Cig30, a Mouse Member of a Novel Membrane Protein Gene Family, Is Involved in the Recruitment of Brown Adipose Tissue*

We have identified a previously uncharacterized gene that is implicated in the thermogenic function of brown adipose tissue of mice. This gene, termed Cig30, is the first mammalian member of a novel gene family comprising several nematode and yeast genes, such as SUR4 and FEN1, mutation of which is associated with highly pleiotropic phenotypes. It codes for a 30-kDa plasma membrane glycoprotein with five putative transmembrane domains. The Cig30 mRNA was readily detected only in brown fat and liver. When animals were exposed to a 3-day cold stress, the Cig30 expression was selectively elevated in brown fat more than 200-fold. Similar increases were brought about in two other conditions of brown fat recruitment, namely during perinatal development and after cafeteria diet. The magnitude ofCig30 mRNA induction in the cold could be mimicked by chronic norepinephrine treatment in vivo. However, in primary cultures of brown adipocytes, a synergistic action of norepinephrine and dexamethasone was required for full expression of the gene, indicating that both catecholamines and glucocorticoids are required for the induction of Cig30. We propose that the CIG30 protein is involved in a pathway connected with brown fat hyperplasia.

We have identified a previously uncharacterized gene that is implicated in the thermogenic function of brown adipose tissue of mice. This gene, termed Cig30, is the first mammalian member of a novel gene family comprising several nematode and yeast genes, such as SUR4 and FEN1, mutation of which is associated with highly pleiotropic phenotypes. It codes for a 30-kDa plasma membrane glycoprotein with five putative transmembrane domains. The Cig30 mRNA was readily detected only in brown fat and liver. When animals were exposed to a 3-day cold stress, the Cig30 expression was selectively elevated in brown fat more than 200-fold. Similar increases were brought about in two other conditions of brown fat recruitment, namely during perinatal development and after cafeteria diet. The magnitude of Cig30 mRNA induction in the cold could be mimicked by chronic norepinephrine treatment in vivo. However, in primary cultures of brown adipocytes, a synergistic action of norepinephrine and dexamethasone was required for full expression of the gene, indicating that both catecholamines and glucocorticoids are required for the induction of Cig30. We propose that the CIG30 protein is involved in a pathway connected with brown fat hyperplasia.
Brown adipose tissue is a mammalian organ specialized for heat production in the process called nonshivering thermogenesis. The proton potential generated by respiration across the inner mitochondrial membrane is dissipated in brown adipocytes through a unique protein channel, the uncoupling protein (UCP), 1 and the electrochemical energy is directly transformed into heat (1,2). Thermogenic activity of brown fat is particularly important in small mammals, such as rodents, when challenged by low ambient temperatures, notably after birth (3) or when arousing from hibernation (4). Energy dissipation by brown fat is also believed to control energy balance in animals fed a cafeteria diet (5).
Activation of nonshivering thermogenesis is controlled cen-trally from the hypothalamus via sympathetic innervation, and, accordingly, the principal agent controlling heat production is norepinephrine (6). Among the acute effects of norepinephrine is an increase in intracellular cAMP concentration and lipolysis of triglycerides to free fatty acids, which are the main substrate for thermogenesis (7). Catecholamines also enhance glucose transport in the brown adipocytes (8 -10) and elicit significant changes in phospholipid composition (11)(12)(13). At the gene expression level, adrenergic stimulation of brown fat rapidly increases the transcription of the Ucp gene (14,15) and elevates mRNA levels of several genes, such as those for LPL (16,17), GPDH (18), and GLUT4 (19). Remarkably, norepinephrine can also induce tissue hyperplasia in brown fat. Increased mitotic activity observed in brown adipose tissue in animals acutely exposed to cold (20) leads to a 3-4-fold increase in DNA content in the tissue about 1 week later (21), and this induction of proliferation can be mimicked by injection of norepinephrine (21,22). Thus, brown adipose tissue is rather unique in responding to a simple, non-invasive environmental stimulus as low ambient temperature by undergoing the recruitment process, i.e. a program of concerted physiological changes leading to accelerated proliferation and differentiation. Also unique is the possibility of using white adipose tissue, which has similar biochemical equipment yet plays a nearly antagonistic role in energy storage and expenditure, as a suitable reference system in a search for genes involved in this recruitment process. Such an approach based on differential screening of cDNA libraries from white adipose tissue and from brown adipose tissue of warmand cold-adapted animals was employed by Jacobsson et al. (15) to isolate the Ucp gene. The authors simultaneously selected several other so-called CIN clones that exhibited a high degree of brown fat specificity and inducibility by cold.
We have investigated here one of the CIN clones, the CIN-2. To this end, we cloned and characterized the corresponding full-length mRNA of CIN-2. It proved to be a novel mammalian gene, which we have termed Cig30 (cold-inducible glycoprotein of 30 kDa) and demonstrated that elevation of Cig30 expression is associated with recruitment of brown adipose tissue. We suggest that the CIG30 protein is involved in a membrane event related to cellular proliferation in brown adipose tissue.

