated in vesicular transport (GenBank^TM accession number CAC01131 or CAC01132). As part of this study, we have confirmed the computer-assisted prediction that adiponutrin is a membrane protein. Its subcellular distribution may extend more broadly to include most, if not all, membranes that recycle between the cell surface and internal compartments. We suggest that this nutrient-sensed protein serves a role in membrane trafficking and/or vesicular transport of some other translocated or secreted adipose-specific component. Adipose tissue secretes a myriad of factors involved in a blood-borne homeostatic mechanism. In vitro experiments of conditional counterexpression of adiponutrin in 3T3 differentiating cells could permit us to directly evaluate its plausible implication in the emergence of the adipocyte phenotype. Otherwise, experiments of gene invalidation through transgenesis in mice should shed further light on these questions.

Acknowledgments-- We thank Dr. Bruno Feve for providing tissue samples from animals subjected to fasting-refeeding protocols. We are also grateful to Dr. Angelica Keller for the gift of -enolase antibody and Dr. Christophe Klein for technical assistance in confocal microscopy. Dr. Philippe Djian is gratefully acknowledged for critical review of the manuscript.

	FOOTNOTES

^* This work was supported by the Centre National de la Recherche Scientifique and the Université Pierre et Marie Curie (Paris VI).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper has been submitted to the GenBank^TM/EMBL Data Bank with accession number(s) AY037763.

Recipient of the Ministère de l'Education Nationale, de la Recherche et de la Technologie.

^§ To whom correspondence should be addressed. Tel.: 33 1 42 34 68 74; Fax: 33 1 46 34 59 73; E-mail: jpairau@ccr.jussieu.fr.

Published, JBC Papers in Press, June 28, 2001, DOI 10.1074/jbc.M105193200

	ABBREVIATIONS

The abbreviations used are: DMEM, Dulbecco's modified Eagle's medium; FCS, fetal calf serum; PCR, polymerase chain reaction; RAP-PCR, RNA arbitrarily primed PCR; bp, base pair(s); EST, expressed sequenced tag; RACE, rapid amplification of cDNA ends; TTS-2, transport secretion protein-2; ICAM-3, intercellular adhesion molecule-3; ADD1/SREBP1c, adipocyte determination and differentiation dependent factor 1/sterol regulatory element binding protein 1c; PBS, phosphate-buffered saline; FAS, fatty acid synthase.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				K. Kobayashi, T. Inoguchi, Y. Maeda, N. Nakashima, A. Kuwano, E. Eto, N. Ueno, S. Sasaki, F. Sawada, M. Fujii, et al. The Lack of the C-Terminal Domain of Adipose Triglyceride Lipase Causes Neutral Lipid Storage Disease through Impaired Interactions with Lipid Droplets J. Clin. Endocrinol. Metab., July 1, 2008; 93(7): 2877 - 2884. [Abstract] [Full Text] [PDF]

				I. J. Bujalska, K. N. Hewitt, D. Hauton, G. G. Lavery, J. W. Tomlinson, E. A. Walker, and P. M. Stewart Lack of Hexose-6-Phosphate Dehydrogenase Impairs Lipid Mobilization from Mouse Adipose Tissue Endocrinology, May 1, 2008; 149(5): 2584 - 2591. [Abstract] [Full Text] [PDF]

				E. E. Kershaw, M. Schupp, H.-P. Guan, N. P. Gardner, M. A. Lazar, and J. S. Flier PPAR{gamma} regulates adipose triglyceride lipase in adipocytes in vitro and in vivo Am J Physiol Endocrinol Metab, December 1, 2007; 293(6): E1736 - E1745. [Abstract] [Full Text] [PDF]

				S. Musaad and E. N. Haynes Biomarkers of Obesity and Subsequent Cardiovascular Events Epidemiol. Rev., May 10, 2007; (2007) mxm005v1. [Abstract] [Full Text] [PDF]

				L. Notari, V. Baladron, J. D. Aroca-Aguilar, N. Balko, R. Heredia, C. Meyer, P. M. Notario, S. Saravanamuthu, M.-L. Nueda, F. Sanchez-Sanchez, et al. Identification of a Lipase-linked Cell Membrane Receptor for Pigment Epithelium-derived Factor J. Biol. Chem., December 8, 2006; 281(49): 38022 - 38037. [Abstract] [Full Text] [PDF]