- Email: [email protected]
Human testicular glutathione S-transferases: insights into tissue-specific expression of the diverse subunit classes
Chemico-Biological Interactions 111 – 112 (1998) 103 – 112
Human testicular glutathione S-transferases: insights into tissue-specific expression of the diverse subunit classes Irving Listowsky *, Jonathan D. Rowe, Yury V. Patskovsky, Tatyana Tchaikovskaya, Naoaki Shintani, Elena Novikova, Edward Nieves Department of Biochemistry, Albert Einstein College of Medicine, 1300 Morris Park A6enue, Bronx, NY 10461, USA
Abstract Cytosolic glutathione S-transferase (GST) subunits from human testis were resolved by HPLC and unambiguously identified by combined use of peptide sequence-specific antisera and electrospray ionization mass spectrometry (ESI MS). Allelic variants of hGSTP1, hGSTM1 and hGSTA2 were distinguished on the basis of observed differences in their molecular masses. Relative amounts of the multiple different subunit types in various human tissues were determined from HPLC profiles. From this type of analysis, tissues from hGSTM1 null allele individuals were readily discerned at the protein level; liver was the only tissue in which the hGSTM1 subunit was the major m-class GST. hGSTM4 and hGSTM5 subunits were found at very low levels in all tissues examined. By far the tissue richest in the unique hGSTM3 subunit was testis, although brain also has significant levels. © 1998 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Glutathione transferase; Mass spectrometry; Testis; HPLC
* Corresponding author. Tel.: +1 718 4302276; fax: + 1 718 8920703. 0009-2797/98/$19.00 © 1998 Elsevier Science Ireland Ltd. All rights reserved. PII S0009-2797(97)00154-3
104
I. Listowsky et al. / Chemico-Biological Interactions 111–112 (1998) 103–112
1. Introduction Members of mammalian glutathione S-transferase (GST) gene superfamilies are expressed in a discrete cell type-specific manner. Accordingly, studies on the multiple GSTs could serve as a paradigm for understanding general mechanisms of tissue-specific gene expression. Although human and rodent GSTs have been studied extensively [1] and the species-independent classification of cytosolic mammalian GSTs into a, m, p and u classes [2] is well accepted, fundamental gaps remain in our knowledge of the system. For instance, it is not yet feasible to establish direct relationships among individual or specific subunit types within a class of GST for different species. This is particularly true for the human GST family where some of the forms predicted to exist on the basis of cDNA sequences have not yet been identified in human tissue extracts. Moreover, although rodent GSTs are inducible by hormones and xenobiotics [3], it is not yet clear whether similar inductions operate for the human enzymes. Accordingly, this study was designed to identify and characterize GST subunits and survey their distribution in human tissues. Evidently subunit profiles vary, but are usually characteristic of their tissue of origin even from different individuals. In particular, specimens of testis were shown to contain multiple different GST types and were especially rich sources of the hGSTM3 subunit.
2. Materials and methods
2.1. Human GSTs Human tissue from donors with no evidence of preexisting diseases was snap frozen less than 12 h post mortem. Reagents were high purity grade from commercial sources. GSTs were purified from the indicated tissues using GSH affinity matrices [4]. Enzymatic activities were determined using 1.0 mM GSH and 1.0 mM 1-chloro 2,4-dinitro-benzene as substrates.
2.2. HPLC analysis A Hewlett Packard HPLC 1090 system was used to analyze human GST subunit distribution. Subunits were resolved on a Vydac (4.6 × 150 mm, 5 m bore) C4 column using acetonitrile (ACN), containing 0.1% trifluoroacetic acid (TFA) and water containing 0.08% TFA as solvents. Unless otherwise indicated, gradients were initiated at 20% ACN, and the percentage of ACN was increased to 40% at 10 min followed by a linear increase to 60% ACN between 20 and 60 min at a flow rate of 1 ml/min. Protein peaks were detected with a diode array detector by monitoring absorbancies at 214 and 280 nm. Peak area integration was carried out using HP Chem Station Software. HPLC fractions were collected by hand and either desiccated in a Speed-Vac or used directly for mass spectrometric analysis (see below).
I. Listowsky et al. / Chemico-Biological Interactions 111–112 (1998) 103–112
105
2.3. Mass spectrometry An API-III triple-quadruple mass spectrometer (PE-SCIEX, Ont., Canada) and the SCIEX Ionspray interface with nitrogen as the nebulizer gas was used for mass spectral analysis. An ionspray voltage of : 3600 V and an orifice voltage of 85 V was used. Mass spectral analysis of samples was carried out by on-line HPLC connected to the mass spectrometer (LC-MS) or by analysis of individual HPLC peaks. Statistical analyses of experimental subunit molecular masses were based on data from at least three different preparations. 3. Results
3.1. Testicular GSTs GSTs were purified from adult human testis which is a rich source for the proteins. The testicular GST subunits were resolved by reversed phase HPLC and
Fig. 1. Human testicular GST subunits. GSTs were isolated from the testis of a 68-year-old black male, and subunits resolved by reversed phase HPLC. Individual subunits were screened by immunoblotting reactions using either polyclonal antibodies for p and a-class GSTs or peptide sequence-specific antisera for the m-class isoforms. A shallow gradient of acetonitrile of 40 – 60% over a period of 80 min was used in this case to resolve the two GSTP1 subunit peaks P1 and P1%. The indicated subunits (M2 = GSTM2, M1(b) =GSTM1b, etc. according to nomenclature in [14] were unambiguously identified by ESI MS.
