Adenylosuccinate Synthetase from Maize

Plant Physiol. (1 997) 11 4: 549-555

Adenylosuccinate Synthetase from Maize

Purification, Properties, and Mechanism of lnhibition by 5’-Phosphohydantocidin

Eric W. Walters, Shy-Fu Lee, Thierry Niderman’, Paul Bernasconi, Mani V. Subramanian, and Daniel 1. Siehl*

Novartis Crop Protection, Research Division, 975 California Avenue, Palo Alto, California 94304-1 104

Adenylosuccinate synthetase (AdSS) is the site of action of hydantocidin, a potent microbial phytotoxin. A kinetic analysis of the mode of inhibition of a plant adenylosuccinate synthetase by the active metabolite 5’-phosphohydantocidin (5’-PH) was the objec- tive of the present study. AdSS was purified 5800-fold from maize (Zea mays), to our knowledge the first purification of the enzyme from a plant source. N-terminal sequencing established the cleavage site of the previously published deduced sequence of the initial transcript. The subunit molecular mass was determined to be 48 kD and the isoelectric point was at pH 6.1. Values of the Michaelis constant for the three substrates IMP, CTP, and aspartate were 21, 16, and 335 p ~ , respectively. lnhibition of AdSS by 5’-PH was measurably time-dependent. The trace of the inactivation curve cpuld not be altered by preincubating the enzyme and inhibitor in the absence of substrates but could be linearized by preincubating the enzyme with inhibitor, aspartate, CTP (or CDP), and inorganic phosphate. lnhibition of AdSS by 5’-PH was competitive with IMP, with an apparent Ki of 22 nM. Apparently, 5’-PH inhibits the enzyme by binding to the IMP site and forming a tight, dead-end complex.

Hydantocidin (Fig. 1) is a potent, nonselective phytotoxin isolated from culture broths of Streptomyces kygro- scopicus (Nakajima et al., 1991). Diagnostic of its mode of action was that it blocked the conversion of IMP to AMP (Heim et al., 1995; Siehl et al., 1996) in the de novo pathway of purine biosynthesis. Neither of the enzymes of this process was inhibited by hydantocidin, but the first, AdSS (1MP:L-aspartate ligase [GDP forming], EC 6.3.4.4), was strongly inhibited by a 5’-phospho derivative of hydantocidin (Siehl et al., 1996). Thus, hydantocidin apparently kills plants by inhibiting AdSS after first being metabolized in vivo to 5‘-PH by an as yet unknown mechanism.

AdSS is known to be the site of action of severa1 antibi- otics, including hadacidin and alanosine (Stayton et al., 1983). The crystal structure of the Esckerickia coli enzyme has been elucidated (Poland et al., 1993; Silva et al., 1995; Poland et al., 1996a, 1996b, 1996c), and the kinetic mechanism has been well described by ongoing investigations (Rudolph and Fromm, 1969; Bass et al., 1984; Cooper et al., 1986; Kang and Fromm, 1995). Knowledge of AdSS from plants consists of partia1 purification and characterization

Present address: Novartis Crop Protection, R-1040.P7A, CH-

* Corresponding author; e-mail [email protected]; fax 4002, Basel, Switzerland.

1- 415- 493-1073.

of the enzyme isolated from wheat germ (Hatch, 1966; Fonné-Pfister et al., 1996) and DNA sequences coding for the Triticum aestivum, Zea mays, and Arabidopsis tkaliana enzymes (Fonné-Pfister et al., 1996; Potter and Ward, 1996).

MATERIALS A N D METHODS

5’-PH and derivatives were synthesized by a variation of the procedure of Mio et al. (1991). A11 other compounds were purchased from Sigma.

Extraction and Purification of AdSS

AdSS was purified from 4- to 5-d-old etiolated seedlings of maize (Zea mays). A11 procedures were performed at 4°C. Seedlings were homogenized in a blender (Waring) with a 1 : l ratio (w/v) of tissue a n d AdSS buffer (50 miv N-[2- hydroxyethyl]piperazine-N’-3-propanesulfonic acid, pH 7.5, 5 mM MgCl,, 1 mM Na, EDTA, 5 mM DTT, and 100 mL/L glycerol) with 10 mM sodium bisulfite and 0.1 mM PMSF. The extract was expressed through Miracloth (Cal- biochem) sandwiched between layers of cheesecloth. Pro- tamine sulfate was slowly added to the extract to a concentration of 4 g/L, and the mixture was stirred for 30 min. The resulting precipitate was centrifuged at 45,0008 for 30 min. The supernatant was fractionated with ammonium sulfate precipitation. Proteins precipitating between 45 and 55% of saturation were retained. The resuspended pellets were desalted with PD-10 columns (Pharmacia) to reduce their conductivity to that of AdSS buffer (1000 pmhos/ cm) and were loaded on a 75-mL bed of DEAE-cellulose (DE-52, Whatman). After a 3-bed-volume wash, the column was eluted with a 10-bed-volume gradient of O to 0.4 M KC1 in AdSS buffer. The active fractions, with conductivities rang- ing from 3500 to 5000 pmhos/cm, were concentrated in a stirred cell (YM-30, Amicon) and then diluted with AdSS buffer containing 0.75 M ammonium sulfate. The preparation was then loaded on a 75-mL bed of phenyl Sepharose HP (Pharmacia) equilibrated with AdSS buffer containing 0.75 M ammonium sulfate, and the column was washed with 3 bed volumes of the same buffer. A 10-bed-volume gradient (0.75-0 M ammonium sulfate) was run, the active fractions were pooled, and the buffer was exchanged (by

