THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1993by The American Society for Biochemietry and Molecular Biology,Inc.
Vol. 268, No. 26, Issue of September 15,pp. 19710-19716,1993 Printed in U S A .
Expression, Purification, and Kinetic Characterization of Recombinant Human AdenylosuccinateLyase* (Received for publication, April 1, 1993, and in revised form, June 8, 1993)
Randy L. Stone$#, Howard ZalkinT, and JackE.DixonS(1 From the $Department of Biological Chemistry, University of Michigan Medical School, Ann Arbor, Michigan 48109-0606and the VDepartment of Biochemistry, Purdue University, West Lafayette, Indiana 47907
Adenylosuccinate adenosine 5’-monophospate lyase (EC 184.108.40.206; ASL) catalyzes two distinct reactions in adenosine 5’-monophosphate (AMP) biosynthesis. A S413P mutation inASL segregates with mentalretardation in an affected family (Stone, R.L., Aimi, J., Barshop, B. A., Jaeken, J., Van den Berghe, G., Zalkin, H., and Dixon, J. E. (1992) Nature Genet. 1, 59-63). ASL and S413P ASL have been expressed, purified, Escheand kineticallycharacterized.Loweringthe richia coli growth temperature to 2 5 “C and the concentration of inducer, [email protected]
,D-galactopyranoside, to 40 PM was necessary for synthesis of soluble, tetrameric enzymes. The recombinant enzymes were purified to homogeneity using anion exchange chromatography followed by chromatography on Blue 2A Sepharose. At pH7.0 and 25 “C, the kafor cleavage of 5-amino4-imidazole-N-succinocarboxamide ribotide (SAICAR) by ASLwas 90 8” with a K , of 2.35 PM. The k,, for adenylosuccinate (SAMP)cleavage was97 s” with a K,,, of 1.79 PM. The catalyticmechanism involved one general base catalyst (pK, = 6.4) and one general acid catalyst (pK. = 7.5). ASL follows an ordered uni-bi reaction mechanism with fumarate released first. 5Amino-4-imidazolecarboxamideribotide (AICAR) and AMP were competitive with SAICAR and SAMP (Ki[AIGAR] = 11.3 PM; K C A M P I = 9.2 PM), whereas fumarate inhibited noncompetitively (Kii = 2.3 mM, Kb = 2.8 mM). The competitive inhibition by AICAR and AMP suggests a single active site that binds both SAICAR and SAMP. The kinetic constants at pH 7.0,25 “C and the ko,,JKmuersu8 pH profiles for ASL and S413PASL were very similar. These results are consistent with a catalytic defect. S413P being a structural rather than
The purine nucleotide biosynthetic and catabolic enzymes are biochemically and medically important (Dixon and Zalkin, 1992;Seegmiller, 1980). Biochemically, the end products of purine biosynthesis serve as precursors for DNA and RNA synthesis,intermediatesin biosynthetic reactions, energy storage depots, and as metabolic regulators. Medically, dis-
* This work was supported inpart by National Institutes of Health Grants NIDDKD 18024 (to J. E. D.) and GM46466 (to H. Z.). The costs of publication of this article were defrayed in part by the payment of page charges. This articlemusttherefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. f Recipient of National Association for Research on Schizophrenia and Depression Young Investigator award. (1 To whom correspondence should be addressed Minor J. Coon Professor and Chair, Dept. of Biological Chemistry, University of Michigan Medical School, 5416 Medical Sciences I, Ann Arbor, MI 48109-0606. Tel.: 313-764-8192;Fax: 313-763-4581.
