Structural organization and chromosomal assignment of the human obese gene.

The obese (ob) gene has been identified through a positional cloning approach; the mutation of this gene causes marked hereditary obesity and diabetes mellitus in mice. We report here the isolation and characterization of the human ob gene. Southern blot analysis demonstrated a single copy of the ob gene in the human genome. The human ob gene spanned approximately 20 kilobases (kb) and contained three exons separated by two introns. The first intron, approximately 10.6 kb in size, occurred in the 5'-untranslated region, 29 base pair (bp) upstream of the ATG start codon. The second intron of 2.3 kb in size was located at glutamine +49. By rapid amplification of 5'-cDNA ends, the transcription initiation sites were mapped 54-57 bp upstream of the ATG start codon. The 172-bp 5'-flanking region of the human ob gene contained a TATA box-like sequence and several cis-acting regulatory elements (three copies of GC boxes, an AP-2-binding site, and a CCAAT/enhancer-binding protein-binding site). By the fluorescence in situ hybridization technique, the ob gene was assigned to human chromosome 7q31.3. This study should establish the genetic basis for ob gene research in humans, thereby leading to the better understanding of the molecular mechanisms underlying the ob gene.

The obese (ob) gene has been identified through a positional cloning approach; the mutation of this gene causes marked hereditary obesity and diabetes mellitus in mice. We report here the isolation and characterization of the human ob gene. Southern blot analysis demonstrated a single copy of the ob gene in the human genome. The human ob gene spanned ϳ20 kilobases (kb) and contained three exons separated by two introns. The first intron, ϳ10.6 kb in size, occurred in the 5untranslated region, 29 base pair (bp) upstream of the ATG start codon. The second intron of 2.3 kb in size was located at glutamine ؉49. By rapid amplification of 5-cDNA ends, the transcription initiation sites were mapped 54ϳ57 bp upstream of the ATG start codon. The 172-bp 5-flanking region of the human ob gene contained a TATA box-like sequence and several cis-acting regulatory elements (three copies of GC boxes, an AP-2binding site, and a CCAAT/enhancer-binding proteinbinding site). By the fluorescence in situ hybridization technique, the ob gene was assigned to human chromosome 7q31.3. This study should establish the genetic basis for ob gene research in humans, thereby leading to the better understanding of the molecular mechanisms underlying the ob gene.
The obese (ob) gene, an autosomal recessive mutation on mouse chromosome 6, arose spontaneously in the mouse colony at the Jackson Laboratory (1). Mice homozygous for the ob mutation, known as ob/ob mice, develop severe hereditary obesity and non-insulin-dependent diabetes mellitus. The molecular identification of the ob gene by Friedman and co-workers (2) has provided new insight into the pathogenesis of obesity and obesity-linked diabetes. The authors identified the mouse ob gene through a positional cloning strategy and determined the structure of the mouse ob protein and also its human homologue (2). The ob protein, a 166/167-amino acid polypeptide with a putative signal sequence, is highly conserved in structure among species, and expression of the ob gene is abundant in and specific to adipose tissue in mice (2). Recently, we and others have also isolated rat and human ob cDNAs (3)(4)(5)(6)(7) and demonstrated that the ob gene is expressed in adipose tissue in a region-specific fashion in rats and humans (3,4,8).
Expression of the ob gene is markedly augmented in adipose tissue in several rodent models of genetic obesity (C57BL/6J ob/ob mice (2) and Zucker fatty (fa/fa) (4,6) and Wistar fatty (fa/fa) rats (8)) and in rodent models of acquired obesity obtained by pure overfeeding of normal rats 1 or by ventromedial lesion to rat hypothalamus (7). The augmentation of ob gene expression in adipose tissue is also region-specific (4,7,8). Furthermore, ob gene expression is also increased in human obesity in proportion to disease severity (5). These observations suggest the pathophysiologic roles of the ob gene in the development of obesity. Indeed, nonsense mutation of the ob gene has been proven to be the obesity-causing mutation in C57BL/6J ob/ob mice (2). On the other hand, no such mutation of the ob gene has been found in human obesity (5).
To understand the physiologic and pathophysiologic roles of the ob gene in humans, it is important to elucidate the structural organization of the human ob gene. Furthermore, molecular characterization of the ob gene from any species has not so far been reported. We report here the isolation and structural organization of the human ob gene. Using the fluorescence in situ hybridization technique, we also determined the chromosomal assignment of the human ob gene.

