点击显示 收起
National Center for Toxicological Research, Jefferson, Arkansas (S.N., B.G., F.F.K.)
Roswell Park Cancer Institute, Buffalo, New York (S.N.)
Cedars-Sinai Medical Center, Medical Genetics Institute, Los Angeles, California (Y.M.T.)
Pierce Biotechnology, Rockford, Illinois (R.W.)
Abstract
Sulfotransferase isoform 1A2 (SULT1A2) is a member of the cytosolic sulfotransferase family of phase II detoxification enzymes. Studies with recombinant enzymes have shown that SULT1A2 can catalyze the bioactivation of several procarcinogens, indicating a potential role in chemical carcinogenesis. However, previous studies have suggested that the SULT1A2 transcript has a splicing defect that might prevent it from becoming translated into protein; therefore, we sought to determine the expression of SULT1A2 in tissues. An antibody directed against a region of human SULT1A2 that differs from other known sulfotransferase isoforms was developed and used to screen a large number of cytosolic fractions from various tissues. Although the SULT1A2 antibody recognized recombinant SULT1A2 and did not cross-react with other SULT isoforms, the expression of SULT1A2 was not detected in any tissue examined. These studies suggest that if SULT1A2 is expressed as protein, the levels are very low and that SULT1A2 probably does not play a physiological role in chemical carcinogenesis.
Human cytosolic sulfotransferases (SULTs) are phase II detoxification enzymes that catalyze the biotransformation of a host of endogenous and exogenous substrates (Jakoby and Ziegler, 1990). Eleven human isoforms that are the products of 10 distinct genes have been identified to date (Glatt and Meinl, 2004a). The cytosolic SULTs are members of a gene superfamily that can be divided into two major subfamilies, the phenol sulfotransferases (SULT1) and the hydroxysteroid (SULT2) sulfotransferases. A SULT4A1 isoform has been described that has high expression and tissue selectivity for brain, although no substrate has been identified for this isoform to date (Falany et al., 2000). All cytosolic SULTs catalyze the transfer of the sulfonoxy group from the obligatory cofactor, 3'-phosphoadenosine-5'-phosphosulfate, to a wide variety of endogenous compounds (steroid and thyroid hormones and bile acids) and a host of structurally diverse xenobiotics. The net result of this reaction, in most cases, is to render the substrate more water-soluble, leading to excretion from the organism. However, in some instances, sulfation of a molecule leads to bioactivation because of the fact that the sulfate group formed by O-sulfonation is electron-withdrawing and can become a good leaving group, generating an electrophile capable of binding to cellular macromolecules (Minchin et al., 1992; Chou et al., 1995).
The SULT1A family consists of three members, SULT1A1, SULT1A2, and SULT1A3, which are 93 to 96% identical at the amino acid level. When recombinant enzymes are examined, SULT1A1 and SULT1A2 share overlapping substrate specificity, but the affinity for substrate and rate of reaction is an order of magnitude lower in SULT1A2.
When Salmonella typhimurium TA1538-derived strains expressing alloenzymes of SULT1A2 were constructed and mutagenicity assays were performed, it was reported that bioactivation of promutagens was evident in the constructs, although the magnitude of mutagenicity varied by alloenzyme and by the promutagen under investigation (Meinl et al., 2002). It was also reported that SULT1A2 was the most efficient enzyme in sulfating some aromatic hydroxylamines.
These data suggest a potential role for SULT1A2 in the mutagenicity and carcinogenicity of substrates that are widespread in our environment. Although RNA for SULT1A2 has been detected in many tissues (Dooley et al., 2000), the protein expression levels of this isoform have not been elucidated. We have generated an anti-peptide antibody that discriminates SULT1A2 from SULT1A1 and SULT1A3 and have screened several human tissues for the protein expression of SULT1A2.
