T. Herbert Manoharan, Andrew M. GulickS, Ralph B. Puchalski, Amy L. Servais, and. William E. FahlO. From the McArdle Laboratory for Cancer Research, ...
THEJOURNAL OF BIOLOGICAL CHEMISTRY
Vol. 267, No. 26, Issue of September 15, pp. 18940-16945,1992 Printed in U.S. A.
0 1992 by The American Society for Biochemistry and Molecular Biology, Inc.
Structural Studies on Human Glutathione S-Transferase n SUBSTITUTION MUTATIONS TO DETERMINE AMINO ACIDS NECESSARY FOR BINDING GLUTATHIONE* (Received for publication, January 21, 1992)
T.Herbert Manoharan,Andrew M. GulickS, Ralph B. Puchalski, Amy L. Servais, and William E. FahlO From the McArdle Laboratory for Cancer Research, University of Wisconsin, Madison, Wisconsin 53706
In order to identify amino acids involved in binding the co-substrate glutathione to the human glutathione S-transferase (GST)A enzyme, we assembled three criteria to implicate amino acids whose role inbinding and catalysis could be tested. Presenceof a residue in the highly conserved exon 4 of the GST gene, positional conservation of a residue in 12 glutathione S-transferase amino acid sequences, and results from published chemical modification studies were used to implicate 1 4 residues. A bacterial expression vector (pUC120~), which enabled abundant production (2-26s of soluble Escherichia coli protein) of wild-type or mutant GST A,was constructed, and, followingnonconservative substitution mutation of the 14 implicated residues, five mutants (R13S,D67K, Q64R, I68Y, L72F) showed a >96%decrease in specific activity. A quantitative assay was developed which rapidly measured the ability of wild-type or mutant glutathione S-transferase tobind to glutathione-agarose. Using this assay, each of the five loss of function mutants showed a >20fold decrease in binding glutathione, an observation consistent with a recent crystal structure analysis showing that several of theseresidues help to form the glutathione-binding cleft.
(* class) to 221 amino acids (aclass). Alignment of deduced amino acid sequences from these glutathione S-transferase cDNAs indicates aregion of the protein (encoded by exon IV) which is highly conserved, as well as specific residues (19 amino acids) located throughout the monomeric protein which are perfectly conserved in position. This positional conservation of theseelements suggests their involvement ina conserved function of these enzymes, such as binding the one substrate which is common to all of the isozymes, glutathione. The glutathione S-transferase-catalyzedconjugation of glutathione to physiological and xenobiotic electrophiles yields products that areusually less reactive, more hydrophilic, and more readily excreted than the substrates(9). Glutathione is present at 0.1-10 mm in mammalian cells, and its sulfhydryl group comprises 10-20% of the non-protein sulfhydryl groups in a cell. The precise mechanism through which the S atom on glutathione is covalently attached to theelectrophilic site ona xenobiotic metabolite is still unclear. The results of kinetic analyses, however, have implied a random order reaction where both substrates bind to theirrespective sites and an enzyme residue then withdraws the proton from the glutathione “SH, and by increasing the S nucleophilicity, greatly accelerates its covalent addition to the adjacent electrophilic atom (4, 10). In thisreport, and in an accompanying report (Gulick et al. Glutathione S-transferases (EC2.5.1.18) comprise a family of isozymes that can catalyze the covalent addition of the (35)), we have used molecular and immunological approaches tripeptide glutathione (7-Glu-Cys-Gly) to a structurally di- to study the structure of human glutathione S-transferase ?r. verse set of physiological and xenobiotic electrophiles (1-4). Our primary goal has been to identify the enzyme domain and At least 10-12 cytosolic glutathione S-transferases from var- the constituentamino acid residues which are responsible for ious rat, human, and mouse tissues have been characterized binding one of the substrates, glutathione,to theenzyme. We and grouped into four classes (alpha, mu, pi, and theta) on applied three criteria to identify amino acids within the huthe basis of enzymatic, immunological, and structural prop- man glutathione S-transferase ?r molecule which were likely erties. The cytosolic glutathione S-transferases arecomposed to be involved in binding the substrate glutathione. We then of two subunits, about 24-28 kDa each, which dimerize by constructed a bacterial expression vector and nonconservanoncovalent interactions. Each subunit of a heterodimer or tively substituted these residues, determining the effects homodimer is kinetically independent of the other subunit which these substitutions had upon the ability of the enzymes and has been shown to dimerize with other subunits of its to bind glutathione and catalyze the conjugation of CDNB.’ class only. Full-length cDNAs encoding many of the described Five residues were identified which had lost their catalytic mammalian, cytosolic glutathione S-transferases have been ability. Using a newly designed assay, we show that none of isolated and sequenced (5-8). The open reading frames of these mutant enzymes is able to significantly bind glutathithese cDNAs encode peptides ranging from 208 amino acids one, thus demonstrating the important role of these residues * This work was supported by National Institutes of Health Grants in forming the glutathione-binding cleft. ~~
Pol-22484 andP30-CA07175. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 3 Supported by a predoctoral fellowship from the National Science Foundation. 5 To whom correspondence and reprint requestsshould be addressed. Tel.: 608-262-1275. Fax: 608-262-2824.
The abbreviations used are: CDNB, l-chloro-2,4-dinitrobenzene; IPTG, isopropyl P-D-thiogalactoside; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; PBS, phosphate-buffered saline. Throughout this paper, single-letter amino acid abbreviations are used, e.g. R = arginine; additionally, the notation R l l S is used to indicate a mutant where the wild-type arginine residue at position 11 is substituted with a serine, etc.