Abbreviations: AdSS, adenylosuccinate synthetase; GTP./S, guanosine 5‘-0-(3-thiotriphosphate); k,,,, rate constant of the enzyme-catalyzed reaction at maximum velocity; Kicappr apparent K,; 5‘-PH, 5’-phosphohydantocidin;

549

550 Walters et al. Plant Physiol. Vol. 1 1 4, 1997

Hydantocidin N-Acetyl-5'-phosphohydantocidin

5'-Phosphohydantocidin N-Methyl-5'-phosphohydantocidin

Figure 1 . Structures of hydantocidin and derivatives.

concentration and dilution with a stirred cell) to AdSS buffer containing 0.2 M KC1. The preparation was loaded on a 2-mL bed of GTP agarose equilibrated with AdSS buffer and 0.2 M KCl. After a 10-bed-volume wash, AdSS was eluted with 3 mM GTP in AdSS buffer containing 0.2 M

KCl. A variation of this procedure was performed in which the phenyl Sepharose HP step was omitted. This truncated procedure produced 6-fold more of nearly homogenous AdSS than the one described here and was used for the kinetic studies. AdSS purified by the short procedure was shown to be free of contaminating GTPase and IMP nucleotidase activities by the HPLC method described below.

Prior to its use in kinetic studies, AdSS purified by the short procedure was stabilized by the addition of chicken egg albumin (Sigma A5503, grade V, minimum 98% purity) to 5 mg/mL. GTP was then removed by passing the enzyme solution through a PD-10 column equilibrated with AdSS buffer. The ovalbumin was assayed by HPLC (see below) to verify its inertness to the substrates and its absence of contaminating nucleotides.

Chromatofocusing was performed on a Mono-P column (Pharmacia) according to the manufacturer's instructions for the pH range 7 to 4, except that starting and elution buffers contained the following amendments: 2 mM MgCl,, 1 mM DTT, and 100 mL L-l ethylene glycol. The collected fractions were quickly tested for pH and then adjusted to neutrality with 1 M N-(2-hydroxyethyl)piperazine-N'-3- propanesulfonic acid at pH 7.5 to protect enzyme activity.

Enzyme Assays

AdSS was assayed at 22°C in AdSS buffer (defined above) with the addition of 10 mM potassium phosphate, pH 7.5. The concentrations of the three substrates (aspartate, IMP, and GTP) varied according to the design of the individual experiment. The production of adenylosuccinate was monitored continuously at 280 nm using a spec- trophotometer (DU-7, Beckman). The factor used to convert the rate of change in absorption into rate of adenylosuccinate production (10,200 M - ~ cm-') was calculated as the difference in extinction that results when 1 mo1 of adenylosuccinate is formed and 1 mo1 of IMP is consumed. Because of the presence of exogenous GTPase activity, for

the crude preparation only, it was necessary to add a GTP-regenerating system consisting of pyruvate kinase and 75 p~ PEP. The amount of pyruvate kinase (P-1903, Sigma) used was adjusted to yield approximately a 20-fold excess of GDP kinase activity (PEP + GDP - pyruvate + GTP, quantified by linkage to NADH oxidation by lactate dehydrogenase) over AdSS activity.

The kinetic analysis of velocity as a function of the concentrations of IMP, GTP, and aspartate was performed by the program EnzymeKinetics (version 1.11, Trinity Soft- ware, Campton, NH) using the Lineweaver-Burk transfor- mation of the Michaelis-Menten equation with fouith- power weighting.

