Genomic Structure and Expression of the Human Gene Encoding Cytochrome b 561 , an Integral Protein of the Chromaffin Granule Membrane*

Cytochrome b 561 is an electron transfer protein unique to neuroendocrine secretory vesicles. The Southern blot hybridization shows that it is a single copy gene highly conserved throughout phylogeny. The transcription unit spans approximately 11 kilobases, and heter- ologous transcription sites are located 404 bases 5 (cid:42) to the translation initiation codon. The sequence of the 5 (cid:42) -flanking region is GC-rich and lacks a typical TATA box at the usual position. However, it has a CAAT se- quence at (cid:50) 132 and potential recognition sequences for several transcription factors including SP1, GR-PR- MMTV, AP4, gERE, JCV repeat, AP2, and NF- (cid:107) B. Each of the five transmembrane segments are encoded by five consecutive exons. This corroborates the five-trans-membrane model proposed for human, mouse, and Xe- nopus rather than six proposed for bovine. The cytochrome was found to be highly expressed in colon cancer cell lines, T cell lymphomas, and K-562 cell lines. However, in B-cell lymphomas such as Burkitt’s and Daudi, the cytochrome b 561 expression was completely shut down. The results in this report are the first to demonstrate the structural organization and regulatory sequences of the cytochrome b 561 gene encoding an in- tegral membrane protein of neuroendocrine storage vesicles of neurotransmitters and peptide hormones. Unexpected results on cytochrome b 561 expression in cells of lymphocytic origin and its complex regulation in tumor cells provide new insights into cytochrome b 561 gene regulation. The

The integral membrane protein, cytochrome b 561 , was first discovered in catecholamine storage granules (1) and was subsequently found to be specific to catecholamine and neuropeptide secretory vesicles of the adrenal medulla, pituitary gland, and other neuroendocrine tissues (2)(3)(4)(5)(6)(7)(8). This 30-kDa protein comprises 15% of the granule membrane protein (9) and is present in both small synaptic vesicles and large dense core vesicles (10,11) where it plays a central role with ascorbic acid in the biosynthesis of several catecholamine (6,12) and peptide neurotransmitters (13,14). Among the cytochromes, cytochrome b 561 is unique in its localization to neuroendocrine tissues and appears to be unusual in its enzymatic mechanisms. Unlike other electron carriers which need other proteins for its oxidation-reduction cycle, the cytochrome b 561 does not interact with any other protein. Instead, ascorbic acid serves as the extravesicular electron donor for cytochrome b 561 , and semidehydroascorbic acid acts as the intravesicular electron acceptor, which replenishes the high intravesicular ascorbic acid concentrations. Ascorbic acid serves as a cofactor for intravesicular dopamine ␤-hydroxylase activity in catecholamine storage granules (15) and peptidyl ␣-amidating monooxygenase activity in neuropeptide storage vesicles (14). Cytochrome b 561 appears to be a simple, functionally symmetric electron transfer protein and acts as a pure electron channel.
Understanding the mechanism of this long range electron transfer process has been facilitated by the cloning of bovine cytochrome b 561 which revealed that it has six transmembrane helices with N-and C-terminal ends of the protein facing the cytoplasm (16). Subsequent biochemical analysis using peptide antibodies raised against N-terminal and C-terminal ends of the protein and protease digestion suggested that the N terminus was unavailable to both (17). In addition, the human cytochrome b 561 lacked the first putative 22 amino acids at the N terminus predicted from the bovine sequence to be in the cytoplasmic side. Based on the sequence comparison and low hydrophobicity of the second transmembrane domain, combined with the biochemical evidence, a revised hypothesis was proposed in which the gene product is likely to be distributed within the membrane with five transmembrane helices (18). Subsequent cloning of the mouse 1 and Xenopus 2 cytochrome b 561 confirmed this structure and further revealed that it was colocalized to specific tissues with either DBH or peptidylglycine ␣-amidating mono-oxygenase. In addition, we found that cytochrome b 561 expression was developmentally regulated in mouse and Xenopus. Furthermore, more recent data on cytochrome b 561 expression in different tissues exhibited uniform distribution in all the neuroendocrine tissues tested.