18940
Glutathione-binding Residues in Glutathione S-Transferase 7~ EXPERIMENTAL PROCEDURES
Materials-Restriction endonucleases and DNA-modifying enzymes wereobtained from New England Biolabs, Bethesda Research Laboratories, Promega Corp., and U. S. Biochemical Corp. Bacterial growth media were purchased from Difco. Supplies for SDS-PAGE were obtained from Boehringer Mannheim. Purified glutathione Stransferase from human placenta, glutathione, and glutathione-agarose (G4510) were from Sigma. Nucleotide triphosphates and IPTG were obtained from Pharmacia LKB Biotechnology Inc. DNA oligonucleotides were synthesized chemically using p-cyanoethylphosphoramidites with an Applied Biosystems Model 391 PCR-Mate E P DNA synthesizer. Nitrocellulose membranes for immunoblotting were obtained from Schleicher and Schuell. All other chemicals were of reagent grade. Construction of Expression Plusmid-Double-stranded phagemid DNA was isolated from bacterial cells by the alkaline lysis method (11). Superinfection of the host cells harboring the phagemid with the helper phage M13K07 (12) was used to induce the excretion of phages. The phages in the growth medium were precipitated using polyethylene glycol, and the single-stranded DNA was isolated by SDS-proteinase K digestion followed by phenol and chloroform extractions and ethanol precipitation. Single-stranded phagemid DNA prepared from Escherichia coli CJ236 and JM109 was used as template for site-directed mutagenesis (13) and dideoxyDNA sequencing, respectively. The DNA sequencing kit and Sequenase were from U. S. Biochemical Corp. The bacterial expression vector (pUC120) used for expression of glutathione S-transferasea was a phagemid (a gift from Dr. J. Vieira, University of Minnesota) which is a derivative of the pUC vector (14). A 714-base pair EcoRI restriction fragment containing the entire region encoding human glutathione S-transferase a and its 5’- and 3”untranslated regions was cleaved from pGpi2 (7) andisolated from a 1%agarose gel. The fragment was ligated into the unique EcoRI site at the5’-end of the lacZ’ gene of pUC120 placing the 5’-end of the glutathione 5’-transferase a gene downstream of the lac promoter. The ligation generated an in-frame 1acZ’-a genefusion. This plasmid was designated pUC12OaEAA (for extra amino acid) and supported expression of glutathione S-transferase a as a 8-Gal-a fusion protein upon induction with IPTG. Amino-terminal amino acid sequencing of the fusion protein (Applied Biosystems protein sequenator)showed the fusion of the first 12 amino acid residues of the lacZ’ protein to the amino terminus methionine residue of glutathione S-transferase R.
In order to delete the 36 nucleotides of lacZ’ gene encoding the first 12 amino acid residues of the fusion protein and express only glutathione S-transferase a from its initiation methionine residue, a site-directed, oligonucleotide-mediated loop-out mutagenesis was undertaken. A primer, 5’-ACAAGGAAACAGCCATGGCGCCCTACACCG-3’, complementary to a portion of the lac promoter, ribosome-binding site, and 5’-end of the glutathione S-transferasea gene beginning with its initiation methionine codon was used to construct the phagemid pUC12Oa which encoded wild-type glutathione S-transferase a now lacking any extra amino-terminal amino acid residues. A unique NcoI restriction site was built into this oligonucleotide to help in screening and selecting the deletion mutant phagemid pUC120a. The oligonucleotide was purified by reverse-phase chromatography on a C-18 Sep-Pak cartridge (Waters, Millipore Corp.), phosphorylated using T4 polynucleotide kinase, and annealed to the single-stranded phagemid DNA prepared from E. coli CJ236. The annealed complex was incubated at 37 “C for 2 h with T7 DNA polymerase, T4 DNA ligase,deoxynucleotides, and ATP. The reaction mixture was then transfected into JM109 cells and plated on 2 X YT agar plates (16 g of Bacto-tryptone, 10 g of Bacto-yeast extract, 5 g of NaCl, 15 g of Bacto-agar per liter of water, pH 7.4) containing ampicillin. The 5”flanking region of the glutathione S-transferase a gene around the initiation methionine codon of pUC120a was sequenced to confirm that additional amino acid residues were no longer encoded at theamino terminus of the glutathione S-transferase a. Single-site mutants of glutathione S-transferase a were prepared by site-directed mutagenesis using single-stranded DNA templates produced from pUC12Oa and the appropriate mutant oligonucleotides. Oligonucleotides used for mutagenesis were generally 20 bases in length, and the desired 1-2-base mismatch was chosen to yield a codon which is commonly represented in E. coli genes. The newly constructed plasmids, containing specific mutations within the glutathione S-transferase a cDNA, are designated by the wild-type amino acid (one-letter symbol), the location of the amino acid residue
18941
