вход по аккаунту


FriedelЦCrafts-Type Mechanism for the Enzymatic Elimination of Ammonia from Histidine and Phenylalanine.

код для вставкиСкачать
J. Rtey and L. Poppe
Enzyme Mechanisms
Friedel–Crafts-Type Mechanism for the Enzymatic
Elimination of Ammonia from Histidine and
Lszl Poppe and Jnos Rtey*
ammonia-lyases · biocatalysis ·
enzymes · histidine · phenylalanine
Dedicated to Professor George Olah
2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/anie.200461377
Angew. Chem. Int. Ed. 2005, 44, 3668 – 3688
The surprisingly high catalytic activity and selectivity of enzymes stem
from their ability to both accelerate the target reaction and suppress
competitive reaction pathways that may even be dominant in the
absence of enzymes. For example, histidine and phenylalanine
ammonia-lyases (HAL and PAL) trigger the abstraction of the
nonacidic b protons of these amino acids while leaving the much more
acidic ammonium hydrogen atoms untouched. Both ammonia-lyases
have a catalytically important electrophilic group, which was believed
to be dehydroalanine for 30 years but has now been revealed by X-ray
crystallography and UV spectroscopy to be a highly electrophilic 5methylene-3,5-dihydroimidazol-4-one (MIO) group. Experiments
suggest that the reaction is initiated by the electrophilic attack of MIO
on the aromatic ring of the substrate. This incomplete Friedel–Craftstype reaction leads to the activation of a b proton and its stereospecific
abstraction, followed by the elimination of ammonia and regeneration
of the MIO group. The plausibility of such a mechanism is supported
by a synthetic model. The application of the PAL reaction in the
biocatalytic synthesis of enantiomerically pure a-amino b-aryl
propionates from aryl acrylates is also discussed.
1. Introduction
Almost all biochemical processes are catalyzed by
enzymes. Reactions for which nonenzymatic analogues exist
can be accelerated by a factor of up to 1020 by these
biocatalysts. Even more challenging for the biochemist are
enzyme-mediated transformations that would never occur in
the absence of the enzyme, because either the corresponding
substrates are completely stable under most conditions or
alternative reactions are much faster. It has been proposed
that enzymes not only catalyze certain transformations but
also suppress alternative, competing reactions. This activity is
especially important in cases in which highly reactive
intermediates are involved, but is a general property of
enzymes which leads to a high reaction selectivity. For this
ability of enzymes to suppress competing reactions the term
“negative catalysis” has been coined.[1] Some proteins function exclusively as negative catalysts. For example, hemoglobin and myoglobin bind dioxygen reversibly at their heme
iron (Feii) center and prevent its oxidation to Feiii, which
occurs spontaneously in free heme.
To generate highly reactive intermediates, the activation
of relatively stable substrates is required; examples are
reactions that involve radical intermediates. Radical initiators
include adenosylcobalamin[1–3] (coenzyme B12) and S-adenosylmethionine (SAM) in combination with iron–sulfur clusters.[4] Both are protected forms of the 5’-deoxyadenosyl
radical, which can be generated by the corresponding
Another method of substrate activation involves the
action of electrophiles. As the side chains of all proteinogenic
Angew. Chem. Int. Ed. 2005, 44, 3668 – 3688
From the Contents
1. Introduction
2. HAL in Histidine Metabolism
3. PAL in Plant and Fungal
4. Mechanism of the HAL and PAL
5. The Active Sites of HAL and
6. HAL and PAL Inhibitors and
Alternative Substrates
7. Use of the Reverse HAL and PAL
Reactions in Biocatalysis
8. Tyrosine 2,3-aminomutase:
Another MIO Enzyme
9. Conclusions and Outlook
amino acids contain only nucleophilic groups, cofactors or
post-translationally modified side chains are required for
electrophilic catalysis. The role of the electrophilic cofactor
pyridoxal phosphate in amino acid metabolism has been
known for a long time.[5] More recently, a number of posttranslational modifications were discovered that transform
nucleophilic groups into electrophilic groups (Table 1).