Materials and Methods
Materials. Recombinant SULT1A1, SULT1A2, SULT1A3, SULT2E1, and SULT2A1 were purchased from Invitrogen (Carlsbad, CA), as well as 12% NuPAGE gels, MOPS electrophoresis running buffer, nitrocellulose, and MOPS transfer buffer. DiscoverLight Arrays and Western blotting detection reagents were purchased from Pierce Biotechnology Inc. (Rockford, IL). Human tissues were obtained from the U.S. Cooperative Human Tissue Network (National Cancer Institute, Bethesda, MD).
Peptide Selection and Antiserum Production. A sequence region of SULT1A2 that is distinguishable from SULT1A1 was identified by sequence comparison. This region began at amino acid residue 147 with the peptide sequence KVYPHPGTWESFC and differed from SULT1A1 and SULT1A3 (the most structurally similar SULTs) sequence by three amino acids. The first 12 amino acid residues are specific to SULT1A2; the C-terminal cysteine residue was designed to couple the peptide to the carrier protein. The antigenicity and the surface possibility of the region was analyzed and predicted by the computer software package MacVector (International Biotechnologies, New Haven, CT). The Kyte-Doolittle method was used in the calculation over a window of seven residues. The peptide was synthesized on an Advanced Chemtech peptide synthesizer by Alpha Diagnostic Inc. (San Antonio, TX) using solid-phase chemistry. The synthesized peptide was then coupled to a carrier protein (keyhole limpet hemocyanin) through the cysteine residue at the C terminus of the peptide using N-succinnidyl bromoacetate as a cross-linking reagent (Bernatowicz and Matsueda, 1986). Specifically, succinimidyl 4-(n-maleimidomethyl)-cyclohexane-1-carboxylate (SMCC) and its water-soluble analog Sulfo-SMCC, heterobifunctional cross-linkers that each contain an N-hydroxysuccinimide (NHS) ester and a maleimide group, were used in the coupling reaction. NHS esters react with primary amines at pH 7 to 9 to form covalent amide bonds. Hydrolysis of the NHS ester, which is a competing reaction, increases with increasing pH and decreasing protein concentrations. Maleimides react with sulfhydryl groups at pH 6.5 to 7.5 to form stable thioether bonds. At pH values >7.5, reactivity toward primary amines and hydrolysis of the maleimide group can occur; however, the maleimide groups of Sulfo-SMCC and SMCC are unusually stable up to pH 7.5. For conjugation, the NHS ester was reacted first, excess reagent was removed, and then the sulfhydryl-containing molecule was added. SMCC was dissolved in dimethyl sulfoxide or and added to the reaction mixture at a final solvent concentration of 10% to minimize detrimental affects to the protein.
Rabbit immunization was performed at Alpha Diagnostics and has been described previously (Tang et al., 1999). For purposes of comparison, an anti-peptide antibody generated against SULT1A1 at the same sequence (beginning at amino acid 147: KVHPEPGTWDSF) was also produced (data not shown). Antibodies were used in assays without further purification.
Cytosol Preparation. Cytosolic fractions were prepared from human tissues using the same method for preparation of microsomal fractions that has been described previously (Tang et al., 1999) with minor modifications. Briefly, the supernatant from the 105,000g spin was aliquoted into tubes, flash-frozen in liquid nitrogen, and stored at eC80°C until Western blot analysis was performed.
Western Blot Analysis. Proteins (100 e蘥 of each sample) were electrophoresed and transferred to nitrocellulose, blocked, and incubated overnight with the anti-SULT1A2 antibody. The blots were developed using reagents from Pierce Biotechnology according to the manufacturer's directions using a horseradish peroxidase-conjugated secondary antibody. Chemiluminescent detection was performed with Chemi-glow (Alpha Innotech, San Leandro, CA) detection reagents. The blots were visualized with an Alpha Innotech model 8900 Imager. Duplicate blots were performed with the SULT1A1 anti-peptide antibody.