in the protein (amino acid residue 1 starts with the amino-terminal proline residue, Ref. 7) and mutated amino acid (Table I). Preparation of Wild-type and MutantGlutathione S-Transferase a Proteins-Plasmids encoding wild-type or mutant forms of glutathione S-transferase a were transfected into E. coli JM109, and cells were grown at 37 “C in 50 ml of 2 X YT medium containing 100 pg/ ml ampicillin and 2 mM IPTG. When the cells reached a turbidity of 0.8-1.0 A at 590 nm, they were harvested by centrifugation at 6,000 X g for 10 min and resuspended in 10 ml of PBS (pH 7.0). Cell suspensions were sonicated twice for 15 s a t a continuous pulse using a micro-tip sonicator (Branson Sonifier Cell Disruptor 185) with an interval of 15 s for cooling the samples. The sonic extract was centrifuged at 10,000 X g for 10 min at 2 “C, and the resulting supernatant was further centrifuged at 20,000 X g for 30 min. The clear supernatant was then dialyzed against PBS for 6 h and used as the source of glutathione S-transferase a. The total protein concentration was determined using the BCA reagent kit(Pierce).The proteins in the bacterial extracts were separated in a 1.5-mm-thick SDS-PAGE gel, and, after Coomassie blue staining, glutathione Stransferase a content in the cell extracts was determined by comparing glutathione S-transferase a band density in cell extracts to the band density of purified glutathione S-transferase a standards using an LKB UltrascanLaser Densitometer with the Gelscan XL Software program. Zmmunoblot Analysis-Immunoblotting was done as previously described (15). Bacterial extracts containing 0.5 pg of glutathione Stransferase a protein or 75pgof total protein for mutants RMD, R70D, R74S, and 179Stop were electrophoresed in 12.5%SDS-PAGE gels, electroblotted onto nitrocellulose paper, and reacted with rabbit polyclonal anti-glutathione S-transferasea or one of a panel of mouse monoclonal anti-glutathione S-transferase a antibodies (Gulick et al. (35)) (1:lOOO dilution) followed by a goat anti-rabbit IgG antibodyalkaline phosphatase conjugate or rabbit anti-mouse IgG and IgM antibody-alkaline phosphatase conjugate for polyclonal and monoclonal antibodies, respectively. Determination of Kinetic Values-Kinetic constants were determined under standard assay conditions (16) for glutathione S-transferase with CDNB as the co-substrate. One-ml reactions were run which contained 1 pgof glutathione S-transferase in 0.1 M NaPO, buffer (pH 6.5) with 1 mM EDTA. To convince ourselves that the native, background proteins from E. coli would not affect the outcome of these kinetic determinations, we made mock lysates by adding purified human glutathione S-transferasea to bacterial extracts. The glutathione S-transferase a content in these extracts ranged from 1 to 15% of the total protein in the lysate. Kinetic assays on these lysates yielded results that did not differ significantly from each other undera variety of substrate concentrations. The reactions were always started by addition of glutathione to the cuvette. For determination of specific activity, glutathione and CDNB were both added to concentrations of 1 mM. Greater than three replicate reactions wereallowed to run for 1 min, and the best linear fit was made according to the kinetics program on the Gilford Response spectrophotometer to determine the reaction rate. Conversion to specific activity was done using an extinction coefficient of 9.6 mM” cm”. For more detailed kinetic analysis on the wild-type glutathione Stransferase a produced in E. coli, the concentrations of both substrates were varied and the initial velocities were used to generate Hanes-Woolf plots (17). Secondary replots of the slopes and intercepts were used to determine the values for VmaX, the Michaelis constants, and the steady state dissociation constant for glutathione as described by Cleland (18). This complete kinetic analysis was performed with four independent extracts of bacterially generated wild-type glutathione S-transferase a. Assay of Glutathione S-Transferase a Ability to Bind to Glutathione-We developed a method for the rapid screening of the ability of the different mutant enzymes to bind to glutathione. Plasmids encoding the wild-type or mutant enzyme forms were added to a bacterial S-30 lysate (Promega) which included [35S]methionine (Trad5Slabel, 1017 Ci/mmol). The reaction was incubated for 1 h according to the manufacturer’s instructions. Glutathione binding was then determined by recovery of the wild-type or mutant enzyme with glutathione-agarose using a slight modification of a described procedure (19).The product of the S-30 reaction was diluted to 600 p1 with MTPBS (150 mM NaC1,16 mM Na2HP04,4 mM NaH2P04 (pH7.3)) and 15 pl of a 1:3 slurry of glutathione-agarose in MTPBS was added. This amount of slurry was shownto be sensitive to changes in enzyme affinity for the substrate. The mixture was rocked for 5 min and then washed four times with 750 pl of MTPBS. Specifically bound protein
18942
Glutathione-binding Residues Glutathione in S-Transferase
was then eluted by the addition of 50 rl of 5 mM glutathione in MTPBS. This was rocked for 2 min and then centrifuged to pellet the glutathione-agarose and all nonspecifically bound proteins. The supernatant was added to an equal volume of 2 X sample buffer and loaded on a 12% acrylamide gel for electrophoresis and quantitation of the glutathione S-transferase T band. RESULTS
?r
1)Hum Ha-1 MAEKPKLH 2)RatYa MSGKPVLH 3)Hum Ha-2 MAEKPKLH 4)Rat Yc MPGKPVLH 5)Hum Uv p MPMILG 6)Hum tes p MSCGSSMVLG 7lRat Yb MPMILG SIRat MPMTLG 9)Rat Y b J MPMTLG 1O)schlStO MSPILG 11)RatP MPPYTIV 12)Hum x MPPYTW
EFEEKFIKS. EFEEKLIQS. EFEEKFIKS. EFEEQFLKT. SYEEKKYTMG SYEEKRYTCG SYEEKRYAMG SYEDKRYSMG SYEEKRYTMG KYEEHLYERD SWKEEWTI. SWKEEWTV.
