Mechanism of 3-methylaspartase probed using deuterium and solvent isotope effects and active-site directed reagents: Identification of an essential cysteine residue

Bioorganic & Medicinal Chemistry 7 (1999) 949±975

Mechanism of 3-Methylaspartase Probed Using Deuterium and

Solvent Isotope Eects and Active-site Directed Reagents:

Identi®cation of An Essential Cysteine Residue^y

John R. Pollard, Susan Richardson, Mahmoud Akhtar, Philippe Lasry, Tracy Neal,

Nigel P. Botting and David Gani*

School of Chemistry and Centre for Biomolecular Sciences, The Purdie Building, The University of St Andrews,

Scotland, KY16 9ST, UK

Received 18 September 1998; accepted 2 February 1999

AbstractÐThe mechanism of the l-threo-3-methylaspartate ammonia-lyase (EC 4.3.1.2) reaction has been probed using deuterium

and solvent isotope eects with three dierent substrates, (2S,3S)-3-methylaspartic acid, (2S)-aspartic acid and (2S,3R)-3-methyl-

aspartic acid. Each substrate appears to form a covalent adduct with the enzyme through the amination of a dehydroalanine

(DehydAla-173) residue. The true substrates are N-protonated and at low pH, the alkylammonium groups are deprotonated

internally in a closed solvent-excluded pocket after K⁺ion, an essential cofactor, has become bound to the enzyme. At high pH, the

amino groups of the substrates are able to react with the dehydroalanine residue prior to K⁺ion binding. This property of the

system gives rise to complex kinetics at pH 9.0 or greater and causes the formation of dead-end complexes which lack Mg²⁺ion,

another essential cofactor. The enzyme±substrate adduct is subsequently deaminated in two elimination processes. Hydrazines act

as alternative substrates in the reverse reaction direction in the presence of fumaric acid derivatives, but cause irreversible inhibition

in their absence. Borohydride and cyanide are not inhibitors. N-Ethylmaleimide also irreversibly inactivates the enzyme and labels

residue Cys-361. The inactivation process is enhanced in the presence of cofactor Mg²⁺ions and Cys-361 appears to serve as a base

for the removal of the C-3 proton from the natural substrate, (2S,3S)-3-methylaspartic acid. The dehydroalanine residue appears to

be protected in the resting form of the enzyme by generation of an internal thioether cross-link. The binding of the substrate and

K⁺ion appear to cause a conformational change which requires hydroxide ion. This is linked to reversal of the thioether protection

step and generation of the base for substrate deprotonation at C-3. The deamination reaction displays high reverse reaction com-

mitments and independent evidence from primary deuterium isotope eect data indicates that a thiolate acts as the base for

Introduction

other species.^3,4The clostridial enzyme is a 91,000 Da

homodimer of 413-residue subunits and requires mono-

valent and divalent cation cofactors for activity.⁵

Potassium and magnesium ions serve as the best metal

ion cofactors. The enzyme has been shown to deaminate

the l-erythro-(2S,3R)-diastereomer of methylaspartic

acid^1,6as well as (2S)-aspartic acid and a number of 3-

alkyl homologues.⁷Moreover, the enzyme was shown to

be able to catalyse the exchange of the C-3 hydrogen

atom of the physiological substrate with hydrogen

derived from the solvent, at rates greater than for the

overall deamination reaction, under some conditions.⁸

No primary deuterium isotope eect was detected for

the deamination reaction over the pH range from 5.5 to

10.5 and, on the basis of these observations, a mechan-

ism involving the intermediacy of a carbanion was pro-

posed,^8,9(see Hansen and Havir for a review of the early

work¹⁰). Since Bright's work, the methylaspartase sys-

tem has been regarded as the archetypal example of an

enzyme which operates via a carbanion elimination

Methylaspartase (3-methylaspartate ammonia-lyase, EC

4.3.1.2) catalyses the reversible a,b-elimination of

ammonia from l-threo-(2S,3S)-3-methylaspartic acid

(1) in an anti-fashion to give mesaconic acid (Scheme 1).¹

The enzyme lies on the main catabolic pathway for glu-

tamate in Clostridium tetanomorphum²and a number of

Key words: Enzymes and enzymic reactions; enzyme inhibitors; kinet-

ics; isotope eects.

Abbreviations: Bis Tris propane, 1,3-bis [tris(hydroxymethyl)-

methylamino]propane; HPLC, high performance liquid chromato-

graphy; NEM, N-ethylmaleimide; NMR, nuclear magnetic resonance;

UV, ultra-violet; TLC, thin layer chromatography; Tris, (Tris[hydroxy-

methyl]aminomethane); GC±MS, gas chromatography±mass spectro-

metry; SIR-GC±MS, single ion recording-gas chromatography±mass

spectrometry.

*Corresponding author. Present address: School of Chemistry, Uni-

versity of Birmingham, Edgbaston, Birmingham, B15 2TT, UK.

^yDedicated to the memory of Sir Derek H. R. Barton and his intellec-

tual impact in the chemical sciences.

PII: S0968-0896(99)00044-9

950

J. Pollard et al. / Bioorg. Med. Chem. 7 (1999) 949±975

More recently evidence was provided to suggest that

methylaspartase might contain an electrophilic dehy-

droalanine prosthetic group derived from Ser residue

173.⁵In accord with this suggestion, some other

ammonia-lyase enzymes^22,23have since been shown to

utilise a modi®ed Ser residue. In methylaspartase the

Ser-173 residue occurs in the sequence Ser-Gly-Asp

which also occurs in histidine and phenylalanine

ammonia-lyase.^24,25For methylaspartase it has been

proposed that the substrate could attack the dehy-

droalanine moiety in a Michael fashion (Scheme 2) to

give a covalent enzyme±substrate complex which is

subsequently deaminated twice via two a,b-elimination

reactions.⁵Model studies with dehydroalanine peptides

have shown that nitrogen nucleophiles can attack the

dehydroalanine residue at C-3, to give a Michael

adduct, or at C-2, which involves trapping an imine (or

iminium) intermediate.²⁶The regioselectivity is depen-

dent upon the electron withdrawing or releasing prop-

erties of the groups attached to both the amino and the

carboxyl group of the dehydroalanine residue. Since

enzymes have the potential to delicately `tune' the

electronic properties of each of the carboxamide moi-

eties in the N-acyldehydroalanyl peptide at their active

site, there are two chemically feasible pathways to

covalent intermediates for the ammonia-lyase reac-

tions (Scheme 3). Regardless of the exact regiochem-

istry, any such covalent adduct formation between the

substrate and the enzyme would involve extra ¹⁵N-sen-

sitive transition states which would complicate the

interpretation of ¹⁵(V/K) isotope eects²¹(see also fol-

lowing article²⁷).

Scheme 1.

mechanism,¹¹and indeed, more recent studies of the

related systems, l-aspartase,¹²phenylalanine ammonia-

lyase¹³and argininosuccinate lyase¹⁴have revealed the

operation of similar elimination mechanisms.

Research in our own laboratories led us to investigate

the kinetics of the methylaspartase system using a range

of substrate analogues^15±17and, furthermore, prompted

an examination of the primary deuterium isotope eects

for the deamination of these substrate analogues.¹⁸In

our hands, (2S,3S)-3-methylaspartic acid, the natural

substrate, showed a signi®cant isotope eect of ꢀ1.7 on

V and on V/K over a wide pH range, in contrast to the

reported ®ndings.⁸Further work established that K⁺

ion binds to the enzyme±substrate±Mg²⁺complex last

in an ordered binding sequence and debinds before the

products or Mg²⁺ion are released, indicating that both

^DV and ^D(V/K) might be suppressed at high K⁺con-

centration.¹⁹It was also noted that double reciprocal

plots of initial rate versus K⁺ion concentration were

not linear but concave down. The repetition of Bright's

experiments at high K⁺concentration con®rmed that

no C-3 primary deuterium isotope eect was apparent

and provided an explanation for the dichotomy.¹⁹In

view of these ®ndings, other evidence in support of the

carbanion mechanism was challenged.

Here, and in the following article, we develop a uni®ed

model for the mechanism of methylaspartase and

describe the results of solvent deuterium isotope eect

and ¹⁵N-fractionation experiments designed to detect

chemical steps that might be involved in forming a

covalent enzyme±substrate complex. It is shown that

the methylaspartase reaction is extremely complicated

because the enzyme releases its products very late and

that large reverse reaction commitments dominate the

kinetics of the system. It is nevertheless demonstrated

that a covalent intermediate complex exists and that

the chemical mechanism for the deamination of the

physiological substrate, (2S,3S)-3-methylaspartic acid,

is concerted. It is shown that the thiolate anion of a

Cys residue serves as the base which removes the C-3

hydrogen. We also present the results of chemical mod-

i®cation studies designed to elucidate details of the

active-site structure, identify the Cys residue which

serves as a base and discover whether the dehy-

droalanine moiety is protected or unprotected from

exogenous nucleophiles in the resting state of the

enzyme.

The mechanism of the C-3 hydrogen exchange reaction

was examined by measuring the primary deuterium iso-

tope eect for tritium wash-in into the substrate from

the solvent.¹⁹The reaction displayed a signi®cant pH-

independent isotope eect on V_maxof 1.5±1.6 which

mirrored the isotope eects determined for the deamin-

ation reaction in the same experiments. However, the

ratios of the rates for the exchange versus deamination

reaction (n_ex/n_deam) varied widely (>®vefold) over the

pH-range of the study, but were not aected by the

presence of deuterium. These results are not in accord

with a simple C-3 carbanion hydrogen exchange

mechanism. The mechanism for the syn-elimination of

ammonia from the l-erythro-(2S,3R)-diastereomer of

methylaspartic acid^6,20was probed using C-3 deuteriated

and unlabelled substrate by measuring the rate of

mesaconic acid formation and C-3 hydrogen exchange

and epimerisation by NMR spectroscopy. The results

indicated that epimerisation and hydrogen exchange

could not occur before the l-erythro-substrate had ®rst

undergone elimination to give mesaconic acid. Again

these ®ndings do not support the operation of a carb-

anion mechanism, but do not exclude the possibility

that the proton removed from the C-3 position is shiel-

ded from the solvent until the products are formed.

Preliminary ¹⁵(V/K) data at low pH also indicated that

the reaction did not follow a simple carbanion mechan-

ism²¹(see Discussion and following article).

Results

Deuterium and solvent isotope eects for deamination

(2S,3S)-3-Methylaspartic acid. The kinetic parameters K_m

(K_Measp), V_maxand V/K for unlabelled and 3-deuteriated

J. Pollard et al. / Bioorg. Med. Chem. 7 (1999) 949±975

951

Scheme 2.

Scheme 3.

(2S,3S)-3-methylaspartic acid for the deamination reac-

tion performed in water and deuterium oxide at pL 6.5,

7.6, 9.0 and 9.4 are shown in Table 1. The derived iso-

tope eects are also given for each study pL value. The

parameter x represents V or V/K and leading superscripts

refer to the varied label; D represents substrate C-3

deuteriation and S represents deuteriated solvent. Trail-

ing subscripts refer to the presence of other non-varied

labelled sites. The primary deuterium isotope eects for

V and V/K, ^DV and ^D(V/K), respectively, are not pH

sensitive and are similar in value at ꢀ1.7 (Table 1). This

indicates that C-3 hydrogen abstraction is kinetically

important across the pL range of the study. The solvent

deuterium isotope eects ^SV and ^S(V/K), respectively, are

not similar in value to each other and increase in magni-

tude with increasing pL. The maximum value of ^S(V/K) is

3.5 (Fig. 1) and this occurs at pL 9.0 which corresponds

to the optimum pH for V/K (Fig. 2). The maximum

value of ^SV is 3.4 and this occurs at pL 9.4 which

corresponds to the optimum pH for V (Figs 1 and 2).

952

J. Pollard et al. / Bioorg. Med. Chem. 7 (1999) 949±975

Table 1. Primary deuterium and solvent isotope eects for the deamination of (2S,3S)-3-methylaspartic acid

pL 6.5

K^a

pL 7.6

K

pL 9.0

K

pL 9.4

K

pL 10.0

K

Solvent, substrate

V^a

V/K

V

V/K

V

V/K

V

V/K

V

V/K

1. H₂O, H-MeAsp

2. H₂O, D-MeAsp

3. D₂O, H-MeAsp

4. D₂O, D-MeAsp

7.51

4.42

4.95

nd

2.3

1.5

3.27

1.92

3.34

70.1

41.1

56.2

nd

2.4

1.9

29.2

17.1

29.6

654

385

210

194

2.4

2.6

2.2

277

164

80.8

88.2

850

434

247

191

6.5

2.6

2.3

130

68

96

83

850

491

nd

9.0

94.4

54.6

nd

Isotope eects

Entry 1/2 ^D(x)

Entry 3/4 ^D(x)_S

Entry 1/3 ^S(x)

Entry 2/4 ^S(x)_D

Entry 1/4 ^SD(x)

V

V/K

1.70 1.70

V V/K

1.71 1.71

V

V/K

V

V/K

V V/K

1.73 1.73

1.70 1.69

1.08 0.91

3.11 3.46

1.95 1.85

3.38 3.10

1.95 1.95

1.29 1.15

3.44 1.35

2.27 0.82

4.45 1.56

1.52 0.98

1.25 0.99

^aErrors in V and K are 10% or less.

Figure 2. Non-®lled dependence of the kinetic parameters for (2S,3S)-

3-methylasparic acid and potassium ion on pH as a logarithm plot. V

(&), V/K (*), K (Á), V_K(&), (V/K)_K(*), K_K(~). Unit slope (- - - -)

is also shown.

Figure 1. Dependence of the kinetic parameters V and V/K on pL for

the deamination of (2S,3S)-3-methylaspartic acid in water and in

deuterium oxide. V-H₂O (*), (V/K)-H₂O (&), V-D₂O (Á), (V/K)-D₂O

(&). Plots are drawn free-hand.