EXPERIMENTAL PROCEDURES
Animals and Treatments-Gene expression studies were carried out with 6 -8-week-old male mice of the NMRI strain purchased from a local supplier (Eklunds). After arrival, the mice were housed separately, preacclimated, and maintained at 28°C, if not otherwise stated. In cold-stress experiments, the animals were exposed to 4°C for specified times, whereas controls remained at 28°C. Mice kept on a cafeteria diet received, besides the standard chow, cookies, chocolates, and cheeses of various kinds for 1 week. For chronic norepinephrine treatment, the mice were implanted subcutaneously (in the interscapular region) with microosmotic pumps (Alzet, model 1003D) delivering 6 g * This work was supported by grants from the Swedish Natural Science Research Council, the Swedish Cancer Society, and the Jeansson Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. This 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 /EBI Data Bank with accession number(s) U97107.
Cell Lines and Primary Cultures of Brown Adipocytes-The HIB-1B hibernoma cells (23) were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, 50 IU/ml penicillin, and 50 g/ml streptomycin. Brown fat precursor cells were isolated from 3-4week-old NMRI mice and cultured as described earlier (24,25). The cells were seeded into six-well dishes and grown for 9 days, with medium exchange on days 1, 3, and 6. Upon confluence (day 6), the cells were chronically treated for 3 days with 0.1 M norepinephrine, 1 M dexamethasone or a combination of both. Control cells were left untreated. Dexamethasone was added in a single dose; freshly prepared norepinephrine was added every 12 h.
cDNA Cloning and Sequencing-The 3Ј-fragment of Cig30 cDNA was originally selected from a cDNA library, which was prepared from cold-stimulated brown fat and cloned into pBR322 (15,26). We subcloned the insert following PCR amplification with primer 1 (CTGGT-TCGGCCCAAACGACGAGCGTGAC) and primer 2 (CTCGCTCGC-CCAGCTAGAGTAAGTAGTT) into the pDIRECT plasmid using the PCR-Direct Cloning System (CLONTECH). The entire insert was then subcloned into both M13mp18 and M13mp19 and sequenced. The missing part of the Cig30 cDNA was recovered using the 5Ј-RACE method. Poly(A) ϩ RNA was first isolated using Dynabeads mRNA Direct kit (Dynal) from one interscapular brown adipose tissue pad (approximately 50 mg of tissue) of a 5-week-old C57Bl/6 male mouse exposed to 4°C for 3 days. 2 g of poly(A) ϩ RNA were used for cDNA synthesis and amplification with 5Ј-AmpliFinder RACE kit (CLONTECH). First strand cDNA synthesis was primed with a Cig30-specific primer 3 (GCTTGGGTCTCTAACAGCAATAAAC), and the cDNA product was amplified with the anchor primer (CLONTECH) and a Cig30-specific nested primer 4 (CTCGCTCGCCCAACATCAAAGTACCAAACCAC-ACACC). The PCR was performed for 35 cycles (94°C, 45 s; 60°C, 45 s; 72°C, 2 min) with AmpliTaq DNA polymerase (Perkin Elmer). The 1.2-kb product was subcloned into pDIRECT and sequenced. To verify the sequence of the 5Ј-untranslated region, we repeated the entire RACE procedure with Pfu DNA polymerase (Stratagene) essentially as before, except for using nested primer 5 (GAACAACGATGAGCAACA-GATAGAC), giving rise to a 230-bp product. The sequence and length of the 5Ј-untranslated region was identical. Moreover, the sequence of the protein-coding region was verified following amplification for 25 cycles with Pfu DNA polymerase with primer 6 (GGTCGACAGCAAAATC-GAAATGGACAC) and primer 7 (CTCTAGACCTAATCATTCTCCCAT-ACCC). DNA sequencing was routinely performed on recombinant M13 phage single-stranded templates using the dideoxynucleotide chain termination method with Sequenase Version 2.0 (U. S. Biochemical Corp.). Reactions were labeled with [ 35 S]dATP (Amersham) and electrophoresed on 5% LongRanger gels (AT Biochem). Dried gels were visualized on a PhosphorImager, and sequence readings were aided with the Image Quant program, version 3.3. University of Wisconsin Genetics Computer Group package (27) was used for sequence data analyses and presentations.
In Vitro Transcription, Translation, and Glycosylation-The Cig30 putative open reading frame (173-985) was subcloned by PCR (using primers 6 and 7) into the pcDNA I vector (Invitrogen). The construct was linearized with HpaI and transcribed with T7 RNA polymerase using mCAP mRNA capping kit (Stratagene). Nuclease-treated rabbit reticulocyte lysate (Promega) was then programmed with approximately 1 g of resulting mRNA in the presence of Redivue L-[ 35 S]methionine (Amersham). Translations were carried out in standard 25-l reactions with various amounts (0 -2.4 l) of canine pancreatic microsomal membranes (Promega). Aliquots were solubilized in SDS sample buffer (final concentrations, 1.5% SDS, 15% glycerol, 100 mM TRIS-HCl, pH 6.8, 1% ␤-mercaptoethanol) for 30 min at 37°C and separated on 12% SDS-PAGE along with the Rainbow 14 C-methylated protein molecular weight markers (Amersham), followed by autoradiography on a PhosphorImager.