.b:
1
Criteria Usedto Implicate Residues Involved in Binding Glutathione S-Transferase-All of the known glutathione Stransferases, of both plantand animal origin, retain the conserved function of binding glutathione to the so-called Gsite and catalyze its covalent addition to hydrophobic substrates (l), making these soluble proteins, by definition, a family of isoenzymes. Since the structure of glutathione is unchanged innature, the binding cleft conformation and residues involved in binding glutathione to glutathione Stransferase might be expected to be conserved in the multiple isoforms of this enzyme. Based on this assumption, we retrieved each GenBank file that described a glutathione Stransferase cDNA and determined the region of greatest DNA sequence homology between it and our reference sequence, the cDNA encoding the human glutathione S-transferase x protein, originally described by Kano etal. (7). As seen inFig. 1, under stringent conditions of homology comparison, each of 11 cDNAs shows the greatest homology within the exon 4 boundary of the human glutathione S-transferase x , and, as expected, the rat x species shows nearly complete homology to thehuman T isozyme (7,20,21). When cDNA sequences encoding cytosolic glutathione Stransferases were first reported (5-8), several papers compared the deduced amino acid sequences of the various glutathione S-transferases (1, 22) looking for residues which, because of their positional conservation, could be involved in a conserved function of these isozymes, such as binding glutathione, binding a common hydrophobic substrate, e.g. CDNB, or perhaps in enablingdimerization of the monomeric gene products to occur. Earlyamino acid alignments (22) showed positional conservation of several residues in six out of six peptides, and recent alignments (Fig. 2) show perfect positional conservation of 18 residues in 12 independent glutathione S-transferasegene products. Over the past decade, some reports have shown that chem-
i jR -1 : -2 i 3 :
6 ) Human testlc. p
.
:
7) Rat Yb, 8 ) Rat Y b 9 ) Rat 10) Schistosome 11) Rat P Human II
41"
i
5: 6 :-
7'
Y6
8 -
9-
12): ID-;
=
EXON
YGKDIKERAL YGKDMKERAL YGKDIKEKAL YGKDMKERAL CGETEEEKIR CGETEEEKIR CGETEEERIR CGETEEERIR CGETEEERIR HNL
F P N L P Y L I ~SHKI ~
10) FPNLPYYID RNKKF LE LGGCPKERAE HNM DVKL 12)
1)
2)
3) 4)
5)
61 7)
8 1 9) 10) 11) 12)
G.. YGKDQKEAAL ..QLPKFb LGL DLTL LKASCL G.. .QLPKFW- DLTL 43 exon I
.
IDMYIEGIAD LGEMILLLPV IDMYSEGILD LTEMIIQLVI IDMYIEGIAD LGEMILLLPF IDMYAEGVAD LDEIVLHYPY VDILENQTMD NH. ..MQLGM VDIIENQVMD FR.. .TQLIR ADIVENQVMD NR ...MQLIM VDVLENQAMD TR ... LQLAM VDILENQLMD NR...MVLAR ISMLEGAVLD IR ... YGVSR VDMVNDGVED LRCKYGTLIY VDMVNDGVED LRCKYISLIY
...
...GKQPWFA
88
CPPEEKDAKL ALIK CPPDQREAKT ALAK TQPEEQDAKL ALIQ IPPGEKEASL A K I K RNR YFPAFEKVLK ICYNPEFEKL KPKY LPE KLKLYSEFL. LCYSSDHEKL KPQT LPG QLKQFSMFL. LCYNPDFEKQ KPEF IPE KMKLYSEFL. VCYSPDFERK KPEY LPE KMKLYSEFL. LCYNPDFEKL KPGY DLPG MMRLYSEFL. IAYSKDFETL KVDFUSkLPE MLKMFEDRL. TNY....ENG KDDY-LPG HLKPFETLLS TNY....EAG K D D Y V ~ L P GQLKPFETLLS 134
E
89
KS.HGQDYLV KS.1IGQDYLV KS.HGQDYLV KS.HGQDYLV GKRPWFA
LGL YGKDQQEAAL
GNKLSRADIH LVELLYYVEE GNRLTRVDIH LLELLLYVEE GNKLSRADIH LVELLYYVEE GNRLSRADVY LVQVLYHVEE GNKITFVDFL VYDVLDLHRI GNKITYVDFL
LDSSLISSTP FDASLLTSTP LDSSLISSTP LDPSALANTP FEPNCLDATP
VYDVLDQHRI
LLKALKT~US LLKAFKSXIS LLKALKTWS LLKALRTWS NLKDFIS~~E
FEPKCLDATP
NLKDFV
...CHKTYLN
GDHVTHPDFM LYDALDWLY MDPMCLDATP KLVCFKKqIE QNQGGKAFIV GNQISFADYN LLDLLLVHQV LAPGCLDFTP LLSAYVABLS QNQGGKTFIV GDQISFADYN LLDLLLIHEV LAPGCLDATP LLSAYVGRZS 135 re 4 NLPTVKKFLQ SLPNVKKFLQ NLPTVKKFLQ NLPTVKKFLQ GLEKISAYMK GLEKIAAYLQ GLKKISAYMN GLKKISDYMK GLKKISDYMK AIPQIDKYLK ARPKIKAFLS ARPKLKAFLA 185
PGSPRKPPMD PGSQRKPAMD PGSPRKPPMD PGSQRKFLED SSRFLPRPVF SDQFCKMPIN CSRYLSTPIF SGRFLSKPIF SSRFLPRPLF SSKYIAWPLQ SPDHLNRPIN SPEYVNLPIN
EKSLEEARKI AKQIEEARKV EKSLEESRKI EKCVESAVKI SKMAVWGNK' NKMAQWGNKP SKLAQWSNK' AKMAFWNPK' TKMAIWGSK* GWQATFGGGD GNGKQ' GNGKQ" 209
FRF' FKF' FRF. FS' VC.