In order to better understand the molecular basis of developmental and tissue-specific cytochrome b 561 expression, it is necessary to characterize the genomic structure and regulatory elements of the gene. This approach could also help establish the relationship between this gene and those of other membrane proteins, thereby providing clues concerning the evolution of the eukaryotic chromaffin granule. Cloning and sequence analysis of the regulatory region of the cytochrome b 561 gene would provide reagents and information necessary to examine the expression of a major structural protein of the chromaffin granule. For these reasons, we have determined the structural organization and promoter sequence of human cytochrome b 561 , and studied gene expression in different cell lineages. * 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

MATERIALS AND METHODS
Library Screening-A genomic DNA library from human peripheral blood leukocytes in Lambda DASH (Stratagene) were plated, and replicas containing 1 ϫ 10 6 clones were screened using the full-length human cytochrome b 561 cDNA from the HCR1A clone (18), which had been radiolabeled by nick translation (Life Technologies, Inc.). Hybridization was performed at 42°C in 6 ϫ SSC, 5 ϫ Denhardt's solution, 0.1% SDS, 50% formamide, and 0.1 mg/ml salmon sperm DNA. Nitrocellulose filters were washed at 65°C in 1 ϫ SSC and 0.1% SDS and exposed to XAR-5 film (Kodak) with an intensifying screen at Ϫ80°C. -DNA was purified using a Qiagen kit (Qiagen, Inc). Four positive clones were sequentially plaque-purified, analyzed, and compared to each other by a combination of restriction mapping and Southern blot analysis. Hybridization with different labeled oligonucleotides corresponding to various regions of the cDNA led to the identification of two clones containing the entire cytochrome b 561 gene.
Determination of the 5Ј-Flanking and Exon 1 Sequences-The 6.5 kb 3 XbaI fragment hybridized with the 5Ј region of the cDNA was isolated and digested with XhoII, and the 2.8-kb, 2.3-kb, and 1.4-kb fragments were subcloned into the pGX2627 vector. These subclones were sequenced using either two primers homologous to sequences flanking the multiple cloning site of pGX2627 or 24 oligonucleotides based on cDNA or derived genomic sequences. The entire gene and the flanking sequences were sequenced by Sanger's dideoxy method (21) using cloned T7 DNA polymerase (Sequenase, U. S. Biochemical Corp.). Oligonucleotides were synthesized in an Applied Biosystems DNA synthesizer Model 380B. DNA sequences were compiled with the PCgene program (Intelligenetics) in order to obtain the 5Ј-flanking sequences and the first exon sequences. The 5Ј-flanking sequences were analyzed for regulatory elements using a computer search routine.
PCR Determination of Exon 2-5 and Intron Sizes-Using different sets of primers (Table I) and -DNA as template, the polymerase chain reaction was performed as described (22) using the GeneAmp kit (Perkin Elmer) with the following parameters: initial denaturation at 94°C for 2 min followed by 35 cycles of denaturation at 94°C for 1 min, annealing at 50°C for 2 min, and extension at 72°C for 3 min with a final extension at 72°C for 7 min. Amplified products were electrophoresed on agarose gels, gel-purified using the Geneclean Kit (Bio 101), and subcloned into the TA cloning vector (Invitrogen) according to the manufacturer's instructions. The plasmids containing the genomic fragments were selected by restriction endonuclease digestion and were sequenced using the M13 universal sequencing primer at the 5Ј-end and the T7 promoter sequence to obtain the 3Ј-end of the genomic fragment's sequence. Some of the fragments containing the exon sequences were further sequenced using the oligonucleotides derived from the cDNA sequences to confirm the cDNA sequence, as well as to find if there were any small intron sequences in between the exons. The sizes of the introns were determined by comparing the PCR amplified products with the migrations of the DNA fragments of known sizes. The combination of these procedures identified five introns.