S

At pL 6.5 and 7.6, V and (V/K) are small, 1.5 or less

and 1.0, respectively, indicating that the solvent deuter-

ium transfer step(s) that are rate limiting at pL 9.0±9.4

are apparently not so important compared to C-3±H

bond cleavage at low pL.^{

9.0 and 9.4. The solvent isotope eects for the deu-

S

teriated substrate V_Dand (V/K)_Dare also suppressed

S

compared to those for the unlabelled material, V and

^S(V/K), again indicating that the primary deuterium and

solvent isotope eects operate on dierent rate limiting

steps in the deamination process.

At pL 9.0 and 9.4, the combined primary deuterium

and solvent isotope eects upon V and V/K, ^SDV and

^SD(V/K) respectively, are very similar to the values for

(2S)-Aspartic acid

S

^SV and (V/K). Examined in a dierent way this analy-

The kinetic parameters K_m(K_Asp), V_maxand V/K for

unlabelled and 3-deuteriated (2S,3R)-[3-²H₁]-aspartic

acid for the deamination reaction performed in water

and deuterium oxide at pL 9.0 and 9.4 are shown in

Table 2. The derived isotope eects are also given and

these are large compared to the eects observed for the

natural substrate. The pH dependencies of the kinetic

parameters for (2S)-aspartic acid over the range from

7.6 to 10.0 are shown in Figure 3.

sis indicates that the primary deuterium isotope eects

in deuteriated solvent, ^DV_Sand ^D(V/K)_S, respectively,

are suppressed compared to those obtained in water.

Indeed, ^DV_Sand ^D(V/K)_Sare very close to unity at pL

^{Note that the pL independent solvent isotope eects are 2.65 for ^SV

and 2.90 for ^S(V/K) and that the pL pro®le in deuteriated solvent is

shifted to higher pL by ca. 0.4 units.

J. Pollard et al. / Bioorg. Med. Chem. 7 (1999) 949±975

953

^D(V/K)=3.39; and, at 50 mM K⁺ion ^DV=6.79 and

^D(V/K)=4.10 indicating that there are very large pri-

mary deuterium isotope eects for the syn-elimination.⁶

Table 2. Primary deuterium and solvent isotope eects for the

deamination of (2S)-aspartic acid

pL 9.0

K^a

pL 9.4

K

Solvent, substrate

V^a

V/K

V

V/K

Solvent isotope eects for amination

1. H₂O, H-Asp

2. H₂O, D-Asp

3. D₂O, H-Asp

4. D₂O, D-Asp

4.80 10.5

2.81 37.9

1.41 32.9

0.46

0.074

0.043

5.07 15.6

4.73 13.0

2.40 23.0

1.18 27.8

0.33

0.36

0.104

0.043

The kinetic parameters K_m(K_Mesa), V_maxand V/K for

the amination of mesaconic acid performed in water

and deuterium oxide at pL 6.5, 9.0 and 9.4 are shown in

Table 4. The derived isotope eects are also given for

each study pL value. Since these re¯ect both the solvent

isotope eects and the C-3 deuterium isotope eect for

Isotope eects

Entry 1/2 ^D(x)

Entry 3/4 ^D(x)_S

Entry 1/3 ^S(x)

Entry 2/4 ^S(x)_D

Entry 1/4 ^SD(x)

V

V/K

1.0

1.7

6.18

10.6

V

V/K

0.92

2.4

3.17

7.7

8.4

1.0

2.0

1.71

3.4

1.07

2.03

2.11

4.3

forming the deuteriated natural substrate, these are

SD

Rev

and ^SD(V/K)_Revand can be compared

4.0

labelled

V

with ^SDV and ^SD(V/K) in Table 1. As for the corre-

sponding eects in the deamination reaction direction,

^aErrors in V and K are 14% or less.

SD

at high pL both

SD

V

and ^SD(V/K)_Revare signi®cant.

Rev

The value of

V

Rev

is 6.7 and maximal at pL 9.4 and

the value of ^SD(V/K)_Revis 1.9 and maximal at pL 9.0.

At lower pL the magnitude of both isotope eects

diminishes to give values similar to those corresponding

to the primary deuterium isotope eects for the forward

reaction, ^DV and ^D(V/K) in Table 1.

The kinetic parameters K_m(K_Fum), V_maxand V/K for

the amination of fumaric acid performed in water and

deuterium oxide at pL 9.0 and 9.4 are shown in Table 5.

Covalent modi®cation of methylaspartase

Inhibition by hydrazines. The relative rates of the inacti-

vation of methylaspartase at pH 8.0 in Tris buer by

various hydrazines are shown in Figure 4 under the

speci®ed conditions. Phenylhydrazine, methylhydrazine

and 4-nitrophenylhydrazine are the most eective irre-

versible inhibitors. Complete protection from inactiva-

tion is aorded by 2 mM (2S,3S)-3-methylaspartic acid.

Sodium borohydride, sodium cyanide, phenylphosphine

and nitromethane are ineective as irreversible inhibi-

tors. The presence of magnesium ions (0.4 mM)

increased the rate of inactivation by 4-nitro- and 2,4-

dinitro-phenylhydrazine, but had little eect on the rate

of inactivation by phenylhydrazine. Potassium ions

(250 mM) and ammonium ions (10 and 20 mM) had no

eect on the rate of the inhibition of the enzyme by

phenylhydrazine at pH 9.0 in 50 mM Tris buer (data

not shown), but the rate of inactivation was faster than

at pH 8.0. At pH 9.0 the inactivation reaction showed

saturation kinetics although the inactivation itself was

biphasic (Fig. 5). The value of K_Scalculated from the

initial faster phase was approximately 2.13 mM.

Figure 3. Non-®lled dependence of the kinetic parameters for (2S)-

aspartic acid on pH as a logarithm plot. V (&), V/K (*), K (Á).

C-3 Hydrogen exchange

Experiments were conducted to simultaneously, in the

same experiment, measure the rate of deamination and

C-3 solvent tritium exchange into (2S)-aspartic acid and

(2S,3R)-[3-²H₁]-aspartic acid at pH 9.0. The rates of

deamination and exchange were similar for each substrate

and the ratio n_Ex/n_Deamwas 16.5 for the unlabelled

material and 15.1 for (2S,3R)-[3-²H₁]-aspartic acid. The

®ndings indicate that the exchange reaction is much

faster than for the natural substrate (see discussion)¹⁹

and that, for aspartic acid, neither the deamination or

the exchange reaction display a primary deuterium iso-

tope eect.

The pH-dependence of the rate for the initial phase of

the inactivation showed increases in log rate versus pH

of greater than unit slope (Fig. 6) indicating that more

than one species must be deprotonated for inactivation

to occur.

(2S,3R)-3-Methylaspartic acid

Inhibition by N-ethylmaleimide. Treatment of the

enzyme with NEM irreversibly inactivated the enzyme

and gave simple ®rst-order kinetics (Fig. 7). The rate of

reaction showed saturation kinetics at pH 9.0. K_Sfor

The kinetic parameters and the primary deuterium iso-

tope eects at pH 9.0 at 1 mM and at 50 mM K⁺ion

are given in Table 3. At 1 mM K⁺ion ^DV=7.15 and

954

J. Pollard et al. / Bioorg. Med. Chem. 7 (1999) 949±975

Table 3. Primary deuterium eects for the deamination of (2S,3R)-3-methylaspartic acid at pH 9.0

pH 9.0, 1.0 mM K⁺

pH 9.0, 50.0 mM K⁺

K

Solvent, substrate

V

K

V/K

V

V/K

1. H₂O, H-MeAsp

2. H₂O, D-MeAsp

17.20.3

2.40.78

400.8

18.88.4

0.43

0.13

41.14.0

6.10.23

5.21.1

3.140.29

7.90

1.93

Isotope eects

Entry 1/2 ^D(x)

V

7.152.74

V/K

3.391.6

V

6.790.92

V/K

4.101.3

Table 4. Solvent isotope eects for the amination of mesaconic acid

in the presence of 0.4 M ammonium chloride

pL 6.5

pL 9.0

K

pL 9.4

K

Solvent

V^aK^aV/K

V

V/K

V

V/K

1. H₂O

2. D₂O

8.18 1.18 6.93 894 1.24 721 492 1.87 263

4.25 0.78 5.48 326 1.10 297 73.7 0.36 205

Isotope eect

Entry 1/2

^SD(x)_Rev

V V/K

1.93 1.27

V V/K

2.12 1.88

V V/K

6.70 1.30

^aErrors in V and K are 10% or less.

Table 5. Solvent isotope eects for the amination of fumaric acid in

the presence of 0.4 M ammonium chloride

pL 9.0

K^a

pL 9.4

K

Solvent

V^a

V/K

V

V/K

1. H₂O

2. D₂O

1702

205

23.2

5.95

73.4

34.5

1565

479

2.96

2.78

524

172

Isotope eect

Entry 1/2

^SD(x)_Rev

V V/K

8.32 2.13

V V/K

3.27 3.04

^aErrors in V and K are 10% or less.

NEM was determined to be 34.55.4 mM at pH 7.0 in

the presence of 5 mM sodium phosphate where k_Inact

was 5.00.6 min ¹. Phosphate served as a linear com-

petitive inhibitor for the inactivation reaction and dis-

played a K_Ivalue of 1.5 mM in the presence of 0.5 mM

NEM at pH 7.0. Acetate had no protective eect. Mg²⁺

ions served as an activator for the irreversible inhibition

of the enzyme and at saturating concentrations

increased the rate by ꢀ10-fold. The activation by Mg²⁺

ions showed simple saturation kinetics (after correction

for the non-catalysed inactivation by NEM), and K_Mg

was 11722 mM. Mg²⁺ions enhanced the anity of

the enzyme for NEM by 10-fold and at saturating con-

centrations of Mg²⁺ions K_Sfor NEM was decreased to

2.760.37 mM and k_Inactwas 3.310.68 min ¹. Thus,

the increase in activity for NEM as an inhibitor in the

presence of Mg²⁺ions results largely from its increased

anity for the enzyme and not from an increase in the

rate of alkylation.

Figure 4. Time course for the inhibition of methylaspartase by phe-

nylhydrazine (^), 4-nitrophenylhydrazine (*), 2,4-dinitrophenylhy-

drazine (*), 3,6-dioxaoctane dioic acid dihydrazide (Á),

acetylhydrazine (&) and a control with no inhibitor (&).

Potassium ions protected the NEM alkylation reaction

and K_Sfor K⁺at pH 9.0 was 4.891.26 mM. Ammo-

nium ion was less eective at protecting the enzyme and

K_Swas 10.173.68 mM.

(2S,3S)-3-Methylaspartic acid and mesaconic acid pro-

tected the enzyme against inactivation by NEM. At pH

7.0 the values of K_Ifor the inhibition of the inactivation

reaction by the two substrates were 0.60.1 mM and

0.50.1 mM, respectively. The K_Ivalue of mesaconic

acid was not altered by the presence of Mg²⁺ions.

The pH-dependence of the inactivation rate is shown in

Figure 8. The enzyme itself tritrates at pH 9.3 to a form

that reacts rapidly with NEM, since NEM cannot

deprotonate under the conditions of the experiment.

J. Pollard et al. / Bioorg. Med. Chem. 7 (1999) 949±975

955

Figure 5. Time course for the inhibition of methylaspartase by phe-

nylhydrazine: 2.5 mM (*), 10 mM (Á), 15 mM (&) and 20 mM (*).

Figure 7. Time course for the inhibition of methylaspartase by N-

ethylmaleimide: 0.05 mM (*), 0.25 mM (&), 0.5 mM (Á), 1.0 mM (*).

Figure 6. Dependence of the inhibition of methylaspartase by phe-

nylhydrazine (*) on pH as a logarithm plot and unit slope (- - - -).

Figure 8. Dependence of the inhibition of methylaspartase by N-

ethylmaleimide (*) on pH as a logarithm plot and unit slope (- - - -).

Proteolysis with trypsin

the endoproteinase enzyme trypsin (which is known to

cleave the peptide chain at the C-terminus of lysine and

arginine residues) at pH 7.8 for 4 h at room temperature

(see Experimental for full details). The resulting peptide

mixture was subsequently lyophilised to aord a white

powder in preparation for analysis by HPLC. In the

case of 2,4-dinitrophenylhydrazine modi®cation, the

native enzyme was incubated with the inhibitor for 45 h,

after which time all of the activity had been lost. The

inactive protein was subjected to repeated dialysis and

was then treated with trypsin using the procedure out-

lined above.

The sites of modi®cation by both N-ethylmaleimide and

2,4-dinitrophenylhydrazine were investigated by a series

of peptide mapping experiments. For N-ethylmaleimide

the enzyme was protected with substrate and was then

treated with unlabelled inhibitor under the speci®ed

conditions. Following repeated dialysis steps, to remove

substrate and excess unreacted N-ethylmaleimide, the

protein (which had retained >95% of the original

activity) was incubated with N-[2-³H]-ethylmaleimide.

All activity was lost after 60 min. The inactive protein

was then subjected to repeated dialysis and treated with

956

J. Pollard et al. / Bioorg. Med. Chem. 7 (1999) 949±975

HPLC Analysis of tryptic peptides

pro®le comprising two peaks absorbing at 360 nm. Both

were isolated and the ®rst ®ve residues of the N-terminal

sequence were determined. The peak which eluted late

in the gradient did not exhibit any amino acid sequence,

and was evidently not a peptide. The single peak with a

lower retention time was subjected to 18 cycles of

N-terminal analysis by Edman degradation. Three pep-

tide sequences were recorded consistent with the bis-

disul®de tripeptide composed of SAEVTTNIGMAC-

GAR, ANGMGAYCGGTCNETNR, and GVDAEL-

VADEWCNTVEDVK. Presumably the ANG peptide

formed one disul®de bridge with each of its two Cys

residues to the single Cys residue in each of the other

peptides. Note no such oxidised peptides were isolated

in digests of enzyme that had not been treated with

hydrazines.

