Abstract
We studied the cell type-specific expression of human β-carotene 15,15′-mono-oxygenase (BCO1), an enzyme that catalyzes the first step in the conversion of dietary provitamin A carotenoids to vitamin A. Immunohistochemical analysis using two monoclonal antibodies against different epitopes of the protein revealed that BCO1 is expressed in epithelial cells in a variety of human tissues, including mucosa and glandular cells of stomach, small intestine, and colon, parenchymal cells in liver, cells that make up the exocrine glands in pancreas, glandular cells in prostate, endometrium, and mammary tissue, kidney tubules, and in keratinocytes of the squamous epithelium of skin. Furthermore, BCO1 is detected in steroidogenic cells in testis, ovary, and adrenal gland, as well as skeletal muscle cells. Epithelia in general are structures that are very sensitive to vitamin A deficiency, and although the extraintestinal function of BCO1 is unclear, the finding that the enzyme is expressed in all epithelia examined thus far leads us to suggest that BCO1 may be important for local synthesis of vitamin A, constituting a back-up pathway of vitamin A synthesis during times of insufficient dietary intake of vitamin A.
β-Carotene 15,15′-mono-oxygenase (BCO1), formerly known as β-carotene 15,15′-dioxygenase, is an enzyme that catalyzes the first step in the conversion of dietary provitamin A carotenoids to vitamin A (retinol and its esters) in the intestinal mucosa (Goodman and Huang 1965; Olson and Hayaishi 1965). Plant carotenoids, primarily β-carotene, are an important dietary source of vitamin A and the sole source of vitamin A for vegetarians. Vitamin A is crucial for vision, and its derivatives, all-trans- and 9-cis-retinoic acid, serve as ligands for the RAR and RXR nuclear hormone receptors that are essential for embryonic development, reproduction, and normal epithelial growth and differentiation.
BCO1, a 62.6-kD cytosolic enzyme, is involved in the symmetrical cleavage of β-carotene, which results in two retinal molecules (Lindqvist and Andersson 2002). This is in contrast to β-carotene 9′,10′-mono-oxygenase (BCO2), which catalyzes eccentric cleavage of β-carotene to β-apo-10′-carotenal and β-ionone. These two enzymes are about 40% identical in amino acid sequence and appear to be expressed in a few of the same tissues. However, the cleavage activity of BCO2 appears to be much lower than that of BCO1 (Kiefer et al. 2001).
In our previous study (Lindqvist and Andersson 2002), we showed that BCO1 mRNA is present not only in the digestive tract in humans but also in liver and kidney and, to a lesser extent, in the steroidogenic tissues such as the testis and ovary, as well as in the prostate and skeletal muscle. High levels of BCO1 mRNA have also been reported in human retinal pigment epithelium, as well as low levels in the brain (Yan et al. 2001). Despite this distribution of BCO1 in extraintestinal tissues, little is known about the physiological function of BCO1 in these tissues.
The majority of absorbed provitamin A carotenoids are converted to retinyl esters in epithelial cells of the intestinal mucosa and then transported in chylomicrons to the liver, the main organ for vitamin A storage. From there, sufficient amounts of vitamin A are normally mobilized and transported in a complex with retinol-binding protein (RBP) to peripheral tissues via the circulation. However, it is important to note that substantial amounts of the absorbed dietary carotenoids are not cleaved by the BCO1 enzyme in the intestine (Blomstrand and Werner 1967; Hicken-bottom et al. 2002; Lemke et al. 2003), and because the majority of carotenoids circulating in blood are associated with very low-density lipoprotein (VLDL) and low-density lipoprotein (LDL) particles (Johnson and Russell 1992; Ziouzenkova et al. 1996), tissues expressing high levels of LDL receptors will, by default, contain high levels of carotenoids.