EXPERIMENTAL PROCEDURES
Genomic Southern Blot Analysis-Human genomic DNA extracted from blood leukocytes was digested with restriction endonucleases SacI, EcoRI, KpnI, SphI, and NcoI; electrophoresed on a 0.7% agarose gel (5 g/lane); and transferred onto a Biodyne A nylon membrane (Pall, Glen Cove, NY) (9). The membrane was prehybridized at 42°C in a solution containing 50 mM sodium phosphate buffer (pH 7.0), 5 ϫ SSC (1 ϫ SSC is 0.16 M NaCl and 0.016 M sodium citrate (pH 7.0)), 50% formamide, 5 ϫ Denhardt's solution, 0.1% SDS, and 200 g/ml salmon testis DNA. Hybridization was performed in the same solution plus the 32 P-labeled human ob cDNA fragment (3) as a probe. The membrane was washed three times at 55°C in 0.1 ϫ SSC and 0.1% SDS. The blot was used to expose an x-ray film with an intensifying screen for 1 week.
Genomic Library Screening-A human genomic DNA library derived from leukocyte DNA in EMBL3 (CLONTECH, Mountain View, CA) was screened with the 32 P-labeled human ob cDNA probe (3). To obtain the 5Ј-flanking region of the human ob gene, a second human genomic DNA library derived from leukocyte DNA in EMBL3 (CLONTECH) was screened with the 32 P-labeled human ob genomic fragment (fragment 1) (see Fig. 2). Prehybridization and hybridization were carried out as described (9,10). The membranes were washed in 2 ϫ SSC and * This work was supported in part by research grants from the Japanese Ministry of Education, Science, and Culture, the Japanese Ministry of Health and Welfare, the Yamanouchi Foundation for Research on Metabolic Disorders, the Salt Science Research Foundation, and the Japan Diabetes 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  0.1% SDS twice at 60°C and in 0.2 ϫ SSC and 0.1% SDS three times at 60°C. Appropriate restriction fragments were subcloned into the pBluescript vector (Stratagene Inc., La Jolla, CA) for further analysis.
Polymerase Chain Reaction-PCR 2 was used to obtain the genomic fragment that contains the first exon and the upstream half of the first intron of the human ob gene (fragment 1) (see Fig. 2). Using a Model 381A DNA synthesizer (Applied Biosystems Inc., Foster City, CA), two oligonucleotide primers (sense, 5Ј-TAGGAATCGCAGCGCCAACGGTT-3Ј; antisense, 5Ј-CTACTTGGGAGGCCAAGGTGGGAGGTTTGC-3Ј) were synthesized based upon the nucleotide sequences of human ob cDNA (3) and the first intron of the human ob gene, respectively (see Fig. 2). Using human genomic DNA as template, PCR was performed with a Takara Shuzo LA PCR kit. The reaction profile was as follows: denaturation at 98°C for 20 s and annealing and extension at 68°C for 3 min for 30 cycles. The amplified DNA fragment of 5.3 kb in size was subcloned into the pGEM-T vector (Promega, Madison, WI) for sequencing.
Tissue Preparation and RNA Extraction-Human adipose tissue was obtained at the time of operation from the subcutaneous abdominal fat pad of a 58-year-old female patient with gastric cancer. Tissues were frozen in liquid nitrogen and stored at Ϫ70°C until use. Total RNA extraction was carried out as described (3,4,8).
Reverse Transcription-PCR-RT-PCR was performed to determine the presence or absence of any introns in the 3Ј-untranslated region of the human ob gene. Approximately 10 g of total RNA from human adipose tissue was reverse-transcribed by random hexamer priming using Superscript Moloney murine leukemia virus reverse transcriptase (Life Technologies, Inc.). The single-stranded cDNA was subjected to PCR as described (11). The human ob cDNA-specific PCR primers were generated using a Model 381A DNA synthesizer. Amplified DNA fragments were subcloned into the pGEM-T vector for sequencing. The presence or absence of any introns in the 3Ј-untranslated region was determined by comparison of the nucleotide sequences of the cloned human ob genomic fragment with those of the RT-PCR products that cover the entire 3Ј-untranslated region of the human ob gene.
