Mechanistic Investigation of Phenylalanine Ammonia Lyase by Using N-Methylated Phenylalanines

Article abstract of DOI:10.1002/hlca.200390305

N-Methyl-L-phenylalanine (5), N-methyl-4-nitro-L-phenylalanine (6), and N,N-dimethyl-4-nitro-L-phenyl-alanine (7·H⁺) were investigated as substrates of inhibitors of phenylalanine ammonia lyase from Petroselinum crispum. Whereas the former was

Full text of DOI:10.1002/hlca.200390305

Helvetica Chimica Acta Vol. 86 (2003)
3601
Mechanistic Investigation of Phenylalanine Ammonia Lyase by Using
N-Methylated Phenylalanines
by Sandra Viergutza), La¬szlo¬ Poppe*b), Anna Tominb), and Ja¬nos Re¬tey*a)
a) Institute of Organic Chemistry, University of Karlsruhe, Richard-Willst‰tter-Allee, D-76128 Karlsruhe
(phone: ‡497216083222; fax: ‡497216084823; e-mail: janos.retey@ioc.uka.de)
b) Institute for Organic Chemistry and Research Group for Alkaloid Chemistry,
¬
¬
Budapest University of Technology and Economics, H-1111 Budapest, Gellert ter 4.
(phone: ‡3614632229; fax: ‡3614633297; e-mail: poppe@mail.bme.hu)
Dedicated to Professor Duilio Arigoni on the occasion of his 75th birthday
N-Methyl-l-phenylalanine (5), N-methyl-4-nitro-l-phenylalanine (6), and N,N-dimethyl-4-nitro-l-phenyl-
‡
alanine (7 ¥ H ) were investigated as substrates or inhibitors of phenylalanine ammonia lyase from Petroselinum
crispum. Whereas the former was a reluctant substrate (Km ˆ 6.6 mm, kcat ˆ 0.22 sÀ1), no reverse reaction could
be detected by using methylamine and (E)-cinnamate (2). The Km value for ammonia in the reverse reaction by
using (E)-cinnamate (2) was determined to be 4.4 and 2.6m at pH 8.8 and 10, respectively. The N-methylated 4-
nitro-l-phenylalanines 6 and 7 showed only strong inhibitory effects (Ki ˆ 130 nm and 8 nm, resp.). These and
former results are discussed in terms of the mechanism of action of phenyalalanine and histidine ammonia
lyases.
Introduction. Phenylalanine ammonia lyase (PAL, EC 4.3.1.5) is an important
plant enzyme that catalyzes the conversion of l-phenylalanine (1) into (E)-cinnamate
(2) (Scheme 1), which is the precursor of a great variety of phenylpropanoids, such as
lignins, flavonoids, and coumarins [1][2]. Because of its central role in plant
metabolism, PAL is a potential target for herbicides [2].
Scheme 1. Reactions of l-Phenylalanine (1) and 4-Nitro-l-phenylalanine (3) with wt-PALand S203A PAL
Mutant

3602
Helvetica Chimica Acta Vol. 86 (2003)
The related enzyme histidine ammonia lyase (HAL, EC 4.3.1.3) catalyzes a very
similar reaction, converting l-histidine into (E)-urocanic acid (ˆ 3-(1H-imidazol-4-
yl)prop-2-enoic acid) [3][4]. In the three decades between the first mechanistically
relevant work published in 1969, which assumed that HAL contains dehydroalanine as
catalytically essential electrophilic group [3], and the recent discovery of the structure
of the electrophilic prosthetic group as 3,5-dihydro-5-methylidene-4H-imidazol-4-one
(MIO) [5], a large number of mechanistic investigations were published (see ref. cit.
in [6][7]). Their results were interpreted mostly in terms of the mechanism proposed
by Hanson and Havir for PAL [8] (Scheme 2, Path I), and later by the mecha-
¬
nisms proposed by Retey and co-workers for HAL [9] and PAL [10] (Scheme 2,
Path II).
The first proposal for the role of the electrophilic group of HAL [3] and PAL [11]
implied a Michael addition of the a-NH2 group of the substrate followed by abstraction
of the b-proton and elimination. In the amino enzyme intermediate, a covalent bond
between the prosthetic electrophile and ammonia was supposed. This seemed to be
consistent with Peterkofsky×s experiments showing a relatively stable amino HAL
intermediate [12]. The latter, in the presence of [14C]urocanate, could be re-converted
into [14C]histidine, while 15NH3 at 10 60 mm concentrations was not incorporated into
the substrate [12]. This led to the erroneous assumption that the HAL and PAL
reactions are irreversible. However, it was demonstrated [13] that the histidase reaction
‡
can be completely reversed at higher (up to 6m) NH4 concentrations.
Later, a kinetic isotope effect (ca. 1%) with [15N]phenylalanine was also interpreted
in favor of the Hanson mechanism [14]. Since the latter could not explain how the
nonacidic b-proton of the substrates can be abstracted by an enzymic base, an
alternative mechanism has been proposed for both HAL [9] and PAL [10] in which the
aromatic ring is attacked by the prosthetic electrophile in a Friedel Crafts-like
reaction. The concomitant electron withdrawal from the ring could significantly
activate the b-proton.
Identification of the electrophilic group as MIO [5] led to the insight that the latter
is much more electrophilic than dehydroalanine and, thus, able to attack the aromatic
ring. The geometry of MIO prevents delocalization of the N lone pairs into the a,b-
unsaturated carbonyl system. Moreover, addition of a nucleophile renders MIO itself
aromatic, thus compensating for the transitional abolition of the aromaticity of the
substrate phenyl ring.
The strongest argument in favor of the Friedel Crafts-type attack was, however, as
follows: It was found that the S203A mutant of PAL, lacking MIO, still catalyzed the
reaction. The Vmax value with l-phenylalanine (1) was, however, ca. 104 times lower
than for wild-type (wt) PAL. In contrast, S203A-PAL reacted relatively fast with 4-
nitro-l-phenylalanine (3) [10] (Scheme 1).
Here we describe further experiments which invalidate some arguments in favor of
the old mechanism and explain why the PAL reaction is practically irreversible under
physiological conditions.
Results. Synthesis of N-Methyl-4-nitro-l-phenylalanine (6) and N,N-Dimethyl-4-
‡
nitro-l-phenylalanine (7 ¥ H ). One route to the synthesis of 6 and 7 started from the
commercially available 4-nitro-l-phenylalanine (3) and involved reductive alkylation