106
I. Listowsky et al. / Chemico-Biological Interactions 111–112 (1998) 103–112
Fig. 2. Electrospray ionization mass spectrum of the GSTP1 components. The HPLC fraction containing the split GSTP1 peak of Fig. 2 was collected and subjected to ESI MS. A reconstructed mass spectrum is shown in the inset to indicate molecular weight differences between the two major components.
individual components identified by immunoblotting with peptide sequence-specific antisera [5]. The subunits were further characterized by electrospray ionization mass spectrometry (ESI MS) to verify their primary structures and to detect variants. A representative HPLC pattern of testis GSTs is shown in Fig. 1. This particular specimen was selected to illustrate allelic variants of hGSTP1 and hGSTM1 subunits. A shallow acetonitrile gradient was used to partially resolve two components of the hGSTP1 peak (P1 and P1% in Fig. 1), which had molecular mass differences of approximately 12 amu (Fig. 2). These results are consistent with the suggestion of Ali-Osman et al. that there are GSTP1 gene variants [6]. The molecular mass of the component identified as the hGSTM1b subunit was 25568 (N172) which was readily distinguished from that of the more frequently occurring allele, hGSTM1a (molecular mass 25582; K172) by ESI MS analysis. An HPLC unit linked on-line to a mass spectral analyzer (LC-MS) was used to determine the molecular masses of the GST subunits and representative LCMS results for m-class GSTs from tissues are shown in Fig. 3. The molecular masses of the major components identified by peptide-specific antisera were: GSTM1a, 255819 3.3; GSTM1b, 25567; GSTM2, 256179 2.4; GSTM3, 264709 1.7; GSTM4, 254749 0.9; GSTM5, 255479 3.2; GSTP1, 232249 1.5; GSTA1, 2554493.0; GSTA2, 255919 4.2 and 255799 2.8. In comparison
Fig. 3. Electrospray ionization mass spectra of m-class GSTs. The ESI MS data were obtained using on-line HPLC resolution of testicular m-class subunits according to procedures of Fig. 2. The GSTM5 subunit was from a brain specimen. The reconstructed mass spectra are shown in the insets.
I. Listowsky et al. / Chemico-Biological Interactions 111–112 (1998) 103–112 107
I. Listowsky et al. / Chemico-Biological Interactions 111–112 (1998) 103–112
Fig. 4.
108
I. Listowsky et al. / Chemico-Biological Interactions 111–112 (1998) 103–112
109
with molecular masses based on deduced sequences of these subunits, subunits GSTM3, GSTM4, GSTA1 and GSTA2 are predicted to have acetylated N-termini.
3.2. GSTs in other human tissues HPLC patterns for a series of GSTs purified from representative matched and individual tissue specimens are shown in Fig. 4. Several components were incompletely resolved under these conditions. For instance, in tissues containing high levels of GSTA1 (peak 6), the GSTM3 subunit (peak 7) could be obscured. Testis is the richest overall source of cytosolic GSTs; previous immunoquantitation studies that reported highest levels in human liver, probably did not detect hGSTM3 subunits. The hGSTM3 subunit levels in testis were usually at least 7–10-fold greater than in any other human tissue. Brain is rich in GSTP1, with GSTM2 and GSTM3 subunits as the next most abundant forms. The a-class GSTs are minor forms in brain, and the presence of low levels of GSTA1 allowed for the prominent display of GSTM3 (Fig. 4). Brain (cerebral cortex) cytosols have approximately 3-fold greater amounts of GST than pituitary which in turn has much lower levels of the GSTM1 subunit. The liver GST isoenzyme composition is in accordance with immunoblots inasmuch as the a-class and GSTM1 subunits predominate and other subunit types occur as very minor forms. Adrenal GSTs consist mostly GSTP1, GSTA1 and GSTM1 subunits. Kidney GSTs are composed mostly of GSTP1, GSTA1 and GSTA2 subunits and have GSTM1, GSTM2 and GSTM3 as minor forms. Testes, which have the greatest overall levels of GSTs, are particularly rich in GSTM3 and GSTA1. In most tissues shown in Fig. 4, GSTM2, GSTM4 and GSTM5 subunits are minor components of the total GST composition. The GSTP1 subunit is the predominant subunit in heart, lung and brain (from 4–10 mg/mg protein). All m-class GSTs (GSTM1 – GSTM5) were also detected at low levels in lung with GSTM1 (when present) as the most abundant of these (Fig. 4). The absence of the GSTM1 peak is striking in the HPLC profiles of GSTs from tissue of persons with the GSTM1 -null genotype. All tissues tested from those individuals were devoid of GSTM1 subunits. This technique may be applied readily for identification of the GSTM1 -null phenotype individuals at the level of protein