Unmodi®ed methylaspartase. In order to compare

directly the HPLC elution pro®les of modi®ed methyl-

aspartase with the native recombinant protein (see

Experimental), unmodi®ed enzyme was digested with

trypsin and the peptide mixture was analysed by

reversed-phase HPLC using a 10 R2 POROS column

(Fig. 9). A number of peptides were isolated and were

subjected to N-terminal amino acid analysis by Edman

degradation. The sequences for all of the isolated pep-

tides were matched with the deduced amino acid

sequence⁵and will be reported on in full elsewhere. The

retention time of peptides corresponding to the phenyl-

hydrazine and NEM modi®ed peptide sequences descri-

bed below were noted.

Methylaspartase labelled with 2,4-dinitrophenylhydrazine.

The peptide mixture resulting from tryptic digestion of

2,4-dinitrophenylhydrazine modi®ed protein was ana-

lysed by reversed phase HPLC, using a 10 R2 POROS

column (100Â4 mM). The eluent was monitored directly

by UV spectroscopy at 220 and 360 nm. The elution

pro®le (Fig. 10) showed the presence of ca. 25 resolved

peaks possessing a UV absorbance at 220 nm and one

major peak possessing an absorbance at 360 nm with a

retention volume of 12.4 column volumes. The material

giving rise to the peak was collected and was further

puri®ed by a repeat HPLC step. In this case the peak

absorbing at 360 nm coincided exactly with a peak

absorbing at 220 nm of retention volume 5.8 column

volumes. A ®nal HPLC puri®cation step using a micro-

bore system (100Â1 mM column) resulted in an elution

Methylaspartase labelled with N-[2-³H]-ethylmaleimide.

The HPLC trace for the peptide mixture resulting from

tryptic digestion of N-ethylmaleimide modi®ed protein

(Fig. 11) showed the presence of ca. 50 UV-absorbing

peaks, consistent with the 52 anticipated peptides.⁵In

addition, one major peak displaying radioactivity was

observed with a retention volume of 7.30.3 column

volumes. Fractions associated with this radioactivity

were pooled, concentrated under reduced pressure and

then subjected to a second HPLC puri®cation step (Fig.

11, inset). In this case just four UV-absorbing peaks and

a single peak possessing radioactivity were observed.

The radioactive peak coincided with a UV-absorbing

peak possessing a retention volume 7.30.3 column

volumes. The material giving rise to the peak was col-

lected and was subjected to Edman N-terminal amino

Figure 9. HPLC chromatograph (10 R2 POROS column) for the tryptic digest of native methylaspartase.

J. Pollard et al. / Bioorg. Med. Chem. 7 (1999) 949±975

957

Figure 10. HPLC chromatograph (10 R2 POROS column) for the tryptic digest of 2,4-dinitrophenylhydrazine inhibited methylaspartase. Detection

at 220 nM (ÐÐ) and 360 nM (- - - -).

Figure 11. HPLC chromatograph (C¹⁸column) for the tryptic digest of N-ethylmaleimide inhibited methylaspartase. Detection at 220 nm (ÐÐ), ³H

content of fraction (- - - -). Inset chromatograph shows the further HPLC puri®cation of the major peak in radioactivity.

958

J. Pollard et al. / Bioorg. Med. Chem. 7 (1999) 949±975

acid sequencing. Aliquots of each sequencing cycle were

3

ion is released ®rst. This situation is very dierent to

that which has been ascribed for phenylalanine ammo-

nia-lyase where the product, cinnamate, can dissociate

from the enzyme before ammonia is released from the

covalent amino±enzyme complex.¹³In the analysis

below we shall assume that mesaconic acid cannot

escape before ammonium ion. In the light of the sug-

gestion that the enzyme operates via the intermediacy of

a dehydroalanine residue⁵(see below), several potential

sites for the addition of a second ammonia molecule to

an enzyme±intermediate complex become apparent. For

example, ammonia might react directly with the enzyme

to give a covalent complex which is still able to bind to

the substrate. Some evidence to support this idea is

oered by the ®nding that the transition state inhibitor,

(1S,2S)-1-methylcyclopropane-1,2-dicarboxylic acid (4)

was a good inhibitor (K_i=20 mM) for the enzyme, but

was even more potent in the presence of ammonium

ion.³²Thus, there is clearly enough room for extra

heavy atoms (Scheme 4). Alternatively, ammonium ion

may be able to trap a covalent enzyme±substrate com-

plex that does not contain a bound Mg²⁺ion. Such

complexes would cause dead-end inhibition and the

importance of the covalent interactions are discussed

below.

retained in order to determine the H-content of indivi-

dual residues. A single peptide sequence ANGMG

AYXGGTZNETNR was observed. Cycle 8 (X) showed

no HPLC elution pro®le, consistent with a Cys residue

and cycle 12 (Z) gave a radioactive eluent and showed

two bands in the HPLC trace of the Edman derivative

running between Pro and Val, indicative of a modi®ed

amino acid. Note that the published sequence of the

enzyme contains Cys residues at positions 8 and 12 in

the only peptide commencing with ANG.⁵

Discussion

Before commencing to attempt to interpret the isotope

eect data it is necessary to construct a model for the

catalytic process. Early investigations of the mechanism

of the l-threo-methylaspartase reaction by Bright and

co-workers indicated that the enzyme operated via a

carbanionic elimination mechanism^8,9,28±31(see Intro-

duction). However, work in our own laboratory revealed

that there is a signi®cant primary deuterium isotope

eect for the deamination reaction with (2S,3S)-3-

methylaspartic acid as the substrate,¹⁸contrary to the

earlier reports. It was also demonstrated that at the

high concentrations of K⁺ion used in Bright's original

experiments the observed isotope eects were sup-

pressed by increasing the forward and reverse reaction

commitments.¹⁹This work also showed that the

exchange of the C-3 hydrogen of the substrate with

solvent derived hydrogen did not occur at the level of

the putative carbanion but after a slow step or steps

involving both C-3 hydrogen and C-2 nitrogen bond

cleavage.

Active site dehydroalanine moiety

At the time of the original work on the metal ion and

product binding and debinding orders,¹⁹we had not

completed a primary structural analysis of the enzyme.

When the deduced amino acid sequence of methylas-

partase became available,⁵it was possible to compare

the sequence with the known amino acid composition

of isolated active site tryptic peptides that had been

labelled with [¹⁴C]-N-ethylmaleimide.³³From this com-

parison it appeared that a serine residue (or its post-

translationally modi®ed product) rather than a cystein

residue, as had been originally proposed, had reacted

with the electrophilic species. In Wu and Williams' ori-

ginal experiments, the enzyme was treated with unla-

belled N-ethylmaleimide in the presence of substrate

and dialysis of the protein solution gave an active chem-

ically modi®ed enzyme.³³When this modi®ed enzyme

was treated with [¹⁴C]-N-ethylmaleimide the resulting

protein was completely inactive. Treatment of this

labelled protein with trypsin gave a single active site

labelled peptide (a ca. 40-mer) which corresponded to

one of the eight uniformly labelled peptides that had

been obtained earlier by the same workers by treating

the enzyme in the absence of substrate with [¹⁴C]-N-

ethylmaleimide.³³Note that there are only seven Cys

residues in native methylaspartase,⁵and presumably

each of these had been covalently modi®ed, to give the

S-succinimide derivatives in the active modi®ed enzyme

referred to above. We suggested that the eighth pep-

tide obtained was a 37-mer generated by tryptic

cleavage at Arg-146 and Lys-183 (see Table 6),

because it was the only single peptide sequence that

displayed a close match to the published amino acid

composition. However, we will show below that the

Wu and Williams' peptide³³may have possessed a more

complex structure.

Product debinding order

Using unlabelled substrate and the labelled products,

[¹⁵N]-ammonia and [methyl-²H₃]-mesaconic acid, it was

shown that the products were released from the enzyme

in an apparently random fashion and that neither of the

species could be trapped on the enzyme and converted

back, through reverse steps, to labelled substrates.¹⁹The

use of methylamine as an inhibitor also suggested that

product release occurred in a random fashion. However,

ammonia itself (in the presence of high concentrations

of potassium ion) acted as a noncompetitive product

inhibitor at high concentration, indicating that it is

released before mesaconic acid, and as a nonlinear

uncompetitive inhibitor at lower concentration (Fig. 12),

indicating that more than one molecule can bind to the

same enzyme species. If ammonia or ammonium ion can

bind to an intermediate enzyme complex in a site other

than the product or activator site, as appears to be the

case (see Fig. 12 and the discussion below), then exo-

genous labelled ammonium ion would not, necessarily,

be able to trap mesaconic acid on the enzyme. Since the

eects of ammonium ion are further complicated by the

fact that it can serve as a cationic activator (as a surro-

gate for potassium ion), the debinding order is dicult

to re®ne further. Nevertheless, it seems certain that

either product release is random, or, that ammonium

J. Pollard et al. / Bioorg. Med. Chem. 7 (1999) 949±975

959

Figure 12. Ammonia inhibition of the deamination reaction at 50 mM potassium ion concentration. Incubations contained; 0.5 M Tris (pH 9.0),

20 mM MgCl₂, (2S,3S)-3-methylaspartic acid and ammonium chloride at (a) zero; (b) 10 mM; (c) 50 mM, and (d) 400 mM concentrations. Reactions

carried out at 30.00.1 ^ꢁC and are corrected for one unit of enzyme.

Scheme 4. (A) Expected transition state for C N bond cleavage in the substrate. (B) Possible interaction of (1S,2S)-1-methyl-1,2-cyclopropanedioic

acid (4) with the enzyme.

As was suggested previously⁵to understand why the

product of the serine codon should react with N-ethyl-

maleimide³³is not easy if serine is the ®nal product

given that the b-hydroxyl group of serine does not

usually react with Michael acceptors to give conjugate

addition products. To accommodate the observation it

was proposed that a dehydroalanine residue could be

formed and then undergo reaction with the amino

group of the substrate to give a covalent enzyme±sub-

strate adduct.⁵The elimination of the 3-amino group of

the 2,3-diaminopropanoic acid moiety from the sub-

strate would then give mesaconic acid and the amino-

960

J. Pollard et al. / Bioorg. Med. Chem. 7 (1999) 949±975

However, hydrazine^35,36and methyl hydrazine serve as

extremely slow co-substrates for the formation of hy-

drazino acids from substituted fumaric acids. Thus, the

enzyme can be rescued from irreversible inactivation by

fumaric acids (Scheme 5).

Table 6. Amino acid composition for peptides derived from 3-

methylaspartase. (A) Determined by acid hydrolysis from the puri®ed

peptide isolated from N-[2-¹⁴C]-ethylmaleimide inhibited enzyme by

Wu and Williams.³³(B) Deduced from the inferred amino acid

sequence of the enzyme for the Val-148-Lys-183 peptide.⁵(C)

Determined by N-terminal amino acid sequencing of the peptide

isolated from N-[2-³H]-ethylmaleimide-inhibited enzyme in this study.

(D) Amino acid content of the putative bis-oligo-peptide thioether

formed by cross-linking Ala-350-Arg-366 (column B) and Asp-156-

Lys-177 (the smallest tryptic peptide containing Ser-173) via Ser-173

and Cys-357. The numbers in brackets refer to the expected outcome of

an amino acid analysis of the bis-oligo-peptide, see column A

Dehydroalanine residues have also been implicated in

the deamination reactions catalysed by phenylala-

nine^13,22and histidine^23,37ammonia-lyase. Remarkably

the precursor serine residue for both enzymes occurs as

part of the triad Ser-Gly-Asp^24,25as it does in methyl-

aspartase.⁵Conversely, the aspartate ammonia-lyase

reaction does not appear to involve the formation of

covalent intermediates.^10,12

Amino acid

A

B

C

D

A

R

N+D

C

Q+E

G

H

I

L

K

M

F

P

S

T

W

Y

V

5

2

8

1

4

8

0

2

1

2

1

2

0

1

0

1

3

4

2

8

0

4

2

0

2

0

2

1

2

1

0

2

5

2

1

3

2

1

4

0

1

0

2

0

1

0

5

2

8

2 (1)

3

6

0

1

0

0 (2)

1

2

1 (0)

2

0

2

Although there is a growing body of evidence and now

also good chemical support for the role of dehy-

droalanine residues in reacting with the amino groups of

substrates in the ammonia-lyase systems, direct proof

has yet to be provided for any system. Indeed, recent

work from Retey's laboratory on phenylalanine ammo-

nia-lyase questions whether the amino group of the

substrate (via a conjugate addition), or, the aromatic

ring (via Friedel±Crafts alkylation) reacts with the

dehydroalanine residue.³⁸These workers, in fact, come

out in favour of the Friedel±Crafts reaction. This new

proposal in which C C bond formation and then C C

bond cleavage occurs either side of a proton abstraction

step, and before C N bond cleavage, does not ®t well

with deuterium and nitrogen isotope eects observed by

Cleland and co-workers.¹³If the same side chain reac-

tion of the substrate with the dehydroalanine residue

did occur for methylaspartase, the b-carboxylate group

would react to give a b-carboxylic ester. While such an

adduct might be expected to possess more acidic C-3

hydrogen atoms, the idea does not ®t in with the results

of several other experiments including the ability of the

enzyme to form a-hydrazino acids rather than b-acyl

hydrazides from fumaric acids and hydrazine.³⁶Thus,

for the purpose of interpreting the data described in this

paper we shall ®rst consider a mechanism in which a

dehydroalanine residue³⁸exists at the active site and

serves to covalently anchor the amino group of the

substrate and the product, through its b-C-3 rather than

it's a-C-2 carbon atom²⁶(Scheme 2).

enzyme. While we have shown that in the native enzyme

the amino-enzyme cannot be trapped,¹⁹the same does

not necessarily apply to the active unlabelled NEM-

modi®ed enzyme which was treated with labelled N-

ethylmaleimide by Wu and Williams.³³In an alternative

mechanism, it is not unreasonable to believe that the

dehydroalanine moiety could act as an enimine and,

therefore, as a nucleophile at C-3 and react with NEM

to form a C C bond to label the protein. Indeed, since

many Cys residues did react with unlabelled NEM, it is

reasonable to expect that the ability of the modi®ed

enzyme to deaminate itself would be impaired compared

to the native enzyme. Thus, the amino-enzyme form

might be able to exist long enough to react with elec-

trophiles such as NEM. In accord with these ideas,

chemical model studies have shown that the addition of

amines including the dimethyl ester of (2S,3S)-3-methyl-

aspartic acid to various N-acyl dehydroalanine esters

and amides is quite rapid at 30 ^ꢁC and occurs at C-3 or

C-2 depending on the electronic properties of the pro-

tecting groups.²⁶It is also established that N²-protected

2,3-diaminopropanoic acid amides react with N-

methylmaleimide to give the conjugate addition pro-

ducts.³⁴Thus, there is now ample evidence to support

the proposals that dehydroalanine residues in proteins

could serve as electrophilic amine group acceptor moi-

eties as has been proposed for ammonia-lyase enzymes

for many years^22,23(see below).