In experiments in which partially purified AdSS was used, we needed to ensure that the preparation was free of extraneous GTPase, IMP-5'-nucleotidase, and adenylosuccinate lyase activity. Anion-exchange HPLC was used to monitor the stability of GTP and IMP, each present at 200 p~ in a mixture of AdSS buffer and enzyme after up to 4 h incubation at 22°C. At the appropriate incubation time 80 pL of the incubation mixture was injected on a column (PRP-X100, Hamilton, Reno, NV) equilibrated with 3% acetonitrile. GTP, GDP, GMP, and IMP were eluted with a convex gradient (curve = -2) to 6% 1 M potassium phosphate, pH 7, 12% 2 M KC1, and 3% acetonitrile delivered over a span of 30 min at a flow rate of 0.8 mL/min by a pump (mode1410 Bio, Perkin-Elmer). The eluate was monitored with a diode array detector (model LC480, Perkin- Elmer) at 270 nm. The identities and concentrations of the GTP, GDP, GMP, and IMP were verified by comparisons of UV spectra and retention times with known standards. Stability of adenylosuccinate was monitored in a reaction stopped by the addition of 5 ~ L M 5'-PH and then chromato- graphed by the same procedure.

Protein concentrations were estimated (within the linear portion of the standard curve) by the Bradford assay (Bio- Rad) using ovalbumin as the standard. Homogeneity of purified AdSS was ascertained by SDS-PAGE. A 10% Tris- Gly resolving gel with a 4% stacking gel (Bio-Rad) was used, and proteins were visualized by silver staining. Gel- filtration chromatography was on Superose 12 (Pharma- cia). The mobile phase was AdSS buffer with various com- binations (a11 yielding the same result) of these amendments: 100 mM KC1, 10 mM potassium phosphate, pH 7.5, 3 mM aspartate, 20 p~ IMP, and 20 p~ GDP. N-terminal sequencing of the purified enzyme was performed at the Protein Structure Laboratory (University of California, Davis).

RESULTS

Purification to homogeneity as determined by SDS- PAGE (Fig. 2A) was achieved as described above. This procedure yielded 3% of the activity present in the crude extract, which was insufficient for the kinetics study but sufficient for N-terminal sequencing. By omitting the phenyl Sepharose step, we purified AdSS to near homogeneity (Fig. 2B) and recovered 20% of the starting activity (Table I). By both procedures, the key step in the purification was affinity chromatography on GTP agarose, which alone pro-

Adenylosuccinate Synthetase from Maize 551

Table II. Properties of AdSS from maize

45-

31-

21-Figure 2. SDS-PAGE of maize AdSS. A, The purification procedureincluded phenyl Sepharose; 100 ng of protein was loaded. B, Puri-fication resulting when phenyl Sepharose was omitted; 80 ng ofprotein was loaded.

Property Value

Activity in maizeSubunit molecular massNative molecular massc

PipH optimumvmix

IMP

GTPAsp

2.8 nmol min ' g ' fresh wt48,000 Dn; 47,933 Db

50,000 Da6.1d

7.83.8 fimol min"1 mg~'3.1 s-'

2116335

0 Determined by SDS-PAGE. b Calculated from the aminoacid sequence of the mature protein. ' Determined by gel-filtration chromatography on Superose 12. '' Measured bychromatofocusing.

vided 400-fold purification. Although the high-yield prep-aration contained trace impurities (Fig. 2B), the absence ofinterfering enzymatic activities was confirmed by theHPLC procedure described in "Materials and Methods."This preparation was utilized for subsequent kinetics anal-ysis. For all preparations used, it was shown that IMP,GTP, and adenylosuccinate were not degraded by otherenzymes in the preparation.

Properties

The enzymological properties of maize AdSS are sum-marized in Table II and are compared in the "Discussion"with those of AdSS from diverse species.

N-Terminal Sequence

The sequence obtained (Table III) was identical to that ofresidues 42 to 75 of the deduced amino acid sequence of themaize AdSS gene (Potter and Ward, 1996), which establishesthat cleavage of pre-AdSS occurs between positions 41 and42. With this knowledge, we calculated the subunit molec-ular mass of the mature sequence, and the value was nearlyequal to that obtained by gel electrophoresis (Table II).

Substrate-Saturation Kinetics

Substrate-saturation kinetics were performed by varyingthe concentrations of each substrate at saturating concen-trations of the other substrates. Lineweaver-Burk plots of

the saturation data for each of the three substrates werelinear, showing no evidence of subunit cooperativity. Ap-parent Km values for IMP, GTP, and aspartate were 21, 16,and 335 JU.M, respectively.