Helvetica Chimica Acta Vol. 86 (2003)
3603
¬
Scheme 2. Mechanisms Proposed by Hanson and Havir [1] (Path I) and by Schuster and Retey [10] (Path II)
for the PALReaction

3604
Helvetica Chimica Acta Vol. 86 (2003)
with aqueous formaldehyde solution and NaBH4 (Scheme 3). In spite of extensive
variation of the reaction conditions (formaldehyde/NaBH4 ratio, reaction time,
temperature, as well as solvent composition), only minor amounts of 6 were detected
by HPLC; in most cases, the main product was 7, even at low conversions. Moreover,
whereas the retention time of the dimethylated product 7 was significantly longer than
that of the educt 3 (reversed-phase HPLC), the retention time of the monomethylated
product 6 was only slightly longer. Therefore, for the synthesis of 6, the commercially
available N-methyl-l-phenylalanine (5) was used as starting material. Although
conventional nitration thereof has been published [15][16], no separation from by-
products and full spectroscopic data have been described. Besides 6 as major product,
ca. 10% of a not fully characterized by-product (most likely N-methyl-2-nitro-l-
phenylalanine as its UV maximum at 274 nm indicates) was also detected by HPLC
in the crude reaction product. The final purification of both N-methylated 4-nitro-l-
1
phenylalanines 6 and 7 was achieved by prep. reversed-phase HPLC. The H- and
13C-NMR spectra as well as HR-MS data of the purified products 6 and 7 were in full
agreement with their proposed structures. It is interesting to note that maximal
solubility of the monomethylated 4-nitro compound 6 in neutral (pH 6 9) water was
only about 2 mm, whereas the dimethylated or non-methylated 4-nitro-l-phenyl-
alanines 7 and 3, respectively, were soluble up to 20 mm. Acidification increased
dramatically the solubility of 6 (over 100 mm in 10% CF3COOH/H2O, 50% HCOOH/
H2O, or in concentrated H2SO4 solution).
Scheme 3. Preparation of N-Methyl-4-nitro-l-phenylalanine (6) and N,N-Dimethyl-4-nitro-l-phenylalanine
‡
(7 ¥ H )
Kinetics of the PALReaction with N-Methyl-l-phenylalanine (5). Highly purified
recombinant PAL [17] was used as catalyst and N-methyl-l-phenylalanine (5) as
substrate (Scheme 4). The concentration of the latter was varied from 4 to 40 mm, and
formation of (E)-cinnamate (2) was monitored at 290 nm. The change of absorbance
was measured at intervals of 1 min for 5 min. On the basis of the Michaelis Menten
and Lineweaver Burk plots, the Km value was determined to be 6.58 mm and the Vmax
value to be 0.042 U/mg (Fig. 1, Table 1). The formation of (E)-cinnamate (2) was also
confirmed by HPLC/UV (retention data and peak spectrum (lmax 278 nm) were

Helvetica Chimica Acta Vol. 86 (2003)
3605
Scheme 4. Kinetic Properties of N-Methyl-l-phenylalanine (5), N-Methyl-4-nitro-l-phenylalanine (6) and N,N-
‡
Dimethyl-4-nitro-l-phenylalanine (7 ¥ H ) with wt-PAL
Fig. 1. Michaelis Menten and Lineweaver Burk plots for N-methyl-l-phenylalanine (5) in its reaction with
wt-PAL
1
identical with those of authentic 2) and H-NMR spectroscopy (the signals of the
isolated product at d(H) 6.48 (d, J ˆ 15.9 Hz, CH (a)), 7.38 (d, J ˆ 15.9 Hz, CH(b), 7.41
(m(t), Hm), and 7.58 (m(dd), Ho) matched with the d(H) of authentic 2 recorded under
similar conditions).