orders such as Down’s syndrome and sensorineural deafness have been associated with purine biosynthetic enzymes (Patterson et al., 1981;Becker et al., 1986). Further, several cases of partial adenylosuccinate AMP-lyase (EC 220.127.116.11;ASL’) deficiency have been correlated with severe mental retardation and secondary autistic features (Jaeken and Van den Berghe, 1984;Jaeken et al., 1988, 1992; Barshop et al., 1989; Van den Bergh et al., 1991a). One case of familial mental retardation segregates with a S413P substitution inASL. The amino acid substitutionresults in astructurallyunstable enzyme (Stone et al., 1992, 1993). ASL is the only enzyme of purine nucleotide biosynthesis to act a t two distinct points in the pathway (Fig. lA).The first ASL-catalyzed reaction is the conversion of SAICAR to AICAR in de mu0 biosynthesis (Fig. 1B).The second ASLcatalyzed reaction is the conversion of SAMP to AMP (Fig. lB),a stepcommon to purine nucleotide synthesis and purine salvage. In Bacillus subtilis, ASL is a regulator of glutamyltRNA synthetase (Gendron et al., 1992). This represents a link between an enzyme capable of monitoring the energy state of the cell and an enzyme catalyzing an early step in protein biosynthesis. It is unknown whether ASL plays a similar role in mammalian cells. ASLalso plays a pivotal role in anti-HIVchemotherapy. The 2’,3’-dideoxynucleoside,ddI, has been shown to inhibit HIV replication significantly (Mitsuya and Broder, 1986). The conversion of ddI to ddATP requires the action of adenylosuccinate synthetase (Fig. U; Powell et al., 1992) and ASL. Adenylosuccinate synthetase converts ddIMP to ddSAMP, and ASL converts ddSAMP to ddAMP, which in turn serves as a substratefor production of ddATP. Importantly, the activity of ASL toward ddSAMP is only 1.85% of the enzyme’s activity toward SAMP (Nair and Sells, 1992). As new mutations in ASL are characterized, it will be useful to compare the properties of the “mutant”ASL enzymes with those of the normal humanenzyme. Additionally, understanding of the ASL catalytic mechanism may provide important insights for the design of more efficaciousanti-HIV nucleotide analogs. For these reasons, we have overexpressed the recombinant human enzyme, obtained it in pure form, and characterized it kinetically. EXPERIMENTALPROCEDURES
Materia&-SAMP, AMP, AICAR, fumarate, PEI-cellulose TLC plates, QEAE-Sepharose, Blue 2A Sepharose, Sephacryl S-300-HR, and gel filtration molecular weight marker (kit MW-GF-1000) were The abbreviations used are: ASL, adenylosuccinate AMP-lyase; HIV, human immunodeficiency virus; AICAR, 5-amino-4-imidazolecarboxamide ribotide; SAICAR, 5-amino-4-imidazole-N-succin0~arboxamide ribotide; dd, dideoxy; ddI, 2’,3’-dideoxyinosine; PEI, polyethyleneimine; PCR, polymerase chain reaction; IPTG, isopropyl-lthio-&D-galactopyranoside.
Expression and Characterization of Human ASL
FIG. 1. Schematic of the purine nucleotide biosynthetic pathway and ASL-catalyzed reactions. Panel A, the figure illustrates themetabolic pathway of purine nucleotide biosynthesis. Solid arrows indicatea single enzymatic step; arrows with dotted lines indicate multiple enzymatic steps. The relevant entry points of the purine salvage pathway are indicated. The position, substrates, and nucleotide products of ASL-catalyzed reactions are indicated. The reaction catalyzed by adenylosuccinate synthetase ( A S ) is also indiproducts cated. Panel B, the chemical structures of the substrates and of the ASL reactions in de novo synthesis and salvage are illustrated. purchased from Sigma. [1-"CIFumarate (12097) was a product of ICN. SDS-polyacrylamide gel electrophoresis equipment and chemicals along with protein assay reagent (Bradford, 1976) were products of Bio-Rad. All other chemicals were of the highest purity available and were used without further purification. All buffers were prepared using deionized, distilled water and were sterilized in an appropriate manner if necessary. Construction of Expression Plasmids-A full-length, human ASL cDNA clone was used as a template for PCR (Stone etal., 1992). The 5' PCR primer had the sequence 5"ACCCTAATGTGCTTCGTGT T T A G and incorporated the initiator methionine within an NdeI restriction endonuclease recognition sequence. The 3' PCR primer had the sequence 5"ACCGAATTCAACAATTGTTCGTTTAA. The 3' end of the resultant PCR fragment terminated with an EcoRI recognition sequence following the termination codon. The PCR cycle conditions were 20 rounds of 94 "C for 30 s, 55 "C for 30 s, and 72 "C for 30 s. The PCR product was subcloned into the expression vector pT7-7 (Tabor and Richardson, 1985). The entire coding sequence of the expression construct, pT7ASL, was verified by DNA sequence analysis. To construct an expression plasmid for the S413P mutant, an NcoI-XhoI restriction fragment housing the point mutation (Stone et al., 1992) was subcloned into the pT7ASL plasmid, replacing the corresponding NcoI-XhoI fragment. The coding sequence was verified in the new plasmid, pT7ASL1751, by nucleotide sequence analysis. Expression and Purification Protocols-The expression plasmids,