Metal ion cofactor binding/debinding orders

The sequence of binding for the two metal ion cofactors

Mg²⁺and K⁺has been investigated in some detail but

again, at the time of the work the possible involvement

of a dehydroalanine residue was not appreciated.¹⁹The

analysis of the results indicated that at all pH values at

which experiments were performed, pH 6.5 through to

pH 9.0, the substrate and Mg²⁺ion added to the

enzyme in a random fashion before K⁺ion. However

several features of the analysis were not easy to accom-

modate including the fact that the patterns for double

reciprocal plots for substrate and Mg²⁺ion binding

changed from an intersecting one below pH 9.0 to a

parallel one at pH 9.0 (Fig. 13). Thus, at pH 9.0 Mg²⁺

acted as an uncompetitive activator and the values of

V/K at dierent Mg²⁺concentrations were the same

Support for the existence of a dehydroalanine residue at

the active site of methylaspartase is also provided by the

®nding that aryl and alkyl hydrazines are potent irre-

versible inhibitors of the enzyme (Fig. 5 and Results).

J. Pollard et al. / Bioorg. Med. Chem. 7 (1999) 949±975

961

Scheme 5.

Figure 13. Eect of Mg²⁺concentrations on deamination at pH 9.0. Incubations contained; 0.5 M Tris±HCl (pH 9.0), MgCl₂, KCl and 3-methyl-

aspartic acid, in a total volume of 3 mL. Reactions carried out at 30.00.1 ^ꢁC and are corrected for one unit of enzyme. Some data points are o

scale. (a) 20 mM MgCl₂, 50 mM KCl; (b) 1.0 mM MgCl₂, 50 mM KCl; (c) 0.4 mM MgCl₂, 50 mM KCl; (d) 20 mM MgCl₂, 1.6 mM KCl; (e) 1.0 mM

MgCl₂, 1.6 mM KCl; (f) 0.4 mM MgCl₂, 1.6 mM KCl; (g) 0.2 mM MgCl₂, 1.6 mM KCl.

and as a consequence the apparent K_Mvalues for the

substrate decreased with decreasing Mg²⁺concentration.

That is, substrate was increasingly tightly bound as the

Mg²⁺ion concentration was decreased. This is certainly

consistent with the formation of a covalent substrate±

enzyme complex. Furthermore, the primary deuterium

isotope eects upon V and V/K were suppressed at low

Mg²⁺concentrations, but not at high concentrations.

Given that the substrate could form a dead-end complex

if it reacts with the dehydroalanine before Mg²⁺binds

to the enzyme as depicted in Scheme 6, it is possible to

account for both of these previously unexplained

observations. The change in the initial velocity double

reciprocal plot patterns with pH would then arise because

the reaction of the substrate with the dehydroalanine

residue would be more rapid at high pH where the

amino group of the substrate would be unprotonated.

Where the substrate was covalently attached to the

enzyme to give adduct A and then trapped on the

enzyme by K⁺ion, to give dead-end complex B, there

would be no escape apart from through reverse steps

since Mg²⁺cannot bind after K⁺ion.¹⁹It would

seem possible that Mg²⁺ion could bind to the cova-

lent aminated adduct (A in Scheme 6) but if it could

not, A would also be a dead-end complex. Thus, the

steady-state concentration of the dead-end complex(es)

at any given concentration of Mg²⁺would be high at

high pH and/or high K⁺ion concentration. If it is

assumed that the rate of breakdown of the covalent

substrate±enzyme complex is slower in the absence of

962

J. Pollard et al. / Bioorg. Med. Chem. 7 (1999) 949±975

Scheme 6.

Mg²⁺ions then high concentrations of Mg²⁺ion would

be required to suppress the formation of the dead-end

complex and rescue the substrate from a pathway pos-

sessing a high forward reaction commitment. This ana-

lysis indicates that for the formation of active

complexes, Mg²⁺must be bound to the enzyme before

the formation of the covalent substrate±enzyme complex

and remain bound. Otherwise, at low Mg²⁺concentra-

tion, the substrate would be sticky and, in its covalent

adduct form, unable to freely dissociate from the

enzyme. This is consistent with the suppressed values of

^D(V/K) for the substrate at low Mg²⁺and the lower

apparent values of K_mfor the substrate.¹⁹

covalent enzyme±substrate complex. Under dierent

conditions, however, some other results indicate that the

addition of K⁺ion to the enzyme occurs before the

formation of the covalent enzyme±substrate complex

(see below).

Some clues to the puzzle come from the knowledge that

K

⁺ion does show simple saturation kinetics well below

pH 9.0. Using the high K⁺ion asymptote of double

reciprocal plots of initial rate versus activator con-

centration at pH 9.0 gives values of V and V/K which lie

on a zero and unit slope plot, respectively, for graphs of

these kinetic parameters versus pH (Fig. 2). The para-

meter V with respect to K⁺is pH independent meaning

that the slow rate of reaction observed at low pH can be

completely overcome by increasing the activator con-

centration. This fact alone indicates that the potassium

ion-bound form of the enzyme is closed and not acces-

sible to solvent derived protons during the course of the

reaction since the pH of the solution is irrelevant. The

substrate amino group must be protonated at low pH

and, since the K_mvalue of the substrate is pH indepen-

dent above pH 6.5 and below pH 9.0 but increases

above pH 9.0 (see Fig. 2), it is evident that the proto-

nated form of the substrate is the true substrate. Thus,

there must also be a pathway in which the amino group

of the substrate attacks the dehydroalanine residue after

K⁺ion has bound (Scheme 7). In this scheme the low

pH site for K⁺would be site A and instead of removing

a proton from the substrate, a proton must be removed

from the enzyme to provide a base for the deprotona-

tion of the alkylammonium group after the enzyme has

closed.

Another consideration that must be accommodated

within the mechanism involving the intermediacy of a

dehydroalanine residue to complete the construction of

a kinetic model concerns the timing of potassium ion

binding. Whilst it is clear that K⁺only binds after the

substrate across the pH range 6.5±9.0,¹⁹whether it

binds before the addition of the amino group of the

substrate to the dehydroalanine residue or after the

formation of the covalent complex, could not be deter-

mined from the results of previous work. In Scheme 6,

for the sake of simplicity, it is assumed that K⁺binds

after the formation of the covalent Mg.E±S complex.

Nevertheless, it was previously established that the

concentration of K⁺ion had a profound eect on the

apparent values of ^DV and ^D(V/K) and suppressed

the isotope eects at high concentration, as is expec-

ted for a species which binds after the labelled one

(see ref 40 for a useful review, and references cited

therein). However, very low concentrations of K⁺ion

also suppressed ^DV and ^D(V/K). Furthermore, the acti-

vator ion did not show simple saturation kinetics at pH

9.0 and two apparent dissociation constants were

detected. From the results presented here (Table 7, Fig.

1 and Results) and in the following article, it is clear

that K⁺does not suppress the value of the solvent

isotope eect or the value of ¹⁵(V/K) at pH 9.0. These

®ndings are more readily consistent with the addition

of K⁺ion to the enzyme after the formation of the

It is also evident from the curved double reciprocal plots

of initial rate versus activator concentration determined

at pH 9.0¹⁹that at low substrate concentration, the

apparent binding constant for K⁺is high (ca. 20 mM).

This value of the apparent binding constant for K⁺is in

accord with values extrapolated for K_Kdetermined at

low pH where the substrate must be protonated. Con-

versely, the apparent value at high substrate concentration

J. Pollard et al. / Bioorg. Med. Chem. 7 (1999) 949±975

963

is low (ca. 0.2 mM). The result suggests that at high pH,

when the substrate can deprotonate, the nucleophilic

amino group can directly attack the dehydroalanine

residue and, in so doing, creates a new binding site for

K⁺(site B in Scheme 7), with a very high anity.

Hence, at pH 9.0 or above, the dependence of the rate of

reaction on K⁺would be dependent upon the substrate

concentration and its reaction with the dehydroalanine

residue and, double reciprocal plots of rate versus K⁺

activator ion concentration would be non-linear, as was

observed.¹⁹

residue. At low Mg²⁺concentration some dead-end

complex is formed. At high K⁺concentration and at

pH 9.0 or below, the substrate binds after the Mg²⁺ion

and is then trapped on the enzyme by K⁺and depro-

tonated internally, without the in¯uence of solvent pH,

to give the species which attacks the dehydroalanine

residue.

The ability of the enzyme to perform its function in a

pH-independent manner at saturating K⁺ion con-

centration over the range pH 6.5 to 9.0 raises some

interesting questions concerning the protonation states

of the various acid±base groups on the enzyme in the

closed form. Clearly, a deprotonated form of the

enzyme and the presence of substrate are required to

create a binding site for K⁺. The increasing anity of

the site A in Scheme 7 with increasing pH with unit

slope in the logarithmic plot (Fig. 2), indicates that a

single acid possessing a pK_avalue of greater than 9.0 is

required in its unprotonated form for K⁺binding.

Above pH 9.3 the plot of log V versus pH plummets

with a negative slope of 3.0 and the reaction appears to

be limited by other steps involving external proton

transfers in both the forward and reverse reaction. The

solvent isotope eects in both directions are large for V,

4.5 and 6.7, respectively, and small for V/K, 1.6 and 1.3,

respectively (see Tables 1 and 4, and Fig. 1). The kinetic

parameters for the slow substrate, (2S)-aspartic acid,

show identical properties with respect to pH below pH

9.0 and V and V/K increase with unit slope and then

level out to zero gradient revealing that an acid posses-

sing a pK_avalue of 9.1 must be deprotonated for activ-

ity (Fig. 3). The dierence in behaviour of the slow

substrate above pH 9.4 compared to the natural sub-

strate (Fig. 2) suggests that the slow substrate can access

an acid at high pH (see Discussion).

The titrations in the deuterium solvent isotope eects

[and ¹⁵(V/K) eects, which are discussed in the follow-

ing article] at or just above pL 9.0 (Fig. 1) are also

consistent with the exposure of the ®rst part of the

reaction sequence to the solvent. The uncatalysed (or at

least less eciently catalysed) addition of the amino

group to the dehydroalanine would be expected to show

a substantial isotope eect for the protonation of the

conjugate base. Moreover, the attack of a deprotonated

amino group would be expected to show a substantially

reduced ¹⁵(V/K) eect by analogy with the phenylala-

nine ammonia-lyase system.¹³Both of these properties

are observed at high pH (see Results and the following

article). Note that the external reaction of the amino

group of the substrate with the open form of the enzyme

in Scheme 7 is entirely analogous to the formation of

the dead-end complexes lacking a bound Mg²⁺ion

referred to above and shown in Scheme 6. Applying the

same reasoning to the reverse reaction with the knowl-

edge that mesaconate, ammonium ion and Mg²⁺bind

in a random fashion¹⁹and that the solvent isotope

eects are largest at high pH (Fig. 1) gives Scheme 8.

Assumptions implicit in kinetic Scheme 8

Scheme 8 embodies the culminated results of all previous

work and analyses. At high Mg²⁺concentration, low

Aspartic acid shows the same type of non-linear double

reciprocal plots for initial rate versus K⁺concentration,

but the curvature is more pronounced (data not shown)

and the dierences in the K⁺binding anities of the

two sites (A and B in Scheme 7) are larger. This may

K

⁺concentration and at pH 9.0 or above, the substrate

binds after the Mg²⁺ion and then is deprotonated to

give the species which attacks the dehydroalanine

Scheme 7.

964

J. Pollard et al. / Bioorg. Med. Chem. 7 (1999) 949±975

Scheme 8.

re¯ect that fact that the substrate is not such a good ®t,

so that complex A is destabilised but, at the same time,

is less sterically hindered and less protonated at low pH

and more able to attack the dehydroalanine in the open

form of the enzyme than the natural substrate. The rate

of the deamination of aspartic acid is not limited above

pH 9.0 in the same way as the natural substrate and

very large solvent isotope eects of 10.8 and 6.4 are

expressed on V/K at pH 9.0 and 9.4, respectively (see

Table 2). The reverse reaction, the amination of fumaric

acid, is faster than the corresponding amination of

mesaconic acid and interestingly, as for mesaconic acid,

the solvent isotope eect is largest for V at pH 9.0 (ca.

8.3) and equally expressed on V and V/K at pH 9.4 (ca.

3.3).

pH limit of the study. If the acid±base groups cannot

exchange protons with each other on opposite faces of

the product, mesaconic acid, in the closed form of the

enzyme, the base B-3 must be unprotonated. Further-

more, the generation of a tight-binding site for K⁺

appears to be associated with the generation of the

other base, B-1 or B-2.