3606
Helvetica Chimica Acta Vol. 86 (2003)
Table 1. Kinetic Constants of N-Methyl-l-phenylalanine (5), N-Methyl-4-nitro-l-phenylalanine (6) and N,N-
‡
dimethyl-4-nitro-l-phenylalanine (7 ¥ H ) with wt-PAL
Km [mm]
Vmax [U/mg]
kcat [sÀ1
]
Ki [nm]
5
6
7
6.6
0.042
0.22
130
8
Kinetics of the Reverse Reaction; Addition of Ammonia to (E)-Cinnamate (2)
Catalyzed by PAL. Inspite of its biocatalytical usefulness [18][19], the kinetics of the
reverse PAL reaction has not yet been quantitatively examined. In particular, the Km
value of ammonia would be of interest. The latter was determined by variation of the
ammonia concentration from 0.1 to 6.4m at constant concentration of (E)-cinnamate
(2) (0.15 mm), and the reaction was monitored by the decrease of absorption at 290 nm.
The Michaelis Menten and Lineweaver Burk plots of the measurements are shown in
Fig. 2. The Km values at pH 8.8 and 10.0 were determined to be 4.4 and 2.6m,
respectively (Table 2). The Vmax values differed only slightly at the above pH values,
being 0.89 and 0.81 U/mg, respectively.
Table 2. Kinetic Constants of Ammonia at Different pH in the Ammonia Addition to (E)-Cinnamate (2)
Catalyzed by wt-PAL
pH
Km [m]
Vmax [U/mg]
kcat [sÀ1
]
8.8
10.0
4.4
2.6
0.89
0.81
4.62
4.20
Inhibition Kinetics of N-Methyl-4-nitro-l-phenylalanine (6) and N,N-Dimethyl-4-
‡
nitro-l-phenylalanine (7 ¥ H ) with wt-PAL. Apparent inhibition constants (Ki) for N-
methylated 4-nitrophenylalanines 6 and 7 were determined at different inhibitor
concentrations (0.1 2 mm of 6 or 7) by varying l-phenylalanine (1) concentrations
(0.5 4 mm). Ki values were 130 mm for 6 and 8 nm for 7, as obtained from linearization
in the Dixon plot (Table 1).
The binding properties of the N-methylated 4-nitro-l-phenylalanines 6 and 7 were
also investigated by using the recently constructed homology model of PAL containing
l-phenylalanine (1) in a Friedel Crafts-like arrangement in its active site [20]. As
shown in Fig. 3, both N-methylated compounds 6 and 7 fit well into the active site in an
almost identical steric arrangement as that of 1: the nitrated aromatic moieties sit in a
relatively hydrophobic pocket close to the exocyclic methylene group of MIO, whereas
the N-methylated amino moieties can occupy the amino binding pocket facing towards
Y110 which proved to be the most important residue in catalysis of PAL. The most
pronounced but minor difference between the l-phenylalanine (1) binding model [20]
and those containing N-methylated 4-nitro-l-phenylalanines 6 and 7 is the modified
orientation of H396, which seems to contribute to H-bonding of the amino moiety of 1.
Discussion. Synthesis of the N-Methylated 4-Nitro-l-phenylalanines 6 and 7. The
synthesis of N-methyl-4-nitro-l-phenylalanine (6) is known [15][16], although without

Helvetica Chimica Acta Vol. 86 (2003)
3607
Fig. 2. Michaelis Menten and Lineweaver Burk plots for ammonia with wt-PAL a) c) at pH 8.8 and c) d) at
pH 10.0 in the addition reaction of ammonia to (E)-cinnamate (2)
details of purification and full spectral data. To the best of our knowledge, N,N-
dimethyl-4-nitro-l-phenylalanine (7) has not been prepared yet. Here two synthetic
pathways were evaluated, and the HPLC separation of the products as well as full
spectroscopic characterization by up-to-date methods are described.
Evaluation of N-Methyl-l-phenylalanine (5) as Substrate or Inhibitor. According to
the Hanson mechanism [8] (Scheme 2, Path I), the a-amino group of the substrate
should attack the prosthetic electrophile of PAL. Therefore, we tested 5 as a substrate.
Indeed, 5 was converted into (E)-cinnamate (2), but its Vmax value was ca. 62 times
lower, while its Michaelis constant (Km) 55 times higher than that found for the natural
substrate [10]. In the concentration range studied, no significant inhibitory effect was
seen as expected from the low binding affinity to PAL.
Noteworthy is the fact that no reverse reaction could be detected with (E)-
cinnamate (2) and molar concentrations of methylamine. Considering the high Km