Mapping of the Transcription Start Sites-The ribonuclease protection assay was carried out to confirm the transcription-initiation site of cytochrome b 561 mRNA previously determined by primer extension and the reverse transcriptase-PCR method (18). The 415-bp fragment from the 5Ј-end of the gene ( Fig. 1; nucleotide 4288 to 4703) was used to synthesize an antisense RNA probe using a RNA transcription kit (Stratagene) with T7 polymerase and [␣-32 P]UTP. The RPA kit from Ambion (Austin, TX) was used for the RNase protection assay with 50 g of human total brain RNA or 50 g of yeast tRNA (Clontech) as negative controls. The riboprobe (1 ϫ 10 6 cpm) was hybridized at 42°C with RNA (in 40 mM Pipes, pH 6.4, containing 1 mM EDTA, 400 mM NaCl, and 80% formamide) and digested with RNase A and RNase T1. The protected fragments were separated by electrophoresis on a 6% polyacrylamide gel. The gel was then dried and exposed to Kodak XAR-5 film.
Reverse Transcriptase-Polymerase Chain Reaction-Reverse transcriptase-PCR was performed as described by the supplier of the enzyme (Perkin-Elmer Cetus Instruments). Briefly, 1 g of poly(A) RNA from HL-60, HeLa, Daudi, K-562, melanoma, aorta, ovary, Burkitt's, colon, lung, and breast cancer cell lines was transcribed into singlestranded DNA using Moloney murine leukemia virus reverse transcriptase (Life Technologies, Inc.), followed by a PCR reaction performed with 20 pmol of cytochrome b 561 specific oligonucleotide antisense primer (5Ј-CTGGGAGCCGGGGCTATCTC-3Ј), 30 pmol of cy-tochrome b 561 specific oligonucleotide sense primer (5Ј-CACAGCACT-GCCTTACTACG-3Ј), and 2.5 units of Thermus aquaticus (Taq) DNA polymerase 1 (Perkin-Elmer Cetus Instruments). The samples were treated for 2 min of initial denaturation at 95°C and 30 cycles of denaturation (1 min at 95°C), annealing, and extension (2 min at 60°C) with a final extension for 7 min at 60°C. The reverse transcriptase was omitted from the PCR reactions in the negative control samples, which were otherwise subjected to the same conditions as above. The PCR products were analyzed electrophoretically using 1.5% agarose gel and visualized by ethidium bromide staining.

Isolation, Restriction Mapping, and Characterization of Human Cytochrome b 561
Gene-Approximately 1 ϫ 10 6 recombinant clones of a Lambda DASH human blood leukocyte genomic library were screened using a human cytochrome b 561 cDNA probe (18). Four clones were isolated, and the entire gene was identified by restriction mapping and Southern hybridization with the cDNA probe and a series of oligonucleotides that encompass the entire region of cDNA.
Sequence Determination and Exon-Intron Organization-Through a combination of restriction mapping, PCR, and sequence analysis, the human cytochrome b 561 gene was determined to contain 6 exons interrupted by 5 introns, defining a gene of approximately 11 kb in size. Fig. 1 shows the organization of the cytochrome b 561 gene in relation to the cDNA sequence and transmembrane domains. The XbaI fragment of the clone identified with the primer from exon 1 was subcloned and sequenced. The DNA sequence of the rest of the exons, the exon/intron splice junctions and the introns were obtained using the clones as templates and the oligonucleotide primer sets A 3 H listed in Table I. The location of each splice junction is shown in Fig. 1. The length of each intron was then determined by PCR using pairs of oligonucleotides from the exons flanking each intron, followed by analysis on agarose gels.
Partial nucleotide sequences, including the sequence of the 5Ј upstream region and less than 200 bp of all introns are shown in Fig. 2. Exons flanked by introns ranged in size from 99 bases (exon 2) to 1374 bases (exon 5) (Fig. 2, Table II). The intron-exon junction sequences are consistent with the reported 5Ј-donor (GTGAG) and the 3Ј-acceptor (C/TAG) consensus sequence (21) ( Table II). The intron splice phase is type 0 (the intron occurs between codons) for intron 3, type 1 (the intron interrupts the first and the second bases of the codon) for introns 1 and 2, and type 2 (the intron interrupts the second and the third base of the codon) for intron 4 (22).