To gain further insight into the physiological role of BCO1, we investigated the cell type-specific expression of the enzyme by immunohistochemical (IHC) staining of human tissues known to express BCO1 mRNA. The results indicate that there is a general localization of BCO1 in epithelial cells, suggesting that BCO1 may be important for local vitamin A synthesis, thereby constituting a back-up pathway of vitamin A synthesis during times of insufficient dietary intake of vitamin A.
Materials and Methods
Production of Monoclonal Antibodies
The monoclonal antibody (MAb) BCO1-11 (previously called Mab-1-11) (subclass IgG1/K), directed against human BCO1, was produced by immunizing mice with a synthetic peptide, [C]RNRKEQLEPVRAKVTGK, corresponding to amino acid residues 7-23 in BCO1 as described (Lindqvist and Andersson 2002). The MAb, BCO1-25 (IgM/K) directed against human BCO1 was produced by immunizing mice with recombinant protein purified from Baculovirus-infected Sf9 cells (Lindqvist and Andersson 2002). The MAb BCO2-9 (subclass IgG1/K) directed against human BCO2 was produced by immunizing mice with a synthetic peptide, [C]NTPQKKAVFGQCRG, corresponding to amino acid residues 3-16 in human BCO2. Mice were immunized with 30-50 μg protein in RIBI MPL+TDM Emulsion Adjuvant System (Ribi Immunochem Research; Hamilton, MT). Three days after the second boost (protein in PBS), spleens were dissected out, and spleen cells were mixed with the myeloma cell line Sp2/0-Ag 14 at a ratio of 4:1 and fused by using polyethylene glycol 1500 as previously described (Andersson and Jornvall 1986). Hybridoma screening by an ELISA and production of MAbs were performed according to methods previously described (Andersson and Jornvall 1986). Mice were maintained and treated in accordance with the guidelines set forth by the Animal Welfare Information Center. Protocols were approved by the Institutional Animal Care and Use Committee at the University of Texas Southwestern Medical Center at Dallas.
Immunoblotting
The human embryonic kidney 293 cell line (HEK293) was maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum, 10 mM HEPES, and 1% penicillin/streptomycin, and transiently transfected with pCMV-hBCO1 or hBCO2 expression plasmids using Fugene 6 (Roche Diagnostics; Indianapolis, IN) as previously described (Lindqvist and Andersson 2002). SDS-PAGE was performed according to Laemmli (1970) and the proteins were transferred onto Immobilon-P PVDF membrane (Millipore; Bedford, MA) using a previously published method (Towbin et al. 1979). For detection of BCO1 and BCO2 proteins, we incubated membrane with the respective MAbs in tissue culture medium (DMEM, 10% fetal calf serum, 10 mM HEPES, 1% penicillin/streptomycin) supplemented with 0.2% Tween-20 for 1 hr at room temperature (RT), followed by a secondary horseradish peroxidase-conjugated goat anti-mouse IgG polyclonal antibody (BIO-RAD; Hercules, CA) diluted 1:10,000, and incubated for 1 hr at RT as described previously (Lindqvist and Andersson 2002). Antibody binding was detected by chemiluminescence using an ECL kit according to the manufacturer's protocol (Amersham Pharmacia Biotech; Piscataway, NJ).