In order to gain further information on the properties

of the active site, a number of experiments were per-

formed using the irreversible inhibitors phenylhydrazine

and N-ethylmaleimide (see Experimental and Results).

A plot of the pH dependence of the log of the rate

inhibition by NEM gave a line with unit slope and a

pK_avalue of 9.3 (Fig. 8). Since NEM does not ionise in

the pH range of the study, the result indicates that an

acid with a pK_aof 9.3 must be deprotonated to react

directly with NEM or that such an ionisation must

occur in order to unmask a nucleophile that is sterically

hindered in the protonated form of the enzyme. The

value of the pK_acorrelates well with the pK_avalue of

9.3 derived from the binding of K⁺ion (see above).

Intriguingly, phenylhydrazine displayed a similar pH-

dependence for inhibition except between pH 9.0 and

9.5 the plot displayed a slope of 2.0 indicating that two

acid groups need to be deprotonated for the reaction

(Fig. 6). This might be expected for a reaction that

occurs in the open form of the enzyme and which

requires a base catalysed conformational change and a

nucleophilic attack on a dehydroalanine. However, the

reaction of hydrazines with the dehydroalanine residue

is reversible initially^35,36and the enzyme can be rescued

with fumaric acids (Scheme 5). Thus, the inactivation

kinetics we observe may represent slow steps involved in

the slower conversion of the b-N-aminoalanine residue

into a permanently inactivated system.

Active site acid±base groups

At least three acid±base groups are required for the

deamination of (2S,3S)-3-methylaspartic acid, two to

deprotonate the amino group of the substrate and amin-

ated adducts of the enzyme, B-1 and B-2, and one to

remove the C-3 proton from the substrate, B-3. The

latter base, B-3, should not be able to access the amino

group of the substrate directly because the C-3 hydro-

gen and amino group must be disposed in an anti-con-

formation upon opposite faces of the double bond in

the nascent product, mesaconic acid. However, a poly-

protic base, possibly B-1, should be able to access the

conjugate base B-3, if the acid form is a monoprotic

thiol in order to account for the exchange reaction (see

below). It should be noted that the requirements for the

syn-deamination of (2S,3R)-3-methylaspartic acid are

dierent and are considered below.

Two of the acid±base groups, B-3 and either of B-1 or

B-2, would need to be unprotonated in the closed sub-

strate bound form (Scheme 9). Interestingly, only V/K

for K⁺shows a pH-dependent increase when all species

are saturating. The unit slope in the log plot of V/K

versus pH indicates that a single acid must deprotonate

in order to bind K⁺and that its pK_ais ca. 9.3 (see dis-

cussion above). This analysis indicates that one of the

acids must be unprotonated at or below pH 6.5, the low

The properties of NEM are more simple to analyse. The

substrate protection aorded by methylaspartic acid

and mesaconic acid and the derived K_Svalues for the

substrates are in complete accord with the independent

®ndings of Bright who used alternative thiol attacking

reagents, including mercury benzoates, to inactivate the

enzyme.⁸

J. Pollard et al. / Bioorg. Med. Chem. 7 (1999) 949±975

965

The discovery that Mg²⁺ions activate the inhibition

process (see Results) is of particular interest because NEM

could bind very close to the position of the base respon-

sible for the removal of the proton from the carbon acid

at C-3 of the substrate. It is believed that the role of Mg²⁺

ion is to activate the C-3 hydrogen to proton abstraction

by the formation of a magnesium 4-carboxylate interac-

tion (Fig. 14). Indeed, it is possible that the protection

from inactivation aorded by phosphate dianion is

attributable to the ability of the dianion to occupy the

site for the 4-carboxylate group of the substrate. Note,

acetate was totally ineective in aording protection at

up to 50 mM. Hence, it was of interest to determine the

nature of the group that was modi®ed by NEM and its

position within the sequence of the enzyme.

tide containing two lysine residues.³³Previously we

suggested that an amino-enzyme form of the protein

derived from a dehydroalanine residue would match the

amino acid composition of the Wu and Williams' pep-

tide⁵(Table 6). However, we were completely unable to

detect a trypsin resistant Lys-Xaa bond in our own

proteolysis studies. Indeed, the short 17 amino acid

sequence (ANGMGAYCGGTCNETNR) obtained for

the peptide labelled with tritiated NEM possessed two

Cys residues (Cys-357 and Cys-361) and no Lys resi-

dues. One of the Cys residues, Cys-361, was modi®ed

and radiolabelled, and the other was unalkylated (see

Results).

From the observed protection against alkylation by

NEM oered by substrates and phosphate dianion,

together with the activation of inhibition caused by

Mg²⁺ions, it is reasonable to propose that the NEM

Wu and Williams claimed that NEM modi®cation of

the enzyme, followed by tryptic digest gave a large pep-

Scheme 9.

Figure 14. Showing A) the expected arrangement of the base (B-3) for deprotonation of C-3 in the sbstrate, and; B) the reaction of enzyme±Mg²⁺

bound NEM with the base in the inactivation reaction.

-

966

J. Pollard et al. / Bioorg. Med. Chem. 7 (1999) 949±975

molecule binds to the enzyme±Mg²⁺complex and

alkylates one of the catalytically essential acid±base

groups (Fig. 14). Given that Mg²⁺ions increase the

binding anity of NEM >10-fold but do not enhance

the maximum rate of the inactivation, k_Inact, by much, it

would seem that the conjugate addition of the thiolate

group of Cys-361 is not catalysed through the Mg²⁺ion

serving as a Lewis acid. Thus, it is likely that the thiolate

group attacks at a position analogous to the C-3 site of

the substrate which, in turn suggests that it is the C-3 base

(B-3 as de®ned in Scheme 9) that is modi®ed (Fig. 14).

possible reaction pathway is depicted in Scheme 10. It is

important to note that neither ourselves or previous

workers have identi®ed disul®de bridges in the native

form of the enzyme.^8,33

If the dehydroalanine residue is protected by Cys-357

then many features of the system become easier to

understand.^xFor example, the failure of sodium boro-

hydride and cyanide to react with the dehydroalanine

residue; the fact that substrates in the presence of K⁺

ion and inhibitors require the same base-catalysed step

to enter the active site, and the fact that the two Cys

residues could be appropriately placed to serve as bases

for the deprotonation of the alkylammonium group

(B-1 or B-2, see above) and C-3 of the substrate (B-3),

respectively, within the closed enzyme form.^k

In order to rationalise our own sequencing results with

the amino acid analyses performed on tryptic peptides

by Wu and Williams (Table 6), it is necessary to sug-

gest that the unlabelled Cys residue was protected

throughout the reaction with NEM, since it remained

unmodi®ed. One attractive manner would be by form-

ing an internal conjugate addition adduct with the

dehydroalanine residue; a thioether bridge linking

Cys-357 and dehydroAla-173. If in our experiments the

re-elimination of thiolate occurred to give a tryptic

digest of the open form of the enzyme, whilst in Wu

and Williams' experiment³³a tryptic digest of the closed

form of the enzyme was obtained, there would be no

dichotomy. One sensible explanation for the dierence

could be that in our proteolytic studies, recombinant

enzyme containing a polyHis tag was used instead of

wild-type enzyme.⁴¹It is possible that the proteins

behave dierently during proteolysis and that the His

tag was capable of catalysing the sulphydryl elimina-

tion. Interestingly, the theoretical conjugate is the cor-

rect size and possesses the correct amino acid residues if

it is accepted that the hydrolysis product of the Cys-

dehydroAla thioether might have appeared to Wu and

Williams to look like two Lys residues (see Table 6).

Primary deuterium isotope eects

It would appear from the analysis above that below pH

9.0, the enzyme catalyses all of its chemistry internally

within a closed structure after K⁺ion has bound. The

fact that ^DV and ^D(V/K) are pH independent across the

entire pH range at 1.70 and the fact that the enzyme

displays solvent hydrogen exchange, measured as tri-

tium incorporation, at rates equal to or faster than the

rate of deamination, and that at dierent pH these

exchange reactions show a primary deuterium isotope

eects of ꢀ1.7, indicates that the C-3 hydrogen is

removed at equilibrium. The ratio of the exchange

reaction to deamination (n_Ex/n_Deam) is pH sensitive^8,19

and varies from much less than 1.0 above pH 9.0 and

below pH 6.0, to about 5.0 at pH 7.6. The rate of dea-

mination increases by 85-fold from pH 6.5 to 9.0.

However, both reactions show similar primary deuterium

isotope eects for V_maxof ꢀ1.7 across the pH range 6.5 to

9.0 and the ratio n_Ex/n_Deamfor the deuteriated substrate

is identical to that for the unlabelled material at any

given pH.¹⁹Although the exchange reaction complicates

the analysis of some of the experiments described here,

the pH-independent magnitude of the isotope eects for

exchange of ꢀ1.6 and the similar pH-independent mag-

nitude of the isotope eects for deamination of ꢀ1.7 for

^DV and ^D(V/K) (which are identical within experimental

error), together with the pH-dependent ratio n_Ex/n_Deam

require rather well de®ned conditions.

A further line of evidence that links the dehydroAla

residue with the NEM labelled peptide comes from stu-

dies of the inhibition of the enzyme with nitrophe-

nylhydrazine. Here it was noted that the tryptic digest

of a modi®ed peptide possessing a 360 nm absorbance

was a bis-disul®de adduct of the ANG peptide contain-

ing the Cys residues 357 and 361. It appeared that these

Cys residues were linked to the Cys-containing peptide

SAEVTTNIGMACGAR, spanning residues 367±382,

and the peptide GVDAELVADEWCNTVEDVK,

which spans residues 299±317. Indeed, all of the Cys

containing peptides sequenced as disul®de adducts with

the ANG peptide indicating that the hydrazine was

oxidising one of the thiol groups in the ANG peptide

and that other Cys containing peptides were then dis-

placing and exchanging with the oxidised species. One

If the exchange reaction requires an external acid±base

group, the exchange must occur after K⁺has been

released from the enzyme, so that the active site can

open to exchange hydrogen with the solvent. The

rebinding of K⁺to the product complex would then

allow the reverse reaction to occur to give C-3 hydrogen

exchanged substrate. There is much additional evidence

to indicate that the enzyme cannot exchange hydrogen

before the deamination reaction is complete including

the ®nding that the slow substrate (2S,3R)-3-methyl-

aspartic acid does not epimerise but rather deaminates

to give mesaconic acid and then re-aminates to give

(2S,3S)-3-methylaspartic acid²⁰(see below). Since tri-

tium incorporation into the substrate pool could display

a very large isotope eect, the exchange rates may be

gross underestimates for the reverse reaction commit-

^xMethylaspartase from Citrobacter amalonaticus strain YG-1002 has

recently been cloned and the gene sequenced (Kato, Y.; Asano, Y.

Appl. Microbiol. Biotechnol. 1998, 50, 468). Comparison of the infer-

red amino acid sequence with that from Clostridium tetanomorphum,

highlights that Cys-357 is not conserved and is glutamine in C. ama-

lonaticus. This suggests that Cys-361 or possibly an alternative Cys

residue is involved in the formation of the thioether cross-link.

^kNote that we are not advocating that the pK_avalue of Cys 361 is ca.

9.3, as it is believed that a conformational change associated with K⁺

binding titrates at pH 9.3 (see above).

J. Pollard et al. / Bioorg. Med. Chem. 7 (1999) 949±975

967

Scheme 10.

ments to C-3 deprotonation. Nevertheless, the reverse

commitments are large and ^DV and ^D(V/K) could only

be equal in magnitude and vary synchronously with

changing cofactor and activator concentrations if there

were no commitments or very large and dominant

reverse commitments and a deuterium isotope equili-

brium constant (^DK_eq) for C-3 deprotonation that is not

equal to 1.0 (eqs (1) and (2)).

direction (to give at least one unbound product) and in

the reverse direction to give C-3 hydrogen exchanged

substrate. Assuming for now that exchange occurs

immediately after the removal of the C-3 proton of the

substrate (i.e. at the earliest time that it could happen in

any mechanism), the values of C⁰_fand C_r⁰would be the

same as those for C_fand C_rin eqs (1) and (3). At pH

9.0, n_Ex=n_Deam(measured as tritium wash-in from tri-

tiated water¹⁹)=1.0 such that C_r⁰ C⁰_f. Substituting for

C_finto eq (1) and assuming that ^Dk=7.0 gives the

expression C_r 5:3=ꢂ2:4 ^DK_eq. Thus, reaction com-

At pH 7.6, n_Ex=n_Deam 5:0 such that C⁰_r 5C_f⁰. Sub-

stituting for C_finto eq (1) now gives the expression

C_r 5:3=ꢂ1:84 ^DK_eq. Both expressions give a com-

mon value of 7.57 for C⁰_fat pH 9.0 and 7.6 when

^DK_eq 1:7 and the values of C⁰_rare then 7.57 and 37.9

at pH 9.0 and 7.6, respectively. Recall that at pH 6.5

^Dk C_f ^DK_eqꢀC_r

D

ꢀV=K 1:7

ꢀ1

ꢀ2

1 C_f C_r

D

mitments can be estimated for assumed values of K_eq.

^Dk C_Vf ^DK_eqꢀC_r

1 C_Vf C_r

^DV 1:7

C_fis the forward reaction commitment for the deuter-

ium sensitive step, C_ris the reverse reaction commit-

ment and C_Vfis the forward reaction commitment ratio

used for V_maxisotope eects as de®ned by Cleland.⁴⁰

n

_Ex/n_Deam=1.0, also, and that n_Ex/n_Deamfalls-o below

pH 6.5 and above pH 9.0.^8,19This analysis clearly indi-

cates that the reverse commitments are large across a

wide pH-range and makes assumptions that would

probably underestimate the size of C_rfor the deuterium

sensitive step in eq (1). This is because an allowance for

the tritium isotope eect was omitted, and because C-3

exchange with the solvent should occur after the release

of K⁺ion from the product complex (i.e. later in the

mechanism than we have assumed). Hence, C_ris so large

that the primary deuterium kinetic isotope eect data

can only be consistent with the operation of a thiolate

base for the removal of the C-3 proton, base B-3. This

®nding accords with the notion that Cys-361 is B-3.