Immunoblotting and Polyclonal Antibodies-Plasma membranes from interscapular brown fat were prepared as in Ref. 28. Protein was measured with the BCA protein assay (Pierce). For Western blotting, 20 g of plasma membrane protein was solubilized in sample buffer (final concentrations, 1.5% SDS, 2.3 M urea, 15 mM Tris-HCl, and 100 mM dithiothreitol, pH 6.8) at 37°C for 30 min, subjected to Tricine-SDS-PAGE (12%) (29), and electroblotted onto nitrocellulose membranes. The membranes were blocked with 1% bovine serum albumin in phosphate-buffered saline and incubated with anti-CIG30 antiserum (1:250) and subsequently with 125 I-labeled anti-rabbit antibody (1:500; 0.2 Ci/ ml). Radioactivity was detected on a PhosphorImager. Rabbit anti-CIG30 antiserum was generated through a commercial service (Neosystem SA, Strasbourg, France) by immunizing rabbits with a peptide corresponding to the C terminus (RPKGKVASKSQ) conjugated to keyhole-limpet hemocyanin. Immunoreactivity was assessed with enzymelinked immunosorbent assay test and by immunoprecipitation of the radiolabeled in vitro translated CIG30 protein.
pcDNA-cig30GFP Construction, Pronuclear Microinjection, and Microscopy-The gpf gene was cut out from the pGFP-N2 plasmid (CLON-TECH) with HindIII and BclI and ligated in HindIII-BamHI sites of the pcDNA I vector, creating pcDNA-GFP. The Cig30 cDNA was PCR amplified with primer 6 and primer 8 (CGGATCCGTTGGCTCTTG-GATGCAACT), which replaces the Cig30 stop codon with a BamHI site. The PCR product was subsequently digested with SalI and BamHI and ligated in XhoI-BamHI sites of the pcDNA-GFP construct resulting in a fusion gene Cig30-gfp, driven by an enhancer/promoter derived from human cytomegalovirus. The CIG30 and GFP polypeptides are connected through a short spacer of 8 amino acids (RIHRPVAT). The expected fusion gene product was verified by immunoblotting with anti-rGFP antiserum (CLONTECH) following in vitro translation of the construct. The pcDNA-cig30GFP construct (20 ng/l) was microinjected into pronuclei of fertilized mouse eggs collected from 4-week-old C57BL/ 6J ϫ CBA/J females (Bommice). Embryos were incubated overnight in M16 medium (Sigma) at 37°C and 5% CO 2 . After the first cleavage, further cell division was inhibited by aphidicolin (4 g/ml), and embryos were cultured until next day, totaling 40 -48 h. Thereafter, the living embryos were mounted in phosphate-buffered saline into a chamber made of a coverslip and object glass spaced with a 200-m-thick tape and subjected to microscopy. In the epifluorescent mode, GFP fluorescence was observed with a Polyvar (Reichert-Jung) microscope using a filter set for excitation and emission of fluorescein. Confocal laserscanning microscopy was performed with a Leica confocal microscope (Leitz Wetzlar). The sample was illuminated with an argon/krypton ion laser with the standard dichroic mirror and a band pass emission filter (BP510 -525 nm).
Establishment of Stable Cell Lines Expressing the Cig30 Gene-A dicistronic expression vector pCI-IRESneo was constructed as follows: First, the internal ribosomal entry site (IRES) from the pCITE-4a vector (Novagen) was amplified by PCR with primer 9 (CGCGGCCGC-GAATTAATTCCGGTTATTTTCCACCA) and primer 10 (GCTCGAAT-TCGGATCCGATATCAGCCA), digested with NotI and BamHI, and ligated into the NotI and BamHI sites of the pCI-neo vector (Promega), replacing the neo gene and generating pCI-IRES. Second, the neo gene with a synthetic poly(A) signal was PCR amplified from the pCI-neo vector with primer 11 (CCATATGATTGAACAAGATGGATTGCA) and primer 12 (CGGATCCTTATCGCTATCGATTCACA), digested with NdeI and BamHI, and ligated in the NdeI and BamHI sites in the pCITE 4a vector in-frame with the major start codon of the IRES signal, creating pCITE-neo. Finally, the IRESneo cassette was completed in the pCI-IRES vector by subcloning the ApaI-BamHI fragment from pCITE-neo into the ApaI and BamHI sites of pCI-IRES, which resulted in the pCI-IRESneo vector. The Cig30 open reading frame was then amplified with primers 6 and 7, digested with SalI and XbaI, and ligated as the first cistron into the XhoI and XbaI sites in the pCI-IRESneo vector. The final construct, pCI-Cig30:IRESneo, was linearized with AhdI before transfections. The HIB-1B cells were grown in 35-mm wells to 50 -75% confluence. Thereafter, 2.5 g of pCI-cig30: IRESneo was transfected per well using the cationic lipid-mediated method with DOTAP (Boehringer Mannheim). For negative controls, the cells were transfected with the pCI-IRESneo vector only. 2 days after transfection, the cells were replated onto 10-cm Petri dishes with the selection medium containing 500 g of G418/ml (Life Technologies, Inc.). After 1-2 weeks, resistant colonies were isolated, expanded, and checked for Cig30 mRNA. Six clones with highest levels of the Cig30 mRNA and three clones of the negative controls were used in further experiments.