Relationship of Splice Junctions to Protein Domains-Exons often represent the functional units of the protein. Indeed, the first, fourth, and fifth transmembrane domains are entirely within exons 1, 4, and 5, respectively. However, the second and third transmembrane domain sequences are interrupted at the end by introns 2 and 3. Exon 1 has the mRNA leader sequence and the first transmembrane domain. Exon 2 contains the conserved His-91 and the putative cyclic AMP binding site and two protein kinase C (PKC) sites. Exon 4 has the conserved His-160 which ligands to the heme moiety and is adjacent to the heme-binding hydrophobic pocket. Exon 5 contains the fifth transmembrane domain with one (PKC) site. Hence, the five transmembrane segments of the protein are structurally dis-crete and apparently have different functional roles. Because this is the first report on the genomic sequence of cytochrome b 561 from any species to be cloned, the degree of evolutionary conservation of the cytochrome b 561 gene organization is yet to be learned.
Repetitive DNA-The human cytochrome b 561 gene was examined for the presence of Alu repeats. Two complete Alu repeats were identified with 82% and 93% identity to published Alu consensus sequences (25). They are found at position 2986 in the 5Ј-flanking sequence and at position 8051 in intron 5. The dinucleotide repeat CACACACACACA was found at position 2949 and at position 3276 which might be a useful microsatellite marker for mapping the cytochrome b 561 gene defects linked to a specific disease on chromosome 17.
Mapping the Transcription Initiation Site-Primer extension analysis and reverse transcriptase-PCR experiments using human brain mRNA identified two major transcription sites and two minor transcription sites (18). To confirm the validity of the above analysis, a ribonuclease protection assay was carried out utilizing human brain RNA and a 415-base pair RNA probe that was generated using the T7 promoter (see "Materials and Methods"). Corresponding to the primer extension study, two protected RNA species of approximately 131 and 139 nucleotides were detected which were not observed in the lane containing yeast tRNA control. Two minor bands were also observed by this assay, suggesting that additional transcripts of the cytochrome b 561 gene might be expressed although at much lower levels (Fig. 3). This result was confirmed by replicate analysis.
Analysis of the 5Ј-Flanking Sequences-To identify sequences that potentially control cytochrome b 561 gene expression, a dynamic computer program (26) was utilized to locate the regulatory elements. The region 5Ј to the identified transcription start sites is GC-rich and contains a possible CAAT box at Ϫ132 nucleotides and potential recognition sequences for several transcription factors (Fig. 4). However, there is no TATA-like element in close proximity. An examination of the 5Ј-flanking sequences revealed the presence of six JCV repeats, six SP-1, two AP-2, MRE, GR-PR-MMTV, AP-4, PEA-3, E-alpha H box, and HTNF.1-hist, all within 1000 nucleotides of the transcription start site. Imperfect sequences, which only closely resemble the reported binding sites, have not been listed in Fig. 4. However, they might have a potential regulatory function, as it has been shown that many binding sites deviate to a certain extent from reported consensus sequences.
Differential Expression of Cytochrome b 561 in Cancer Cell Lines-Despite extensive studies on the ubiquitous expression of cytochrome b 561 in normal neuroendocrine tissues, little is known about its expression in tumor tissues or regulation of cytochrome b 561 expression underlying its induction or suppression at the molecular level, which will provide the molecular basis of the physiological function of cytochrome b 561 . Therefore, Northern blot analyses were performed using poly(A) RNA from tumor cell lines derived from melanoma (G 361), lung carcinoma (A 549), colorectal carcinoma (SW 480), Burkitt's lymphoma (Raji) T cell lymphoblastic leukemia cell line (MOLT-4) and chronic myelogenous leukemia cell line (K-562), HeLa cell, 53 and promyelocytic leukemia (HL-60), and peripheral blood leukocyte as normal cells. The colon carcinoma cells (SW 480), T cell lymphoma (HL-60) and K-562, expressed high levels of cytochrome b 561 (Fig. 5A, a) compared to peripheral blood leukocytes (data not shown). Again in Burkitt's which is a B cell lymphoma cell type, the cytochrome b 561 expression was not detected, although the amount of mRNA was almost same in all the cell lines tested using actin as the probe (Fig. 5A, b). To further verify the results obtained by Northern blot analysis, reverse transcriptase-PCR analysis was carried out using mRNA from diverse tumor cell  Table I. c, arrows show portions of the various subclones sequenced.