Fig. 4. HPLC profiles of GSTs isolated from human tissues. GSTs were purified from the individual tissues by GSH-affinity chromatography (see Section 2). The purified GSTs were applied to a Vydac C4 HPLC column and individual GST subunits were eluted from the column over an increasing gradient of ACN as described in Section 2. Peaks were assigned numbers 1 – 8 where positive assignment of GST subunits based upon immunological criteria using specific antisera, mobility on SDS gels, and ESI MS data could be obtained. Peak 1, found in some samples, reacted with the GSTM5-specific antibody; peak 2, GSTP1; peak 3, GSTM2; peak 4, GSTM1; peak 5, GSTM5 and GSTM4; peak 6, GSTA1; peak 7, GSTM3; and peak 8, GSTA2. The symbols indicate tissue from the same individual. (*, 48-year-old black male; **, 1 month old Caucasian male; ", 31-year-old Caucasian female and the liver in the upper left panel was from a 34-year-old black female; , 48-year-old Caucasian male). GSTM1 ( + ) and ( − ) denote the presence or absence of GSTM1 subunits (peak 4).
110
I. Listowsky et al. / Chemico-Biological Interactions 111–112 (1998) 103–112
expression. Liver was the only tissue with substantial levels of the GSTM1 subunit, (approximately 10 mg/mg protein), and in tissues such as liver and adrenal, the GST subunit profiles are drastically altered by the absence of GSTM1 subunits (Fig. 4).
4. Discussion
4.1. Tissue distribution Because of the complexity of the system, it has been a formidable task to compile a detailed description of human GST subunits in tissue specimens [7]. In many earlier studies on human GSTs, subunit assignments were fragmentary and based on imprecise criteria. For instance, early on GSTs were categorized as acidic (p), near neutral (m) and basic (a) forms [8], or simply on the basis of electrophoretic mobilities in SDS gels with p as the fastest migrating subunit followed by a and then m. Such classifications could be misleading since the electrophoretic mobilities of GSTM4 and GSTM2 overlap with those of the a-class subunits, and since m-class subunits in general have pl’s that encompass acidic and basic ranges. Moreover, some of the modifications such as oxidation and glutathionylation that occur in the course of purification and storage of the proteins may further confound analyses based on charge differences among the subunits. Recent advances in technology have facilitated the use of ESI MS for characterization of proteins [9], and several studies have examined primary structures of rat and mouse liver GSTs [10,11]. In the present study, application of ESI MS for this system has clarified several complex issues about human GSTs and provides a basis to plan and interpret future studies of GST functions. Of greater significance, the ESI MS data in conjunction with information derived from the use of peptide-specific antisera, provide unequivocal identification of human GST subunits resolved by HPLC methods. The on-line HPLC-ESI MS system detects minor subunits that probably occur in the cell primarily in heterodimeric combination with other subunit types. Most of the previous high resolution HPLC analyses of human GSTs have focused on liver forms, with some studies on other tissue. Some GST subunit peaks in those studies were unidentified. On the basis of such analysis, appropriate strategies may therefore be devised to determine relative amounts of particular subunit types and their tissue distribution. Specimens devoid of GSTM1 subunits from ‘null’ phenotype individuals are readily identified (Fig. 4). Thus, it is now possible to conduct rapid and reliable surveys of GST subunit distribution, and to observe their profiles under various conditions including changes that may occur during development, in malignant and other diseased tissue, and to determine whether select GSTs are induced by drugs or xenobiotics.
4.2. Inter-species comparisons Combinations of several criteria may be applied for comparison of subclasses of GST subunits from different mammalian species. These include sequence homolo-
I. Listowsky et al. / Chemico-Biological Interactions 111–112 (1998) 103–112
111
gies, similarities in tissue expression patterns, catalytic specificities and other properties. Each of these criteria alone may not suffice to establish which subunits in the different species are analogous, however in some cases inferences may be made from combinations of criteria. For example, the human hGSTM2 subunit is primarily found in muscle, brain and testis but not in liver or kidney (Fig. 4). A parallel tissue distribution pattern has been observed for the rat rGSTM3 (Yb3) subunit [12]. Moreover hGSTM2 and rGSTM3 have 84% sequence identities and both contain a unique signature pentapeptide sequence, ERNQV at residues 164–168. Similarly, the novel hGSTM3 subunit is probably comparable to the rat Yo and mouse mGSTM5 subunits on the basis of sequence homologies and tissue distribution patterns (Listowsky, Rowe, and Tchaikovskaya, unpublished observations).