Inhibition by 5'-PH

Previously, we reported that N-acetyl-5'-PH (Fig. 1)was a potent inhibitor of AdSS and suggested that the trueinhibitory form of hydantocidin was 5'-PH (Siehl et al.,1996). We have since obtained several lines of evidencethat the N-acetyl group is very labile in aqueous solutionand that our data were actually obtained with 5'-PH.First, when 5'-PH became synthetically accessible, its 50%inhibition concentration against the synthetase was ex-actly the same as that obtained for N-acetyl-5'-PH,namely, 150 nM at 200 /AM IMP. Later, an analog with theacetyl group replaced by a stable methyl group (Fig. 1)was found not to inhibit the enzyme. It seems highlyimprobable that an N-methyl but not an N-acetyl substi-tution would render 5'-PH noninhibitory. Finally, thecrystal structure of a complex of bacterial AdSS, GDP, Pi,succinate, and N-acetyl-5'-PH was solved, and althoughthe hydantoin portion was not unambiguously defined,there was no evidence for the presence of the acetyl group(B.W. Poland and R.B. Honzatko, unpublished data).Therefore, we consider the inhibitor used in the former

Table I. Purification of AdSS by the procedure with which phenylSepharose is omitted

Purification Step

Crude extractProtamine sulfateAmmonium sulfateDEAE-celluloseGTP agarose

Total Activity

nmol min" '

284014201323861568

Yield

%

10050463020

SpecificActivity

nmolmin" ' mg~ '

0.660.331.929.51

3840

PurificationFactor

10.52.9

14.45810

Table III. N-terminal amino acid sequence of maize AdSS

40MSLSTLSHPA AAAAGSGKSL FPAGPAAQSV HFPKARLPVP

80AAVSAATAAV HAEDRVSSLT QVSGVLGSQW GDEGKGKLVD3

AVSAATAAVH AEDRVSSLTQ VSGVLGSQXG DEGh __________a Deduced amino acid sequence of maize AdSS cDNA (Potter and

Ward, 1996). b N-terminal sequence of AdSS purified from maize(italic, present study).

552 Walters et al. Plant Physiol. Vol. 114, 1997

GTPIIMP NO 5’-PH

2.01 I / GTPIIMP I

O 30 Minutes

Figure 3. Time c o m e of inhibition by 5’-PH of AdSS (partially purified through the DEAE-cellulose step, Table I ) . At -15 min, enzyme was combined with 5 p~ 5’-PH, 10 mM potassium phosphate (pH 7.5), and the substrate(s) shown, in a volume of 190 pL . At O min reactions were started by the addition of 10 p L of the substrate(s) absent from t h e preincubation. For example, the bottom curve was obtained by preincubating enzyme with 10 mM potassium phosphate, 5 y~ 5’-PH, 3.16 mM aspartate, and 421 mM IMP. After 15 min the reaction was started by the addition of 1 O p L of 4 mM CTP. Final substrate concentrations were aspartate, 3 mM; CTP, 200 p ~ ; and IMP, 400 p ~ . Some of the lines were repositioned to avoid overlap.

study to be 5‘-PH, which has since been synthesized and its structure confirmed by NMR.’

Previously, we reported that inhibition of AdSS by 5’-PH was time-dependent if the inhibitor was added to the enzyme simultaneously with a mixture of the three substrates or if the enzyme was incubated for 30 min in the presence of 5’-PH prior to starting the reaction with the addition of the substrate mixture (Siehl et al., 1996). These observations suggest that one or more of the substrates must be present for the inhibitor to inactivate the enzyme. To determine which substrate(s) must be present, enzyme was preincubated with 5 p~ 5’-PH for 15 min with a11 possible combi- nations of one or two of the substrates, and then reactions were started by the addition of the missing substrate(s). As shown in Figure 3, only the combination of aspartate and GTP during the preincubation could eliminate the slow on- set of inhibition. Under these conditions the initial rate was linear and nearly as slow as the final inhibited rate eventually attained by a11 reaction mixtures. Linearization could not be achieved in the absence of substrates, even when the enzyme was incubated with 2 p~ 5’-PH for up to 24 h.

The minimum effective concentration of GTP that, if present during the preincubation resulted in linear, inhibited reaction rates, was 1 p~ (data not shown), which was much lower than its K, value. To determine whether linearization specifically required GTP, enzyme was preincubated for 15 min with 5 p~ 5’-PH, 3 mM aspartate, AdSS buffer containing 10 mM potassium phosphate, and various nucleotides at 10 p ~ . Reactions were started at time O with the addition of IMP and GTP for final concentrations of 400

‘Spectral data supporting the structure of 5‘-PH: ‘H NMR (D20) S 3.84 (2 H, dd), 4.29 (1 H, dd), 4.41 (1 H, m), and 4.58 (1 H, d); I3C NMR (D,O) 6 63.7, 72.1, 74.3, 86.7, 95.8, 163.1, and 179.5.

and 200 p ~ , respectively. Preincubation of enzyme with 5’-PH, aspartate, Pi, and either GDP or GTPyS (a nonhy- drolyzable GTP analog in which sulfur takes the place of oxygen between the P- and y-phosphorus) resulted in a linear, inhibited reaction (data not shown). When the nucleotide present in the preincubation was GMP, ATP, xanthosine-5‘-triphosphate, or adenylosuccinate, time- dependent inhibition was observed. Thus, linearization during the preincubation specifically required GDP or GTP yS.