Only one type of base, a thiolate, which possesses a

fractionation factor of 0.4 to 0.65,^42±44could give rise to

such a large equilibrium isotope eect.⁴⁵The fractiona-

tion for the removal of the C-3 proton from methyl-

aspartic acid is expected to be ca. 1.15⁴⁶and therefore,

D

the equilibrium isotope eect K_eqwould be f_C-3(the

fractionation factor estimated for the C-3 proton of the

substrate) divided by f_BH(the fractionation factor for

the C-3 base); that is ca. 1.15/(0.5 to 0.65)=1.77 to 2.30.

In order to estimate the size of the commitments in eqs

(1) and (2) an expression was derived to account for the

exchange reaction (eq (3)) where C⁰_fand C_r⁰are equal to

C_fand C_rin eq (1) if exchange occurs immediately after

the deprotonation at C-3 of the substrate.

Timing of C-3 solvent hydrogen exchange

The products ammonia and mesaconic acid appear to

debind in a random order, or ammonium ion debinds

®rst (see above), after the deamination of the inter-

mediate 2,3-diaminopropanoic acid residue, since an

amino-enzyme intermediate cannot be trapped with

labelled mesaconic acid.¹⁹K⁺ion must debind and/or

n_Ex=n_Deam C⁰_r=C⁰_f

ꢀ3

Now the commitments describe the partitioning of the

deprotonated enzyme-bound substrate in the forward

968

J. Pollard et al. / Bioorg. Med. Chem. 7 (1999) 949±975

there must be a conformational change before there is

enough room for mesaconic acid to escape from the

enzyme. Note that this aspect of the mechanism diers

from that for phenylalanine ammonia-lyase system

where the amino-enzyme form can exist after the other

product, cinnamic acid, has escaped.¹³

forward deamination reaction direction (see above).

Presumably this is because the increase in K⁺ion con-

centration must also increase the forward reaction

commitments (and retards the release of methylas-

partate) which cancels the eect where the binding a-

nities of the enzyme±product and enzyme±substrate

complexes for K⁺ion are similar. However, ammonium

ion displays a pronounced eect in increasing the C-3

deuterium wash-out rate at low and at high K⁺ion

concentrations and also inhibits the deamination reac-

tion. At high K⁺concentration, the increase in wash-

out rate is accompanied by non-linear uncompetitive

inhibition at low concentrations of ammonium ion and

by mixed uncompetitive±noncompetitive inhibition at

high concentrations of ammonium ion (Fig. 12). If ¹⁵N-

ammonium ion is used, no ¹⁵N-label is incorporated

into the substrate pool (through reverse amination) and

it is evident that the ammonium ion is not binding sim-

ply as a product.¹⁹Most consistent with these results is

the notion that the activator site for K⁺ion has a high

anity for ammonium ion in the product complex and

simply traps the products and re-commits them to par-

tition in the reverse reaction direction. Since K⁺ion

must have been released to open up the active site to

solvent, the proton derived from C-3 of the substrate

would be able to exchange. This site for ammonium ion

is, of course, the same site that ammonium ion occupies

as an activator in the reverse (amination) reaction where

there is a squared dependence on the ammonium ion

concentration.¹⁹

Since for the methylaspartase system K⁺ion should be

released before solvent can enter the active site and

exchange protons with acid±base groups, the cause of

the exchange reaction should derive from the competi-

tion between product release in the forward direction

and K⁺rebinding (after proton exchange between B-3

and the solvent) in the reverse direction. The ratio of the

exchange reaction to deamination (n_Ex=n_Deam) is pH

sensitive but both reactions show similar primary deu-

terium isotope eects for V_maxof ꢀ1.7 across the pH

range from 6.5 to 9.0.¹⁹In the presence of 150 mM K⁺ion

the rate of deuterium wash-out from (2S,3S)-3-methyl-

aspartic acid into the substrate pool shows a bell-

shaped dependence on pH.⁸The wash-out rate increases

with the deprotonation of an enzyme-bound acid group

possessing a molecular pK_avalue of ca. 6.5 and decrea-

ses with the deprotonation of a species possessing a pK_a

value of ca. 9.2. Since the base B-3 (Cys-361) must be

protonated after the removal of the C-3 proton in the

product complex, it would appear that the lower pK_a

value corresponds to the deprotonation of Cys-361 in

the product complex. Note that earlier it was suggested

that the base responsible for the deprotonation of the

C-3 position of the substrate should possess a pK_avalue

of 6.5 or below because, there is only one ionisation

required for deamination reaction in the range 6.5±9.4

(Fig. 2) and this corresponded to K⁺ion binding.

The interpretation above indicates that the reverse

commitments estimated for the purpose of evaluating

^DK_eqare probably very much larger than the values

deduced from the C-3 exchange reaction. However, this

does not spoil the analysis. The values, as before, cor-

respond to the reaction commitments for the hydrogen

exchange step but, because the exchange takes place

later in the mechanism, C_rfor the exchange is now

equal to the rate constant for the ®rst product to debind

divided by the second order rate constant for K⁺ion

binding to the product complex. If ammonium ion

binds faster than K⁺, C_rand the observed ratio

The high pK_aof ca. 9.2 associated with a decrease in the

deuterium wash-out rate may represent the ionisation of

the group which protonates the B-3 thiolate. Several

candidate acids exist including the protonated form of

either of B-1 or B-2, or the protonated product ammo-

nium ion. It is interesting to note that the deprotonation

of the B-3 thiolate and the deprotonation of either B-1

or B-2 must occur before a new substrate molecule

binds in the active site because, in the product complex,

two of these bases would be in the wrong ionisation

state.

n

duct of the net rate constants leading to the release of

_Ex=n_Deamwill increase. C_ffor the exchange is the pro-

K⁺ion in the forward direction.

Eect of ammonium ion

Conformational changes associated with K⁺ion binding

In the discussion concerning product release (above), it

was noted that the involvement of a dehydroalanine

residue provides, potentially, many sites to which a sec-

ond ammonium ion could bind. All experiments per-

formed to date are consistent with the random binding

of ammonium ion and mesaconate in the reverse reac-

tion¹⁹where two molecules of ammonium ion are

required in the absence of K⁺. One serves as the acti-

vator and double reciprocal plots of initial rate versus

the square of ammonium ion concentration are linear.¹⁹

Increasing potassium ion concentration does not

increase the exchange rate for C-3 deuterium wash-out

even though K⁺ion binding in the reverse reaction

direction must compete with product release in the

The requirements for K⁺binding are a substrate-bound

enzyme complex, whether or not Mg²⁺ions are bound,

and hydroxide ions. At neutral pH the dissociation

constant for K⁺is large and decreases with increasing

pH but, the value of V is pH independent at saturating

K⁺concentrations. Thus, at low pH, higher concentra-

tions of K⁺are required to compensate for the low

concentration of hydroxide ions. In our steady-state

kinetic experiments we have used 0.1±1.0 nM enzyme

solution so that even at pH 6.5 there is a several

100-fold excess of hydroxide ions. The binding of

K⁺ion brings about a conformational change which

appears to release a base, B-1 or B-2, and close up

the active site so that it is no longer accessible to

J. Pollard et al. / Bioorg. Med. Chem. 7 (1999) 949±975

969

solvent-derived protons. Several other monovalent

cations will serve as surrogates for potassium ion,

including rubidium, lithium, ammonium and sodium, in

order of decreasing eectiveness.¹Thus, the enzyme is

not speci®c for K⁺and the role of the activator ions

seems to be to stabilise a closed conformation. The true

substrates are N-protonated aspartates and the ®rst

step, following the closure of the K⁺-bound complex, is

to deprotonate the alkylammonium group with B-1 or

B-2 (Scheme 9).

above) or by some other means. The pK_avalue of the

base in the open form of the enzyme must be less than

6.5 or higher than 9.3 since there are no titrations in

that range other than for the binding of K⁺ion.

Another attractive possibility is that the binding of K⁺

ion stabilises the unprotected form of the dehy-

droalanine residue such that the titration of the binding

anity of K⁺causes the elimination of, for example,

the thiolate moiety of Cys-357 to generate a base (B-2)

in the closed form of the enzyme which removes the

proton from the alkylammonium group of the substrate

(Scheme 11). The nucleophilic attack by the amino

group would then give a covalent substrate±enzyme

complex possessing a dialkylammonium moiety. Both

B-1 and B-2 would now exist in their protonated forms.

As has been alluded to above, B-1 and B-2 should be in

opposite protonation states. One base is required to

deprotonate the alkylammonium group of the substrate

to generate the nucleophile to attack the dehydroalanine

residue. Thus, the dehydroalanine residue must be

formed and ready to react with the N-nucleophile if it is

pre-protected as a thioether (with Cys-357, as suggested

The elimination of the C-3 proton and ammonium

group from the substrate (step 11) could now proceed

Scheme 11.

970

J. Pollard et al. / Bioorg. Med. Chem. 7 (1999) 949±975

aided by the base B-3 (Cys-361) to give an unprotonated

2,3-diaminopropionic acid moiety in place of dehy-

droAla-173, enzyme-bound mesaconate and three pro-

tonated bases. The acidic form of B-2 would then

protonate the 3-amino group of residue 173 either

directly, or indirectly, via B-1. The protonated residue

would undergo elimination (step 15) to regenerate the

dehydroalanine residue and form enzyme-bound

ammonia. Still in the closed form of the enzyme, the

protonated base B-1 would protonate the ammonia

molecule to give enzyme bound ammonium ion, such

that both B-1 and B-2 are now free bases. B-2 (Cys-357)

would then attack the dehydroalanine residue to give a

thioether and allow K⁺ion to debind. Following the

release of the metal ion, water could enter the active site

and exchange hydrogen with the bases B-2 (unprotonated)

and B-3 (protonated), each of which would need to

swap ionisation states in order to rebind and process a

new substrate molecule. The reverse reaction would lead

to hydrogen exchange, as has been discussed above, and

forward steps would allow the products, mesaconate

and ammonium ion, to escape. The entire proposed

mechanism is shown in Scheme 11.

It is interesting to note that at pH 9.0 at equilibrium, the

enzymic amination of mesaconate by ammonia in the

presence of Mg²⁺ions gives an almost 1:1 (actually

1.05:1) mixture of the threo:erythro methylaspartic acid

diastereomers. Thus, their free energies are very similar.

The relative values of V/K for the two substrates are

277 and 0.43, for the threo- and erythro-diastereomers,

respectively, and their ratio is 644. This would represent

the dierences in the correctly protonated states of the

enzyme for the reaction with the two substrates if their

k_catvalues were equal, but V_Thris 38 times larger than

V_Ery. Hence 38/644=5.9% of the enzyme at pH 9.0

exists in the correctly protonated form for the deami-

nation of the erythro-substrate and incorrectly proto-

nated for the deamination of the threo-substrate. For

the syn-addition/elimination to work B-3 must be

retained in its unprotonated form and both B-1 and B-2

must be unprotonated after K⁺ion binds and closes the

active site in the deamination reaction direction. Either

B-1 or B-2 (or, indeed, the amino group of the amino-

enzyme)^{should serve as the base for the removal of the

C-3 proton in the elimination reaction (Scheme 12) and

the pK_avalue of the conjugate acid must be higher than

that of 6.5 for B-3. The analysis above indicates that the

pK_avalue should be 1.14 pK units higher than the pH

of the experiment (i.e. ca. 10.1). Interestingly, at pH 9.76

Barker found that the ratio of the values of V/K for the

two substrates were 106 and the ratio for the values of V

were also 106.¹This would indicate that 100% of the

enzyme was in the correct form for processing the ery-

thro-isomer. However, we now know that the activity of

the threo-isomer falls-o dramatically at pH 9.4 (Fig. 2),

and therefore, that the values of V and V/K for the

threo-substrate would not correspond to those for the

correctly protonated form of the enzyme. Thus, from

Figure 2, corrections applied to Barker's data give a

reduction for V/K of 3.3-fold. That is 33% of the

enzyme would be in the correct form for processing the

erythro-isomer and the pK_avalue of the acid would be

0.48 pK units higher than the pH of the experiment (i.e.

10.24), which is close to the value of 10.1 estimated

above. For the threo-substrate, the fall-o in activity

with increasing pH is almost certainly assisted by the

deprotonation of the conjugate acid which must be able

to protonate the ammonia molecule after its elimination

from the 2,3-diaminopropionic acid residue, thereby

preventing it from re-aminating the nascent dehy-

droalanine residue. We ascribe this pK_aof ca. 10.1 to

the conjugate acid of B-1 and tentatively suggest that

the functional group could be the e-ammonium group

of a Lys residue in close proximity to the B-2 base,

which we believe might be a Cys-357 thiolate. Interest-

ingly we have isolated an active-site tryptic peptide

(AIK) and an active-site chymotryptic peptide

(DDQRAIKKGAGHDGF) containing the same Lys

residue upon treatment with [¹⁴C]-acetic anhydride. The

side-chain modi®cations are completely protected by

substrate. Further work will be required to determine

the pK_avalue for its modi®cation. The only other mod-

i®ed peptide isolated from the active site upon treatment

with [¹⁴C]-acetic anhydride was the dipeptide LR. If

potassium ion is not associated with the 1-carboxy

group of the substrate, an Arg residue would be a

Methylaspartase can also process the erythro-substrate,

(2S,3R)-3-methylaspartic acid via a syn-elimination

process, at low rates. The reaction displays very large

primary deuterium isotope eects upon V of ca. 7.0 and

upon V/K of ca. 4.0 at 1 mM and 50 mM K⁺ion at pH

9.0. The reaction shows a similar dependence on K⁺

ions as for the natural substrate and (2S)-aspartate, but

does not show an exchange reaction for the C-3 proton

with solvent hydrogen.^6,20Note that the natural sub-

strate shows a n_Ex=n_Deamratio of 1.0 at pH 9.0 and

recall from Results that (2S)-aspartic acid displays a

n_Ex=n_Deamratio of ꢀ16 at pH 9.0.