Immunohistochemistry
Normal human tissues were obtained during routine surgical biopsies or autopsies (ProPath Laboratory; Dallas, TX). Representative sections of each specimen stained with hematoxylin and eosin were examined by experienced diagnostic pathologists to confirm the preservation of the histological structure and to validate the normalcy. The tissues, in paraffin blocks, were cut into 4-μm sections, mounted on charged adhesive slides, and dried in a 1000-W microwave oven (set on “high”) for 3 min, followed by 10 min in a 56C oven. Slides were then deparaffinized in xylene and rehydrated in graded alcohols to distilled water. Tissues were quenched for endogenous peroxidase activity for 10 min at RT (using 0.3% H2O2 with 0.1% sodium azide added). Epitope retrieval was performed by placing the slides in 0.1 M citrate buffer, pH 6.0, in a pressure cooker (BioCare Medical; Walnut Creek, CA) for 5 min at full pressure, followed by cool-down (total time in pressure cooker from start to finish is 48 min). After rinsing three times with PBS with 0.1% Tween-20 added, slides were incubated for 30 min in a 25C incubation oven using gentle orbital rotation at 40 rpm with MAb BCO1-11 or MAb BCO1-25 in tissue culture medium. For negative controls, irrelevant mouse IgG1/K MAb (cat no. M9269; Sigma, St Louis, MO) or mouse IgM/K MAb (cat no. 8-1501; Assay Design, Ann Arbor, MI) in tissue culture medium, or medium without primary antibodies, was used. After rinsing three times in PBS (without Tween-20), incubation with the appropriate horseradish peroxidase-conjugated polymer (PowerVision reagent) (ImmunoVision Technologies; Daly City, CA) was performed for 30 min at 25C. Finally, the slides were immersed for 5 min in 25C diaminobenzidine (DAB) (Research Genetics; Huntsville, AL), enhanced with 0.5% copper sulfate in PBS for 5 min at 25C, counterstained in hematoxylin, dehydrated in graded alcohols, cleared in xylene, and coverslipped.
Results
MAbs that recognize distinct epitopes of BCO1, designated BCO1-11 and BCO1-25, were generated for IHC analysis of BCO1 protein in various human tissues. The fact that the amino-terminal peptide that was used to generate MAb BCO1-11 was unable to abolish binding of BCO1-25 to the BCO1 protein, and that BCO1-25 was unable to bind to this peptide in an ELISA, strongly suggests that MAb BCO1-25 recognizes an epitope of BCO1 distinct from that of BCO1-11 (data not shown). The two antibodies were tested in an immunoblotting system for specificity and to confirm that they do not react with BCO2. Both antibodies reacted strongly with purified recombinant human BCO1 (Figure 1, Lanes A and D) and also with BCO1 transiently expressed in HEK293 cells transfected with a plasmid containing the BCO1 cDNA (Figure 1, Lanes B and E). None of the antibodies showed any reactivity with lysate from HEK293 cells expressing recombinant BCO2 (Figure 1, Lanes C and F). To confirm BCO2 expression in the HEK293 cells, an MAb raised against a synthetic amino-terminal peptide in human BCO2 was used. This antibody reacted with a 63.5-kD protein in the lysate from BCO2 expressing HEK293 cells (Figure 1, Lane H). However, no signal was observed in the lane with BCO1 lysate (Figure 1, Lane G) demonstrating that the BCO1-11 and BCO1-25 antibodies are specific for the BCO1 enzyme.
Detection of BCO1 and BCO2 by immunoblotting. The specificity of the two antibodies used in the IHC experiments, BCO1-11 and BCO1-25, was compared using purified recombinant BCO1 (Lanes A and D) and lysates from HEK293 cells transfected with a plasmid containing either human BCO1 cDNA (Lanes B, E, and G) or human BCO2 cDNA (Lanes C, F, and H). The presence of BCO2 in HEK293 lysates was confirmed using an anti-BCO2 MAb (Lane H). The positions of prestained molecular size markers (kD) are shown at left.
To determine the cell type-specific expression of BCO1, we performed IHC using the two anti-BCO1 MAbs. BCO1 immunostaining was scored as positive only if both BCO1-11 and BCO1-25, each recognizing different epitopes of the protein, stained the same cell type in a given tissue. None of the negative IgG1 and IgM controls stained any of the cell types in which BCO1 staining was detected (data not shown).
BCO1 activity has been described in intestinal mucosa from a number of different animal species (Devery and Milborrow 1994; Goodman and Huang 1965; Olson and Hayaishi 1965; Goodman et al. 1967). In addition, in a previous study we showed that BCO1 mRNA was present all along the human digestive tract from stomach to rectum, with the highest level found in the jejunum (Lindqvist and Andersson 2002). Therefore, we first examined the BCO1 distribution in small intestine. As shown in Figures 2A and 2B, BCO1 was detected only in the epithelium of the mucosa, both in the absorptive cells found in the villi and in the intestinal glands, in the Paneth cells in the bottom of the glands, and in goblet cells, when the two different antibodies against BCO1 were used. No BCO1 was observed in the lamina propria.