HPPK*
FIG. 2. Alignment of glutathione S-transferase isoenzyme amino acid sequences deduced from respective cDNA sequences. (The isozyme designation and thecorresponding GenBank
11 -12
[L-
I II
bpb
4 w
VC
~ I 5 ) Human liv. p
i
.....
files are the same as those used in Fig. 1.) Each amino acid sequence was aligned in comparison to the reference human glutathione Stransferase a sequence (1). Boldletters represent those residues positionally conserved either in 6 glutathione S-transferase isozymes in an early alignment (22) or in all 12 glutathione S-transferase isozymes present in this later, expanded alignment. Solid underline indicates those residues within the boundaries of exon IV in human glutathione S-transferase T.Boxed residues are those which met two selection criteria and, thus, were selected for substitution mutation.
I)Human Ha-1 21 Rat Ya 3)~ u m n nna-2 GST
:PJM$LD
.
.AE.DL.DK ..PE.DL.EK ..AE.DL.DK ..RD.DL.AR DAPDYDRSQW DAPDYDRSQW DAPDYDRSQW DAPDYDRSQW DAPDFDRSQW EGDK W ...DVWLQGS ...ETWQEGS 42
111
IV 2 w
V
VI
J'UTS
VI1
6bo
I
FIG. 1. Best-fit alignment of most homologous regions of cDNA sequences encoding glutathione S-transferase isozymes from human, rat, or Schistosornajaponicurnorigin. Full-length cDNA sequences reported in 11 GenBank files (Genbank file names are: 1 = Humgstb, 2 = Ratgstlya, 3 = Humgstc, 4 = Ratgstyc, 5 = Humgstd, 6 = Humgstmua, 7 = Ratgstb, 8 = RatybZa, 9 = Ratgstyb, 10 = Scmag, 11 = Ratgstp, 12 = Humgstpi) were individually aligned to the reference sequence, human glutathione S-transferase T (Genbank file = Humgstpi) using the Bestfit program (University of Wisconsin Genetics Computer Group software; parameters: gap weight = 5-10, gap weight length = 0.3-1.0). Each individual solid line represents the region of greatest nucleotide sequence homology. The boundary of each of the exons in the human glutathione Stransferase 7r gene (21) is indicated by the alternating boxes.
ical modification of certain glutathioneS-transferase residues resulted in enzymes which were catalytically inhibited, and that in some instances, competition by glutathione blocked the protein alkylation, presumably of a residue at or near the G-site (23). Earlyin thisproject there were reports suggesting the importance of His (24), Arg (l),Lys? and Cys (25) residues in glutathione S-transferase catalysis and/or glutathione binding. These, then, were the three criteria which we assembled and used to identify residues putatively involved in binding glutathione to the enzyme: (i) those amino acids positioned in the most highly conserved glutathione S-transferase exon C. L. Xia and H. L. Chen, Meetingabstract,Glutathione STransferase and Drug Resistance, Edinburgh, Scotland, August,1988.
Glutathione-binding Residues
in Glutathione S- Transferase a
18943
most mutants were produced at levels at or above the wildtype; however, for some mutants, such as the R70D substitution or the179Stop truncated protein, there was no detectable level of glutathione S-transferase A in the cells. The ratio of glutathione S-transferase A band intensity in theCoomassiestained gel (Fig. 3A) relative to theWestern immunoblot (Fig. 3B) was the same, within error,for each peptide. Reversion mutation of the R70D GAT codon to the wildtype CGT codon or to the mutantR70T ACT codon resulted in abundantproduction of glutathione S-transferase A protein (Table I). Expression of the R70D mutant in the proteasedeficient PR746 strain of E. coli still resulted inno detectable production of glutathione S-transferase A (not shown). As seen in Fig. 4, the RNA species encoding the R18D, R70D, R74S, and 179Stop mutants were as abundant as the wildtype glutathione S-transferase A RNA species in E. coli when normalized to the cellular content of the 8-lactamase RNA species. The mutant codons which we used for substitution were chosen because of their common usage in E. coli genes; therefore, the combined data suggest that thepoor production of these glutathione S-transferase A mutants was related to the instability of the translatedprotein. Kinetic Analysis of Wild-type and Mutant Enzymes-Extensive analysis was done on four independent preparations of the bacterially produced, wild-type glutathione S-transferase A (Table I, top panel) to demonstrate: (i) that thespecific TABLEI activity of the bacterially produced enzyme (52.0 & 9.8 units/ Kinetic constants for wild-type and mutant formsof human glutathione S-transferase K mg) was the same as that for the purified human placental A what the batch to batch Cleared extracts were prepared from IPTG-induced JM109 cells enzyme (47.6 f 6.4 units/mg), and (ii) expressing either the wild-type or a mutant form of glutathione S- variation was, so that we could recognize a mutant with a transferase T . Concentrations of glutathione and CDNB were system- significant decrease or increase in activity when it appeared. atically varied in 1-ml incubations containing 1 pg of glutaThe specific activity was determined for each mutant gluthione S-transferase K protein, and initial reaction rates were detertathione S-transferaseA, and these values are listed in Table mined. I (bottom panel). Although many mutants showed a reduced Specific Enzyme Vmax Km(GSW Km(CDNB) Kt (GSH) activity specific activity, only five mutants (R13S, D57K, Q64R, I68Y, and L72F) showed reductions to 6 % of the wild-type, a cutnmolfmin mM mM FM unitslmg off level for significance