4.3. Testis GSTs The testicular GSTs provide insight into potential differential tissue-specific functions of the proteins. Thus testis is by far the tissue richest in the hGSTM3 subunit (Figs. 1 and 4). It is likely that GSTs in testis serve in a protective capacity for germ cells from the action of reactive chemicals, hydroperoxides and other products of oxidative stress. In particular, fatty acid oxidations induced by leukocytes and/or defective sperm in semen, results in disruption of sperm motility, capacitation and failure of sperm – oocyte fusion reactions [13]. It is logical to assume that the presence of high levels of hGSTM3 could prevent or ameliorate consequences of these noxious reactions by virtue of particular GST catalytic and binding functions. This GST form may be a subject for future studies on targeting agents that could enhance fertility or as a drug target for male contraception.
Acknowledgements This work was supported by grant CA42448 from the National Cancer Institute, Cancer Center Core 5P30-CA 13330 and Liver Center Core 5P30-DK41296. JDR was supported by genetics and biochemistry graduate training grant T32 GM07128 from the National Institutes of Health. ESI MS, peptide synthesis and protein sequence analysis were carried out in the Laboratory for Macromolecular Analysis of the Albert Einstein College of Medicine. We appreciate the valuable suggestions of Dr John D. Hayes.
References [1] J.D. Hayes, D.J. Pulford, The glutathione S-transferase supergene family: regulation of GST and the contribution of the isoenzymes to cancer chemoprotection and drug resistance, Crit. Rev. Biochem. Mol Biol. 30 (1995) 445–600. [2] B. Mannervik, P. Alin, C. Guthenberg, Identification of three classes of cytosolic glutathione transferases common to several mammalian species: Correlation between structural data and enzymatic properties, Proc. Natl. Acad. Sci. 82 (1985) 7202 – 7206.
112
I. Listowsky et al. / Chemico-Biological Interactions 111–112 (1998) 103–112
[3] T.H. Rushmore, C.B. Pickett, Glutathione S-transferases, structure, regulation, and therapeutic implications, J. Biol. Chem. 268 (1993) 11475 – 11478. [4] H. Jenson, P. Alin, B. Mannervik, Glutathione transferase isoenzymes from rat liver cytosol, Methods Enzymol. 113 (1985) 504–507. [5] Y. Takahashi, E.A. Campbell, Y. Hirata, T. Takayama, I. Listowsky, A basis for differentiating among the multiple human mu-glutathione S-transferases and molecular cloning of brain GSTM5, J. Biol. Chem. 268 (1993) 8893–8898. [6] F. Ali-Osman, O. Akande, G. Antoun, J.X. Mao, J. Buolamuini, Molecular cloning, characterization, and expression in Escherichia coli of full-length cDNAs of three human glutathione S-transferase Pi gene variants, J. Biol. Chem. 272 (1997) 10004 – 10012. [7] B. Mannervik, M. Widersten, Human glutathione transferases: classification, tissue distribution, structure and functional properties, in: G.M. Pacifici, G.N. Fracchia (Eds.), Advances in Drug Metabolism in Man, Europe Com. (UK), Bernan assoc. publisher, 1995, pp. 407 – 459. [8] B. Mannervik, U. Danielson, Glutathione transferases-structure and catalytic activity, Crit. Rev. Biochem. Mol. Biol. 23 (1988) 283– 337. [9] K. Biemann, The coming of age of mass spectrometry in peptide and protein chemistry, Protein Science 4 (1995) 1920–1927. [10] P. Rouim, L. Debrauwer, J. Tulliez, Electrospray ionization-mass spectrometry as a tool for characterization of glutathione S-transferase isozymes, Anal. Biochem. 229 (1995) 304 – 312. [11] H.I. Yeh, C.H. Hsieh, L.Y. Wang, S.P. Tsai, H.Y. Hsu, M.F. Tam, Mass spectrometric analysis of rat liver cytosolic glutathione S-transferases: modifications are limited to N-terminal processing, Biochem. J. 308 (1995) 69–75. [12] M. Abramovitz, I. Listowsky, Selective expression of a unique glutathione S-transferase Yb3 gene in rat brain, J. Biol. Chem. 262 (1987) 7770 – 7773. [13] R.J. Aitken, A free radical theory of male infertility, Reprod. Fertil. Dev. 6 (1994) 19 – 24. [14] B. Mannervik, Y.C. Awashi, P.G. Board, J.D. Hayes, C.D. Dilio, B. Ketterer, I. Listowsky, R. Morgenstein, M. Marumatsu, W.R. Pearson, C.B. Pickett, K. Sato, M. Widersten, C.R. Wolf, Nomenclature for human glutathione transferases, Biochem. J. 282 (1992) 305 – 306.