mRNA Expression Analysis and cDNA Probes-Total RNA was prepared from tissues and cultured cells using the Ultraspec RNA isolation protocol (Biotecx Laboratories). RNA aliquots (5-20 g) were denatured, separated on 1.2% formaldehyde gels, and blotted onto Hybond N membranes (Amersham). Prehybridization and hybridization were carried out in 50% formamide, 5 ϫ SSC, 5 ϫ Denhardt's solution, 50 mM sodium phosphate, pH 6.5, 0.5% SDS, and 100 g/ml degraded herring sperm DNA at 45°C overnight. Membranes were washed twice in 2 ϫ SSC, 0.1% SDS at room temperature for 15 min and once in 0.1 ϫ SSC, 0.1% SDS at 50°C for 15-30 min. Following exposure to DuPont Cronex x-ray films or PhosphorImager screens, membranes were stripped in 0.1% SDS for 15 min at 95°C and reprobed. A 980-bp MaeIII fragment from 5Ј-end of the Cig30 cDNA was used as a cig30 hybridization probe.
The UCP and ␤-actin probes were as described previously (30); the LPL probe was as described in Ref. 17. The GLUT4 was obtained from Dr. M. Mueckler, and cDNA was excised with EcoRI. The probes were labeled using random-primed DNA labeling kit (Boehringer Mannheim) with [ 32 P]dCTP.
Phospholipid Synthesis and Glucose Uptake Measurements-Phospholipid synthesis was measured in stably transfected HIB-1B cells as the incorporation of [ 32 P]P i into the phospholipid fraction. Preconfluent cells cultured in 10-cm Petri dishes were incubated in 5 ml of Krebs-Ringer bicarbonate buffer with 2% fatty acid-free bovine serum albumin in the presence of 100 Ci of [ 32 P]P i at 37°C and 8% CO 2 for 1 h. The buffer was then quickly removed, the cells were scraped into 3 ml of phosphate-buffered saline, and the phospholipids were extracted as described earlier (31). Final samples were dissolved in 20 l of chloroform, 6 l of which were applied as 10-mm streaks on 10 ϫ 10 cm HPTLC Silica gel 60 F 254 plates (Merck). The plates were developed according to Ref. 32 in methyl acetate-2-propanol-chloroform-methanol-0.25% KCl (25:25:25:10:9, v/v), dried, and exposed to a PhosphorImager screen. For glucose uptake measurements, primary cultures of brown adipocytes or cultures of HIB-1B cell lines were grown on 35-mm dishes. Where indicated, the cells were treated with 1 M dexamethasone or 0.1 M norepinephrine (or both) for 3 days and with 0.1 M insulin for 1 h before assay, and 2-deoxy-D-[ 3 H]glucose transport was measured according to Ref. 33. Harvested cells were thoroughly sonicated in 300 l of phosphate-buffered saline and counted on a Beckman liquid scintillation system 3801. The net uptake of 2-deoxy-D-glucose was expressed in pmol/min/mg protein Ϯ S.E.

RESULTS
Cloning and Analysis of a Full-length Cig30 cDNA-The CIN-2 clone was initially isolated from a brown adipose tissue cDNA library (26) by differential screening for cold-inducible genes (15). By DNA sequencing, we found a poly(A) ϩ tract in this cDNA, but no open reading frame, implying that the clone covered the 3Ј-untranslated region. To obtain the full-length cDNA, we used the 5Ј-RACE method as described under "Experimental Procedures." The resulting 1.9-kb cDNA, corresponding to a novel gene named Cig30, was sequenced (Fig.  1A). Sequence analysis revealed a single open reading frame of 813 bp, flanked by a 172-bp 5Ј-untranslated region and a ϳ900-bp 3Ј-untranslated region. Fickett's prediction of protein coding sequences (34) coincides very well with the putative open reading frame (Fig. 1B). The deduced polypeptide consists of 271 amino acid residues with a theoretical molecular mass of 32 kDa and a very basic pI of 10.6. Black boxes indicate amino acids conserved in more than 50% of the homologs. D, phylogenetic analysis of Cig30 family members. The UPGMA method was used to reconstruct the phylogenetic tree from a distance matrix created by the GCG program.
The inferred polypeptide was compared with those in protein banks using the BLAST network service (35). Nine proteins were found to have significant homology with CIG30, six of which are hypothetical gene products from Caenorhabditis elegans and the other three are yeast proteins. Yeast mutant strains exist for all three of these genes (36), but complex phenotypes make determination of their primary function very difficult. Among the reported defects are modified phospholipid composition (37), altered expression of some glucose-regulated genes, such as the plasma membrane ATPase gene (38), bud localization defect (39), resistance to inhibitors of sterol synthesis, particularly to fenpropimorph (40), and a large reduction in 1,3-␤-glucan synthase activity (41). Some physical properties of these proteins are summarized in Table I. This group emerges as a novel protein family (36) with remarkable similarities from yeast to mice (Fig. 1, C and D).