which is commonly used (31) in 48.8 Wild-type 1 0.083 0.142 0.626 0.0 studies of enzyme mutagenesis. One mutant, K120E, showed 63.2 Wild-type 2 0.107 0.150 0.541 28.3 0.088 0.207 0.763 8.3 40.2 a significant increase ( p < 0.05) in its specific activity. Wild-type 3 0.151 0.722 19.9 55.8 Wild-type 4 0.102 Ability of Mutant Enzymes to Bind Glutathione-Because 0.148 0.703 18.8 52.0 0.095 Average we were trying to ablateglutathione binding by changing key f 0.011 f 0.005 f 0.070 f 10.0 f 9.8 residues in the glutathione-binding site, it was important to Relative determine whether this was the reason why a given mutant Enzyme GST Level Specific GSH-binding activitv enzyme had lost its catalytic ability. Early in these experiments, we determined the glutathione K, (1.9 X M) for the wild-type enzyme by extrapolation from Hanes-Woolf 52.00 f 9.80 3.67 Wild-type plots of kinetic data (not shown) or by distribution of [3H] 4.01 26.30 f 3.00 RllS 6.55 R13S 0.14 f 0.04 glutathione ( K d = 6.4 X M) between microdialysis chamC0.4‘ R18D bers (not shown). However, neither of these approaches was L48F 4.77 f 1.42 26.20 acceptable to study mutant enzymes which either had no 6.41 D57K NMd activity or were at insufficient concentration to allow accurate 9.26 G58F 10.66 f 0.92 equilibrium dialysis. A third approach was devised using an 1.44 Q64R 0.28 f 0.22 in vitro bacterial S-30 transcription/translation system to 9.90 I68Y NMd C0.4 R70D make reproducible amounts of [35S]methionine-labeled,wild5.78 R70T 46.14 f 3.58 type or mutant, glutathione S-transferase a (Fig. 5 ) . OrigiH71R 1.60 5.74 k 0.41 nally designed to produce partial fragments of glutathione SL72F 18.35 0.32 f 0.23 transferase A for mapping monoclonal antibody epitopes, the 12.85 L72C 14.49 f 1.60 lac i promoter also enabled in vitro production of full-length C0.4 R74S 5.93 K120E 61.05 f 2.20 or defined-length fragments to study their retained ability to C0.4 179Stop bind to glutathione-agarose (Gulick et al. (35)). As seen in 1.87 R182S 5.64 f 0.70 Fig. 5A, each of the five loss of function mutants, although Values in parentheses represent specific activity represented as a efficiently synthesized in the S-30 reaction (Fig. 5B), showed percentage of the wild-type activity. an approximately 20-fold decrease (Table I) in its ability to ’Less than 0.4 pg of GST r protein was not detectable on bind to glutathione-agarose, relative to the wild-type glutaCoomassie-stained gel. thione S-transferase A, underlining the importance of these e Not determined. Not measurable. residues.
(IV), (ii) those residues which were highly conserved in their position, either 6 out of 6 in anearly alignment (22) or 12 out of 12 in an alignment based upon a later, larger data base (Fig. 2), and (iii) those residues which were implicated as a result of chemical modification studies published in the literature. As shown in Fig. 2, any residue in humanglutathione S-transferase A which met two of these three criteria was chosen for a nonconservative substitution mutation. Production and Quantification of Wild-type and Mutant Enzymes-The pUC12Oa plasmid, which is a derivative of the ~ICBORA vector used in our earlier experiments(26, 27), was designed to enable efficient production of glutathione Stransferase A in E. coli at abundant levels (2-30% of protein in cleared bacterial lysate, Table I),so that mutant enzymes could be studied without need for purification. In these experiments, a“successful” mutation (i.e. ablating G-site)would by definition eliminate purification of the mutant enzyme using glutathione-agarose. Nonconservative substitutionmutations were made for each of the 14 boxed residues indicated in Fig. 2, and cleared lysates from IPTG-induced cells expressing the glutathione S-transferase A constructs were electrophoresed to quantitate the level of glutathione S-transferase A in each cell extract (Fig. 3, A and B, summarized in Table I). As seen in Fig. 3,
Glutathione-binding Residues
18944
in Glutathione S-Transferase
T
0,
a
1 1 12 13141516 1718192021
1p1I.
.8 .6 .4:2
I
2 3 4
B$
5 6 7 8
-x
9 10 1 1 12 13 14 15 16 17 18 19 2021 . .
" r l P r g
-x
FIG. 3. Quantitation of recombinant glutathione S-transferase n in extracts of E. coli JM109 cells containing pUCl2On plasmids; cells were induced for12 h with IPTG, sonicated, and centrifuged (see "Experimental Procedures").Glutathione Stransferase x content in extracts was determined by laser densitometry of glutathione S-transferase x band versus glutathione S-transferase x standards. A , extracts from cells expressing wild-type or mutant glutathione S-transferase x molecules were previously analyzed, and, here, either that amount of extract which contained 1 pg of glutathione S-transferase x, or 75 pg of extract (lanes I , 5, 11, 13, 18, 20) where no glutathione S-transferase x was detectable, were electrophoresed and stained with Coomassie Blue. B, either that amount of extract which contained 0.5 pgof glutathione S-transferase x or 75pgof extract (lanes I , 5, 11, 13, 18, 20) where no GST x was detectable, were electrophoresed. Following transfer to a nitrocellulose membrane, blots were incubated with polyclonal antisera uersus human glutathione Stransferase x, and a secondary alkaline phosphatase-conjugated antibody, nitro blue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate, was used for color development of glutathione S-transferase x bands.