.
Human testicular glutathione S-transferases: insights into tissue-specific expression of the diverse subunit classes Irving Listowsky *, Jonathan D. Rowe, Yury V. Patskovsky, Tatyana Tchaikovskaya, Naoaki Shintani, Elena Novikova, Edward Nieves Department of Biochemistry, Albert Einstein College of Medicine, 1300 Morris Park A6enue, Bronx, NY 10461, USA
Abstract Cytosolic glutathione S-transferase (GST) subunits from human testis were resolved by HPLC and unambiguously identified by combined use of peptide sequence-specific antisera and electrospray ionization mass spectrometry (ESI MS). Allelic variants of hGSTP1, hGSTM1 and hGSTA2 were distinguished on the basis of observed differences in their molecular masses. Relative amounts of the multiple different subunit types in various human tissues were determined from HPLC profiles. From this type of analysis, tissues from hGSTM1 null allele individuals were readily discerned at the protein level; liver was the only tissue in which the hGSTM1 subunit was the major m-class GST. hGSTM4 and hGSTM5 subunits were found at very low levels in all tissues examined. By far the tissue richest in the unique hGSTM3 subunit was testis, although brain also has significant levels. © 1998 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Glutathione transferase; Mass spectrometry; Testis; HPLC
* Corresponding author. Tel.: +1 718 4302276; fax: + 1 718 8920703. 0009-2797/98/$19.00 © 1998 Elsevier Science Ireland Ltd. All rights reserved. PII S0009-2797(97)00154-3
104
I. Listowsky et al. / Chemico-Biological Interactions 111–112 (1998) 103–112
1. Introduction Members of mammalian glutathione S-transferase (GST) gene superfamilies are expressed in a discrete cell type-specific manner. Accordingly, studies on the multiple GSTs could serve as a paradigm for understanding general mechanisms of tissue-specific gene expression. Although human and rodent GSTs have been studied extensively [1] and the species-independent classification of cytosolic mammalian GSTs into a, m, p and u classes [2] is well accepted, fundamental gaps remain in our knowledge of the system. For instance, it is not yet feasible to establish direct relationships among individual or specific subunit types within a class of GST for different species. This is particularly true for the human GST family where some of the forms predicted to exist on the basis of cDNA sequences have not yet been identified in human tissue extracts. Moreover, although rodent GSTs are inducible by hormones and xenobiotics [3], it is not yet clear whether similar inductions operate for the human enzymes. Accordingly, this study was designed to identify and characterize GST subunits and survey their distribution in human tissues. Evidently subunit profiles vary, but are usually characteristic of their tissue of origin even from different individuals. In particular, specimens of testis were shown to contain multiple different GST types and were especially rich sources of the hGSTM3 subunit.
2. Materials and methods
2.1. Human GSTs Human tissue from donors with no evidence of preexisting diseases was snap frozen less than 12 h post mortem. Reagents were high purity grade from commercial sources. GSTs were purified from the indicated tissues using GSH affinity matrices [4]. Enzymatic activities were determined using 1.0 mM GSH and 1.0 mM 1-chloro 2,4-dinitro-benzene as substrates.
2.2. HPLC analysis A Hewlett Packard HPLC 1090 system was used to analyze human GST subunit distribution. Subunits were resolved on a Vydac (4.6 × 150 mm, 5 m bore) C4 column using acetonitrile (ACN), containing 0.1% trifluoroacetic acid (TFA) and water containing 0.08% TFA as solvents. Unless otherwise indicated, gradients were initiated at 20% ACN, and the percentage of ACN was increased to 40% at 10 min followed by a linear increase to 60% ACN between 20 and 60 min at a flow rate of 1 ml/min. Protein peaks were detected with a diode array detector by monitoring absorbancies at 214 and 280 nm. Peak area integration was carried out using HP Chem Station Software. HPLC fractions were collected by hand and either desiccated in a Speed-Vac or used directly for mass spectrometric analysis (see below).
I. Listowsky et al. / Chemico-Biological Interactions 111–112 (1998) 103–112
105
2.3. Mass spectrometry An API-III triple-quadruple mass spectrometer (PE-SCIEX, Ont., Canada) and the SCIEX Ionspray interface with nitrogen as the nebulizer gas was used for mass spectral analysis. An ionspray voltage of : 3600 V and an orifice voltage of 85 V was used. Mass spectral analysis of samples was carried out by on-line HPLC connected to the mass spectrometer (LC-MS) or by analysis of individual HPLC peaks. Statistical analyses of experimental subunit molecular masses were based on data from at least three different preparations. 3. Results
3.1. Testicular GSTs GSTs were purified from adult human testis which is a rich source for the proteins. The testicular GST subunits were resolved by reversed phase HPLC and
Fig. 1. Human testicular GST subunits. GSTs were isolated from the testis of a 68-year-old black male, and subunits resolved by reversed phase HPLC. Individual subunits were screened by immunoblotting reactions using either polyclonal antibodies for p and a-class GSTs or peptide sequence-specific antisera for the m-class isoforms. A shallow gradient of acetonitrile of 40 – 60% over a period of 80 min was used in this case to resolve the two GSTP1 subunit peaks P1 and P1%. The indicated subunits (M2 = GSTM2, M1(b) =GSTM1b, etc. according to nomenclature in [14] were unambiguously identified by ESI MS.