pT7ASL and pT7ASL1751, were transformed into Escherichia coli strain BLPl(DE3)lpLysS(Studier and Moffatt, 1986). 10-ml cultures of 2 X YT supplemented with 100 pg/ml ampicillin and 12.5 pg/ml chloramphenicol were inoculated with single colonies from fresh transformations and incubated with vigorous shaking at 37 "C for 2.5 h. The 10-ml cultureswere subsequently transferred to 1-liter cultures = 0.5. The of the same medium and antibiotics and grown to Acultures were immediately cooled to 25 "C in a water bath, and icecold ethanol was added to a final concentration of 2% (v/v). The cultures were induced by the addition of IPTG toa final concentration of 40 p ~ Vigorous . shaking at 25 "C was continued overnight (10-12 h). Thecells were harvested by centrifugation a t 5,000 X g for 5 min. The cell pellet was resuspended in 20 ml of 10 mM Tris-HC1, 10 mM NaCl, 2 mM EDTA, 1 mM dithiothreitol, 20% (v/v) glycerol, pH 7.4 (25 "C; buffer A). This and all subsequent steps were performed at 4 "C. The cells were lysed by three passages through a French press, maintaining pressure above 1,200 p.s.i.A soluble extract was obtained by centrifugation of the cell lysate at 30,000 X g for 20 min. Polyethyleneimine (PEI) was added from a 50% (w/v) stock solution (Sigma) to a final concentrationof 0.35% to precipitate nucleic acids (Burgess and Jendriask, 1975). Following slow stirring for 5 min, the nucleic acid precipitate was removed by centrifugation at 30,000 X g for 20 min. The soluble extract was passed over a 20-ml QEAE-Sepharose column preequilibrated with buffer A. The column was washed with 4 column volumes of buffer A before elution with a 200-ml linear gradient from 10 mM to 1M NaCl (buffer A to buffer A with additional NaCl). 2.5-ml fractions were collected. Column fractions were assayed in buffer A for ASL activity using the standard spectrophotometric assay for SAMP (Parsell and Sauer, 1989). Column fractions containing ASL activity were pooled and concentrated to a final volume of approximately 2 ml using Centriprep-30 concentrators (Amicon). The concentratedfractions were desalted on 10-ml bedvolume, prepacked Sephadex G-25 columns (Pharmacia LKB Biotechnology Inc.) equilibrated in buffer A. The desalted concentrates were subsequently loaded onto a 40-ml, Blue 2A Sepharose column equilibrated in buffer A. The column was washed with 4 bed volumes of buffer A before elution with 80 ml of 10 mM Tris-HC1, 2.5 mM fumarate, 2.5 mM AMP, 2 mM EDTA, pH 7.4 (25 "C). Fractions were monitored for protein content using the Bradford assay (Life Technologies, Inc.; Bradford, 1976). Fractions containing protein were concentrated as before. The purity of the resultant, recombinant ASL was assessed by SDS-polyacrylamide gel electrophoresis on 10% gels followed by Coomassie Brilliant Blue staining and by activity assay following desalting (as described). The recombinant proteins, wild-type and mutant, were stored at 4 'C in elution buffer with the addition of 1 mM dithiothreitol. Size Exclusion Chromatography-Apparent molecular size was determined by gel filtration on a Sephacryl S-300-HR column (5 X 98 cm) in 10 mM Tris-HC1, 150 mM NaC1, 2 mM EDTA, 1 mM dithiothreitol, pH 7.4 (25 "C) at 4 "C. The column was standardized using mass standards a t a concentration of 2.5 mg/ml each. The standards were blue dextran (2,000 kDa), bovine thyroglobulin (669 kDa), horse spleen apoferritin (443 kDa), sweet potato p-amylase (200 kDa), yeast alcohol dehydrogenase (150 kDa), bovine serum albumin (66 kDa), and bovine erythrocyte carbonic anhydrase (29 kDa). The flow rate of the column was 0.2 ml/min. Elution of active ASL was monitored with the standardspectrophotometric assay (Parsell and Sauer, 1989). Purified ASL and S413P ASL (100 pg) were loaded in column buffer plus 10% (v/v) glycerol. Synthesis of P'CISAICAR and SAZCAR Assay-[14C]SAICAR was synthesized from AMP and [1-"Clfumarate essentially as described (Van den Bergh et al., 1991b) with the following modifications. Lyophilized [l-''C]fumarate (6.62 mCi/mmol; ICN) was resuspended in gel filtration buffer (described above) to a final concentration 100 mM, with adjustment of the overall pH to 7.0by the addition of NaOH. 100 mM stock solutions of unlabeled fumarate and AICAR were prepared in the same manner. All stock solutions were stored aliquoted a t -80 "C. In a typical 200-pl reaction, AICAR at a final concentration of 60 mM was allowed to react with fumarate at a final concentration of 10 mM and with [1-"C]fumarate at a final concentration of 10 mM in the presence of60pgof purified ASL. The incubation was continued overnight at 25 "C. The reaction was terminated by spotting on PEI-cellulose a plate. The plate was developed with 1 M ammonium acetate serving as the mobile phase. The products and reactants were visualized on the plate with a short wavelength, hand-held UV light. In this system an AICAR standard had an RF value of 0.36, fumarate had an RF of 0.45, and SAICAR had an RF of 0.07. The SAICAR spot was scraped from the plate and eluted
Characterization Expression and
of Human ASL
with 1 M NH,OH. The PEI-cellulose was then removed by centrifugation a t 30,000 X g for 30 min a t 4 “C in a microcentrifuge (Tomy). [ “CISAICAR was recovered by lyophilization. Typical yields of product were roughly 700 pg/reaction with a specific activity of 5,000 dpm/nmol. For kinetic analyses utilizing [“CISAICAR as a substrate, reaction time points were collected by spotting on PEI-cellulose. Plates were developed and products identified as described above. Both fumarate and SAICAR spots were excised and counted for radioactivity a t each time point. All dpm determinations were made with a Beckman model LS-3801 liquid scintillation counter. [ l-I4C]Fumaratewas used as a dpm standard.