Serine, whose OH group is an important nucleophile in
many enzymatic reactions, can be transformed into a pyruvyl
group.[6] The pyruvyl enzymes catalyze similar reactions to
those in which pyridoxal phosphate is used. As they only
occur in certain bacteria, one can speculate that the pyruvyl
group was the evolutionary precursor of pyridoxal phosphate.
Another modification involves the internal tripeptide AlaSer-Gly and leads to the very strong electrophile 5-methylene-3,5-dihydroimidazol-4-one (MIO). The corresponding
enzymes are the topic of this Review. Another recently
discovered modification of serine or cysteine side chains leads
to formylglycine, which occurs in prokaryotic and eukaryotic
aryl sulfatases.[7] Further modifications affect the aromatic
[*] Prof. Dr. J. Rtey
Institut fr Organische Chemie
Universitt Karlsruhe
Fritz-Haber-Weg 6, 76128 Karlsruhe (Germany)
Fax: (+ 49) 721-608-4823
Prof. Dr. L. Poppe
Institute of Organic Chemistry
Research Group for Alkaloid Chemistry
Budapest University of Technology and Economics
1111 Budapest, Gellrt tr 4 (Hungary)
DOI: 10.1002/anie.200461377
2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
J. Rtey and L. Poppe
Table 1: Conversion of nucleophilic amino acid side chains into electrophiles by post-translational modifications.
prosthetic group
histidine decarboxylase,
SAM decarboxylase
or cysteine
histidine ammonia-lyase (HAL),
phenylalanine ammonia-lyase (PAL)
or cysteine
prokaryotic or eukaryotic
aryl sulfatases
amino acids tyrosine and tryptophan, which are oxidized to
quinones. The mechanism of action of quinoenzymes is
attracting a great deal of attention.[8]
Finally, the modification of lysine to pyrrolysine was
discovered recently.[9] It was first proposed that a specific
lysyl-tRNA is modified at the 6-amino group through the
formation of an amide with 4R,5R 4-substituted pyrroline-5carboxylic acid.[10] A more recent report stated that pyrrolysine is formed before the binding to tRNA occurs.[11] The
codon of the resulting pyrrolysyl-tRNA is the stop codon
UAG. Precedence for such a pretranslational modification is
the formation of selenocysteinyl-tRNA from a specific seryltRNA.[12]
2. HAL in Histidine Metabolism
Whereas the catabolism of most amino acids starts with
transamination to the corresponding 2-keto acid, histidine is
degraded in a different way. The first step of histidine
metabolism consists of the elimination of ammonia catalyzed
by histidine ammonia-lyase (HAL; Scheme 1). The product,
(E)-urocanate, is further processed to imidazolonepropionate, a transformation catalyzed by urocanase (urocanate
hydratase, imidazolonepropionate hydrolase).
Each subunit of this enzyme contains a tightly bound
NAD+ molecule, which functions as an electrophile and forms
a covalent adduct with the imidazole ring of the substrate.[13]
Scheme 1. Degradation pathway of histidine in various organisms;
enzymes involved: 1) histidine ammonia-lyase, 2) urocanase, 3) imidazolonepropionate hydrolase, 4) formiminoglutamate hydrolase, 5) Nformylglutamate aminohydrolase, 6) glutamate formiminotransferase.
thf = tetrahydrofolate.
The result is an “umpolung”, which is followed by the
addition of water to form 4-hydroxyimidazolylpropionate.[13]
This compound tautomerizes spontaneously to racemic imidazolonepropionate. The next step is catalyzed by imidazo-
Lszl Poppe graduated as a chemical engineer from Budapest University of Technology
and Economics (BUTE) and completed his
PhD in 1987 with L. Novk on pheromone
synthesis and biocatalysis. He was a research
fellow at the Central Research Institute for
Chemistry of the Hungarian Academy of Sciences before becoming a research professor
at BUTE in 2001. He also spent over a year
(1991–1992) as an Alexander von Humboldt Fellow with J. Rtey at the Universitt
Karlsruhe. His areas of interest are the stereoselective synthesis of biologically active
compounds, biocatalysis, and enzyme mechanisms.