~ U C 1 2 0 wild-type
R18D
R70D
R74S
179Stop
FIG. 4. Quantitation of glutathione S-transferasen and 8-lactamase RNA i n E. coli cells containing pUC12On vectors with wild-type or mutant glutathione S-transferase x cDNAs. Following a 12-h induction with IPTG, cells were lysed and RNA was isolated as described (28). Dilutions of samples were slot-blotted, and the filter was hybridized first with a glutathione S-transferase x probe (342base pair BstEII-BglII fragment of pUC12Ox), stripped, and then rehybridized with a p-lactamase probe (364-base pair BglII-ScaI fragment of pUC12Ox).
families, each retaining the ability to bind a common subDISCUSSION In these experiments, our goal was to identify amino acid strate,glutathione,with a consistent affinity (1) between isoenzymes. This implied that theisoenzymes had retained a residues in human glutathone S-transferase A which play a common domain which enabled bindingof glutathione. necessary role in binding the co-substrate glutathione to the A recent report by Reinemer et al. (29) implicated residues enzyme. Theresults of ourexperimentsindicatethat:(i) R13, Q64, and D96 as importantfor binding glutathionebased selection criteria based upon positional conservativeness and upon x-ray crystallographic analysis of a pig glutathione Schemical modification studies of glutathione S-transferase A transferase cglutathione-sulfonate co-crystal. A 4th residue, residues can accurately predict many of the residues necessary Y7, was viewed as important in catalysis(29-31) because the for binding glutathione, (ii) abundantly produced, recombiclose proximity of the Y7 hydroxyl group to the glutathione nant glutathione S-transferase A in bacterial lysates is ame- sulfhydryl group could help to increase the nucleophilicity of nable tokinetic analysis of its wild-type or mutant phenotype, the S atom and acceleratecatalysis. A separate set of substiand (iii) five substitution mutants (R13S, D57K, Q64R, I68Y, tution mutants hasbeen constructed and characterized by us, and L72F), whichshowed a >95% loss of catalysis,also and the resultsperfectly support thehypothesized role of the showed a near totalloss of their ability to bind glutathione in R13, Q64, and D98 residues in binding GSH and the role of a quantitative assay. Many of these latter 5 residues have the Y7 residue in binding glutathione and chemically activatsince beenidentified asinteractingwithglutathioneside ing the S atom on glutathione (36). groups in a recently published crystal structure analysis(29). With thesignificance of the Y7, R13, Q64, and D98 residues Our first effort in this study was to devise criteria with now accepted as a result of crystallography and our substituwhich to identify residues for substitution mutation. At that tion mutation analysis, it is instructive to look back to see time, there were no structural data, such as crystal structure which of the three criteria (1, exon IV; 2, 12/12 positional or NMR data, to serve as reference, only compiled cDNA conservation; 3, chemical modification)was best at identifying sequences (22) and a few papers or abstracts describing chem-residues involved in binding glutathione: R13 (criteria 2, 3); ical modification studies (1, 24, 25). As we looked at the Q64 (criteria 1, 2); D98 (criterion 2); Y7 (criterion 2). In the deduced amino acidsequences (Fig. 2), the glutathione S- absence of crystal structure results, clearly criterion 2, positransferases showed the characteristics of a primordial gene tionalconservativenessin a large database, was the best whichhad diverged throughevolutionintomultiple gene indicator. With regard to chemical modification as an indi-
Glutathione-binding Residues
in Glutathione S-Transferase T
18945
residues 1-77 are fully capable of binding to glutathioneagarose. The fact that this glutathione-binding minimal fragment represents the residues comprising theconserved region of exon IV (Fig. 1) reinforces thenotion of a genetically conserved glutathione-binding domain. Future efforts will hopefully define those residues in glutathione S-transferase isozymes that enable on one hand the binding of a wide array of structurally diverse xenobiotics while retaining the almost invariate ability between isozymes to bind certainhydrophobic substrates such as CDNB. Definition of these key residues may then enable the design of glutathioneS-transferase isoformswith medically, genetically, or environmentally useful phenotypes.
0
5
10
p 1 GSH-AGAROSE
15 20 ADDED
FIG.5. Analysis of glutathione (GSH)-binding capacity of wild-type and mutant glutathione S-transferase(GST) s molecules. "S-Labeled glutathione S-transferase A molecules were synthesized using a bacterial S-30 reaction (see "Experimental Procedures").A, one-half of theproduct of each S-30reaction was incubated with 15 pl of a 1:3 slurry of glutathione-agarose (34nmol of glutathione) for 5 min at 20 "C, spun, washed three times, and eluted with 5 m M glutathione; eluates were electrophoresed and the glutathione Stransferase A bound to the glutathione-agarose was quantified by densitometry of the glutathione S-transferase A bond. I?, one-half of the product of each S-30reaction was electrophoresed and immunoblotted using monoclonal antibody N5D12/10 (35)to determine that equivalent amounts of glutathione S-transferase s protein were synrecovery of "S-labeled, wild-type thesized in each S-30 reaction. glutathione S-transferase A protein using increasing volumes of a 1:3 glutathione-agarose slurry.