3608
Helvetica Chimica Acta Vol. 86 (2003)
Fig. 3. Substrate (1: in red) and inhibitor (6: in green; 7: in blue) binding active site models of wt-PAL
value of ammonia for the reverse reaction, this is understandable. From the large
increase of the Km for 5 as compared to that of l-phenylalanine, one can estimate that
the required concentration of methylamine can hardly be realized.
Inhibition of the PALReaction by the N-Methylated 4-Nitro-l-phenylalanines 6 and
7. Interestingly, N-methyl-4-nitro-l-phenylalanine (6) and N,N-dimethyl-4-nitro-l-
phenylalanine (7) were bound stronger (Ki ˆ 130 and 8 nm, resp.) than 5 (Km ˆ
6.6 mm) and behaved as inhibitors but did not serve as substrate. Although this can
not fully exclude or favor either the Hanson mechanism or the Friedel Crafts-type
mechanism, the relatively strong inhibition can be better explained with the latter.
Considering the slow but definitive reaction of N-methyl-l-phenylalanine (5) with wt-
PAL, a faster reaction with N-methyl-4-nitro-l-phenylalanine (6) is expected via the
Hanson route because an electron-withdrawing group at the 4-position of the phenyl
ring would enhance the proton abstraction from the EA' state (Scheme 2, Path I) and
also stabilize the forming carbocation in the EC state.

Helvetica Chimica Acta Vol. 86 (2003)
3609
It was shown that both wt-PAL and the MIO-less S203A-PAL mutant reacted with
4-nitro-l-phenylalanine (3) [10] (Scheme 1). Whereas the binding ability of 3 was
found to be comparable to the substrate l-phenylalanine (1) (3: Km ˆ 0.32; and 1: Km ˆ
0.17) [10], it reacts 21 times slower than the natural substrate indicating that the 4-nitro
substituent can not fully take over the function of MIO. The slower reaction of 3 can be
interpreted in terms of the Friedel Crafts-like mechanism [10] (Fig. 2, Path II) as
follows. Although the electron-withdrawing 4-nitro group acidifies the HSiÀC(b) bond,
it also deactivates the aromatic ring of 3 and, thus, it may not add to the exocyclic
methylene of the MIO system. However, the amino function of 3 can be similarly
protonated at the active site as that of 1 and, thus, can become a similarly good leaving
group. In this context, the inhibitory behavior of the N-methylated 4-nitro-l-phenyl-
alanines 6 and 7 can also be rationalized. Molecular modelling indicated that the 4-
nitro-bearing aromatic moieties of 6 and 7 fit well into the apolar binding pocket. On
the other hand, gas-phase proton affinities of phenylalanine (1) and N-methyl- and
N,N-dimethylphenylalanine [21] determined to be 220.3, 223.6, and 224.5 kcal/mol,
respectively (relative to the proton affinity of NH3 (204.0 kcal/mol) indicate that
protonation of the N-methylamino and N,N-dimethylamino moieties are favored over
the non-methylated one×s. Molecular modelling also imply that the N-methylated
amino moieties may fit well into the binding pocket. If the N-methylamino moieties are
strongly anchored to this site (e.g., to Y110) but the aromatic portion does not add to
MIO, the enzymic base (presumably Y351) cannot get close enough to HSiÀC(b). In
addition, some effects (e.g., change in interaction with H396) may influence proper
protonation and/or positioning of the leaving moiety and, thus, also diminish the
elimination.
Why is the PALReaction under Physiological Conditions Irreversible? The kinetics
of the reverse reaction catalyzed by PAL revealed an unprecedented high Km value for
ammonia, being 4.4m at pH 8.8 and 2.6m at pH 10.0. The dependence on pH indicates
that ammonia and not ammonium ion is the active species, since the molarity was
calculated from the total amount of both added to the reaction mixture. It is astonishing
that PAL works at such a high ammonia concentration. On the other hand, it makes it
clear that no reverse reaction occurs under physiological conditions, e.g., at mm
concentrations of ammonia.
Although it has been known for some time that enantioselective synthesis of various
arylalanines can be achieved by using 6m concentrations of ammonia in the PAL
reaction [18][19], the Km value for ammonia has not yet been determined. To our
knowledge, there is only one report on an even higher Km value for ammonia in the
related HAL reaction [22]. However, no experimental details are given in this
communication [22].
General Conclusions about the Mechanism of Action of Phenylalanine Ammonia
Lyase.
Although the mechanism implying a Friedel Crafts-type attack at the
aromatic ring [10] is supported by a number of experiments, there are a few results in
the old literature that have been interpreted in favor of the first proposed mechanism
implying the addition of the substrate a-amino group to the prosthetic electrophile. The
relative stability of the amino enzyme [12] and the 15N kinetic isotope effects [14],
have been discussed in the introduction.