Tissue Lysate Arrays. DiscoverLight array kits were used to screen for proteins showing immunoreactivity with antibodies to either SULT1A1 or SULT1A2, according to the manufacturer's directions. In cases in which there was potential immunoreactivity with anti-SULT1A2, individual tissue lysates were purchased from Pierce Biotechnology and Western blotted to confirm immunoreactivity.
Results
SULT1A2-Specific Peptide Selection and Characterization of Antisera. Sequence alignment data from Fig. 1 was used to select an antigenic epitope with a sequence significantly different from either SULT1A1 or SULT1A3, the two isoforms with structural similarity to SULT1A2. The peptide was predicted to have a high degree of hydrophilicity, surface probability, and antigenicity according to computer modeling (Fig. 2). The purity of the peptide, as determined by high-performance liquid chromatography, was greater than 85% (data not shown). The peptide was coupled to the carrier protein (keyhole limpet hemocyanin) via the cysteine residue at the C terminus of the peptide and was used to immunize New Zealand White rabbits. Antisera produced against this peptide had a titer of approximately 105, as determined by enzyme-linked immunosorbent assay.
Specificity and Sensitivity of the SULT1A2 Antibody. Potential cross-reactivity of the antibody with other known SULT isoforms was investigated using recombinant enzymes. Immunoblotting was performed using recombinant SULT1A1, SULT1A2, SULT1A3, SULT1E1, SULT2A1, and control protein supplied by the manufacturer. The dilution of the primary antibody was 1:200, and the blots were incubated overnight at 4°C; the secondary antibody dilution was 1:10,000. Figure 3A shows that when 50 e蘥 of recombinant enzyme is used, there is no apparent cross-reactivity with these other SULT isoforms.
A 5- to 50-e蘥 sample of recombinant SULT1A2 was transferred to nitrocellulose and exposed to the SULT1A2 antibody for 2 h at room temperature. As shown in Fig. 3B, immunoreactive bands were detected at each concentration of recombinant enzyme tested. Other experiments using lesser amounts of enzyme (0.1eC5 e蘥) indicated that SULT1A2 recombinant protein could be detected at levels as low as 0.5 e蘥. When recombinant SULT1A1 was probed with the anti-SULT1A1 antibody, the limit of detection of this antibody was identical with that found for the SULT1A2 antibody/recombinant enzyme (data not shown).
Screening of Tissues for SULT1A2 Expression. A 100-e蘥 sample of each cytosol prepared from several human tissues was immunoblotted and probed for the presence of SULT1A2. Table 1 lists the different tissues and the numbers of individual samples from each type of tissue examined. Examination of 52 individual liver samples failed to show immunoreactivity for SULT1A2. Likewise, there was no immunoreactivity in 35 colon samples, 34 prostate samples, and 32 small intestine samples. Although fewer numbers of individual samples from other tissues were available, examination of these tissues also failed to detect SULT1A2 expression.
SULT1A2 Expression Screening Using the DiscoverLight Array Kit. To expand the number of tissues examined and to include tumor tissue, the DiscoverLight tissue array was probed for immunoreactivity with the SULT1A2 antibody (Fig. 4B). The array consists of pooled samples from 17 different organs and includes both tumor and normal tissue from each organ. These samples are arrayed on a nitrocellulose membrane that contains positive control elements in addition to the samples. As a control, a duplicate array was probed for immunoreactivity with an antibody specific for SULT1A1 (Fig. 4A). Table 2 lists the array positions, tissue types, and a summary of immunoreactivity with either SULT1A1 or SULT1A2 antibodies. SULT1A1 expression was found, to varying degrees, in most of the tissues on the array. SULT1A2 expression, on the other hand, was not high in any tissue, but potential positives were detected with samples from the cervix, skin, and stomach. To confirm these potential positives, Western blot analysis was performed using 30 e蘥 of the tissue lysates and the most sensitive detection reagent available (SuperSignal West Femto; Pierce Biotechnology), which can detect low femtogram amounts of proteins. A very faint immunoreactive band was evident in the skin tumor lysate, but the other lysates tested showed no immunoreactivity (Fig. 4C).