2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Jnos Rtey studied chemistry at the ETH
Zrich and completed his PhD on the stereospecificity of oxidoreductases in 1963 with
V. Prelog. After postdoctoral periods with F.
Lynen (Mnchen) and D. Arigoni (ETH
Zrich) he became a lecturer at the ETH
Zrich. In 1972 he moved as Professor of
Biochemistry to the Universitt Karlsruhe.
His research interests include the mechanisms and stereospecificity of enzyme reactions, synthetic enzyme models, bioorganic
and bioinorganic chemistry, and the investigation of enzyme structures by molecular
biological methods.
Angew. Chem. Int. Ed. 2005, 44, 3668 – 3688
lonepropionate hydrolase, which is enantioselective for the
S substrate. As imidazolonepropionate racemizes spontaneously, the result is a kinetic resolution that leads quantitatively to (S)-formiminoglutamate. The formimino group may
be removed in one or two steps, depending on the organism.
In mammals it is transferred to tetrahydrofolate by glutamate
formiminotransferase. In certain bacteria, such as Bacillus
subtilis,[14] Klebsiella aerogenes,[15] and Salmonella typhimurium,[16] hydrolysis to formamide occurs; in Pseudomonas
spp.,[17] hydrolysis first to formylglutamate then, in a second
step, to formate and glutamate takes place. Histidinaemia, a
rare but usually lethal disease in humans, is caused by a defect
in HAL.[18]
Urocanate is a component of human sweat and protects
the skin from UV radiation. When exposed to UV radiation,
the E form of urocanate is converted into the Z form, which
can initiate an immunosuppressive process.[19] Urocanase
deficiency in the liver is believed to be a cause of mental
retardation.[20] The recently elucidated structure of urocanase
is consistent with the previously proposed mechanism.[21]
The genes of Pseudomonas putida that code for HAL
(hutH) and urocanase (hutU) have been cloned and
sequenced.[22] The HAL sequences from human, rat, and
murine tissues are also described in the literature.[23]
In certain bacteria the structural and regulatory genes that
code for the enzymes of the histidine utilization pathways are
clustered in the hut operon (Figure 1).[24] For example, in
3. PAL in Plant and Fungal Metabolism
Phenylalanine can be degraded in two different ways,
depending on the organism. Whereas in animals and most
bacteria transformation to the corresponding 2-keto acid is
the initial step, in plants,[25] fungi,[26] and at least one
bacterium[27] the initial step is the elimination of ammonia
catalyzed by phenylalanine ammonia-lyase (PAL).[28] The
product of the elimination reaction is (E)-cinnamate, which is
the precursor of a large number of plant metabolites,
including lignin, coumarins, and flavonoids.[25] Lignin is the
major constituent of wood; the flavonoids are the colorful
components of many flowers (Figure 2). By genetic modification it was possible to alter lignin or flavonoid levels and
thus flower color.[29] PAL lies at the branching point between
the primary and secondary metabolism of plants, which makes
it a target for herbicides.[30, 31]
Figure 2. Colors of the plant world: anthocyanins (flavonoid pigments).
Figure 1. Structure and regulation of histidine utilization (hut) operons
in a) P. putida, b) Klebsiella aerogenes and Salmonella typhimurium,
c) Bacillus subtilis. Promotors are marked as shaded boxes; the gene
boxes indicate the direction of transcription.
P. putida the hut operon has six open reading frames and four
regions whose transcription is regulated by three promotors.
The three structural genes, hutU, hutH, and hutI, are
negatively regulated by a promoter–repressor system. The
expression of all three genes is induced by urocanate, not by
histidine. Therefore, a basal expression of HAL is necessary.