This indicated that the initiator methionine had been removed in E. coli but that the remainder of the amino acid sequence was as anticipated. No contaminants were detected from the sequence data. Table I summarizes the purification scheme for the human enzyme. The PEI precipitation step (Burgess and Jendriask, 1975) removes nucleic acids without affecting the solubility or yield of ASL (Table I, rows 1 and 2). The removal of nucleic acids is particularly helpful in facilitating protein binding to the QEAE-Sepharose column in the subsequent step. The purification factor obtained by anion exchange chromatography was somewhat variable, ranging from 2.5 to 9-fold. However, the impurities in QEAE RESULTS pools were always effeciently removed by elution of ASL from Expression and Purification of Normal and Mutant Human Blue 2A Sepharose with AMP and fumarate (Fig. 2, lanes 2 ASL-Sequences encoding ASL and S413P ASL were tran- and 4 ) . Elution from Blue 2A Sepharose with a salt gradient scribed from the bacteriophage T7, d l 0 promoterin the did not achieve the same degree of purification, suggesting plasmid pT7-7 (Tabor and Richardson, 1985). E. coli strain that theASL active site interactswith the Blue 2A dye (data BL21(DE3) was transformed with each construct for expres- not shown). Recombinant human ASL has a specific activity sion (Studier and Moffatt, 1986). Induction at 37 “C by the nearly 10-fold higher than that of the purified rat skeletal addition of 400 PM IPTG to exponentially growing cultures muscle enzyme (Casey and Lowenstein, 1986) when assayed resulted in theproduction of insoluble, inactive recombinant under identical conditions. The number of purification steps enzyme (data notshown). Previous studieshad suggested that involved in the preparation of the rodent enzyme would alteration of growth conditions could increase the solubility appear to generate a substantial amount of inactive enzyme. of heterologous proteins in E. coli (Steczko et al., 1991). With Recombinant ASL was very unstable when stored at -80 or the pT7 expression system, this could be approached in two -20 “C regardless of salt orglycerol content. The protein was ways. Introduction of the T7bacteriophage lysozyme gene on stable for a period of approximately 1 month when stored in a plasmid (plasmids pLysS and pLysE; Studier andMoffatt, Blue 2A Sepharose elution buffer (see “Experimental Proce1986) will inactivate the noninduced T 7 RNA polymerase dures’’) at 4 “C. (Moffatt andStudier, 1987), thus inhibiting basal expression Tetrameric Nature of Recombinant Human ASL-The acof recombinant protein. Lowering the temperature of induc- tive form of ASL from a varietyof species is tetrameric (Casey tion and concentration of IPTG used also has the effect of and Lowenstein, 1986). To estimate the oligomeric state of moderatingrecombinantproteinsynthesis(Steczko et al., active recombinant ASL and s413PASL, a Sephacryl S-3001991). A systematicarray of inducer concentrations, induction HR size exclusion chromatography column wasemployed. temperatures, and BLPl(DE3)lpLys host strains was tested The column was equilibrated and calibrated with proteins of for production of soluble, active ASL and S413P ASL. The known apparent molecular mass as described under “Experigreatest improvement in solubility was produced in strain mental Procedures.” The elution position of ASL activity was BLBl(DE3)lpLysS with induction for 12 h a t 25 “C by 40 PM determined by the SAMP spectrophotometric assay (Parsell IPTG in the presence of ethanol (2% (v/v); data not shown). and Sauer, 1989). ASL activity eluted from the column just All ASL activity resided in the cleared supernatant following after theelution position of sweet potato @-amylase(200 kDa) lysis and centrifugation. Homogeneous enzyme was obtained but well in advance of yeast alcohol dehydrogenase (150 kDa; by anion exchange chromatography followed by dye ligand Fig. 3). Based on a plot of log(M,) uersus VJV0 for the size affinity chromatography,yielding approximately 3mg of pro- standards,theapparent mass of active human ASLwas tein/liter of bacterial culture for both ASL (Fig. 2) and S413P calculated to be 196 kDa. The apparent mass of S413P ASL ASL. was calculated to be 201 kDa (data notshown). These results Identity and purity of recombinant ASL and S413P ASL were consistent with the active forms of recombinant human were assessed by automated Edmandegradation. The first10 ASL and S413P ASL being tetrameric. No ASL activity was cycles from each protein gave the sequence CFVFSDRYKF. detected at theexpected elution position of monomer, dimer, or trimer, and no protein absorbance (280 nm) was detected a t positions other than those at which activity eluted (data 1 2 3 4 5 not shown). kDa Kinetic Properties of ASL-Detailed analyses of the cata106 lytic properties of human ASL and S413P ASL were con80 ducted over a range of pH values from 5.5 to 9.0 utilizing SAICAR and SAMP as substrates. The results for the wild50 33
TABLE I Purification of human adenylosuccinatelyase
FIG. 2. SDS-polyacrylamide gel electrophoretic analysis of human ASL purification. Lane I , PEI supernatant.Lane 2, QEAESepharose pool. Lane 3, concentrated QEAE pool. Lane 4, Blue 2A Sepharose pool. Lane 5, concentrated Blue 2A pool. The position of molecular weight standards (Life Technologies, Inc. prestained, high molecular weight standards) are indicated. Proteinfractions were separated on a 10% acrylamide gel and subsequently stained with Coomassie Brilliant Blue. 3 pgof protein was loaded in lanes 4 and 5.