Discussion
SULT1A2 is a member of the phenol sulfotransferase subfamily of cytosolic sulfotransferases. Transcript for this isoform has been identified, but to date, detection of SULT1A2 as a protein in human tissues has not been reported. Using an anti-peptide antibody specific for SULT1A2, we examined more than 200 cytosolic fractions from 10 different tissues and found no evidence of immunoreactive protein. When a commercially available tissue lysate array containing 17 normal and 17 tumor pooled tissue lysate elements was screened for SULT1A2, potential immunopositive tissues were identified. Western blotting was performed to confirm this finding, but immunostaining in the potentially positive tissues was too low to be conclusive.
Previous studies have shown that SULT1A2 shares similar substrate specificity with SULT1A1, although Km values are generally an order of magnitude higher with SULT1A2 compared with SULT1A1 (Raftogianis et al., 1999). Raftogianis et al. determined that in the liver, SULT1A2 did not contribute to the "thermostable phenol sulfotransferase" phenotype because of the lack of detectable activity when assay conditions were optimal for SULT1A2. When SULT1A2 alloenzymes were expressed in S. typhimurium and metabolic activation of several aromatic amines was examined, in some cases, SULT1A2-expressing cells were more efficient at activation than SULT1A1 (Meinl et al., 2002; Glatt and Meinl, 2004b). Additionally, when SULT1A2 is expressed in Chinese hamster V79 cells, it is capable of bioactivating 3-nitrobenzanthrone and its metabolites (Arlt et al., 2003). These findings suggest a potential role of SULT1A2, if it is expressed as a functional protein, in the chemical carcinogenesis of these agents. However, it has been suggested that a splicing defect in SULT1A2 prevents its expression as a functional protein, in which case SULT1A2 would effectively be considered a pseudogene (Dooley et al., 2000).
SULT1A2 genetic polymorphisms have been identified (Raftogianis et al., 1999), and recently, these polymorphisms have been investigated in relation to cancer risk. Studies by Peng et al. (2003) showed that, as previously reported, SULT1A11 and SULT1A21 alleles were in positive linkage disequilibrium but showed no association with risk of liver, colon, lung, oral, gastric, renal, or cervical cancer in a Taiwanese population. Another study examining breast cancer risk in Chinese women indicated that SULT1A2 polymorphisms were not associated with breast cancer risk but might be associated with early onset of disease (Hou et al., 2002). However, because polymorphisms in SULT1A1 and SULT1A2 are linked, it is possible that the observed association is actually caused by SULT1A1 alleles.
Given these emerging molecular epidemiological studies, it is important to determine whether SULT1A2 is expressed as a protein in human tissues. We therefore designed an antipeptide antibody specific for human SULT1A2 and surveyed several normal and tumor tissues. In cytosols prepared from human tissues, we did not detect SULT1A2 immunoreactivity. However, when a tissue lysate array was probed with anti-SULT1A2 and developed with a chemiluminescent substrate designed to detect immunoreactive proteins present in the femtogram range, we identified potential positives in cervix, skin, and stomach. To confirm these findings, we then performed Western blot analysis of both tumor and normal lysates from these tissues. We found that there was a weak immunopositive reaction in skin tumor but not in the other lysates. However, the low degree of reaction was not convincing of expression. In any case, the antibody failed to detect expression of SULT1A2 in all other cytosols and tissue lysates examined. Anti-peptide antibodies in general are not as avid as antibodies directed against whole proteins, so it is possible that some very low levels of SULT1A2 could be expressed under certain circumstances. However, the antipeptide antibody directed against SULT1A1 revealed abundant expression of this isoform. Little is known about the regulation of SULT1A2; therefore, it is possible that expression is induced in some physiological states. It does not seem to be hormonally regulated because we detected no gender-related differences in expression.