Following a similar protocol, we substituted various amino acids for aspartate in the preincubation to assess the specificity of aspartate in promoting the inactivation of AdSS by 5’-PH. Only when aspartate, GTP, and Pi were present in the preincubation was the enzyme inactivated by 5’-PH (data not shown). Glutamate, Gln, and Ala were ineffective. Succinate, which is a modest competitive inhibitor versus aspartate (Stayton et al., 1983), was also effective in allowing inactivation when present at 3 mM (data not shown).

The final factor that must be present in a 15-min preincubation mixture to yield linear, inhibited reaction rates was found to be Pi. Enzyme was preincubated with 5 p~ 5‘-PH, 3 mM aspartate, 10 p~ GTP, and 10 mM or no potassium phosphate, and then reactions were initiated with the addition of 400 and 200 p ~ , respectively, IMP and GTP. When Pi was absent from the preincubation, the ensuing enzymatic reaction was not linear but exhibited the same dependency of inhibition on time that was seen with no preincubation (data not shown). It was interesting that the development of inhibition did not depend on the addition of Pi upon initiation of the reaction. This presum- ably was because the hydrolysis of GTP resulting from substrate turnover eventually generated sufficient Pi to facilitate the inactivation of the enzyme by 5’-PH. In addition to that effect, 10 mM potassium phosphate activated AdSS by 70% compared with reaction mixtures lacking Pi.

We reported previously that binding of 5’-PH to AdSS could be reversed by passing the enzymelinhibitor complex through a gel-filtration column (Siehl et al., 1996). To obtain a view of the speed of release, a preformed enzymel inhibitor complex was diluted 40-fold and the reaction rate was observed. Restoration of the reaction rate to near that of the control required 6 to 10 min (data not shown), showing that release of the inhibitor is also slow.

Kinetics of lnhibition by 5’-PH

with knowledge of the conditions that allow binding of 5’-PH, we could design procedures for measuring linear, inhibited reaction rates. These are given in the legend to Figure 4. PH inhibited AdSS competitively with IMP, with an apparent Ki of 22 nM (Fig. 4A). The conditions for analysis of the kinetics of inhibition versus IMP as defined in Figure 4 were critical. If IMP and 5‘-PH were not simultaneously present during preincubation, the slow release of 5’-PH from the tight enzymel inhibitor complex prevented IMP from gaining access to its binding site. The effect was simply to titrate some of the enzyme out of the reaction, resulting in a noncompetitive plot.


-0.04 -0.02 0.02 0.04 0.01 0.02 0.03 0.04 0.05 0,001 0.002 0.003 0.004

l/[Asp], l/mM lI[IMP], l/mM l/[GTP].l/pM

Figure 4. Kinetics of inhibition of AdSS by 5’-PH. The inhibitor and varying substrate were allowed to interact with AdSS during a 1 O-min preincubation. Product formation was initiated by the addition of a final component as follows: A, lnhibition versus IMP. The preincubation mixture contained 3 mM Asp, 5 p~ GDP, the indicated concentration of 5’-PH, and varying concentrations of IMP. After 10 min, the reaction was started with the addition of 200 p~ GTP. B, lnhibition versus GTP. The preincubation mixture contained 3 mM Asp, varying concentrations of CTP, and the indicated concentration of 5’-PH. After 10 min, the reaction was started with the addition of 400 p~ IMP. C, lnhibition versus Asp. The preincubation mixture contained varying concentrations of Asp, 200 p~ GTP, and the indicated concentration of 5’-PH. After 1 O min, the reaction was started with the addition of 400 p~ IMP. The concentrations indicated for nonvarying substrates were saturating. All reaction mixtures contained 67 ng of purified AdSS in a final volume of 200 pL. lnitial reaction rates were monitored for 5 min.