The presence of hydrophobic stretches indicated that CIG30 might be a membrane protein. As shown in Fig. 2A, a hydropathy plot based on the GES algorithm (42) predicts five transbilayer helices in the CIG30 protein. In good agreement with the GES algorithm, the TMAP program (43) also predicts five transmembrane regions (Figs. 1A and 2A). An arbitrary stretch of 35 amino acid residues in the middle part of the CIG30 polypeptide shares 88% identity with the C40H1.4 protein and 31% identity throughout the protein family (Fig. 1A, underlined with a dotted line, and evident also in Fig. 1C). This region of highest homology lies in a putative loop and seems to be rather amphiphilic in an ␣-helical projection with conserved positions on the opposite sides of the helix (Fig. 2B).
In positions 6 -9 of the CIG30 protein, we found a consensus sequence for an N-glycosylation site (N-F-S-R). Examination of the sequence for an N-terminal cleavable presequence gave negative results. A perfect consensus for a leucine zipper was found in the fourth transmembrane region. The protein also contains potential phosphorylation sites for protein kinase C lyzed by SDS-PAGE. All clones gave rise to a 28-kDa protein product, and the shortest clone was translated with highest efficiency (data not shown). For further study, we therefore chose the 162-1056 cDNA fragment covering the putative open reading frame (nt 173-985). When translated in rabbit reticulocyte lysate with canine pancreatic microsomal membranes, the apparent size of the CIG30 polypeptide was shifted from 28 kDa to 30 kDa, which is proposed to be due to a glycosylation event (Fig. 3A), thus implying that CIG30 is a glycoprotein. We assume that the start codon is located at nt 173 of the Cig30 cDNA, both because it is the first ATG codon and because it is embedded in a minimal Kozak consensus sequence ((A/G)NNATGG) (44).
Intracellular Localization-To address the question of the intracellular localization of the CIG30 protein, we constructed a fusion gene, in which the green fluorescent protein (GFP) was attached to the C terminus of the CIG30. The Cig30-gfp gene was cloned downstream of the enhancer/promoter from human cytomegalovirus and assayed by transient expression in mouse preimplantation embryos as described under "Experimental Procedures." Despite the relatively weak fluorescence of the wild type GFP protein, this expression system allowed us to demonstrate that the CIG30 fusion protein is primarily located to the plasma membrane. Nonetheless, we cannot exclude that CIG30 alone is also localized to other membranes, such as those of the endoplasmic reticulum (Fig. 3B).
Polyclonal antibodies were raised against a synthetic peptide corresponding to the C-terminal part of the protein. In plasma  1A). In an ␣-helical projection, the distribution of hydrophobic residues (outline letters) and hydrophilic residues (bold letters) is clearly amphiphilic. The conserved amino acids are marked by circles. membrane fractions prepared from brown fat of warm-acclimated, 1-week, and 1-month cold-acclimated mice, the antibody detected a faint band of 30 kDa in stimulated tissues. This band was of similar size to the in vitro translated and glycosylated CIG30 (Fig. 3C). No signal was found in total homogenates (data not shown). As an additional indication that the protein is a membrane protein, we noticed that CIG30 tended to aggregate, especially when boiled, and had to be solubilized at low temperatures.

HIB-1B Cells Stably Transfected with the Cig30 Gene Have neither Altered Phospholipid Synthesis nor Abnormal Glucose
Uptake-We tested the HIB-1B immortalized cell line derived from brown adipose tissue (23) for the presence of the Cig30 mRNA and found no signal under various experimental conditions (not shown). We proposed that overexpression of Cig30 in these model cells could affect some cellular processes in which the corresponding yeast mutants were defective, e.g. phospholipid synthesis (37) or glucose uptake (38). To study these parameters in vitro, we established several lines of HIB-1B cells stably transfected with the Cig30 gene. These lines expressed high levels of the Cig30 mRNA, whereas control HIB-1B cells showed no detectable Cig30 expression (Fig. 4A).
Phospholipid synthesis was measured as incorporation of [ 32 P]P i into phospholipids for 1 h. The phospholipid fraction was then isolated and separated on TLC. However, as shown in Fig. 4B, the pattern of radioactivity incorporation into phosphatidylcholine, phosphatidylserine, phosphatidylethanolamine, and phosphatidylinositol was unchanged between Cig30-transfected cell lines and control lines.
HIB-1B cells are poorly responsive to insulin with respect to glucose uptake. To test whether Cig30 overexpression could improve the poor glucose uptake in these cells, we employed one stable cell line and measured glucose uptake with 2-deoxyglucose. Data presented in Table II indicate that expression of the brane was sequentially incubated with anti-CIG30 antiserum and 125 Ilabeled anti-rabbit antibody, and the immunocomplexes were visualized on a PhosphorImager.

FIG. 3. CIG30 appears as a plasma membrane glycoprotein.