IHC detection of BCO1 in normal human tissues. The BCO1-11 antibody is used in
When we expanded our studies to include other parts of the digestive tract, e.g., stomach and colon, the same pattern was seen, with BCO1 immunostaining present only in the epithelium. In the stomach, the mucus-secreting columnar epithelium that covers the surface and lines the pits, and the cells in the glands emptying into the pits, were all positive for BCO1 (Figure 2D). In the colon, both the water-absorbing columnar epithelial lining and the abundant goblet cells were positive (Figure 2E). No immunostaining in lamina propria, muscularis mucosae, or submucosa was observed in either tissue.
An organ associated with the digestive tract is the liver. BCO1 activity was demonstrated in rat liver preparations (Olson and Hayaishi 1965; Napoli and Race 1988; Olson and Lakshman 1990) and BCO1 mRNA was present in human liver (Lindqvist and Andersson 2002; Yan et al. 2001). Our IHC study demonstrated that human BCO1 was expressed in hepatocytes (Figure 2F). However, no immunostaining was observed in the sinusoid endothelial cells.
To our knowledge, there are no data available concerning BCO1 in pancreas. However, because it is a gland associated with the digestive tract we included it in our study. The results showed that the exocrine portion of the pancreas, the acinar glands, stain strongly positive for BCO1 (Figure 2J). No staining was detected in the islets of Langerhans.
BCO1 activity was demonstrated in rat kidney (Napoli and Race 1988) and BCO1 mRNA was found in human kidney by RNA blotting (Lindqvist and Andersson 2002) and RT-PCR (Yan et al. 2001) analyses. IHC using the anti-BCO1 antibodies showed immunostaining in the cuboidal/columnar epithelium of the proximal convoluted tubule and distal convoluted tubule (Figures 2K-2L). The BCO1-11 antibody also stained the glomeruli (Figure 2K). However, we believe this to be an artifact because the glomeruli were negative when the BCO1-25 antibody was used (Figure 2L). This is the only tissue in which a discrepancy between BCO1-11 and BCO1-25 staining was observed, and is probably due to a glomerulus-specific protein crossreacting with the anti-peptide BCO1-11 antibody. Alternatively, it is conceivable that a glomerulus-specific protein may mask the epitope recognized by the BCO1-25 antibody.
As seen with kidney, rat testes homogenates display BCO1 activity (Napoli and Race 1988) and BCO1 mRNA was detected in human testes (Lindqvist and Andersson 2002; Yan et al. 2001). The present study indicated that BCO1 was expressed in the Sertoli cells in the seminiferous tubules and the androgen-producing Leydig cells in the interstitial tissue (Figure 3A). BCO1 mRNA was detected in human ovaries as well (Lindqvist and Andersson 2002). By using IHC, BCO1 protein was found to be expressed in the follicle, both in the estrogen-secreting granulosa cells in the membrana granulosa, and in the cells of the theca interna (Figure 3B). Because the steroidogenic cells in the gonads were positive for BCO1, we examined the highly steroidogenic adrenal gland for expression of BCO1, despite the fact that there have been no reports of BCO1 activity in this organ. The results showed that BCO1 was indeed expressed in the cortex of the adrenal gland, both in the mineralocorticoid-producing glomerulus and in the glucocorticoid-producing fasciculata (Figure 3C), as well as in the reticularis (data not shown). The capsule was negative for BCO1 immunostaining.