106
I. Listowsky et al. / Chemico-Biological Interactions 111–112 (1998) 103–112
Fig. 2. Electrospray ionization mass spectrum of the GSTP1 components. The HPLC fraction containing the split GSTP1 peak of Fig. 2 was collected and subjected to ESI MS. A reconstructed mass spectrum is shown in the inset to indicate molecular weight differences between the two major components.
individual components identified by immunoblotting with peptide sequence-specific antisera [5]. The subunits were further characterized by electrospray ionization mass spectrometry (ESI MS) to verify their primary structures and to detect variants. A representative HPLC pattern of testis GSTs is shown in Fig. 1. This particular specimen was selected to illustrate allelic variants of hGSTP1 and hGSTM1 subunits. A shallow acetonitrile gradient was used to partially resolve two components of the hGSTP1 peak (P1 and P1% in Fig. 1), which had molecular mass differences of approximately 12 amu (Fig. 2). These results are consistent with the suggestion of Ali-Osman et al. that there are GSTP1 gene variants [6]. The molecular mass of the component identified as the hGSTM1b subunit was 25568 (N172) which was readily distinguished from that of the more frequently occurring allele, hGSTM1a (molecular mass 25582; K172) by ESI MS analysis. An HPLC unit linked on-line to a mass spectral analyzer (LC-MS) was used to determine the molecular masses of the GST subunits and representative LCMS results for m-class GSTs from tissues are shown in Fig. 3. The molecular masses of the major components identified by peptide-specific antisera were: GSTM1a, 255819 3.3; GSTM1b, 25567; GSTM2, 256179 2.4; GSTM3, 264709 1.7; GSTM4, 254749 0.9; GSTM5, 255479 3.2; GSTP1, 232249 1.5; GSTA1, 2554493.0; GSTA2, 255919 4.2 and 255799 2.8. In comparison
Fig. 3. Electrospray ionization mass spectra of m-class GSTs. The ESI MS data were obtained using on-line HPLC resolution of testicular m-class subunits according to procedures of Fig. 2. The GSTM5 subunit was from a brain specimen. The reconstructed mass spectra are shown in the insets.
I. Listowsky et al. / Chemico-Biological Interactions 111–112 (1998) 103–112 107
I. Listowsky et al. / Chemico-Biological Interactions 111–112 (1998) 103–112
Fig. 4.
108
I. Listowsky et al. / Chemico-Biological Interactions 111–112 (1998) 103–112
109
with molecular masses based on deduced sequences of these subunits, subunits GSTM3, GSTM4, GSTA1 and GSTA2 are predicted to have acetylated N-termini.
3.2. GSTs in other human tissues HPLC patterns for a series of GSTs purified from representative matched and individual tissue specimens are shown in Fig. 4. Several components were incompletely resolved under these conditions. For instance, in tissues containing high levels of GSTA1 (peak 6), the GSTM3 subunit (peak 7) could be obscured. Testis is the richest overall source of cytosolic GSTs; previous immunoquantitation studies that reported highest levels in human liver, probably did not detect hGSTM3 subunits. The hGSTM3 subunit levels in testis were usually at least 7–10-fold greater than in any other human tissue. Brain is rich in GSTP1, with GSTM2 and GSTM3 subunits as the next most abundant forms. The a-class GSTs are minor forms in brain, and the presence of low levels of GSTA1 allowed for the prominent display of GSTM3 (Fig. 4). Brain (cerebral cortex) cytosols have approximately 3-fold greater amounts of GST than pituitary which in turn has much lower levels of the GSTM1 subunit. The liver GST isoenzyme composition is in accordance with immunoblots inasmuch as the a-class and GSTM1 subunits predominate and other subunit types occur as very minor forms. Adrenal GSTs consist mostly GSTP1, GSTA1 and GSTM1 subunits. Kidney GSTs are composed mostly of GSTP1, GSTA1 and GSTA2 subunits and have GSTM1, GSTM2 and GSTM3 as minor forms. Testes, which have the greatest overall levels of GSTs, are particularly rich in GSTM3 and GSTA1. In most tissues shown in Fig. 4, GSTM2, GSTM4 and GSTM5 subunits are minor components of the total GST composition. The GSTP1 subunit is the predominant subunit in heart, lung and brain (from 4–10 mg/mg protein). All m-class GSTs (GSTM1 – GSTM5) were also detected at low levels in lung with GSTM1 (when present) as the most abundant of these (Fig. 4). The absence of the GSTM1 peak is striking in the HPLC profiles of GSTs from tissue of persons with the GSTM1 -null genotype. All tissues tested from those individuals were devoid of GSTM1 subunits. This technique may be applied readily for identification of the GSTM1 -null phenotype individuals at the level of protein