A, in vitro translation and glycosylation of CIG30. The Cig30 mRNA was prepared by in vitro transcription, and equal amounts (1 g) were translated in rabbit reticulocyte lysate in the presence of indicated amounts of canine pancreatic microsomal membranes (Promega). Aliquots (3 l) were then analyzed by SDS-PAGE and autoradiography of the 35 S-labeled proteins. The indicated molecular weights were those of Rainbow 14 C-labeled molecular weight markers (Amersham). B, transient expression of the GFP-tagged Cig30 gene in preimplantation mouse embryos. Fertilized mouse eggs were pronuclearly microinjected with pcDNA-cig30GFP or pcDNA-GFP and photographed 48 h later in differential interference contrast microscopy (Noma) and fluorescence microscopy (Fluo). In another experiment, similarly treated eggs were visualized by confocal laser scanning microscopy (Conf). C, immunoblot of plasma membrane fractions from brown adipose tissue. Protein aliquots of plasma membrane fractions prepared from brown adipose tissue of thermoneutral controls (0) and mice exposed to cold for 1 week (1w) and 1 month (1m) were separated on SDS-PAGE together with unglycosylated (-mm) and glycosylated (ϩmm) form of the in vitro translated CIG30 protein. Following electroblotting, the nitrocellulose mem- Cig30 gene did not alter glucose transport into HIB-1B cells in the basal state, nor did it significantly enhance their responsiveness to insulin. Thus, the CIG30 protein is unlikely to be directly involved in the transport of glucose into mammalian cells.
Increase of Cig30 Expression Coincides with the Recruitment of Brown Fat-We examined expression of the Cig30 gene in 12 tissues of mice kept either at 28°C or at 4°C for 3 days (Fig.  5A). By Northern blotting of total RNA from warm-acclimated animals, we detected Cig30 mRNA in brown fat and liver, whereas no detectable signal was seen in white fat, heart, lung, kidney, testis, muscle, spleen, brain, thymus, and intestine.
Following overly prolonged autoradiography, a weak signal could be observed in kidney, white fat, heart, and skin (not shown). The Cig30 mRNA was consistently detected as two bands. The major hybridization band was about 1.9 -2 kb; the size of the minor species was approximately 1.6 -1.8 kb. The short mRNA form could be accounted for by an alternative use of the polyadenylation site at nt 1472 (see Fig. 1A). There is, however, no poly(A) consensus site in the 3Ј-end of the long Cig30 mRNA form.
Of particular interest was the fact that the Cig30 mRNA level in brown fat of cold-exposed animals was 200-fold elevated, whereas the liver Cig30 mRNA level remained unchanged. Therefore, we measured Cig30 mRNA in brown fat of mice exposed to cold for various periods of time. As shown in Fig. 5B, only a very small increase was seen after 4 h of cold exposure, and the signal peaked after 3 days. It then gradually declined to 50 -60% of the maximum after 1 month, being, however, still about 100-fold increased above controls.
Brown adipose tissue is recruited not only during cold exposure but also during early postnatal development to meet increased heat production requirements. We therefore investigated whether the Cig30 gene was activated at this stage of development in mice. As shown in Fig. 5C, the Cig30 mRNA RNA samples from brown fat of cold-exposed mice were subjected to a Northern blot analysis and quantitated on a PhosphorImager; the maximum value was set to 100. The results are means Ϯ S.E. from two mice per point. The samples were taken at 0 (28°C thermoneutral controls), 4 h, 1 day, 3 days, 1 week, and 1 month of the cold exposure. C, developmental Northern blot analysis of Cig30 expression in brown fat of mice. Total RNA was prepared from pups approximately 3 days before term (Ϸ17 days p.c.), 1 day before term (Ϸ19 days p.c.), and 3 h to 20 days postpartum. 20 g of RNA of the perinatal samples were electrophoresed and blotted together with 10 g of brown fat RNA from thermoneutral (28°C) and cold-stimulated (4°C) adult animals. To check the loading, the membrane was reprobed with the ␤-actin probe. Similar results were obtained in a parallel experiment. D, cafeteria diet induction of Cig30 expression. Total RNA isolated from brown adipose tissue of mice kept at various nutrient conditions were analyzed by Northern blotting using the cig30 cDNA probe. 28°C, thermoneutral controls; 4°C, cold stimulation for 3 days; CAFE, cafeteria diet (7 days); FAST, 24-h fasting; FASTϩCOLD, 24-h fasting followed by a 4-h cold exposure; FASTϩREF, 24-h fasting followed by a 4-h refeeding. Membranes were stripped and rehybridized with the ␤-actin probe.

TABLE II
Glucose uptake in Cig30-transfected IIIB-1B cells 2-Deoxyglucose uptake was measured in HIB-1B cells stably transfected with either pCI-Cig30:IRESneo (HIB-CIG30) or pCI-IRESneo (HIB control) following a treatment with 0.1 M insulin for 1 h (ϩinsulin) or no treatment (Ϫinsulin). Data are given in pmol/min/mg protein, and expressed as means Ϯ S.E. from triplicate measurements. *, different from control at p Ͻ 0.05.
ϪInsulin ϩInsulin HIB control 0.91 Ϯ 0.022 1.03 Ϯ 0.017* HIB-CIG30 0.87 Ϯ 0.024 0.95 Ϯ 0.020 level was transiently elevated in brown fat in the perinatal period, reaching a maximum level immediately after birth and declining to basal levels before weaning. However, the stimulation was of much lower amplitude than that observed in brown fat of cold-exposed mature mice. Cafeteria diet has also been recognized as a factor inducing brown fat recruitment. We found a significant effect of cafeteria diet on Cig30 expression in mice kept at 28°C (Fig. 5D), although only very slightly when compared with a cold stimulus. Fasting did not have any stimulatory effect, neither did it prevent the early effect of cold. Refeeding after 24 h of fasting did not appreciably increase the Cig30 mRNA level, suggesting that neither glucose nor insulin is capable of eliciting Cig30 mRNA expression.