In summary, we did not detect SULT1A2 expression in any tissue using a specific anti-peptide SULT1A2 antibody. These findings suggest that although SULT1A2 expression might occur very occasionally, given the fact that SULT1A1 is expressed at much higher levels and is much more metabolically active than SULT1A2, this enzyme is not likely to be physiologically relevant under normal conditions. Therefore, caution should be used in interpreting SULT1A2 genotype/disease association studies.
doi:10.1124/mol.104.006171.
References
Arlt VM, Glatt H, Muckel E, Pabel U, Sorg BL, Seidel A, Frank H, Schmeiser HH, and Phillips DH (2003) Activation of 3-nitrobenzanthrone and its metabolites by human acetyltransferases, sulfotransferases and cytochrome P450 expressed in Chinese hamster V79 cells. Int J Cancer 105: 583eC592.
Bernatowicz MS and Matsueda GR (1986) Preparation of peptide-protein immunogens using N-succinimidyl bromoacetate as a heterobifunctional crosslinking reagent. Anal Biochem 155: 95eC102.
Chou HC, Lang NP, and Kadlubar FF (1995) Metabolic activation of N-hydroxy arylamines and N-hydroxy heterocyclic amines by human sulfotransferase(s). Cancer Res 55: 525eC529.
Dooley TP, Haldeman-Cahill R, Joiner J, and Wilborn TW (2000) Expression profiling of human sulfotransferase and sulfatase gene superfamilies in epithelial tissues and cultured cells. Biochem Biophys Res Commun 277: 236eC245.
Falany CN, Xie X, Wang J, Ferrer J, and Falany JL (2000) Molecular cloning and expression of novel sulphotransferase-like cDNAs from human and rat brain. Biochem J 346: 857eC864.
Glatt H and Meinl W (2004a) Pharmacogenetics of soluble sulfotransferases (SULTs). Naunyn Schmiedeberg's Arch Pharmacol 369: 55eC68.
Glatt H and Meinl W (2004b) Use of genetically manipulated Salmonella typhimurium strains to evaluate the role of sulfotransferases and acetyltransferases in nitrofen mutagenicity. Carcinogenesis 25: 779eC786.
Hou MF, Chen ST, Chen JC, Yeh KT, Lee TP, Chen CM, Shih MC, Lin CC, and Chang JG (2002) Sulfotransferase 1A22 is a risk factor for early-onset breast cancer. Int J Mol Med 10: 609eC612.
Jakoby WB and Ziegler DM (1990) The enzymes of detoxication. J Biol Chem 265: 20715eC20718.
Meinl W, Meerman JH, and Glatt H (2002) Differential activation of promutagens by alloenzymes of human sulfotransferase 1A2 expressed in Salmonella typhimurium. Pharmacogenetics 12: 677eC689.
Minchin RF, Ilett KF, Teitel CH, Reeves PT, and Kadlubar FF (1992) Direct O-acetylation of N-hydroxy arylamines by acetylsalicylic acid to form carcinogen-DNA adducts. Carcinogenesis 13: 663eC667.
Peng CT, Chen JC, Yeh KT, Wang YF, Hou MF, Lee TP, Shih MC, Chang JY, and Chang JG (2003) The relationship among the polymorphisms of SULT1A1, 1A2 and different types of cancers in Taiwanese. Int J Mol Med 11: 85eC89.
Raftogianis RB, Wood TC, and Weinshilboum RM (1999) Human phenol sulfotransferases SULT1A2 and SULT1A1: genetic polymorphisms, allozyme properties and human liver genotype-phenotype correlations. Biochem Pharmacol 58: 605eC616.
Tang YM, Chen GF, Thompson PA, Lin DX, Lang NP, and Kadlubar FF (1999) Development of an antipeptide antibody that binds to the C-terminal region of human CYP1B1. Drug Metab Dispos 27: 274eC280.