At ®rst, the lack of exchange for the erythro-substrate

might seem to be trivial since the substrate shows a near

intrinsic primary deuterium isotope eect upon V, and

therefore, the reverse commitments must be very low,

indeed, less than 1.0. However, the product of the syn-

elimination of the erythro-substrate is also mesaconic

acid and if the protonation state of the product complex

was the same as for the deamination of the natural

substrate, the reverse reaction would occur very rapidly

to give the threo-isomer possessing a C-3 proton derived

from the solvent. Such a reaction does not occur and the

®rst product to form is mesaconic acid.^6,20This is a

highly informative result as it indicates that the enzyme

is tightly closed during the reaction.

The results also indicate that there is a base suitably

disposed to remove the C-3 hydrogen from the erythro-

substrate which cannot interact with the C-3 base (B-3)

for the threo-substrate when the substrate and K⁺ion

are bound. If it could, a C-3 epimerisation reaction

would not be observed.

^{Note that in a syn-elimination process, the unprotected amino group of

the amino-enzyme residues, on the same face of the product as the C-3

hydrogen, could also serve as a base in a carbonium ion substrate elim-

ination mechanism (see Bioorg. Med. Chem., 1999, 7, 977, this issue).

J. Pollard et al. / Bioorg. Med. Chem. 7 (1999) 949±975

971

strong contender for its binding site. Full details of

these protein modi®cation experiments will be reported

elsewhere.

Experimental

Materials

Tris(hydroxymethyl)aminomethane (Tris), magnesium

chloride hexahydrate and deuterium oxide (99.8

atom%) were obtained from Sigma Chemical Co. (St.

Louis, MO). Potassium chloride and ammonium chlo-

ride were obtained from British Drug Houses (Poole,

Dorset, U.K.). (2S,3R)-[3-²H₁]-Aspartic acid, (2S,3S)-3-

methylaspartic acid and (2S,3S)-[3-²H]-3-methylaspartic

acid were prepared as previously reported.^16,48(2S,3R)-

3-Methylaspartic acid and (2S,3R)-[3-²H]-3-methylas-

partic acid were prepared as described by Archer et al.²⁰

All batches of the deuteriated substrate contained >95

atom% deuterium at the 3-position. All other chemicals

were of analytical grade or were puri®ed before use.

Substrates, buers, ammonium chloride and hydrated

salts for use in rate determinations performed in deu-

terium oxide were separately pre-dissolved in deuterium

The explanation for the role of the acid±base group B-1

in the mechanism for both the threo- and erythro-

methylaspartic acid substrates also explains why the V

and V/K pro®les for aspartic acid do not plummet

above pH 9.4 (compare Figs 2 and 3). It is suggested

that when B-1 is not protonated, the base can remove

the 3-pro-S hydrogen from aspartic acid to give a

syn-elimination similar to the erythro-diastereomer of

methylaspartic acid (Scheme 12).

The results described here and the proposed mechanistic

schemes are used to help interpret ¹⁵N-isotope eects

measured for (2S,3S)- and (2S,3R)-3-methylaspartic

acid and (2S)-aspartic acid in the following article.

Scheme 12.

972

J. Pollard et al. / Bioorg. Med. Chem. 7 (1999) 949±975

1

oxide and were lyophilised prior to use. The pD of all

deuteriated buer solutions was adjusted using either

deuterium chloride solution or sodium deuteroxide.

formation of 1 mmol of mesaconic acid min at pH 9.0

at 30 ^ꢁC^17,19as determined by the increase in OD₂₄₀

under the assay conditions.

Enzyme

Determination of kinetic parameters

The enzyme used in all kinetic experiments was puri®ed

from Clostridium tetanomorphum strain H1 (ATCC 15920),

obtained from the American Typed Culture Collection,

grown according to the method of Barker et al. using a

modi®cation of literature procedures^1,49as previously

Deamination of (2S,3S)-3-methylaspartic acid and (2S)-

aspartic acid. Experiments to determine V_maxand K_m

for the substrates, where pH and K⁺concentration

were varied, were performed in the presence of 20 mM

MgCl₂as described previously.¹⁹Some of these deter-

minations were repeated with (2S,3S)-[3-²H]-3-methyl-

aspartic acid and (2S,3R)-[3-²H]-aspartic acid. Experiments

were also performed with each substrate using deuter-

ium oxide as the solvent at pD values which corresponded

to the pH values for those experiments conducted in

water. pD values were determined using a standard pH

electrode according to the equation pD 0.4=pH.⁴⁴

The ®nal deuterium content of all incubation solutions

was estimated to be greater than 95 atom% and for all

rate determinations the addition of the enzyme (typically

10±25 mL, in water, to a cuvette containing 2.8 mL of the

buered substrate solution) did not signi®cantly reduce

this ®gure. Experiments performed at pH 6.5 contained

0.1 M PIPES buer, those performed at pH 7.6 con-

tained 0.1 M Bis Tris propane, those performed at pH

9.0 and 9.4 contained 0.5 M Tris buer and those

performed at pH 9.5 and 10.0 contained 0.1 M CAPS

buer.

described.^5,19The speci®c activity of the enzyme used in

1

these studies was at least 40±200 units (mg of protein)

.

Irreversible inhibition studies using NEM and hydrazines

and peptide mapping experiments were carried out using

a polyhistidine-tagged recombinant enzyme expressed in

Escherichia coli strain BL21 (DE3).⁴¹This protein dif-

fered to the wild type enzyme in possessing additional

residues (MGHHHHHHHHHHSSGHIEGRHMLEDP)

upstream and additional residues (YGSGC) downstream

of the native sequence.⁵The recombinant enzyme was

prepared in 2-L batches and puri®ed as described below.

A 2-L culture of E. coli BL21 (DE3) was grown at 37 ^ꢁC

in Luria broth, to an OD₆₀₀of 0.6 prior to induction

with 1 mM IPTG and then grown for a further 5 h at

37 ^ꢁC. The cells were harvested by centrifugation for

20 min at 9000Âg and stored at 20 ^ꢁC. Frozen cell

pellet was resuspended in 50 mM sodium phosphate

buer at pH 8.0 containing 50 mM imidazole, 100 mM

NaCl, 1 mM PMSF and 1 mM 2-mercaptoethanol at a

concentration of 1 g of cell paste/mL of buer. Cell lysis

was carried out by sonication (140 W, 20 kHz) in

10Â30 s bursts at 4 ^ꢁC and the cell debris removed by

centrifugation for 30 min at 20 000Âg. Cell free extract

was applied to a Ni-chelate column (10Â1.6 cm), pre-

equilibrated with the puri®cation buer and washed

with 2 column volumes of the puri®cation buer. A 0±1

M imidazole concentration gradient was then applied

(10 column volumes) and collected as 5 mL fractions. 3-

Methylaspartase activity (11 000 units, 126 u/mg) which

was eluted in fractions 10±15 (300 mM imidazole con-

centration) was pooled and concentrated by ultra-®ltra-

tion (Amicon ultra-®ltration apparatus) under a 45 psi

pressure of nitrogen. This was then dialysed overnight

at 4 ^ꢁC against 1 L of the puri®cation buer, retaining

50% activity. The dialyased pool was then applied to an

anion exchange HPLC column (Bio-Cad HQ Poros 20

media, 10Â3 mM) in aliquots of 1 mL and washed with

3 column volumes of the puri®cation buer. A 50±

1000 mM NaCl concentration gradient was then applied

(20 column volumes) where upon 3-methylaspartase

activity was found in those fractions eluted with 200 mM

NaCl concentration (4125 units, 257 u/mg). Analysis by

silver stain 12% SDS-polyacrylamide gel electrophoresis

showed the presence of a single protein band at 46 kDa.

This protein showed very similar (but not identical)

kinetic properties to the wild-type enzyme.^5,17,19

Amination of mesaconic and fumaric acid. Experiments

to determine V_maxand K_mfor mesaconic acid were

conducted at saturating concentrations of ammonium

ion as described previously^17,18at pH 6.5, 9.0 and 9.4.

Experiments were also performed using deuterium oxide

as the solvent at corresponding pD values. Deuteriated

solutions of speci®c pD were prepared as described

above. Similar experiments were performed using

fumaric acid at pL 9.0 and 9.4.

Initial rate assays were conducted as described pre-

viously¹⁷and each rate measurement was carried out in

triplicate. All incubations were performed at 300.1 ^ꢁC

and reactions were followed directly spectrophotomet-

rically at 240 nm, on a Pye-Unicam SP8-500 or Shimadzu

2101 instrument, under the stated conditions. Reactions

were linear over the time course measured (up to 5% of

total conversion). The initial rate data obtained was

analysed by non-linear regression analyses using the

computer program Enz®tter, by Leatherbarrow.⁵⁰Values

for V and K were obtained and primary deuterium iso-

tope eects and solvent deuterium isotope eects for V

and V/K were deduced.

C-3 Hydrogen exchange

Experiments were conducted to simultaneously measure

the rate of deamination and C-3 solvent tritium exchange

into (2S)-aspartic acid using protocols previously

described for (2S,3S)-methylaspartic acid.¹⁹Incubations

contained 0.1 M substrate, 0.5 M Tris (pH 9.0), 20 mM

MgCl₂, 1.0 mM KCl and tritium oxide 1Â10⁹dpm

containing 5 units of enzyme in a total volume of 5 mL.

Enzyme assay

Methylaspartase activity was assayed according to the

method of Barker¹where 1 unit of enzyme catalyses the

J. Pollard et al. / Bioorg. Med. Chem. 7 (1999) 949±975

973

Reactions were started by the addition of enzyme,

the extent of deamination was measured spectro-

photometrically at 240 nm and the extent of exchange

was measured by the periodic removal of aliquots of the

incubation solution, acid denaturation, and lyophilisa-

tion to constant radioactivity, as described previously.

The experiment was repeated using (2S,3R)-[3-²H]-

aspartic acid as the substrate.

concentration of 0.02 mM was used at pH 8.5 and 10.0.

CAPS buer (50 mM) was used in experiments con-

ducted at pH 9.75, 10.25 and 10.75 with a concentration

of 0.01 mM N-ethlymaleimide.

The eect on the rate of inhibition of l-threo-(2S,3S)-3-

methylaspartic acid and mesaconic acid was investi-

gated using the methods described above in potassium

phosphate buer (pH 7.0, 5 mM) in the presence and

absence of 0.01 mM MgCl₂. The concentration of both

l-threo-(2S,3S)-3-methylaspartic acid and mesaconic

acid was varied over the range 0±5 mM with 0.5 mM N-

ethylmaleimide concentration. The intial rate of inhibi-

tion data was analysed by plotting reciprocal rate versus

the concentration of either l-threo-(2S,3S)-3-methylas-

partic acid or mesaconic acid.

Covalent modi®cation of methylaspartase

Inhibition by hydrazines. Inhibition of 3-methylaspartase

by hydrazines was performed by treating buered

enzyme solution (0.5 units in 50 mM Tris±HCl, pH 9.0)

with various hydrazines and hydroxylamines including

phenyl-, 4-nitrophenyl-, 2,4-dinitrophenyl-, methyl- and

acetylhydrazine and unsubstituted hydrazine, over a

concentration range of 0.01 to 1 mM. The solutions

were incubated at 30 ^ꢁC and aliquots (5 mL) were

removed at regular time intervals and assayed for

enzyme activity. The rate of inhibition followed biphasic

®rst order kinetics and the initial faster rates were

estimated by plotting the logarithm of the activity

(relative to the activity at zero time) for early times in

the inactivation process (up to 50% inactivation)

against time and calculating the gradient of the best ®t

line. At pH 9.0 the inhibition reaction showed con-

centration dependent saturation kinetics and value of

K_Sfor phenylhydrazine was calculated from the initial

faster phase. The inhibition rate data was analysed

using non-linear regression analysis.⁵⁰The pH-depen-

dence of the rate for the initial phase of the inactivation

was determined using phenylhydrazine in 50 mM Tris

buer (pH 8.0±9.5) and 50 mM CAPS buer at higher

pH.

Selective active site labelling

With 2,4-dinitrophenylhydrazine. To 2 mL of potassium

phosphate buer (50 mM, pH 9.0) containing 2 mg of 3-

methylaspartase (500 units) was added 0.2 mL of a

50 mM solution of 2,4-dinitrophenylhydrazine in ace-

tone (5 mM ®nal concentration) and the solution was

maintained at 30 ^ꢁC. Aliquots (5 mL) were removed at

regular time intervals and the enzyme activity deter-

mined using the standard assay, see above. After 45 h

almost all of the original activity had been inactivated.

The small amount of precipitate that had formed was

removed by centrifugation for 5 min at 13 000Âg and

the supernatent solution was dialysed at 4 ^ꢁC against

two 1 L changes of water. The inactivated enzyme

was subjected to digestion with trypsin as described

below.