The plots of inhibition with respect to GTP and aspartate (Fig. 4, B and C) indicate that 5‘-PH binds poorly to the enzyme-IMP-Pi-aspartate complex (Fig. 4B; Ki(app) = 1.1 FM) or to the enzyme-IMP-ri,-GTP complex (Fig. 4C; = 1.1 PM). However , because binding of 5‘-PH is compet - itive with IMP, these Ki values were actually a function of the concentration of IMP, which was 400 FM in plots shown in Figure 4, B and C. Binding to the enzyme-IMP-Pi- aspartate-GTP complex was much better, as indicated by the Ki(,pp) values calculated from the data in Figure 4, B and C. These were 296 and 10 nm, respectively. These data provide a kinetic explanation for the qualitative observation that aspartate and GTP must be present for 5’-PH to inactivate the enzyme, as described above.

DISCUSSION

In this paper we describe the first purification, to our knowledge, of AdSS from a plant source, as well as its mode of inhibition by 5’-PH, the presumed active derivative of the potent phytotoxin hydantocidin, a natural product isolated from Streptomyces hygroscopicus. Affinity chromatography on GTP agarose, which provided 400-fold purification, was the key step in the procedure. The loss of 50% of the activity in the protamine sulfate step was puzzling, because the treat- ment is typically done to remove nucleic acids. AdSS is known to bind to the specific sequences of DNA in yeast that constitute the T-rich strand of the consensus core of the autonomously replicating sequence (Zeidler et al., 1993). Under the premise that maize AdSS has similar DNA- binding properties and that these would be ionic in nature, we included 0.4 M KCl in the buffer during the protamine sulfate precipitation. This resulted in complete recovery of AdSS in the supernatant but introduced insurmountable problems in later stages of the purification and was therefore omitted. Whether or not the improved recovery of AdSS

activity was due to release of enzyme from nucleic acid was not investigated further.

The subunit molecular mass of 48,000 D (47,933 calculated) is close to that of AdSS from severa1 other sources, i n c l u d i n g 47,300 for the E . coli enzyme (Moe et al., 1996). Others ranged from 51 kD for the Yoshida sarcoma enzyme to 60 kD for the Novikoff ascites enzyme (Stayton et al., 1983). In the present study AdSS from both maize and E. coli were recovered from the gel-filtration column at an elution volume corresponding to 50 kD. However, the crystal structure of the E . coli enzyme shows salt links between subunits related by molecular symmetry and an Arg resi- due at position 143 that projects into the active site of a neighboring subunit (Poland et al., 1993). The involvement of Arg-143 in the binding of IMP and GTP was shown by site-directed mutagenesis (Moe et al., 1996; Wang et al., 1997). These data suggest a dimeric quaternary structure. An Arg equal in context to Arg-143 is present in the sequence of maize AdSS (Fonné-Pfister et al., 1996). Our attempts to alter the elution volume of AdSS from the gel-filtration column by including the substrates in the mobile phase were unsuccessful.

The k,,, of maize AdSS, calculated assuming one catalytic center per 48-kD subunit, was 3.1 s-’. This is comparable to the k,,, of AdSS from Yoshida sarcoma ascites, which was calculated by Matsuda et al. (1980) to be 2 spl, as well as to the acidic isozyme from rat liver (2.63 s-’, calculated by Rudolph and Clark [1982]) and the E . coli enzyme (1.56 s-’, Liu et al. [1992]).

The pH optimum of maize AdSS (7.8) falls outside the average range (6.5-7.5) tabulated by Stayton et al. (1983) for most proteins with the exception of the AdSS from Leish- mania donovani (7.7-8.7). The pI was at pH 6.1, which is close to that of the acidic form of the mammalian liver enzyme, i.e. pH 5.9 (Stayton et al., 1983).

554 Walters et al. Plant Physiol. Vol. 114, 1997

The apparent K , values obtained for IMP, GTP, and aspartate, 21, 16, and 335 pM, respectively, were very close to those reported for the E. coli enzyme by Rudolph and Fromm (1969), namely, 20, 10, and 350 p ~ , respectively. The values for the E. coli enzyme were reproduced in the present study (not shown) using enzyme obtained from Dr. Richard Honzatko (Iowa State University, Ames).