Rapid Amplification of 5Ј-cDNA Ends (5Ј-RACE)-The 5Ј-RACE experiment was performed essentially as described (4) using the 5Ј-Am-pliFINDER TM RACE kit (CLONTECH). Approximately 10 g of total RNA from human adipose tissue was reverse-transcribed by a human ob cDNA-specific antisense primer (5Ј-ATGGGGTGGAGCCCAGGAAT-3Ј). The single-stranded cDNA was ligated to the AmpliFINDER anchor and amplified by PCR using the AmpliFINDER anchor primer and a second upstream human ob cDNA-specific antisense primer (5Ј-TTG-GATGGGCACAGCTTG-3Ј). A single fragment of ϳ200 bp in size was obtained, which was subcloned into the pGEM-T vector for sequencing.
Rapid Amplification of 3Ј-cDNA Ends (3Ј-RACE)-The 3Ј-RACE experiment was carried out as described (9) to determine the 3Ј-end of the human ob gene. Approximately 10 g of total RNA from human adipose tissue was reverse-transcribed by adaptor oligo(dT) 15 priming (5Ј-GGCAGTCCGAATTCCTCGAGTTTTTTTTTTTTTTT-3Ј) using Superscript Moloney murine leukemia virus reverse transcriptase. After synthesis of the second strand cDNA by a 5Ј-gene-specific primer (5Ј-GGCCAGAAGAATTGAGATTC-3Ј), PCR was carried out using the primer and the adaptor oligonucleotide (without 13 dT nucleotides on the 3Ј-end) (7). An aliquot of the reaction was further subjected to PCR using a downstream 5Ј-gene-specific primer (5Ј-TAGGCTGAGGCAG-GAGAATC-3Ј) and the adaptor primer. The 3Ј-RACE product was analyzed by a 1.5% agarose gel, and amplified DNA was subcloned into the pGEM-T vector for sequencing.
DNA Sequencing-Nucleotide sequences were determined by the dideoxy chain termination method (12) using Sequenase version 2.0 (U. S. Biochemical Corp.) and a DyeDeoxy TM Terminator Cycle sequencing kit (Applied Biosystems Inc.). Sequence-specific primers were synthesized using a Model 381A DNA synthesizer. All DNA sequences were confirmed by reading both DNA strands.
Fluorescence In Situ Hybridization-Metaphase spreads were prepared from phytohemagglutinin-stimulated lymphocyte culture by a thymidine synchronization, 5-bromodeoxyuridine release technique for the delineation of G-bands. Before hybridization in situ, chromosomes were stained in Hoechst 33258 and irradiated with UV (13). The BamHI/SalI-digested fragments of the isolated genomic clone (OB1-8) (see Fig. 2) were labeled with biotin-16-dUTP (Boehringer Mannheim GmbH, Mannheim, Germany) by nick translation. Hybridization sig-nals were detected with fluorescein isothiocyanate-avidin (Boehringer Mannheim GmbH), and chromosomes were counterstained with propidium iodide (1 g/ml). The precise signal positions were determined by the delineation of G-band patterns as described (14,15). Microscopy was performed with a Nikon FXA fluorescent microscope. Propidium iodide-stained chromosomes and fluorescein isothiocyanate signals were visualized through a Nikon B-2A filter, and G-band patterns on the same metaphase chromosomes were delineated through a Nikon UV-2A filter.

RESULTS
Genomic Southern Blot Analysis-Southern blot analysis of human genomic DNA with the human ob cDNA probe identified a single hybridizing band upon digestion with restriction endonucleases EcoRI, KpnI, and SphI (3.7, 18, and 4.3 kb in size, respectively). On the other hand, digestion with SacI and NcoI gave two hybridizing bands of 7.2 and 3.8 kb in size and of 5.8 and 5.1 kb in size, respectively (Fig. 1).
Isolation and Characterization of the Human ob Genomic Fragments-To isolate the human ob gene, ϳ6 ϫ 10 5 recombinants from a human genomic DNA library in EMBL3 were screened with the 32 P-labeled human ob cDNA probe (3). A single positive clone (OB1-8) harbored an ϳ14-kb human ob genomic fragment, which contained the 5.3-kb downstream half of the first intron and the second and third exons of the human ob gene (Fig. 2). The 5.3-kb genomic fragment (fragment 1) was amplified by PCR and contained the first exon (ϳ29 bp) and the 5.3-kb upstream half of the first intron (Fig.  2). To obtain the 5Ј-flanking region of the human ob gene, ϳ5 ϫ 10 5 clones from a second human genomic DNA library in EMBL3 were screened with 32 P-labeled fragment 1. Six positive clones were identified and plaque-purified. DNA from one clone (OB3-1) harbored an ϳ16-kb genomic DNA fragment that contained the 5.0-kb 5Ј-flanking region of the human ob gene (Fig. 2).