With N-[2-³H]-ethylmaleimide. To 1 mL of 50 mM

potassium phosphate buer (pH 7.8) containing 50 mg

of 3-methylaspartase (12 500 units) was added 1 mL of a

1 M l-threo-(2S,3S)-3-methylaspartic acid solution (pH

7.8) to protect the active site. The solution was stored at

30 ^ꢁC for 30 min prior to the addition of 0.1 mL of a 0.2

M N-ethylmaleimide solution (14 mM ®nal concentra-

tion). After a further 30 min at 30 ^ꢁC, 2-mercaptoetha-

nol was added to a ®nal concentration of 20 mM in

order to remove any excess, unreacted N-ethylmalei-

mide. The solution was subsequently dialysed at 4 ^ꢁC

against 250 mL of 50 mM potassium phosphate buer

at pH 7.8 containing 5 mM l-threo-(2S,3S)-3-methyl-

aspartic acid and then 2Â500 mL of 50 mM, pH 7.8

potassium phosphate buer and, ®nally, 500 mL of

10 mM potassium phosphate buer at pH 7.0 contain-

ing 1 mM MgCl₂. To the product of dialysis, which

had retained 95% of the original activity, was added

0.1 mL of a 40 mM solution of N-[2-³H]-ethylmaleimide

(14 mCi/mmol speci®c activity). A 10 mL aliquot was

removed at regular time intervals and the enzyme activ-

ity was determined. After 60 min almost all activity

had been lost. Excess unreacted inhibitor was removed

by the addition of 2-mercaptoethanol to a ®nal con-

centration of 1.25 mM and the solution was dialysed at

4 ^ꢁC against two 500 mL changes 50 mM potassium

phosphate buer at pH 7.0 and four 1 L changes of

water.

The eect on the rate of inactivation at pH 9.0 in

50 mM Tris buer of potassium ions (250 mM) and

ammonium ions (10, 20 and 40 mM) was also tested.

Inhibition by N-ethylmaleimide. Inhibition of 3-methyl-

aspartase by N-ethylmaleimide was performed by

treating buered enzyme solution (0.5 units in 50 mM

Tris±HCl, pH 9.0) with various amounts of N-ethylma-

leimide to give ®nal concentrations of 20 mM±0.5 mM.

The solutions were incubated at 30 ^ꢁC and aliquots

(5 mL) were removed at regular time intervals and

assayed for enzyme activity. The rate of inhibition fol-

lowed ®rst order kinetics and the rates were determined

by plotting the logarithm of the activity (relative to the

activity at zero time) against time and calculating the

gradient of the best ®t line.

The value of K_Sfor N-ethylmaleimide was determined

in potassium phosphate buer (pH 7.0, 5 mM) using the

above methods over the concentration range 0±10 mM

with and without 0.01 mM MgCl₂. The inhibition rate

data was analysed using non-linear regression analy-

sis.⁴⁹The pH dependence of the inhibition reaction was

studied using the above method in Tris±HCl buer

(50 mM) in 0.5 pH intervals over the range 7.0±10.0. An

N-ethylmaleimide concentration of 0.5 mM was used for

the experiments performed at pH 7.0 and 8.5 and a

974

J. Pollard et al. / Bioorg. Med. Chem. 7 (1999) 949±975

Proteolytic digestion of methylaspartase

MeCN:TFA (99.95:0.05) and H₂O:TFA (99.9:0.1) over

20 column volumes). A single peak in radioactivity was

located in fractions 25 and 26 (retention volume of

7.30.6 column volumes) which coincided with one of

4 major peaks in absorbance at 220 nm. The 2 frac-

tions were pooled and subjected to N-terminal amino

acid sequencing. An aliquot of each cycle from the

sequencer (about 100 mL) was retained and ³H-incor-

poration determined by scintillation counting.

An aqueous solution of methylaspartase or inactivated

methylaspartase was adjusted to pH 7.8, denatured in a

water bath at 100 ^ꢁC for 15 min and then allowed to

1

cool to room temperature. An aliquot of a 4 mg mL

solution of trypsin in 1 mM CaCl₂solution was added,

in an amount equal to 2.5% by weight of 3-methylasp-

artase. The pH of the solution was maintained at pH 7.8

by the addition of small aliquots of a 0.1 M NaOH

solution. After 2 h at room temperature a second

equivalent amount of trypsin was added and the solution

incubated at room temperature for a further 2 h. The

digestion reaction was then quenched by the addition of

glacial acetic acid to pH 2±3 and any precipitated pro-

tein was removed by centrifugation at 13 000Âg for

5 min. The supernatant solution was lyophilised and the

white amorphous tryptic digest was stored at 20 ^ꢁC.

Acknowledgements

The authors thank Dr. G. D. Kemp (University of St

Andrews) for Edman sequencing facilities, and the Bio-

molecular Sciences committee of the BBSRC and

EPSRC for grant B05857.

References

1. Barker, H. A.; Smyth, R. D.; Wilson, R. M.; Weissbach, H.

J. Biol. Chem. 1959, 234, 320.

2. Barker, H. A.; Smyth, R. D.; Wawszkiewicz, E. J.; Lee, M.

N.; Wilson, R. M. Arch. Biochem. Biophys. 1958, 78, 468.

3. Williams, V. R.; Traynham, J. G. Federation Proc. 1962,

21, 247.

4. Ueda, S.; Sato, K.; Shimizu, S. J. Nutri. Sci. Vitaminol.

1982, 28, 21.

5. Goda, S.; Minton, N. P.; Botting, N. P.; Gani, D. Bio-

chemistry 1992, 31, 10 747.

6. Archer, C. H.; Gani, D. J. Chem. Soc. Chem. Commun.

1993, 140.

7. Winkler, M. F.; Williams, V. R. Biochim. Biophys. Acta

1967, 146, 287.

8. Bright, H. J. J. Biol. Chem. 1964, 239, 2307.

9. Bright, H. J.; Ingraham, L. L.; Lundin, R. E. Biochim.

Biophys. Acta. 1964, 81, 576.

10. Hanson, K. R.; Havir, E. A. Enzymes (3rd ed.) 1972, 7, pp

75±163.

11. Abeles, R. H.; Frey, P. A.; Jencks, W. P. Biochemistry;

Jones and Bartlett Press: Boston MA: 1992; pp 508±510.

12. Nuiry, I. I.; Hermes, J. D.; Weiss, P. M.; Chen, C.; Cook,

P. F. Biochemistry 1984, 23, 5168.

13. Hermes, J. D.; Weiss, P. M.; Cleland, W. W. Biochemistry

1985, 24, 2959.

14. Kim, S. C.; Raushel, F. M. Biochemistry 1986, 25, 4744.

15. Akhtar, M.; Cohen, M. A.; Gani, D. J. Chem. Soc., Chem.

Commun. 1986, 1290.

16. Akhtar, M.; Botting, N. P.; Cohen, M. A.; Gani, D. Tet-

rahedron 1987, 43, 5899.

HPLC Analysis of tryptic peptides

Methylaspartase labelled with 2,4-dinitrophenylhydrazine.

The product of tryptic digestion was dissolved in 0.1 mL

of water ®ltered through a Millipore ®lter and then

analysed by reversed-phase HPLC using a Perseptive

Bio-Cad HPLC system ®tted with a 10 R2 POROS

column (4Â100 mM). The column was eluted with a

gradient of 100% H₂O:TFA (99.9:0.1) to a 60:40 mix-

ture of MeCN:TFA (99.95:0.05) and H₂O:TFA (99.9:0.1)

over 20 column volumes. The eluent was monitored

directly by UV spectroscopy at 220 and 360 nm collect-

ing 0.75 mL fractions. Fraction 21 coincided with the

major peak in absorbance at 360 nm (retention volume

of 12.40.3 column volumes) and was further puri®ed

by the same method (elution gradient from a 75:25 mix-

ture of H₂O:TFA (99.9:0.1) and MeCN:TFA (99.95:0.05)

to a 60:40 mixture of MeCN:TFA (99.95:0.05) and

H₂O:TFA (99.9:0.1) over 20 column volumes). Fraction

36, which coincided with the major peaks in absorbance

at both 220 and 360 nm, was concentrated under

vacuum and further puri®ed by a ®nal microbore HPLC

step (1 mM column). Two peaks in absorbance at

360 nm were observed and collected. These were sub-

mitted for N-terminal amino acid sequencing using a

Procise Edman sequencer.

17. Botting, N. P.; Akhtar, M.; Cohen, M. A.; Gani, D. Bio-

chemistry 1988, 27, 2953.

18. Botting, N. P.; Cohen, M. A.; Akhtar, M.; Gani, D. Bio-

chemistry 1988, 27, 2956.

19. Botting, N. P.; Gani, D. Biochemistry 1992, 31, 1509.

20. Archer, C. H.; Thomas, N. R.; Gani, D. Tetrahedron

Asym. 1993, 4, 1141.

21. Botting, N. P.; Jackson, A. A.; Gani, D. J. Chem. Soc.,

Chem. Commun. 1989, 1583.

22. Havir, E. A.; Hanson, K. R. Biochemistry 1968, 7, 1904.

23. Wickner, R. B. J. Biol. Chem. 1969, 244, 6550.

24. Langer, M.; Reck, G.; Reed, J.; Retey, J. Biochemistry

1994, 33, 6462.

25. Langer, M.; Rother, D.; Retey, J. Biochemistry 1997, 36,

10 867.

26. Gulzar, M. S.; Morris, K.; Gani, D. J. Chem. Soc., Chem.

Commun. 1995, 1061.

27. Gani, D.; Archer, C. H.; Botting, N. P.; Pollard, J. R.

Bioorg. Med. Chem. 1999, 7, 977.

Methylaspartase labelled with N-[2-³H]-ethylmaleimide.

The product of tryptic digestion was taken up in

0.1 mL of water and analysed by reversed-phase HPLC

using a Perseptive Bio-Cad HPLC system ®tted with a

C¹⁸pep-map column (4Â250 mM). The column was

eluted with a gradient of 100% H₂O:TFA (99.9:0.1)

to a 60:40 mixture of MeCN:TFA (99.95:0.05) and

H₂O:TFA (99.9:0.1) over 15 column volumes. The elu-

ent was monitored directly by UV spectroscopy at

220 nm and 0.75 mL fractions were collected, from

which 50 mL aliquots were removed in order to deter-

mine the ³H-incorporation by scintillation counting.

The major peak in radioactivity was observed in frac-

tion 28 (retention volume of 7.30.3 column volumes)

and was further puri®ed by the same method (elution

gradient from a 75:25 mixture of H₂O:TFA (99.9:0.1)

and MeCN:TFA (99.95:0.05) to a 60:40 mixture of

J. Pollard et al. / Bioorg. Med. Chem. 7 (1999) 949±975

975

28. Bright, H. J.; Ingraham, L. L.; Lundin, R. E. Biochemistry

1964, 3, 1224.

29. Bright, H. J.; Silverman, R. Biochim. Biophys. Acta. 1964,

81, 175.

30. Bright, H. J. J. Biol. Chem. 1965, 240, 1198.

31. Bright, H. J. Biochemistry 1967, 6, 1191.

32. Badiani, K.; Lightfoot, P.; Gani, D. J. Chem. Soc., Chem.

Commun. 1996, 675.

40. Cleland, W. W. In Investigation of Rates and Mechanisms

of Reactions (4th ed.); Bernasconi, C. F., Ed.; John Wiley:

New York, 1986; Part 1, pp 791±870.

41. Neal, T. Ph.D. Thesis, University of St Andrews, 1997.

42. Belasco, J. G.; Bruice, T. W.; Fisher, L. M.; Albery, W. J.;

Knowles, J. R. Biochemistry 1986, 25, 2552.

43. Szawelski, R. J.; Wharton, C. W.; White, S. Biochem. Soc.

Trans. 1982, 10, 232.

33. Wu, W. T.; Williams, V. R. J. Biol. Chem. 1968, 243, 5644.

34. Gulzar, M. S. Ph.D. Thesis, University of St Andrews,

1997.

35. Gulzar, M. S.; Akhtar, M.; Gani, D. J. Chem. Soc., Chem.

Commun. 1994, 1601.

44. Weiss, P. M.; Cook, P. F.; Hermes, J. D.; Cleland, W. W.

Biochemistry 1987, 26, 7378.

45. Schowen, R. L. In Isotope Eects on Enzyme-Catalysed

Reactions; Cleland, W. W.; O'Leary, M. H.; Northrop, D. B.,

Eds.; University Park Press: Baltimore, MD, 1977; pp 64±99.

46. Fisher, L. M.; Belasco, J. G.; Bruice, T. W.; Albery, W. J.;

Knowles, J. R. Biochemistry 1986, 25, 2543.

36. Gulzar, M. S.; Akhtar, M.; Gani, D. J. Chem. Soc., Perkin

Trans. 1 1997, 649.

37. Langer, B.; Starck, J.; Retey, J. Bioorg. Med. Chem. Lett.

1997, 7, 1077.

47. Clark, J. D.; O'Keefe, S. J.; Knowles, J. R. Biochemistry

1988, 27, 5961.

38. Schuster, B.; Retey, J. Proc. Natl. Acad. Sci., USA, 1995,

92, 8433.

48. Botting, N. P.; Akhtar, M.; Cohen, M. A.; Gani, D. J.

Chem. Soc., Chem. Commun. 1987, 1371.

39. Botting, N. P.; Akhtar, M.; Archer, C. H.; Cohen, M. A.;

Thomas, N. R.; Goda, S.; Minton, N. P.; Gani, D. In Mole-

cular Recognition: Chemical and Biochemical Problems II;

Roberts, S. M., Ed.; Royal Soc. Chem: RSC Special Publica-

tion 111: London, 1992; pp 95±109.

49. Hsiang, M. W.; Bright, H. J. Methods Enzymol. 1969, 13,

347.

50. Leatherbarrow, R. J. Enz®tter, A non-linear regression

data analysis program; Published and distributed by Biosoft:

Cambridge, UK, 1987.

Article Doi

Mechanism of 3-methylaspartase probed using deuterium and solvent isotope effects and active-site directed reagents: Identification of an essential cysteine residue

DOI: 10.1016/S0968-0896(99)00044-9

Source and publish data:

Authors:

Article abstract of DOI:10.1016/S0968-0896(99)00044-9

Full text of DOI:10.1016/S0968-0896(99)00044-9

Products guided by the article

R&D Labs maybe for 110-17-8

Relevant to this article

Hot Product