Norepinephrine Is Sufficient for Induction of Cig30 mRNA in Vivo-Recruitment of brown adipose tissue is brought about by norepinephrine released from sympathetic nerves. To investigate whether norepinephrine could mimic the effect of cold in terms of Cig30 expression, we treated mice kept at 28°C with norepinephrine for 3 days. The hormone was administered continuously using microosmotic pumps placed subcutaneously in the interscapular region. The mice received either vehicle alone or norepinephrine for 72 h. As shown in Fig. 6, norepinephrine induced the Cig30 mRNA level by about two orders of magnitude, suggesting that the cold effect on Cig30 expression is mainly due to norepinephrine. The marginal induction of Cig30 expression in the saline controls was probably due to catecholamine release following the mechanical stress of the minipumps.
Synergistic Action of Dexamethasone and Norepinephrine Is Required for Full Expression of Cig30 in Primary Cultures of Brown Adipocytes-Brown adipocyte precursors isolated from the stromal vascular fraction of brown adipose tissue can differentiate in primary culture in vitro into morphologically and physiologically adequate mature cells (45). To investigate whether brown adipocytes were specifically the site of Cig30 expression and whether the induction of Cig30 expression in vivo was a primary or secondary effect of norepinephrine, we analyzed Cig30 mRNA levels in primary cultures of brown fat cells by Northern blotting. The results are presented in Fig. 7A.
The Cig30 mRNA was virtually absent in brown adipocytes spontaneously differentiated in standard culture medium. Chronic treatment with 0.1 M norepinephrine for 3 days from day 6 resulted in appreciable expression, the level of which was, however, much lower than that in cold-stimulated brown adipose tissue. Because glucocorticoids have been implicated in the promotion of brown adipocyte differentiation (33), we added 1 M dexamethasone to the cells and found a comparably small increase in Cig30 expression. Unexpectedly, when both hormones were present together for 3 days, the Cig30 expression was strongly potentiated, being 10-fold greater than either of the stimuli alone. Norepinephrine stimulation for 1 h after chronic treatment with dexamethasone had no additional effect, which was in keeping with the very slow induction of Cig30 transcription in vivo. Stimulation with insulin (0.1 M) for 1 h did not further increase the highly stimulated Cig30 mRNA level brought about by chronic norepinephrine plus dexamethasone.
Comparison of Cig30 Expression Pattern with Other Differentiation Markers of Cultured Brown Adipocytes-To further correlate Cig30 expression and brown adipocyte differentiation, we chose to measure glucose uptake capacity in cultured cells expressing Cig30 mRNA and to determine mRNA levels of genes for UCP, GLUT4, and LPL in parallel.
Brown fat precursor cells were cultured for 9 days. For the last 3 days, they were treated with norepinephrine, dexamethasone, or both as in the previous experiment. Glucose uptake was measured as 2-deoxyglucose influx either directly or after a 1-h pretreatment with 0.1 M insulin. As shown in Table III, under all conditions the cultured cells displayed good responsiveness to insulin, which activated glucose uptake 2.5-4.5-fold. The absolute value of insulin-stimulated glucose uptake varied depending on the prior treatment, being significantly higher in the cells chronically stimulated with both norepinephrine and dexamethasone. Thus, these cells would seem to most closely resemble mature brown adipocytes, the glucose uptake capacity of which is enhanced during the recruitment process (9,46).
We measured the mRNA levels for three differentiationrelated proteins (UCP, GLUT4, and LPL), in addition to the Cig30 mRNA level, in cells cultured and treated as in the glucose uptake experiment, except that no acute insulin stimulus was given (Fig. 7B). Consistently with the above results (Fig. 7A), the Cig30 gene required the simultaneous presence of norepinephrine and dexamethasone for full expression; either of the hormones alone was less than 20% as potent. The expression of the Ucp gene was also maximally stimulated in the norepinephrine plus dexamethasone-treated cells; norepinephrine in itself was about 2-fold less potent, and dexamethasone had virtually no effect. The genes for GLUT4 and LPL, both primarily involved in anabolic processes, exhibited a different expression profile, reaching a maximum level in the dexamethasone-treated cells, and a similar level in the cells treated with norepinephrine plus dexamethasone. Norepinephrine in itself was less potent. Taken together, these results indicate that the cells strongly expressing the Cig30 gene in vitro have the traits of mature brown adipocytes with a high thermogenic capacity. DISCUSSION In this report, we have described cloning and characterization of a cDNA from mouse brown adipose tissue that corresponds to a novel mammalian gene, which we have termed Cig30. The gene codes for a 30-kDa glycoprotein of yet unknown function. However, our data strongly suggest that this function is associated with recruitment and thermogenesis in brown adipose tissue.