Fig. 4. HPLC profiles of GSTs isolated from human tissues. GSTs were purified from the individual tissues by GSH-affinity chromatography (see Section 2). The purified GSTs were applied to a Vydac C4 HPLC column and individual GST subunits were eluted from the column over an increasing gradient of ACN as described in Section 2. Peaks were assigned numbers 1 – 8 where positive assignment of GST subunits based upon immunological criteria using specific antisera, mobility on SDS gels, and ESI MS data could be obtained. Peak 1, found in some samples, reacted with the GSTM5-specific antibody; peak 2, GSTP1; peak 3, GSTM2; peak 4, GSTM1; peak 5, GSTM5 and GSTM4; peak 6, GSTA1; peak 7, GSTM3; and peak 8, GSTA2. The symbols indicate tissue from the same individual. (*, 48-year-old black male; **, 1 month old Caucasian male; ", 31-year-old Caucasian female and the liver in the upper left panel was from a 34-year-old black female; , 48-year-old Caucasian male). GSTM1 ( + ) and ( − ) denote the presence or absence of GSTM1 subunits (peak 4).
110
I. Listowsky et al. / Chemico-Biological Interactions 111–112 (1998) 103–112
expression. Liver was the only tissue with substantial levels of the GSTM1 subunit, (approximately 10 mg/mg protein), and in tissues such as liver and adrenal, the GST subunit profiles are drastically altered by the absence of GSTM1 subunits (Fig. 4).
4. Discussion
4.1. Tissue distribution Because of the complexity of the system, it has been a formidable task to compile a detailed description of human GST subunits in tissue specimens [7]. In many earlier studies on human GSTs, subunit assignments were fragmentary and based on imprecise criteria. For instance, early on GSTs were categorized as acidic (p), near neutral (m) and basic (a) forms [8], or simply on the basis of electrophoretic mobilities in SDS gels with p as the fastest migrating subunit followed by a and then m. Such classifications could be misleading since the electrophoretic mobilities of GSTM4 and GSTM2 overlap with those of the a-class subunits, and since m-class subunits in general have pl’s that encompass acidic and basic ranges. Moreover, some of the modifications such as oxidation and glutathionylation that occur in the course of purification and storage of the proteins may further confound analyses based on charge differences among the subunits. Recent advances in technology have facilitated the use of ESI MS for characterization of proteins [9], and several studies have examined primary structures of rat and mouse liver GSTs [10,11]. In the present study, application of ESI MS for this system has clarified several complex issues about human GSTs and provides a basis to plan and interpret future studies of GST functions. Of greater significance, the ESI MS data in conjunction with information derived from the use of peptide-specific antisera, provide unequivocal identification of human GST subunits resolved by HPLC methods. The on-line HPLC-ESI MS system detects minor subunits that probably occur in the cell primarily in heterodimeric combination with other subunit types. Most of the previous high resolution HPLC analyses of human GSTs have focused on liver forms, with some studies on other tissue. Some GST subunit peaks in those studies were unidentified. On the basis of such analysis, appropriate strategies may therefore be devised to determine relative amounts of particular subunit types and their tissue distribution. Specimens devoid of GSTM1 subunits from ‘null’ phenotype individuals are readily identified (Fig. 4). Thus, it is now possible to conduct rapid and reliable surveys of GST subunit distribution, and to observe their profiles under various conditions including changes that may occur during development, in malignant and other diseased tissue, and to determine whether select GSTs are induced by drugs or xenobiotics.
4.2. Inter-species comparisons Combinations of several criteria may be applied for comparison of subclasses of GST subunits from different mammalian species. These include sequence homolo-
I. Listowsky et al. / Chemico-Biological Interactions 111–112 (1998) 103–112
111
gies, similarities in tissue expression patterns, catalytic specificities and other properties. Each of these criteria alone may not suffice to establish which subunits in the different species are analogous, however in some cases inferences may be made from combinations of criteria. For example, the human hGSTM2 subunit is primarily found in muscle, brain and testis but not in liver or kidney (Fig. 4). A parallel tissue distribution pattern has been observed for the rat rGSTM3 (Yb3) subunit [12]. Moreover hGSTM2 and rGSTM3 have 84% sequence identities and both contain a unique signature pentapeptide sequence, ERNQV at residues 164–168. Similarly, the novel hGSTM3 subunit is probably comparable to the rat Yo and mouse mGSTM5 subunits on the basis of sequence homologies and tissue distribution patterns (Listowsky, Rowe, and Tchaikovskaya, unpublished observations).