c,
cator, no one had published the necessary chemical modification study for the Y7 residue in order to implicate it, and for the 3 His residueswhich had been implicated (24) by chemical modification, these residues did not meet criteria 1 and 2 and they have been shown not tobe critically involved in glutathione S-transferase catalysis (32, 33). If criterion 2 is truly instructive, then one must wonder what role(s) the remaining, untested 8 perfectly conserved (Fig. 2) residues play in glutathione S-transferasecatalysis. The importance of the first 80 residues of glutathione Stransferase A in binding glutathione (referred toas domain I, Ref. 29) is also corroborated by results within the accompanying paper (Gulick et al. (35)), where a series of partial, defined fragments of glutathione S-transferaseA sharing only
REFERENCES 1. Mannervik, B. & Danielson, U. H. (1988)CRC Crit. Reu. Biochem. 23, 283-337 2. Pickett, C. B. & Lu, A.Y. H.(1989)Annu. Reu. Biochem. 58,743-764 3. Ketterer, B. (1988)Mutat. Res. 202,343-361 4. Armstrong, R. N. (1991)Chem. Res. Toxicol. 4, 131-140 5. Pickett, C. B., Telakowski-Hopkins, C. A,, Ding, G. J.-F., Argenbright, L. & Lu, A. Y. H. (1984)J. Riol. Chem. 259,5182-5188 6. Lai, H.-C. J. & Tu, C.-P. D. (1986)J. Biol. Chem. 261,13793-13799 7. Kano, T., Sakai, M. & Muramatsu, M. (1987)Cancer Res. 47,5626-5630 8. Vorachek, W. R., Pearson, W. R. & Rule, G. S. (1991)Proc. Natl. Acad. Sci. U. S. A. 88,4443-4447 9. Townsend, A. J. & Cowan, K. H. (1989)Cancer Bull. 41,31-32 10. Graminski, G. F., Kubo, Y . & Armstrong, R. N. (1989)Biochemistry 28, 3562-3568 11. Sambrook,J.,Fritsch, E. F. & Maniatis, T. (1989)Molecular Cloni A Laboratory Manual, 2nd Ed,Cold Spring Harbor LaboratoryPress,?old Spring Harbor, NY 12. Vieira, J. & Messing, J. (1987)Methods Enzymol. 153, 3-11 13. Kunkel, T. A,, Roberts, J. D. & ZaKour, R. A. (1987)Methods Enzymol. 154,367-382 14. Yanisch-Perron, C., Vieira, J. & Messing, J. (1985)Gene (Amst.) 33, 103109 15. Puchalski, R. B. & Fahl, W. E. (1990)Proc. Natl. Acud. Sci. U.S. A. 87, 2443-2447 16. Habig, W. H., Pabst, M. J. & Jakoby, W. B. (1974)J. Biol. Chem. 249, 7130-7139 . - .- . - - .
17. Hanes, C. S. (1932)Biochem. J. 26, 1406-1411 18. Cleland, W. W. (1970)in The Enzymes (Boyer, P. D., ed) 3rd Ed, Vol. 2, pp. 1-65,Academic Press, New York 19. Smith, D. B. & Johnson, K. S. (1988)Gene (Amst.) 67,31-40 20. Okuda, A,, Sakai, M. & Muramatsu, M. (1987)J. Biol. Chem. 262,38587Qfi7
"l,""
21. Cowell, I. G., Dixon, K. H., Pemble, S. E., Ketterer. B. & Taylor, J. B. (1988)Biochem. J. 255,79-83 22. Tu, C.-P. D., Lai, H.-C. J. & Reddy, C.C. (1987) inGlutathione STramferuse and Carcinogenesis (Mantle, T. J., Pickett, C. B. & Hayes, J. P. eds) pp. 87-110,Taylor and Francis, London 23. Katusz, R. M. & Colman, R. F. (1991)Biochemistry 30,11230-11238 24. Awasthi, Y . C., Bhatnagar, A. & Singh, S. V. (1987)Biochem. Biophys. Res. Commun. 143,965-970 25. vanOmmen, B., Ploemen, J. H. T. M., Ruven, H. J., Vos, R. M. E., Bogaards, J. J. P., van Berkel, W. J. H. & van Bladeren, P. J. (1989) Eur. J. Biochem. 18 1,423-429 26. Fahl, W. E., Manoharan, T. H. & Puchalski, R. B. (1990)in Glutathione STransferases and Drug Resistance (Hayes, J. D., Mantle, T. & Pickett, C. B., eds) pp. 186-195,Taylor and Francis, London 27. Manoharan, T. H., Puchalski, R. B., Burgess, J. A,, Pickett, C. B. & Fahl, W. E. (1987)J. Biol. Chem. 262,3739-3745 28. Reddy, K. J., & Gilman, M. (1987)In Current Protocols in Molecular Bwlogy (Ausubel, F. M., Brent, R., Kingston, R. E., Moore, S. D., Seidman, S. G., Smith, J. A,, and Struhl, K., eds) pp. 4.4.1-4.4.7,John Wiley & Sons, Nnur Vork .._.. _"...
29. Reinemer, P., Dirr, H.W., Ladenstein, R., Schiiffer, J., Gallay, 0.& Huber, R. (1991)EMBOJ. 10,1997-2005 30. Liu, X., Zhang, P., Ji, X., Johnson, W. W., Gilliland, G. L. & Armstrong, R. N. (1992)J . Biol. Chem. 267,4296-4299 31. Stelberg, G., Board, P. G. & Mannervik, B. (1991)FEBS Lett. 293,1531ss
32. Wang, R. W.,,Newton, D. J., Pickett, C.B. & Lu, A.Y. H. (1991)Arch. Biochem. Brophys. 286,574-578 33. Zhang, P., Graminski, G. F. & Armstrong, R. N. (1991)J.Biol. Chem. 266,
_"
lQA75-1QA7Q ". . -
34. Chirgwin, J. J.. Przbyla, A. E., Macdonald, R. J. & Rutter, W. J. (1979) Biochemistry 18,5294 35. Gulick, A. M., Goihl, A. L. & Fahl, W. E. (1992)J. Biol. Chem. 267,189461 QOKP lllV"l
36. Manoharan, T. H.,Gulick, A. M., Reinemer, P., Dirr, H. W., Huber, R., & Fahl, W. E. (1992)J . Mol. B~ol.,in press