All nine proteins homologous to CIG30 that are currently present in the protein banks are fairly well conserved both in being hydrophobic and basic, in the amino acid sequence of the central core (Table I and Fig. 1C), and probably in the overall topology. Similarly to the CIG30 protein, an N-linked glycosylation site has been reported in the N-terminal part of the Fen1p protein (41), and leucine zipper structures found in Sur4p and Fen1p proteins have been suggested to trigger dimerization in the plasma membrane (38,41). The number of predicted transmembrane stretches reported for different proteins varies from four to six, which is probably the result of some ambiguity in the central part of the consensus. The core region is the most highly conserved amino acid sequence (Fig.  1C) implicating this segment as being crucial for protein function. We have noticed that the conserved region displays amphiphilic properties (Fig. 2B), which might explain the uncertainty in estimating transmembrane topology. It is possible that this structure is associated with or inserted into the membrane producing an active site. In accordance with the structure published by El-Sherbeini and Clemas (41) for the yeast Gns1p protein (which is identical with Fen1p), we suggest that the common topology of this protein family is a five-membranespanning structure (Fig. 8).
Using the Northern blotting approach, we have found that the most prominent feature of Cig30 gene regulation is its high inducibility in brown adipose tissue by cold (Fig. 5A). The time course of the cold-induced expression is delayed compared with genes directly involved in thermogenesis, such as Ucp and Lpl (14 -17), and is more reminiscent of induction of the genes for GPDH (18) or GLUT4 (19). As the Cig30 mRNA peaks after 3 days in the cold (Fig. 5B), it is plausible that CIG30 is primarily related to hyperplastic changes in the tissue during cold acclima-  tion, rather than being involved in immediate heat production. This would be in line with the slow proliferation onset observed in the stimulated brown adipose tissue (20,21).
In the present paper, we show that diet-induced thermogenesis is also associated with stimulation of Cig30 gene expression (Fig. 5D). There are two possible effects of a cafeteria diet on brown adipose tissue, either a central effect mediated via the hypothalamus or a direct effect of high blood glucose level on brown adipocytes. We were not able to distinguish these two pathways in our experiments. However, since refeeding after 24-h fasting did not lead to a detectable stimulation of Cig30 mRNA, we speculate that neither glucose nor insulin is sufficient to induce Cig30 gene expression. Similarly, fasting could not prevent the early onset of Cig30 mRNA induction on cold exposure (Fig. 5D). The Cig30 mRNA level in liver was not significantly influenced by any treatment in this experiment (not shown). It may be concluded that the diet-induced increase of Cig30 gene expression is most likely due to norepinephrine release caused by stimulation of the hypothalamus. This is also in agreement with our results demonstrating that chronically administered norepinephrine is sufficient for marked induction of Cig30 expression in vivo (Fig. 6).
Following chronic treatment with norepinephrine and dexamethasone, we were able to detect high Cig30 mRNA levels in brown adipocytes differentiated in vitro (Fig. 7A). The requirement for a synergistic effect of these two agents for high expression in vitro was unexpected and did not fully conform with the in vivo results, which indicated that norepinephrine alone is a sufficient stimulus to greatly induce Cig30 gene expression. This discrepancy could perhaps be explained by endogenous glucocorticoids, the peak plasma levels of which are only about 10-fold lower (10 Ϫ7 M) than those used in vitro (10 Ϫ6 M). Thus, we could hypothesize that glucocorticoids have only a permissive role in Cig30 gene expression and that the actual gene regulation is physiologically modulated by norepinephrine. Also, the requirement of both glucocorticoid and catecholamine stimuli for full Cig30 gene expression possibly means that the gene may be involved in an anabolic function, which is recruited during tissue hyperplasia.
Although the nature of the Cig30 function is at present unknown, significant homology of CIG30 to several yeast proteins indicates that this function has been well conserved during evolution. The homologous yeast genes SUR4 (also called APA1) and FEN1 (also called GNS1) have been implicated in a bewildering array of biochemical functions, including phospholipid metabolism (37), sterol synthesis (47), budding (39), activation of several glucose-regulated genes (38), glucose uptake (38), and glucan synthesis (41), which makes prediction of their primary functions very precarious. Yet, the data available seem to implicate this gene family in remodeling of the plasma membrane and actin cytoskeleton in response to growth signals, perhaps by modulating interaction between membrane phospholipids and cytoskeleton.
In our hands, however, overexpression of Cig30 in HIB-1B cells did not elicit any significant changes in phospholipid synthesis (Fig. 4) nor did we see any marked abnormalities in cell size, shape, or growth. Since Cig30 mRNA is not normally detectable in HIB-1B cells, it is possible that ectopic expression of Cig30 is redundant to an already existing mechanism, or Cig30 needs other factors to provoke any measurable difference in our assay. Nevertheless, in the cold-exposed brown adipose tissue, there is a remarkable increase in turnover of specific phospholipids, particularly that of phosphatidylinositol (11), and this increase is attributable to the chronic stimulation of ␣ 1 -adrenergic receptors by norepinephrine (13,31). Accordingly, it seems of importance to investigate whether the in-crease in Cig30 expression during tissue hypertrophy could be involved in the elevated phospholipid turnover.