3610
Helvetica Chimica Acta Vol. 86 (2003)
In the paper of Hermes et al. [14], which describes the 15N kinetic isotope effects, the
problem of the Hanson mechanism is formulated as follows: −For now the facility of
enzymes in forming carbanions remains a challenging puzzle×. The Friedel Crafts-type
attack [10] is the adequate solution to this puzzle.
The recent discovery of a MIO-containing tyrosine amino mutase adds a further
member to the MIO-enzyme family [23]. Here the ammonia is not released but
returned to the b-position of the intermediate (E)-4-hydroxycinnamate. Nevertheless,
the lyase products were also detected in small amounts [23]. The mechanism of the
mutase reaction is formulated [23] as proposed by the Karlsruhe group for PAL. It will
be interesting to see which mechanism is used by other tyrosine [24] and phenylalanine
[25] a/b aminomutases.
We consider the results presented here in favor of the mechanism involving the
Friedel Crafts-like attack of MIO at the aromatic portion of the substrates for more
conclusive than the results in older papers [8][14] in favor of the alternative mechanism
[8].
We thank Prof. G. E. Schulz and M. Baedeker (University of Freiburg) for providing us with the new PAL
expression system and also Dr. M. Stieger (Hoffmann-La Roche Ltd., Basel, Switzerland) for vector pREP4-
GroESL carrying the HSP-60 system. The work was supported by the Deutsche Forschungsgemeinschaft and the
Fonds der Chemischen Industrie. L . P.thanks the Hungarian OTKA Foundation (T-033112 and M4-041764) for
financial support.
Experimental Part
General. l-Phenylalanine (1), N-methyl-l-phenylalanine (5), 4-nitro-l-phenylalanine monohydrate (3 ¥
H2O), NaBH4, 36% aq. formaldehyde soln., 40% aq. methylamine soln., and 25% NH4OH soln. were products
of Fluka or Aldrich. All solvents were HPLC grade from Merck. Anal. HPLC: 150 Â 3.9-mm, 5-mm NovaPack-
C18 column (Waters), Waters 600E quaternary gradient pump/Rheodyne 7725i injector/Waters 2996 photodiode-
array detector/Waters Millenium-32 software system or Hewlett-Packard 1050 gradient pump/autosampler/
photodiode-array detector/Hewlett-Packard ChemStation for NT software system; tR in min. Prep. HPLC: 250 Â
21.2-mm, 12-mm Supelcosil-PLC-18 column (Supelco), with the same HPLC systems. Melting points: hot-plate
method; uncorrected. UV/VIS Spectra: Cary-3E spectrometer (Varian), lmax in nm, 1-cm quarz cuvettes. NMR
Spectra: Bruker WM-250 or Bruker DRX-500 spectrometers for 1H and Bruker DRX-500 spectrometer operating
at 125 MHz for 13C and APT experiments; D2O solns., unless otherwise stated; d in ppm, J in Hz. EI-MS and
HR-MS: Finnigan MAT-90 high-resolution instrument; EI (electron ionization) at 70 eV, m/z(rel. intensity).
N-Methyl-4-nitro-l-phenylalanine (6). To an ice-cooled and well stirred soln. of N-methyl-l-phenylalanine
(5; 179 mg, 1 mmol) in conc. sulfuric acid (700 ml, 96 %, d ˆ 1.85), nitric acid (70 ml, 65%, d ˆ 1.40) was added
portionwise, and the soln. was stirred at r.t. for 40 min. The resulting yellow mixture was poured into ice-water
(7 g) and the pH was set to 6.5 with 25% NH4OH soln. After cooling to À58, the forming precipitate was filtered
off and washed with H2O (2 Â 3 ml). Drying in air (overnight) resulted in apparently dry crude product (165 mg)
as light yellow solid. A part (82 mg) of this material was dissolved in 50% aq. formic acid (2 ml) and, after
filtration through a 30-kDnitrocellulose membrane, was purified by prep. HPLC ( Supelcosil PLC-18, isocratic
0.5% HCOOH/H2O and 10% MeOH/H2O, flow 5.0 ml/min, l 280 nm): pure (by HPLC) 6 (32 mg) and a later
eluting crude fraction (45 mg). 6: White solid. M.p. 250 2538 (dec.) ([15]: 1558 (dec.); [16]: 254 2578 (dec.)).
Anal. HPLC (0 ! 10% B in A from 0 to 20 min, A ˆ 0.5% HCOOH/H2O, B ˆ 0.5% HCOOH/MeOH flow
1.0 ml/min, l 275 nm): tR 10.5, >98%. UV (0.5% HCOOH/H2O): 274 ([15]: 270 (8750)). 1H-NMR (500 MHz,
10% CF3COOH/D2O): 2.576 (s, Me); 3.205 (dd, J(a,bRe) ˆ 7.1, J(bSi,bRe) ˆ 14.6, HReÀC(b); 3.293 (dd, J(a,bSi) ˆ
5.8, J(bSi,bRe) ˆ 14.6, HSiÀC(b)); 4.149 (dd, J(a,bSi) ˆ 5.8, J(a,bRe) ˆ 7.1, H ÀC(a)); 7.317 (d, J ˆ 8.7, 2 Ho), 8.017
(d, J ˆ 8.7, 2 Hm). 13C-NMR (125 MHz, 10% CF3COOH/D2O): 31.75 (MeN); 34.40 (C(a)); 61.28 (C(b)); 124.04
(Cm); 130.36 (Co); 141.51 (Cp); 147.22 (Cipso), 169.88 (CˆO). MS (2008): 42(28), 70(4), 80(100), 89(8), 117(2),
‡
133(16), 137(2), 79(27), 224(1). HR-MS: 224.0788 (C10H12N2O4 ; calc. 224.0797).
N,N-Dimethyl-4-nitro-l-phenylalanine (7). To a well stirred soln. of 4-nitro-l-phenylalanine monohydrate
(3 ¥ H2O; 114 mg, 0.5 mmol) and 35% aq. formaldehyde soln. (185 ml, ca. 2 mmol) in 30% MeOH/H2O (2 ml),