IHC detection of BCO1 in normal human tissues using MAb BCO1-11. (
BCO1 mRNA was reported in human prostate (Lindqvist and Andersson 2002), and the immunostaining in the prostate indicated that BCO1 was expressed solely in the epithelium of the glands, with no immunostaining in the fibromuscular stroma (Figure 3G). Examination of other exocrine gland-containing tissues, such as endometrium (Figure 3H) and mammary gland (data not shown), showed the same pattern, i.e., BCO1 expression in the epithelium of the glands exclusively.
Low levels of BCO1 mRNA were detected in human skeletal muscle (Lindqvist and Andersson 2002), and immunostaining was observed in a subset of skeletal muscle fibers (Figure 3I). Finally, the epidermis of the skin displayed strong immunostaining for BCO1 (Figure 3M).
The results of the IHC detection of BCO1 are summarized in Table 1.
Immunohistochemical detection of BCO1 in normal human tissues
Immunostaining graded as weak (+), moderate (+ +), or strong (+ + +).
Discussion
In this study we demonstrated the presence of BCO1 protein in cell types of various normal human tissues. Tissue distribution of human BCO1 was studied previously by RT-PCR (Yan et al. 2001), and RNA blotting analyses (Lindqvist and Andersson 2002). In agreement with the results of the present study, both publications report the presence of BCO1 mRNA in small intestine, liver, kidney, and testis. In addition, using RNA blotting analysis, BCO1 mRNA was found along the entire alimentary tract and in prostate, ovary, and skeletal muscle (Lindqvist and Andersson 2002), findings that were supported here by the detection of BCO1 protein in those same tissues. In contrast, Yan and colleagues could not find any BCO1 mRNA in pancreas, whereas this study reports strong staining for BCO1 in this organ.
It was first reported in the mid-1960s that the epithelial mucosa is the site of BCO1 activity in the small intestine (Goodman and Huang 1965). Here BCO1 cleaves provitamin A carotenoids to vitamin A, which are then transported in chylomicrons to the liver for storage as retinyl esters. Therefore, the role of BCO1 in the small intestine is to assist in the shuttling of vitamin A to the main storage facility, from whence it is released to peripheral tissues when needed. However, the cell type-specific results from this study indicate that there is a more widespread expression of BCO1 in epithelia, not restricted to the small intestine. The mucosal lining, including the glands, of the entire gastrointestinal tract, perhaps with the exception of the esophagus, contains BCO1. Furthermore, the epithelia in the glands associated with the digestive tract, liver, and the exocrine part of pancreas all stain positive for BCO1. In glands of the breast, prostate, and endometrium, BCO1 is also found in the epithelium. In addition, BCO1 is found in epithelia of the proximal convoluted tubule and distal convoluted tubule in the kidney, and in squamous keratinized epithelium constituting epidermis in the skin. Even though the physiological role of vitamin A in many of these tissues during fetal development is well established, e.g., renal organogenesis (Gilbert 2002), little is known about specific roles of vitamin A in many of these tissues in the adult. What is known, however, is that epithelia are sensitive to the vitamin A status. Deterioration of mucous membranes in the form of keratinization or atrophy is one of the characteristics of vitamin A deficiency (Wolbach and Howe 1925; Moore 1965). Many publications have described histological changes in the epithelia of different tissues during mild to severe vitamin A deficiency, e.g., atrophy of epithelium of rat esophagus (Mak et al. 1987), mild atrophy, later accompanied by hyperkeratinization, of rat forestomach epithelium (Klein-Szanto et al. 1982), a decrease in numbers of goblet cells in duodenal crypts in the rat (De Luca et al. 1972; Rojanapo et al. 1980), reduction of villous height and reduction of goblet cells and enterocytes in small intestine in rat (Reifen et al. 1998), reduction of villous height and crypt depth in the small intestine of the chicken (Uni et al. 1998), hyperkeratinization of rat tongue epithelium with fewer and severely atrophied taste buds in the epithelium (Chole and Charpied 1983), keratinization of tracheal epithelium and atrophy of acini in submaxillary glands in rat (Anzano et al. 1980), and keratinization of epithelium in rat urinary bladder (Hicks 1957). Functional changes such as decreased activities of the small intestine brush border enzymes, e.g., disaccharides, have also been observed in rat and chicken (Warden et al. 1996; Uni et al. 1998), as well as the hypersecretion and reduced electrogenic distal absorption seen in the intestine of rat (Nzegwu and Levin 1991,1992). All these changes can be reversed by adding vitamin A or its derivative, retinoic acid, to the animals' diets. Hence, epithelia are strongly dependent on an adequate supply of vitamin A to maintain their structure and subsequently their function. There is an obvious positive correlation between cell types sensitive to vitamin A deficiency and cell types expressing BCO1. Therefore, the β-carotene-cleaving enzyme may have dual physiological functions: first, postprandial synthesis of vitamin A in the small intestine from newly ingested provitamin A carotenoids to load the liver storages and, second, serving as a local supplier of vitamin A by cleavage of locally stored provitamin A carotenoids in epithelia in a variety of tissues at times of need.