Total Total Volume activity protein
High speed superna- 746 40 tant 40 PEI supernatant Dilute QEAE pool 665 50 2 Concentrated QEAE pool Blue 2A pool 15
177 704 356
163 27.3 15.3
337 45.2112.4 3.0
Specific activity pmof/minmg
4.21 4.33 24.4 23.2
Recovery % ’
100 94.4 89.2 41.7
Expression and Characterization of Human ASL
lo&) = log(c/(l
+ H/K, + KdH))
In Equation 1, y is kat/K,,,, c is the pH independent value of H is the concentration of hydrogen ion, and KI and 41 0 Kz are ionization constants for the ionizable groups involved in the reaction mechanism. The plot for each substrate (Fig. 4) has a slope of +1below pH 6.5 and a slope of -1 above pH 3, 7.5. Thus, these profiles indicate that the catalytic mechanisms ofASL involve one general base catalyst and one general acid catalyst. The pK, values for each ionizable group 2. are summarized in Table111. Similar analyses were performed with S413P ASL. It is clear from inspection of Table 111 that the catalytic mechanisms, like the kinetic constants, for ASL 1. and S413P ASL are virtually identical at 25 "C. Product Inhibition Studies and the ASL Reaction Mecha0 nism-The reaction mechanism of an enzyme can be determined by a diagnostic analysis of initial rate kinetics and 1.4 1.2 1.6 1.8 2.0 product inhibition patterns (Cleland, 1963). Purified ASL from yeast (Bridger and Cohen, 1968) and crude ASL from Ve/Vo murine cells (Brox, 1973) seem to follow a simple, ordered FIG.3. Sephacryl S-300-HR elution profile of human ASL. uni-bi reaction mechanism in which fumarate is the first The figure illustrates the elution position of human ASL activity product released, followed by the release of the nucleotide from the column (cj. "Experimental Procedures" for conditions). The size standards are: A , bovine thyroglobulin (669 kDa); B , horse spleen product. The reaction mechanism of human ASL was invesapoferritin (443 kDa); C, sweet potato @-amylase(200 kDa); D,yeast tigated by conducting product inhibition studieswith AICAR, alcohol dehydrogenase (150 kDa); E , bovine serum albumin (66 kDa); AMP, and fumarate. K, values are reported in Table IV. The F, bovine erythrocyte carbonic anhydrase (29 kDa). Inset, calibration effects of AMP and fumarateon the rate of cleavage of SAMP curve showing elution positions of standards (0)and ASL activity are shown in Fig.5, A and B. The inhibition by AMP is (I.* clearly competitive over a wide range of concentrations (Fig. 5A). When the slopes of the reciprocal plots are replotted TABLEI1 against AMP concentration (Fig. 5A, inset), a straight line is pH dependence of A S L catalysis obtained. Thenature of the inhibition and of the replot All assays were performed a t 25 "C with buffer ionic strength held indicates that AMP binds to the free form of the enzyme. In a t 50 mM and overall ionic strength adjusted to 150 mM with NaC1. contrast, the inhibition by fumarate is noncompetitive with
5.504.7 f 0.5 0.8 f 0.1 6.00 14.1 f 2.1 1.4 f 0.3 6.5045.7 f 1.4 1.8 f 0.3 7.00 90.1 f 5.62.4 f 0.5 7.35 181.8f 8.45.2 f 0.9 7.50 259.8f 7.8 12.8 f 2 f 19 18.4 f 3 7.80297 8.00 286 f 9.6 21.2 f 13.5 3 8.20 8.50 702 f 24 108 f6.5 7740 9.00 400 f 9.7 131 f3.1 9425
5.9 5.7 f 0.6 0.7 k 0.3 8.2 10.1 16.3 f 1.7 1.1f 0.4 14.8 25.3 48.6 f 1.6 1.4 f 0.2 34.7 37.597.0 f 5.2 1.8 f 0.353.9 f 12 4.1f 0.7 50.2 35.0 206 f 19 10.6 f 1 26.3 20.3 287 3 16.1 316 f 24 14.0 f 22.6 308 f 9.0 16.0 f 1 19.3 359 f 16 40.3 f 8.9 5 f 26 103 f 5 7.2 f 7.6 125 f 8 3.5
PH type enzyme are summarized in Table 11. The overall ionic strength of each reaction was adjusted to 150 mM with NaC1. FIG.4. pH Profiles of the forward reactions of ASL. All All reactions were conducted at 25 "C because of the instabil- reactions were conducted at 25 "C with buffer ionic strength held at ity of the S413P mutant athigher temperatures (Stone et al., 50 mM and overall ionic strength adjusted to 150 mM with NaCl. The 1992). In theSAMP assays, the Beckman DU-64 spectropho- curves through the data points are computer fits to Equation 1.Data tometer was blanked a t 282 nm against enzyme and buffer points obtained using SAICAR as a substrate are labeled by open boxes (0); data points obtainedusing SAMP as a substrate arelabeled before initiation of the reaction with the addition of SAMP. by open circles (0). In the SAICAR assays, reactions were initiated by the addition of enzyme, and time points were terminated by spotting TABLEI11 on PEI-cellulose (cf. "Experimental Procedures"). kcatand K,,, Ionizations and k,.JK,,, maximum values of ASL and S413P ASL values and kCat/K,,,limits for each substrateare virtually catalysis identical to one another. The catalytic efficacy of ASL apEnzyme and proaches the theoretical diffusional limit(Table 11). The b.clKm(max~ substrate pK.2 pK.1 optimum pH for each form of ASL is pH 7.0. Within experipM" S-' mental error, ASL and S413P ASL were kinetically indistinWild-type guishable at 25 "C (data not shown). 6.29 f 0.16 7.85 f 0.16 36.4f 7.6 SAICAR TO determine the number and pKa values of ionizable 7.54 f 0.19 65.4 f 20.2 6.43 f 0.21 SAMP S413P groups involved in ASL catalysis, curves which fit the data SAICAR 6.32 f 0.29 7.53 f 0.29 47.7 f 20.9 from %,/K,,, uersw pH were generated by computer-assisted, 7.40 f 0.18 71.2 f 20.9 SAMP 6.41 f 0.19 nonlinear regression (Equation 1).