Helvetica Chimica Acta Vol. 86 (2003)
3611
NaBH4 (95 mg, 2.5 mmol) was added portionwise, and the soln. was stirred at r.t. for 3 h. After acidification to
pH 2 with conc. HCl soln. and filtration through a 30-kDnitrocellulose membrane, the mixture was separated by
prep. HPLC (Supelcosil PLC-18, 0 ! 50% B in A from 0 to 60 min, A ˆ 0.5% HCOOH/H2O, B ˆ 0.5%
HCOOH/MeOH, flow 5.0 ml/min, l 280 nm): pure (by HPLC) 7 (62 mg, 52%). Slightly brownish solid. M.p.
165 1708 (dec.). Anal. HPLC (0 ! 10% B in A from 0 to 20 min; A ˆ 0.5% HCOOH/H2O, B ˆ 0.5% HCOOH/
MeOH, flow 1.0 ml/min, l 275 nm): tR 13.7, > 98%. UV (0.5% HCOOH/H2O): 274. 1H-NMR (500 MHz, D2O):
2.906 (s, 2 Me); 3.127 (dd, J(a,bRe) ˆ 9.6, J(bSi,bRe) ˆ 13.1, HReÀC(b)); 3.401 (dd, J(a,bSi) ˆ 5.3, J(bSi,bRe) ˆ 13.1,
HSiÀC(b)); 3.816 (dd, J(a,bSi) ˆ 5.3, J(a,bRe) ˆ 9.6, HÀC(a)); 7.439 (d, J ˆ 8.0, 2 Ho); 8.120 (d, J ˆ 8.0, 2 Hm).
13C-NMR (125 MHz, D2O): 34.46 (C(a)); 42.34 (br, MeN); 72.19 (C(b)); 124.33 (Cm); 130.62 (Co); 143.94 (Cp);
147.17 (Cipso); 171.85 (CˆO). MS (1408): 42(9), 56(5), 88(11), 102(100), 147(18), 179(4), 193(29), 238(1). HR-
‡
MS: 238.0950 (C11H14N2O4 ); calc. 238.0954).
Bacterial Strain and Enzyme. Wild-type (wt) PAL was overexpressed in E. coli BL21(DE3) cells. The gene
coding for phenylalanine ammonia lyase from Petroselinum crispum was changed to the codon usage of E. coli
and cloned in vector pT7-7 followed by a transformation in E. coli BL21(DE3) cells containing vector pREP4-
GroESL [17]. The E. coli BL21 (DE3) cells carrying the plasmids with the gene of wt-PAL was cultured, and wt-
PAL was purified as described in detail in [17]. The isolated enzyme was electrophoretically pure as verified by
staining with Coomassie Brillant Blue R250, and, therefore, it was possible to measure the turnover numbers
(kcat) with the molecular mass Mr of 311.313 for the tetrameric wt-PAL. Determination of protein concentration
was carried out according to Warburg and Christian [26][27], Murphy and Kies [28], and Groves et al. [29].
Bovine serum albumin (BSA) was used as reference protein for the measurements.
Enzyme Kinetics Assays. wt-PAL Activity and inhibition measurements were made spectrophotometrically
at 308 in 0.1m Tris ¥ HCl (pH 8.8) in 1-cm quartz cuvettes by modification of the method described in [30].
Substrate Activity Assays. wt-PAL (20 mg) was preincubated in buffer (900 ml) at 308 for 5 min. The reaction
was started by adding 100 ml of a substrate soln., and absorption data (at 290 nm for reactions yielding (E)-
cinnamate (2); at 340 nm for reactions yielding (E)-4-nitrocinnamate (4)) were recorded at intervals of 1 min
for 5 min, or for less active substrates at intervals of 5 min for 30 min. The assays of N-methylated 4-nitro-l-
phenylalanines (6 up to 2 mm; 7 up to 4 mm) with wt-PAL (20 mg) in Tris ¥ HCl (pH 8.8; 1 ml) at 308 showed no
change in extinction over 30 min.