Structural Organization of the Human ob Gene- Fig. 3 shows the nucleotide and deduced amino acid sequences of the human ob gene. The exon/intron borders were determined by comparison of the nucleotide sequences of the human ob gene with those of human ob cDNA (3). The human ob gene spanned ϳ20 kb and was organized into three exons separated by two introns. Splicing donor and acceptor consensus sequences (16) were located at the putative exon/intron borders. The first intron was ϳ10.6 kb in size and occurred in the 5Ј-untranslated region, 29 bp upstream of the ATG start codon. The second intron, ϳ2.3 kb in size, was located at glutamine ϩ49. The third exon contained the downstream coding region and the 3Ј-untranslated region of the human ob gene. Since complete nucleotide sequences of the 3Ј-untranslated region of the human ob cDNA have not yet been reported, nucleotide sequences of the third exon were determined by sequencing the 3Ј-RACE product and the corresponding genomic regions. Comparisons of the nucleotide sequences of the human ob genomic regions with those of the 3Ј-RACE/RT-PCR products that cover the entire 3Ј-untranslated region revealed the absence of any introns in the 3Ј-untranslated region of the human ob gene (data not shown).
Determination of the Transcription Initiation Sites of the Human ob Gene-To determine the transcription initiation sites of the human ob gene, the 5Ј-RACE experiment was carried out. To exclude the nucleotide misincorporation during the PCR amplification, a total of 10 clones were sequenced. Sequence analysis identified the transcription initiation sites 54ϳ57 bp upstream of the ATG start codon (G at position Ϫ57, three clones; T at position Ϫ56, one clone; A at position Ϫ55, one clone; G at position Ϫ54, five clones) (Fig. 3). The 5Ј-end of the cloned human ob cDNA (3) was located 44ϳ47 bp downstream of the transcription initiation sites (Fig. 3). The 5Ј-ends of mouse and rat ob cDNAs have been located 57 and 60 bp upstream of the ATG start codon, respectively (2,4,17). Although there is a high nucleotide sequence similarity in the 5Ј-untranslated region between mouse and rat ob cDNAs (93%), nucleotide sequences of the 5Ј-untranslated region of the human ob gene were less homologous to those of mouse and rat ob cDNAs (51 and 47%, respectively).
Analysis of the 5Ј-Flanking Region of the Human ob Gene-The 172-bp 5Ј-flanking region of the human ob gene sequenced in this study contained a TATA box-like sequence (TATAWAW, W ϭ A/T; positions Ϫ87 to Ϫ81) (16) 27ϳ30 bp upstream of the transcription initiation sites (Fig. 4). A computer search of the 5Ј-flanking region for cis-acting regulatory elements also revealed the presence of three copies of GC boxes (GGGCGG) (18) at positions Ϫ79 to Ϫ74, Ϫ155 to Ϫ150, and Ϫ160 to Ϫ155; a binding site for CCAAT/enhancer-binding protein (C/EBP) (TKNNGYAAK, K ϭ G/T, N ϭ A/C/G/T, and Y ϭ C/T) (19) at positions Ϫ111 to Ϫ103; an E box (CANNTG, N ϭ A/C/G/T) (20) at positions Ϫ114 to Ϫ109; and an AP-2-binding site (CCCAGGGC) (21) at positions Ϫ199 to Ϫ192.
Analysis of the 3Ј-Flanking Region of the Human ob Gene-Molecular cloning studies of mouse, rat, and human ob cDNAs (2-7) have revealed no polyadenylation sites for their mRNAs, and the 3Ј-end of ob cDNA from any species has never been elucidated. To determine the 3Ј-end of human ob cDNA, the 3Ј-RACE experiment was carried out. Using total RNA from human adipose tissue, a single band of ϳ130 bp in size was obtained. Sequence analysis of the 3Ј-RACE products revealed that the cytosine nucleotide at position ϩ4183 is followed by the poly(A) stretch (data not shown), suggesting that the cytosine nucleotide at position ϩ4183 is the 3Ј-end of human ob cDNA or the 3Ј-end of the third exon of the human ob gene (Fig.  3). The overall size of the three exons (4240 bp) and the potential poly(A) stretch (usually ϳ200 bp) is consistent with that of human ob mRNA (ϳ4.5 kb) as revealed by Northern blot analysis (3). No typical polyadenylation signal (AATAAA) (22) was found near the putative poly(A) addition site. Nucleotide sequences of the 3Ј-untranslated region of the human ob gene were ϳ50% homologous to those of mouse ob cDNA (2). In the 3Ј-flanking region of the human ob gene, there was a characteristic CT-rich sequence at positions ϩ4417 to ϩ4538 (Fig. 3).