Expression and Characterization of Human ASL
respect to SAMP (Fig. 5 B ) . Replots of slopes and intercepts from the original Lineweaver-Burk plot versus concentration of fumarate are linear (Fig. 5B, inset). These results indicate that fumarate does not combine with the free form of the enzyme to a measurable degree. The product inhibition data using SAMP as a substrate and fumarate or AMP as inhibitor are consistent with an ordered uni-bi reaction mechanism in which fumarate is the first product released. The product inhibition data obtained utilizing SAICAR as a substrate and AICAR or fumarate asan inhibitor are quite similar to those obtained with SAMP, AMP, and fumarate (Fig, 5, C and D). The recombinant human enzyme seems to follow the same ordered reaction mechanism as ASL from other species. DISCUSSION
Using the expression and purification system described in this report, 3 mgof homogeneous human ASL (or S413P ASL) can be produced from a liter of bacterial culture. IniTABLE IV ASL catalytic constants All values are reported for pH 7.0 and 25 "C with overall ionic strength adjusted to 150 mM with NaC1. Constant KSAICAR)
K;(AICAR) &(fumarate) &fumarate) &) &)
2.4 p M 11.3 p M 2.3 mM 1.9 mM 1.8 p M 9.2 p M 2.4 mM 2.0 mM
tially, the recombinant protein produced in the pT7 expression system (Tabor and Richardson, 1985)was completely insoluble and inactive. However, bylowering the temperature of induction to 25 "C, increasing the time of induction to 12 h andreducing the amount of inducer used to 40 WM,a soluble, active protein was obtained. Addition of ethanol during induction further increased the yield of active recombinant, presumably by inducing E. coli heat shock protein synthesis (Steczko et al., 1991). The specific activity of recombinant human ASL is 10-fold higher than that of the purified rat enzyme (Casey and Lowenstein, 1986). Based on this comparison, production of the recombinant human enzyme followed by purification using the scheme outlined in Table I would seem to have several advantages. The rapidity and ease of this expression and purification system should facilitate future mechanistic and structural studieswith ASL. Recombinant human ASL and the S413P variant are kinetically indistinguishable at 25 "C. Previously, we have demonstrated differences between the activities of the two enzymes at elevated temperatures and in low concentrations of denaturant (Stone et al., 1992). Based on our analyses, it seems likely that the S413P mutation generates a structural change in the protein which leads to anunstable but catalytically competent enzyme. There are two ionizable groups in the ASL reaction mechanism which are relevant to catalysis (Fig. 4). The first ionization observed in a plot of kat/K,,, versus pH (pKo= 6.4) could potentially be caused by ionization of a nucleotide phorphoryl group proton or of a succinate moiety carboxylic acid proton. However, the natureof the reaction catalyzed by the enzyme would seem to require a general base catalyst to
FIG.5. Product inhibition plots for the ASL reaction. All reactions were conducted at 25 "Cwith buffer ionic strength held at 50 mM and overall ionic strength adjusted to 150 mM with NaCl. The lines through the datapoints in each Lineweaver-Burk plot are computer fits by linear regression. Panel A , AMP inhibition of the SAMP cleavage reaction. The datapoints represent AMPconcentrations of 0 p~ (O), 21.6 p~ (O), 43.2 p M (A), 64.8 p M (a),and 86.4 p M (m). The inset is areplot of slope with respect to [AMP]. Panel B, fumarate inhibition of the SAMP cleavage reaction. The data pointsrepresentfumarateconcentrations of 0 mM (O), 2.5 mM (O), 5.0 mM (A), and 7.5 mM (0).The insets are replots of slope and intercept with respect to [fumarate]. Panel C, AICAR inhibition of the SAICAR cleavage reaction. Thedata pointsrepresent AICAR concentrations of 0 p M (O), 25 p M (O), 50 p M (A), 75 p M (e),and 100 p M (m). The inset is a replot of slope with respect to [AICAR]. Panel D,fumarate inhibition of the SAICAR cleavage reaction. The datapoints represent fumarate concentrations of 0 mM (0),2.5 mM (0),5.0 mM (A), and 7.5 mM (0).The insets are replots of slope and intercept with respect to [fumarate].