If the physiological function of BCO1 is to promote a local source of retinoids to the cells, then the necessary prerequisite is availability of substrate for BCO1 in the epithelial cells. Only about 55-75% of the β-carotene absorbed in the human intestine is cleaved on site (Goodman et al. 1966; Blomstrand and Werner 1967; Hickenbottom et al. 2002; Lemke et al. 2003), so contrary to species such as the rat, rabbit, chicken, pig, and sheep, humans absorb small amounts of unchanged β-carotene. β-Carotene and other carotenoids must be incorporated into chylomicrons in order to enter the circulation. The majority of the highly lipophilic β-carotene then ends up being transported in the core of VLDL and LDL particles (Johnson and Russell 1992; Ziouzenkova et al. 1996) and ultimately is delivered to peripheral tissues via the LDL receptor. LDL receptors are present at various amounts not only in different tissues but in different cell types (Kovanen et al. 1979; Spady et al. 1983), with the adrenal gland and liver having the highest levels of LDL receptors. Therefore, the β-carotene content in various tissues may, at least in part, reflect of the degree of local LDL receptor activity. Interestingly, β-carotene levels were reported for seven human tissues and the adrenal gland had the highest β-carotene content, 5.6 nmol/g wet tissue, followed by liver, testis, kidney, ovary, and adipose tissue, with about 0.4 nmol/g wet tissue (Stahl et al. 1992). Hence, the β-carotene level appears to correlate directly with the LDL receptor activity of these tissues.
Similar to the findings of Stahl et al. (1992) with β-carotene concentrations, the present study also indicates that BCO1 is expressed in steroidogenic cells. The enzyme appears to be localized to androgen-producing Leydig cells in testis, estrogen-producing granulosa cells, and the C19 steroid-producing theca interna in the ovary, as well as to cells in the adrenal cortex, i.e., the zona glomerulosa (mineralocorticoids), zona fasciculata (glucocorticoids), and zona reticularis (androgen precursors). These cell types have not been implicated as sensitive to vitamin A, nor have skeletal muscle cells. Both Sertoli cells and granulosa cells are important for the germ cells in their respective gonad, having protective and nourishing roles. Vitamin A deficiency has been reported to cause a loss of advanced germ cells in rat testis, a condition corrected by supplementation with retinol, which reinitiates spermatogenesis (Wang and Kim 1993). Therefore, local conversion of provitamin A carotenoids to vitamin A by BCO1 may directly serve an important role in steroidogenesis.
Our current knowledge of BCO1 is based primarily on activity studies performed with intestinal and liver homogenates. This study indicates that BCO1 potentially could also be responsible for local cleavage activity in structures known to depend heavily on vitamin A. Further studies, including determination of cell type-specific expression of the eccentrically cleaving enzyme BCO2 may help to clarify the function of BCO1 in extraintestinal tissue.
Footnotes
Acknowledgements
Supported by Grant DK62192 from the National Institutes of Health.