lines and tumor tissues, including HL-60, HeLa, Daudi, K-562, melanoma, aorta, ovary, Burkitt's, colon, lung, and breast. Using primers from the coding region of human cytochrome b 561 , the expected 727-bp band was detected in all the cancer cell lines except Burkitt's and Daudi where the cytochrome b 561 expression was completely absent (Fig. 5B, a). In order to verify this observation further, an oligonucleotide from the internal sequence was end-labeled with 32 P and used for Southern blot hybridization. The results in Fig. 5B, b, also show that there is no signal detected in Burkitt's and Daudi cell lines, although the amount of mRNA used was the same as confirmed by the amplification of mRNAs from all the cancer lines using primers from a different gene (Fig. 5B, c). This limited analysis suggests that expression of cytochrome b 561 can be selective in hematopoietic cells.

DISCUSSION
To understand the molecular basis of the tissue-specific expression of cytochrome b 561 , we have now cloned and charac- terized the genomic sequence of cytochrome b 561 including the promoter sequences possibly regulating the specific expression of cytochrome b 561 . Although the structure, function, and expression of cytochrome b 561 protein has been studied in detail, the biological significance of its uniform expression in neuroendocrine tissues have not been understood. The results described here, therefore, are of particular significance in as much as they represent first characterization of the gene encoding the integral membrane protein of the chromaffin granule and its differential expression in hematopoietic cells.
Southern hybridization analysis established that cytochrome b 561 gene is highly conserved (data reviewed, but not included) and occurs as a single copy of the cytochrome b 561 gene in the human genome, spanning approximately 11 kb, and is split into five exons. The coding sequence of the gene is 753 bp, representing 7.1% of the gene, and perfect agreement between the nucleotide sequences of the chromosomal gene exons and the cDNA sequences was observed. The introns interrupt the cytochrome b 561 protein coding sequence in such a way that many of the protein segments are revealed as products of individual exons. A similar correspondence has been observed in other highly hydrophobic proteins such as human myelin (27), human T cell-specific proteolipid protein (28), and human vacuolar ATPase (29). Studies on the structure of the bovine cytochrome b 561 have been interpreted to indicate that the protein might traverse the membrane six times with N-and    3. Site of transcription initiation as determined by ribonuclease protection analysis. Human total brain RNA (20 g) or yeast tRNA was annealed to a 32 P-labeled antisense RNA probe (prepared by T7 polymerase transcription) and digested with RNase A and T1 as described under "Materials and Methods." The probe fragments protected by total RNA from human brain was analyzed on 6% polyacrylamide-8 M urea gel electrophoresis and were compared with the adjacent sequencing ladder. C-terminal ends facing the cytoplasm. However, subsequent biochemical analysis using antibodies raised against N-and C-terminal peptides along with topological orientation suggested that N-terminal is not facing the cytoplasm. Comparative amino acid sequences derived from human, mouse, and Xenopus, along with low hydrophobicity and unusual amino acid composition of the second transmembrane segment, suggested that bovine is 22 amino acids shorter, and a model with five transmembrane segments was proposed (18). The analysis of the cytochrome b 561 gene in the present study corroborates the above view in that five transmembrane segments occur in five different exons, and thus each exon encodes a potential unit of the cytochrome b 561 gene. The two histidine residues His-91 and His-160, which form putative ligands to heme, are separated into different exons. In addition, the amino acid residues Try, Gly, and Phe, which form the heme binding pocket, are in exon 4 along with His-160.