4.3. Testis GSTs The testicular GSTs provide insight into potential differential tissue-specific functions of the proteins. Thus testis is by far the tissue richest in the hGSTM3 subunit (Figs. 1 and 4). It is likely that GSTs in testis serve in a protective capacity for germ cells from the action of reactive chemicals, hydroperoxides and other products of oxidative stress. In particular, fatty acid oxidations induced by leukocytes and/or defective sperm in semen, results in disruption of sperm motility, capacitation and failure of sperm – oocyte fusion reactions [13]. It is logical to assume that the presence of high levels of hGSTM3 could prevent or ameliorate consequences of these noxious reactions by virtue of particular GST catalytic and binding functions. This GST form may be a subject for future studies on targeting agents that could enhance fertility or as a drug target for male contraception.
Acknowledgements This work was supported by grant CA42448 from the National Cancer Institute, Cancer Center Core 5P30-CA 13330 and Liver Center Core 5P30-DK41296. JDR was supported by genetics and biochemistry graduate training grant T32 GM07128 from the National Institutes of Health. ESI MS, peptide synthesis and protein sequence analysis were carried out in the Laboratory for Macromolecular Analysis of the Albert Einstein College of Medicine. We appreciate the valuable suggestions of Dr John D. Hayes.
References [1] J.D. Hayes, D.J. Pulford, The glutathione S-transferase supergene family: regulation of GST and the contribution of the isoenzymes to cancer chemoprotection and drug resistance, Crit. Rev. Biochem. Mol Biol. 30 (1995) 445–600. [2] B. Mannervik, P. Alin, C. Guthenberg, Identification of three classes of cytosolic glutathione transferases common to several mammalian species: Correlation between structural data and enzymatic properties, Proc. Natl. Acad. Sci. 82 (1985) 7202 – 7206.
112
I. Listowsky et al. / Chemico-Biological Interactions 111–112 (1998) 103–112
[3] T.H. Rushmore, C.B. Pickett, Glutathione S-transferases, structure, regulation, and therapeutic implications, J. Biol. Chem. 268 (1993) 11475 – 11478. [4] H. Jenson, P. Alin, B. Mannervik, Glutathione transferase isoenzymes from rat liver cytosol, Methods Enzymol. 113 (1985) 504–507. [5] Y. Takahashi, E.A. Campbell, Y. Hirata, T. Takayama, I. Listowsky, A basis for differentiating among the multiple human mu-glutathione S-transferases and molecular cloning of brain GSTM5, J. Biol. Chem. 268 (1993) 8893–8898. [6] F. Ali-Osman, O. Akande, G. Antoun, J.X. Mao, J. Buolamuini, Molecular cloning, characterization, and expression in Escherichia coli of full-length cDNAs of three human glutathione S-transferase Pi gene variants, J. Biol. Chem. 272 (1997) 10004 – 10012. [7] B. Mannervik, M. Widersten, Human glutathione transferases: classification, tissue distribution, structure and functional properties, in: G.M. Pacifici, G.N. Fracchia (Eds.), Advances in Drug Metabolism in Man, Europe Com. (UK), Bernan assoc. publisher, 1995, pp. 407 – 459. [8] B. Mannervik, U. Danielson, Glutathione transferases-structure and catalytic activity, Crit. Rev. Biochem. Mol. Biol. 23 (1988) 283– 337. [9] K. Biemann, The coming of age of mass spectrometry in peptide and protein chemistry, Protein Science 4 (1995) 1920–1927. [10] P. Rouim, L. Debrauwer, J. Tulliez, Electrospray ionization-mass spectrometry as a tool for characterization of glutathione S-transferase isozymes, Anal. Biochem. 229 (1995) 304 – 312. [11] H.I. Yeh, C.H. Hsieh, L.Y. Wang, S.P. Tsai, H.Y. Hsu, M.F. Tam, Mass spectrometric analysis of rat liver cytosolic glutathione S-transferases: modifications are limited to N-terminal processing, Biochem. J. 308 (1995) 69–75. [12] M. Abramovitz, I. Listowsky, Selective expression of a unique glutathione S-transferase Yb3 gene in rat brain, J. Biol. Chem. 262 (1987) 7770 – 7773. [13] R.J. Aitken, A free radical theory of male infertility, Reprod. Fertil. Dev. 6 (1994) 19 – 24. [14] B. Mannervik, Y.C. Awashi, P.G. Board, J.D. Hayes, C.D. Dilio, B. Ketterer, I. Listowsky, R. Morgenstein, M. Marumatsu, W.R. Pearson, C.B. Pickett, K. Sato, M. Widersten, C.R. Wolf, Nomenclature for human glutathione transferases, Biochem. J. 282 (1992) 305 – 306.
.