Expression and Characterization of Human ASL
assist in removal of a methylene proton from the succinate marate as a product (Aimi et al., 1990), contains a highly moiety of the substrate. Therefore, it is likely that the ioni- conserved lysyl residue (Lys-270) which could participate in zation in question is that of an enzyme group. Human ASL coordination of a carboxylic acid group. Our product inhibition studies indicate that SAMP and contains a number of phylogenetically conserved amino acid residues that have potential pK, values near 6.4. In the human SAICAR cleavageproceeds via an ordered, uni-bi mechanism. enzyme, these include aspartic acid residues 119 and 243, The nature of AMP and AICAR inhibition is particularly glutamic acid residues 35, 225, and 277, and histidyl residue important. AICAR is a competitive inhibitor of both SAICAR 134. The second ionization associated with catalysis (pK, = cleavage (Fig. 5, A and B ) and SAMP cleavage (data not 7.5) may also be enzyme-associated. In this case, an ASL shown). AMP is competitive a inhibitor of both SAMP cleavgeneral acid catalyst might protonate theamino group left on age (Fig. 5, C and D )and SAICAR cleavage(data notshown). the nucleotide product during C-N bond scission. Candidate Therefore, all nucleotide substrates and products bindto the human ASL amino acid residues include lysyl residues 10 and same form of the enzyme. This form of the enzyme isthe free 270 and histidyl residue 134. Our results are consistent with enzyme. Further, concentration dependence plots of SAMP the mechanism illustrated in Fig. 6 in which general base and SAICAR cleavage by ASLdemonstrate classical Michaeabstraction of the methylene proton is concerted with general lis-Menten kinetics (data notshown). There is no cooperativacid catalysis and scission of the C-N bond. All of the residues itiy. Based on the combined evidence, it seems highly likely discussed here are potential candidates for site-directed mu- that ASL has asingle active site for binding of either SAICAR tagenesis and selective chemical modification studies aimed or SAMP. Moreover, the similarities between SAICAR and at further elucidating the catalytic mechanism. SAMPstructures (Fig. 1B)andthe kinetic constants for From calculations based on the initial rate expression for cleavage of each substrate (TableIV) tend to suggest that the an ordered uni-bi reaction (Cleland, 1963), the association ASL active site contacts each substrate at positions common constants for the reactants and products can be calculated. to both. The importance of further understanding the sites of At pH 7.0, &) = 1.47 pM, K(SAICAR) = 1.92 p M , K(AMP) = the ASL interaction with its nucleotide substrates is under9.23 pM, &CAR) = 11.3 pM, and K(hmarate) = 2.30 mM. The scored by the role ASL plays in anti-HIV chemotherapy and differences in the association constants for each pair of nu- in some forms of mental retardation. cleotide substrates and products canbe used to calculate the The purine salvage pathway, of which ASL is a part, is binding free energy contributed by the succinate moiety of critical in the conversion of the anti-HIV pro-drug ddI to its each substrate. active form, ddATP(Nairand Sells, 1992). The specific Kwc.moiety = KprodKsubn. (Eq. 2) conversion catalyzed by ASL, cleavage of ddSAMP to form ddAMP, is very inefficient. For this therapeutic approach, a AAG,,. moiety = -RT ln(Ksuc. (Eq. 3) better understanding of the contact requirements for ASL and itscompanion enzyme, adenylosuccinate synthetase (Fig. In each case, AAG(BUC. moiety) = 1.2 kcal/mol, implying that the succinate moiety contributes one or two hydrogen bonds to 1B; Powell et al., 1992), would clearly be desirable. The binding. This binding is most likely accomplished through the availability of purified human ASL and a humanadenylosuccarboxylate groups. The so-called “fumarate lyase” signature cinate synthetase cDNA will greatly aid in this type of work. sequence, found in ASL and other lyases which liberate fu- Further, more work is required to determine the underlying bases of the ASL deficiency. We are currently investigating the molecular defects in several patients. The expression and purification scheme andthe baseline catalytic parameters outlined here will also greatly aid this effort. Acknowledgments-We thank Dr. Zhong-Yin Zhang for help and guidance with the kinetic analyses and Dr. Georges Van den Berghe and Dr. Jaak Jaeken for continuing interest inthe project.