The cause of the slow inhibition and reactivation of AdSS by 5’-PH is unknown, but it is common for tight-binding inhibitors (Schloss, 1988). Finding conditions in which reactions containing 5‘-PH exhibited linear initial reaction rates was critica1 for obtaining the plots that characterize the mode of inhibition of AdSS by 5’-PH. This was achieved by including a 10-min preincubation of mixtures, which did not allow substrate turnover but allowed binding of the inhibitor. Thus, our binding constant is for the final tight complex, and does not express the affinity of 5’-PH for the enzyme at the moment of encounter. Necessary components of the preincubation were enzyme, inhibitor, aspartate, GTP (or GDP), and Pi. We considered the possibility that the requirement for GTP relates to the proposed mechanism of the AdSS reaction. Lieberman (1956) and Fromm (1958) suggested that IMP is phosphorylated at the 6-position of the purine ring, which then subjects IMP to nucleophilic attack by the a-nitrogen of aspartate. By analogy, the requirement for GTP could be for phosphorylation of 5‘-PH at the 2-position (carbonyl carbon) of the hydantoin ring, and 2-5‘- diphosphohydantocidin is the true tight-binding inhibitor. However, the fact that GDP and GTPyS could substitute for GTP to function in combination with aspartate in allowing binding of 5‘-PH to the enzyme argues against any requirement for hydrolysis of the y-phosphate in the inhibitory mechanism. Furthermore, the radiographic structure of a complex of E. coli AdSS, GDP, HP0,2p, Mg2+, hadacidin (an aspartate analog), and 5’-PH showed no phosophoryl group on the hydantoin ring of the 5’-PH (Poland et al., 1996~). This was also the case in the singly ligated 5‘-PH-AdSS crystal (Fonné-Pfister et al., 1996).

Our kinetic data show competitive inhibition of AdSS with respect to IMP. That the two ligands actually share the same site was shown by radiographic crystallography, in which electron density for 5’-PH mapped to the IMP site (Poland et al., 1996~). The same study also provided a struc- tural correlate for our observation that Pi must be present for 5‘-PH to inactivate the enzyme. Mg2’ and HP02- were observed to form a bridge between GDP and 5’-PH, forming contacts with the substrates, the enzyme, and each other.

Severa1 results suggest that binding of GTP and 5’-PH to AdSS is mutually synergistic. First, we saw no evidence for binding of 5’-PH when GTP (or GDP) was absent, whereas with GTP present, an apparent Ki of 22 nM was measured. Second, GTP was effective for promoting inactivation of AdSS by 5’-PH at a concentration (1 p ~ ) one-tenth or less of its apparent K , (16 p ~ ) . Further evidence for synergistic binding of these ‘ligands comes from radiographic crystallography of the E. coli enzyme ligated with GDP, HPO,’-, Mg2+, hadacidin, and 5’-PH (Poland et al., 1996~). Ligand binding caused movement in specific loops and ordering of loops, which were disordered in the unligated enzyme.

Changes induced by one ligand could result in facilitated binding of another ligand.

Although we could not obtain linear, inhibited reaction rates by preincubating AdSS with low micromolar concentrations of 5‘-PH in the absence of aspartate and GTP, the kinetic plots of inhibition versus GTP and aspartate (Fig. 4, B and C) were not precisely parallel, indicating that 5’-PH can bind to complexes of AdSS lacking GTP or aspartate. This is confirmed by a study in which crystals of E . coli AdSS in complex with 5’-PH were obtained by co- crystallization from 0.4 mM AdSS and 2 mM 5‘-PH (Fonné- Pfister et al., 1996).

In conclusion, inhibition of maize AdSS by 5‘-PH ap- pears to result from binding of the inhibitor to the IMP site, forming a dead-end complex that is 1000-fold tighter than that with IMP.

ACKNOWLEDCMENTS

We thank Dr. Richard Honzatko for providing purified E. coli AdSS, for critically reading the manuscript, and for providing preprints of relevant literature.

Received November 12, 1996; accepted March 4, 1997. Copyright Clearance Center: 0032-0889/97/ 114/0549/07.

t

LITERATURE ClTED

Bass MB, Fromm HJ, Rudolph FB (1984) The mechanism of the adenylosuccinate synthetase reaction as studied by positional isotope exchange. J Biol Chem 259: 12330-12333

Cooper BF, Fromm HJ, Rudolph FB (1986) Isotope exchange at equilibrium studies with rat muscle adenylosuccinate synthetase. Biochemistry 25: 7323-7327

Fonné-Pfister R, Chemla P, Ward E, Girardet M, Kreuz KE, Honzatko RB, Fromm HJ, Schar H-P, Griitter MG, Cowan- Jacob SW (1996) The mode of action and the structure of a herbicide in complex with its target: binding of activated hydantocidin to the feedback regulation site of adenylosuccinate synthetase. Proc Natl Acad Sci USA 93: 9431-9436

Fromm HJ (1958) On the equilibrium and mechanism of adenylo- succinic acid synthesis. Biochim Biophys Acta 29: 255-262

Hatch MD (1966) Adenylosuccinate synthetase and adenylosuccinate lyase from plant tissues. Biochem J 98: 198-203

Heim DR, Cseke C, Gerwick BC, Murdoch MG, Green SB (1995) Hydantocidin: a possible proherbicide inhibiting purine biosynthesis at the site of adenylosuccinate synthetase. Pestic Biochem Physiol53: 138-145