Substrate Activity of N-Methyl-l-phenylalanine (5) with wt-PAL. Formation of (E)-cinnamate (2) in the
reaction of 5 with wt-PAL was determined at 290 nm by recording the extinction at intervals of 1 min for 5 min.
Kinetics constants for 5 with wt-PAL were determined from these data (4 40 mm of 5, 8 data points) according
to Lineweaver and Burk [31] (Fig. 1). For the identification of the product of this reaction, 5 (62 mg, 0.35 mmol)
was incubated with wt-PAL (400 mg) in 0.1m Tris ¥ HCl (pH 8.8) at 378 for 16 h. After concentration of the
resulting mixture to 2 ml, the pH was set to 10 by 5m NaOH, the forming precipitate removed by filtration
through a 30-kDnitrocellulose membrane, and the filtrate separated by prep. HPLC ( Supelcosil PLC-18,
isocratic 0.5% HCOOH/H2O and 10% MeOH/H2O flow 5.0 ml/min, l 280 nm): tR 14.1 for 5 (lmax 253 nm) and
30.7 for 2 (lmax 278 nm)). The solvent was evaporated from the product fraction (29 33 min), and the residue
1
(2.5 mg) was redissolved in D2O and analyzed by H-NMR (500 MHz).
Reaction of (E)-Cinnamate (2) and Methylamine with wt-PAL. To 2 (5 mm) in methylamine solns. (2 and
4.5m, pH set to 10 with conc. HCl soln.; 1 ml, each), wt-PAL (20 mg in 200 ml 50% glycerol soln.) was added at
308, and the reaction was monitored at 290 nm. After 1, 8, and 16 h, the mixtures were also analyzed by anal.
HPLC (10-ml samples, 0 10% B in A from 0 to 20 min, A ˆ 0.5% HCOOH/H2O, B ˆ 0.5% HCOOH/MeOH,
flow 1.0 ml/min, l ˆ 254 and 280 nm). No significant decrease of extinction at 290 nm in the UV or peak
matching with authentic 5 by HPLC (tR 6.7 for 5 (lmax 254 nm)) were observed.
Kinetics Properties of Ammonia at Different pH on Addition of Ammonia to (E)-Cinnamate (2) Catalyzed
by wt-PAL. Consumption of 2 in the reaction of 2 and ammonia with wt-PAL was determined spectrophoto-
metrically at 290 nm by recording the extinction at 1-min intervals for 10 min. Kinetics constants for ammonia
with wt-PAL were determined from these data (0.15 mm of 2, 0.1 6.4M of ammonia, pH 8.8 and 10.0, 7 data
points at each pH) according to Lineweaver and Burk [31] (Fig. 2).
Inhibition Kinetics of N-Methyl-4-nitro-l-phenylalanine (6) and N,N-Dimethyl-4-nitro-l-phenylalanine (7 ¥
H ) with wt-PAL. Apparent inhibition constants (Ki) for N-methylated 4-nitrophenylalanines 6 and 7 were
‡
determined at different inhibitor concentrations (0.25 1 mm of 6 or 7) by varying l-phenylalanine (1)
concentrations (0.5 2 mm; 4 6 data points for each inhibitor concentration): A soln. of 1 (200 ml, 0.5 2 mm
calc. to the final 1000-ml volume) was added to a soln. of the inhibitor 6 or 7 (780 ml in 0.1m Tris ¥ HCl (pH 8.8),
0.25 1 mm calc. to the final 1000ml volume), and the resulting mixture was preincubated at 308 for 2 min. The
reaction was started by addition of wt-PAL (20 ml of a 50% glycerol soln.) and monitored at 290 nm at intervals