Inspection of the genomic sequences in the vicinity of the transcription start sites indicates the absence of an apparent TATA box in this region. The 5Ј-flanking region includes the dinucleotide CpG in its recognition sequence, suggesting that the cytochrome b 561 promoter is G ϩ C-rich. These features place the cytochrome b 561 gene in an emerging group of tissuespecific genes with TATA-less and G ϩ C-rich promoters, which also includes other genes differentially expressed in T lymphocytes such as CD7 (30) and Thy-1 (31). The present results also provide initial information as to what factors potentially regulate the expression of human cytochrome b 561 . Analysis of the transcription factor binding sites in the human gene promoter suggests that seven potential signaling pathways could regulate cytochrome transcription: SP-1, involved in maintaining appropriate basal transcription levels (32); AP-2, which medi-ates induction involving cyclic AMP (33); AP-4, whose function is unknown (34); glucocorticoid and progesterone receptor, which binds to the promoter of mouse mammary tumor virus (MMTV) (35); MRE, a putative metal regulatory factor which induces transcription in response to heavy metals (36); NF-B, which regulates gene expression via cytokines (37); and the JCV repeat, a sequence specific to human polyoma virus which is highly neuro-oncogenic in hamsters (38). Another feature in the cytochrome b 561 gene that could be related to tissue-or development-specific or cell type-specific expression is the presence of regulatory elements such as PEA-3, H4TF-1, hist, E-alpha H box, and HinF-hist (20,39,40), along with the heterologous transcription initiation sites. One example of tissuespecific mRNAs with alternative 5Ј-untranslated sequences is the different transcripts that arise from the ␣-amylase gene in salivary gland and pancreas (19). Since cytochrome b 561 is known to be specifically expressed in neuroendocrine tissues and in this study differentially expressed in hematopoietic tumor cells, additional experimentation is needed to elucidate the elements responsible for the regulation of tissue-and cellspecific expression.
We have demonstrated previously that cytochrome b 561 is constitutively expressed in all the neuroendocrine tissues studied, at high levels in brain, placenta, lung, and pancreas (18). As a first step toward investigating the biological function of cytochrome b 561 , the expression patterns of the corresponding transcripts were analyzed in a wide variety of cancer lines using reverse transcriptase-PCR and Northern blot analysis. Significant levels of cytochrome b 561 mRNA expression were detected in colon cancer cell lines, T cell lymphomas, and undifferentiated cell lines such as K-562 compared to peripheral blood leukocyte. In addition, the cytochrome b 561 expression was completely shut down in B cell lymphomas. These findings suggest that it may play a specialized role in growth and differentiation pathways of the hematopoietic cells. The differential expression of cytochrome b 561 mRNA confirmed by reverse transcriptase-PCR and Northern blot analysis in hematopoietic cell lines will provide an excellent model system for identification of cis-acting elements required for gene expression in cells from T and B lymphomas.
Cytochrome b 561 plays a pivotal role in regenerating ascorbate in neuroendocrine vesicles for dopamine ␤-hydroxylase and peptidyl ␣-amidase, and, hence, understanding the regulation of cytochrome b 561 may yield insights for development and neuroendocrine disorders. Because it constitutes 15% of the membrane protein in chromaffin granules, cytochrome b 561 may have evolved concurrently with the formation of the vesicles. The studies reported here provide key surprising findings on the expression of cytochrome b 561 in hematopoietic cells. Thus, it is obvious that the cytochrome b 561 expression is not exclusive to neuroendocrine cells. Further studies in the process of tumorigenesis will give clues about the biological function of cytochrome b 561 in the newly identified cells. Cloning of the human cytochrome b 561 gene now enables us to use molecular genetic techniques to further study its expression patterns, gene regulation, and function, particularly its role in processes such as cell differentiation, development, and carcinogenesis.
FIG. 5. A, a, Northern blot analysis of poly(A) ϩ RNA from different cancer lines and tissues from human. The blot was hybridized to the human cytochrome b 561 coding region. A, b, the same blot was hybridized to ␤-actin after stripping. B, reverse transcriptase-PCR and Southern blot analysis of human cytochrome b 561 mRNA. Poly(A) ϩ RNA from different cancer cell lines was reverse-transcribed and run on a 1.5% agarose gel followed by blotting onto nitrocellulose paper (see "Materials and Methods"). B, a, agarose gel electrophoresis of reverse transcriptase-PCR amplified mRNA from different cancer lines from human. B, b, agarose gel electrophoresis of reverse transcriptase-PCR amplified mRNA from cancer cell lines using oligonucleotide primers from a different gene. B, c, Southern blot of the same gel probed with 32 P-labeled antisense primer from set D in Table I. Control lane has the markers.