Kang C, Fromm HJ (1995) Identification of an essential second metal ion in the reaction mechanism of Escherichia cozi adenylosuccinate synthetase. J Biol Chem 270 15539-15544

Lieberman I (1956) Enzymatic synthesis of adenylosine-5’- phosphate from inosine-5’-phosphate. J Biol Chem 223: 327-339

Liu F, Dong Q, Fromm HJ (1992) Site-directed mutagenesis of the phosphate-binding consensus sequence in Escherichia coli adenylosuccinate synthetase. J Biol Chem 267: 2388-2392

Matsuda Y, Shimura K, Shiraki H, Nakagawa H (1980) Purifica- tion and properties of adenylosuccinate synthetase from Yo- shida sarcoma ascites tumor cells. Biochim Biophys Acta 616:

Mio S, Kumagawa Y, Sugai S (1991) Synthetic studies on (+)- hydantocidin (3): a new synthetic method for construction of the spiro-hydantoin ring at the anomeric position of D-ribofuranose. Tetrahedron 47: 2133-2144

Moe OA, Baker-Malcolm JF, Wang W, Kang C, Fromm HJ, Col- man RF (1996) Involvement of arginine 143 in nucleotide substrate binding at the active site of adenylosuccinate synthetase from Eschevichia coli. Biochemistry 35: 9024-9033

340-350


Nakajima M, Itoi K, Takamatsu Y, Kinoshita T, Okazaki T, Kawakubo K, Shindo M, Honma T, Tohjigamori M, Haneishi T (1991) Hydantocidin: a new compound with herbicidal activity from Streptomyces hygroscopicus. J Antibiot 44: 293-300

Poland BW, Fromm HJ, Honzatko RB (1996a) Crystal structures of the complex of IMP, GDP, nitrate and hadacidin with adenylosuccinate synthetase from Escherichia coli. J Mo1 Biol 264: 1013- 1027

Poland BW, Hou Z, Bruns C, Fromm HJ, Honzatko RB (199613) Refined crystal structures of guanine nucleotide complexes of adenylosuccinate synthetase from Escherichia coli. J Biol Chem 271: 15407-15413

Poland BW, Lee S-F, Subramanian MV, Siehl DL, Anderson RJ, Fromm HJ, Honzatko RB (1996~) Refined crystal structure of adenylosuccinate synthetase from Escherichia coli complexed with hydantocidin 5’-phosphate, GDP, HPO,*-, Mg2+ and hadacidin. Biochemistry 35: 15753-15759

Poland BW, Silva MM, Serra MA, Cho Y, Kim KH, Harris EM, Honzatko RB (1993) Crystal structure of adenylosuccinate synthetase from Escherichia coli: evidence for convergent evolution of GTP-binding domains. J Biol Chem 268: 25334-25342

Potter SL, Ward ER, inventors. 1996 May 21. Plant adenylosuccinate synthetase and DNA coding therefor. US Patent 5, 519, 125.

Rudolph FB, Clark SW (1982) High-performance liquid chromatography of proteins: purification of the acidic isozyme of ad-

enylosuccinate synthetase from rat liver. Ana1 Biochem 127

Rudolph FB, Fromm HJ (1969) Initial rate studies of adenylosuccinate synthetase with product and competitive inhibitors. J Biol Chem 244 3832-3839

Schloss JV (1988) Significance of slow-binding enzyme inhibition and its relationship to reaction-intermediate analogues. Acc Chem Res 21: 348-353

Siehl DL, Subramanian MV, Walters EW, Lee S-F, Anderson RJ, Toschi AG (1996) Adenylosuccinate synthetase: site of action of hydantocidin, a microbial phytotoxin. Plant Physiol 110: 753-758

Silva MM, Poland BW, Hoffman CR, Fromm HJ, Honzatko RB (1995) Refined crystal structures of unligated adenylosuccinate synthetase from Escherichia coli. J Mo1 Biol 254: 431446

Stayton MM, Rudolph FB, Fromm HJ (1983) Regulation, genetics and properties of adenylosuccinate synthetase: a review. Curr Top Cell Regul 22: 103-141

Wang W, Gorrell A, Honzatko RB, Fromm HJ (1997) A study of Escherichia coli adenylosuccinate synthetase association states and the interface residues of the homodimer. J Biol Chem 272:

Zeidler R, Hobert O, Johannes L, Faulhammer H, Krauss G (1993) Characterization of two nove1 single-stranded DNA-specific autonomously replicating sequence-binding proteins from Saccha- romyces cerevisiae, one of which is adenylosuccinate synthetase. J Biol Chem 268: 20191-20197

193-197

7078-7084

Documents

Adenylosuccinate Synthetase from Maize