3612
Helvetica Chimica Acta Vol. 86 (2003)
of 1 min for 5 min. Inhibition constants were calculated from the Ki (slope ¥ [l-Phe] ¥ Vmax)/Km equation (Ki, [l-
Phe], Km [mm]; Vmax [U/mg]; linear regression in Dixon plot; final Ki taken as an average of the particular Ki
values of the data sets at different [l-Phe] values).
Inhibitor Fit and Optimization within the Active Site of wt-PAL. Conformations of 6 and 7 resembling best
to 1 bound into the wt-PAL model in their zwitterionic state [20] were optimized by PM3 calculations [32], and
the l-phenylalanine (1) in the substrate-binding wt-PAL model [20] was replaced with these conformations of 6
and 7. These constructs were used as starting structures of the inhibitor-binding wt-PAL models. The inhibitor-
binding models of the active site of wt-PAL were then optimized by MM ‡ calculations [32] within 10-ä radii
around C(a) of the inhibitors 6 and 7. The outside sphere between 10 and 25 ä of the whole 25-ä-radii portion
of the wt-PAL model was kept −frozen× during the calculations.
REFERENCES
[1] K. R. Hanson, E. A. Havir, Rec. Adv. Phytochem. 1978, 12, 91.
[2] K. Hahlbrock, D. Scheel, Annu. Rev. Plant Physiol. Plant Mol. Biol. 1989, 40, 347.
[3] I. L. Givot, T. A. Smith, R. H. Abeles, J. Biol. Chem. 1969, 244, 6341.
[4] R. B. Wickner, J. Biol. Chem. 1969, 244, 6550.
¬
[5] T. F. Schwede, J. Retey, G. E. Schulz, Biochemistry 1999, 38, 5355.
¬
[6] B. Langer, M. Langer, J. Retey, in −Advances in Protein Chemistry×, Eds. J. P. Klinman and J. E. Dove,
Academic Press, New York, 2001, Vol. 58.
[7] L. Poppe, Curr. Opin. Chem. Biol. 2001, 5, 512.
[8] K. R. Hanson, E. A. Havir, Arch. Biochem. Biophys. 1970, 141, 1.
¬
[9] M. Langer, A. Pauling, J. Retey, Angew. Chem., Int. Ed. 1995, 34, 1464.
¬
[10] B. Schuster, J. Retey, Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 8433.
[11] D. S. Hodgins, J. Biol. Chem. 1971, 246, 2977.
[12] A. Peterkofsky, J. Biol. Chem. 1962, 237, 787.
[13] V. R. Williams, J. M. Hiroms, Biochim. Biophys. Acta 1967, 139, 214.
[14] J. D. Hermes, P. M. Weiss, W. W. Cleland, Biochemistry 1985, 24, 2959.
[15] J. Pless, W. Bauer, F. Cardinaux, A. Closse, D. Hauser, R. Huguenin, D. Rˆmer, H.-H. Buescher, R. C. Hill,
Helv. Chim. Acta 1979, 62, 398.
[16] J. E. Rivier, G. Jiang, J. Porter, C. A. Hoeger, A. G. Craig, A. Corrigan, W. Vale, C. L. Rivier, J. Med. Chem.
1995, 38, 2649.
[17] M. Baedeker, G. E. Schulz, FEBS Lett. 1999, 457, 57.
[18] S. Yamada, K. Nabe, N. Izuo, K. Nakamichi, I. Chibata, Appl. Environ. Microbiol. 1981, 42, 773; C. T.
Evans, K. Hanna, C. Payne, D. Conrad, M. Misawa, Enzyme Microb. Technol. 1987, 9, 417; M. Yanaka, D.
Ura, A. Takahashi, N. Fukuhara, Jap. Pat. 06-113870, 1994 (Chem. Abstr. 1994, 121, 155941); W. Liu, US
Pat. 5981239, 1999 (Chem. Abstr. 1999, 131, 321632).
¬
¬
Õ ¬
[19] A. Gloge, B. Langer, L. Poppe, J. Retey, Arch. Biochem. Biophys. 1998, 359, 1; A. Gloge, J. Zon, A. Kovari,
¬
L. Poppe, J. Retey, Chem. Eur. J. 2000, 6, 3386.
¬
[20] D. Rˆther, L. Poppe, G. Morlock, S. Viergutz, J. Retey, Eur. J. Biochem. 2002, 268, 6011.
[21] S. Campbell, E. M. Marzluff, M. T. Rodgers, J. L. Beauchamp, M. E. Rempe, K. F. Schwinck, D. L.
Lichtenberger, J. Am. Chem. Soc. 1994, 116, 5251.
[22] R. L. Fuchs, J. F. Kane, J. Bacteriol. 1985, 162, 98.
[23] S. D. Christensen, W. Liu, M. D. Toney, B. Shen, J. Am. Chem. Soc. 2003, 125, 6062.
[24] P. Spiteller, M. Ruth, F. von Nussbaum, W. Steglich, Angew. Chem., Int. Ed. 2000, 39, 2754.
[25] K. D. Walker, H. G. Floss, J. Am. Chem. Soc. 1998, 120, 5333.
[26] O. Warburg, W. Christian, Biochem. Z. 1942, 310, 384.
[27] E. Layne, Methods Enzymol. 1957, 3, 447.
[28] J. B. Murphy, M. W. Kies, Biochim. Biophys. Acta 1960, 45, 382.
[29] W. E. Groves, F. C. Davis, B. H. Sells, Anal. Biochem. 1968, 22, 195.
[30] A. Zimmermann, K. Hahlbrock, Arch. Biochem. Biophys. 1975, 166, 54.
[31] H. Lineweaver, D. Burk, J. Am. Chem. Soc. 1934, 56, 658.
[32] HyperChem 7.0: Hypercube, Inc., 1115 NW 4th Street, Gainesville, FL 32601, USA.
Received August 22, 2003