FIG. 6. Proposed concerted catalytic mechanism of ASL. The figure shows, in diagram form, the active site of ASL during cleavage of SAMP. General acid and base catalysts on the enzyme are indicated along with their potential identity. Hydrogen bonding is indicated with dotted lines, and the flow of electrons during the concerted cleavage reaction is indicated with arrows.
REFERENCES Aimi, J., Badylak, J., Williams, J., Chen, Z., Zalkin, H., and Dixon, J. E. (1990) J. Biol. Chem. 266,9011-9014 Barshop, B. A., Alberts, A., and Gruber, H. (1989) Biochim. Biophys. Acta 999, 19-23 Becker, M. A., Losman, M. J., Rosenberg, A. L., Mehlman, I, Levinson, D. J., and Holmes, E. W. (1986) Arthritis Rheum. 29,880-888 Bradford, M. M. (1976) A m l . Biochem. 72, 248-254 Bridger, W. A., and Cohen, L. H. (1968) J.Biol. Chem. 243,644-650 Brox, L. W. (1973) Can. J. Biochem. 61,1072-1076 Burgess, R., and Jendriask,J. (1975) Biochemistry 14,4634-4638 Casey, P. J., Abeles, R. H., and Lowenstein, J. (1986) J. Bml. Chem. 261, 13637-13642 Cleland, W. W., (1963) Biochim. Biophys. Acta 67,173-187 Dixon, J. E., and Zalkin, H.(1992) Prog. Nucleic Acids Res. 42, 262-288 Gendron, N., Breton, R., Champagne, N., and LaPointe, J. (1992) Proc. Natl. Acad. Sci. U. S. A. 89,5389-5392 Jaeken, J., and Van denBerghe, G. (1984) Lancet ii, 1058-1061 Jaeken, J., Wadman, S. K., Duran, M., van Sprang, F. J., Beemer, F. A,, Holl, R. A,, Theunissen, P. M., de Cock, P., Van den Bergh, F., and Van den Berghe, G. (1988) Eur. J.Pediatr. 148, 126-131 Jaeken, J., Van den Bergh, F., Vincent, M. F., Casaer, P., and Van den Berghe, G. (1992) J. Inherited Metab. Dis. 16,416-418 Mitsuya, H., and Broder, S. (1986) Proc. Natl. Acad. Sci. U.S. A. 83, 19111915 Moffatt, B. A., and Studier, F.W. (1987) Cell 49,221-227 Nair, V., and Sells,T. B. (1992) Biachim. Biophys. Acta 1119,201-204 Parsell, D. A., and Sauer, R.T.(1989) J. Biol. Chem. 264,7590-7595 Patterson, D., Graw, S., and Jones, C. (1981) Prac. Natl. Acad. Sei.U.S. A. 78, 405-409 Powell, S. M., Zalkin, H., and Dixon, J. E. (1992) FEBS Lett. 303, 4-10 Seegmiller, J. E. (1980) in Diseases of Purine and Pyrimidine Metabolism:
Expression and Characterization of Human ASL
Metabolic Control and Disease (Bondv. . _ . P. K..and Rosenbere. -. L. E.. e&) W. B. Saunders, Philadelphia Shen, B. W., Tsao, F. H., Law, J. H., and Kezdy, F. J. (1975) J. Am. Chem. SOC.97.1205-1208 Spector, T. (1977) Bwchim. Biophys. Acta 481, 741-745 Steczko, J., Donoho, G. P., Dixon, J. E., Sugimoto, T., and Axelrod, B. (1991) Protein Expression Purif. 2, 221-227 Stone, R. L., Aimi, J., Barshop, B. A., Jaeken, J., Van den Berghe, G., Zalkin, H.,and Dixon, J. E. (1992) Nature Genet. 1,59-63 Stone, R. L., Aimi, J., Barshop, B. A., Jaeken, J., Van den Berghe, G., Zalkin,
H., and Dixon, J. E. (1993) Second Annual Mumhener Aduentssymposium: Molecular Genetics, Biochemistry and Clinical Aspects of Inherited Disorders of Purine and Pyrimidine Metabolism (Gresser, U., ed) Springer Verlag, Heidelber ,in press Studier, F. and Moffatt, €3.A. (1986)J. Mol. Biol. 189,113-130 Tabor, S., and Richardson, C . C . (1985) P m . Natl. Acad. Sci. U. S.A. 82, 1074-1078 Van den BerQ F., Vincent, M. F., Jaeken, J., andVan den Berghe, G. (1991a) ed. Bwl. 309B, 277-200 Van den ,er h, F , Vincent, M. F., Jaeken, J., and Van den Berghe, G. (1991b) fkf%&m. '193,287-291