Chromosomal Assignment of the Human ob Gene-The chromosomal localization of the human ob gene was determined by the fluorescence in situ hybridization technique (Fig. 5). A total of 50 metaphase cells were examined. Of these, nine cells (14%) exhibited twin-spot signals on both homologous 7q31.3 chromosomes, and the other 17 cells (34%) had twin-spot signals on one 7q31.3 chromosome and a single spot on another 7q31.3 chromosome. Such specific accumulation of the signals could not be detected on any other chromosomes. These results indicate that the human ob gene is localized on chromosome 7q31.3.

DISCUSSION
In this study, we succeeded in the isolation and characterization of the human ob gene. Southern blot analysis of human genomic DNA identified a single hybridizing band upon digestion with EcoRI, KpnI, or SphI and two hybridizing bands upon digestion with SacI or NcoI (Fig. 1). These results are consistent with the restriction endonuclease map showing that a single site of SacI and NcoI is observed in the second intron of the genomic region that covers the ob cDNA sequence used as a probe, while EcoRI, KpnI, and SphI sites are not present (data not shown). These results indicate that the ob gene is present as a single-copy gene in the human genome. Using the mouse ob cDNA fragment as a probe, Zhang et al. (2) identified, by Southern blot analysis of human genomic DNA, a single hybridizing band of Ͼ11 kb in size upon digestion with EcoRI. Differences in the size of the hybridizing band observed may represent the restriction fragment length polymorphisms between the human genomic DNAs used.
This study demonstrates that the human ob gene is com- posed of three exons separated by two introns. The first intron occurred in the 5Ј-untranslated region, and the coding region was separated by a single intron at glutamine ϩ49. It has been demonstrated that in mice and humans, two different cDNAs encode the 166/167-amino acid ob proteins, which differ in the presence or absence of glutamine ϩ49 (2, 3). On close inspection, there are one donor and two acceptor sites (18) around the junction region (Fig. 3). Furthermore, we have also observed that there is an internal alternative splice site (23) at glutamine ϩ49 of the mouse ob gene. 3 These observations suggest that the two ob proteins in mice and humans are generated by the alternative mRNA splicing mechanism.
The 5Ј-flanking region of the human ob gene contained a TATA box-like sequence and several cis-acting regulatory elements (three copies of GC boxes, a C/EBP-binding site, an E box, and an AP-2-binding site). The C/EBP transcription factor has been implicated in the coordinate transcriptional activation of adipocyte-specific genes during the course of adipocyte differentiation (24,25). We 4 and others (26) have observed that ob gene expression is induced in stromal-vascular cells or 3T3-F442A preadipocytes during the course of adipocyte development and/or maturation, although no significant amount of ob mRNA is present in undifferentiated cells. Therefore, the C/EBP-binding site in the 5Ј-flanking region of the human ob gene might be involved in the transcriptional activation of the ob gene during adipocyte differentiation. Further studies are needed to elucidate the functional significance of these cisacting regulatory elements.
We demonstrated by the fluorescence in situ hybridization technique that the ob gene is mapped on human chromosome 7q31.3. Of particular note is that the cystic fibrosis transmembrane conductance regulator gene has been assigned to human chromosome 7q31.3 (27). It has been demonstrated that the ob gene is localized on the proximal region of mouse chromosome 6 (2, 28). The mouse chromosomal region on which the ob gene is located is part of a known segment with genes that are conserved between mice and humans and is syntenic to human chromosome 7q (29). This study has provided direct evidence that the ob gene is a member of the conserved syntenic group in mice and humans and further helps gene mapping in both species.
In conclusion, we succeeded in the isolation and characterization of the human ob gene. Using the fluorescence in situ hybridization technique, we also determined the chromosomal assignment of the human ob gene. This study helps to establish the genetic basis of ob gene research in humans, thereby leading to a better understanding of the physiologic and pathophysiologic implications of the ob gene. FIG. 5. Chromosomal assignment of the human ob gene by the fluorescence in situ hybridization technique. Left, partial metaphase chromosomes stained with propidium iodide showing the twin-spot signals on the long arms of both homologous chromosomes 7 (arrowheads); right, the G-band pattern of the same chromosomes delineated through a Nikon UV-2A filter. These results clearly indicate that the human ob gene is localized on the region of chromosome 7q31.3.