Interestingly, transcription by hutF occurs in the opposite
direction to transcription by the other genes. The organization
of the hut operon in other bacteria is different from that in
P. putida.[24] Only P. putida possesses hutF, which codes for
formylglutamate amidohydrolase.[14, 24a,b]
Angew. Chem. Int. Ed. 2005, 44, 3668 – 3688
The metabolism of phenylalanine in plants is shown in
Scheme 2. The biosynthesis of phenylalanine follows the wellknown shikimate pathway.[31a] Hydroxylation at the para
position leads to tyrosine, another essential amino acid.
Rearrangement through the action of a 2,3-aminomutase
affords b-phenylalanine, which is a precursor of taxol, an
important anticancer agent.[32] The hydroxylation of (E)cinnamic acid, the immediate product of the PAL reaction,
can take place at either the ortho or the para position and
leads to coumarins or 4-hydroxybenzoic acid, respectively.
The latter is a building block of ubiquinone.[33, 34]
In some plants, substituted coumaric acids can be
converted into the corresponding coenzyme A esters. Ferulyl-CoA, for example, is transformed by hydroxycinnamoylCoA hydratase/lyase.[34] This mechanistically interesting reaction consists of the addition of water to the Michael system,
followed by a retroaldol cleavage to give vanillin, an
2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
J. Rtey and L. Poppe
4. Mechanism of the HAL and PAL
4.1 Early Mechanistic Proposals for HAL and
PAL: Discovery of a Catalytically Important
Electrophilic Group
Scheme 2. Metabolism of l-phenylalanine.
important flavor substance, which is eventually oxidized to
vanillic acid.
A partially purified PAL from Rhodotorula glutinis (also
known as Rhodosporidium toruloides) is commercially available.[35] It is the most thoroughly investigated fungal PAL
enzyme.[26] Recently, the isolation and properties of PAL from
Streptomyces maritimus were also described.[27] This is the
only bacterial PAL known to date. It is involved in the
biosynthesis of the antibiotic enterocin via (E)-cinnamate and
benzoyl-CoA. The rarity of PAL in bacteria may be explained
by the absence of phenylpropanoids in these species. On the
other hand, (E)-cinnamate may be the precursor of some
specific bacterial products. A similar case is the discovery of
tyrosine ammonia-lyase (TAL) in Rhodobacter capsulatus.[36]
The recombinant enzyme reacts 150 times faster with tyrosine
than with phenylalanine and corresponds to an alternative
pathway to p-coumaryl-CoA. It is involved in the biosynthesis
of the photoactive yellow protein chromophore of this
2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
After the elucidation of the structure of
urocanic acid by Hunter in 1912,[37] its intermediacy in the histidine degradation pathway
was proposed. Cell-free systems from various
organisms were able to transform histidine
into glutamate. Edlbacher named such systems “histidases”.[38] Later the name histidase
was restricted to one of the component
enzymes, which catalyzes the conversion of
histidine into urocanic acid and ammonia.
Tabor and Mehler described the purification
of the enzyme from P. putida in 1954;[17a] the
partial characterization of the enzyme was
described by Peterkofsky.[39]
A mechanism for the HAL reaction was
first proposed in 1967 by Abeles and coworkers,[40] who found that histidase contains a
catalytically essential electrophilic group. Evidence for the presence of this group was the
inhibition of HAL by strong nucleophiles,
such as KCN, CH3NO2, and NaBH4. When the
radioactively labeled inhibitors [14C]CN[41]
and [3H]NaBH4[42] were used, total hydrolysis
of the inhibited protein afforded [14C]aspartate and [3H]alanine, respectively. From these
results it was concluded that the prosthetic
electrophile is dehydroalanine, which behaves
as a Michael acceptor. It was suggested that
the amino group of the substrate reacts with
the dehydroalanine residue, whereby the leaving ability of the now positively charged
ammonium group would be enhanced.[40] In
the following this mechanism is referred to as the E1cB
Although the E1cB mechanism via the EC state (see
Scheme 3) does not explain how the nonacidic b proton of
histidine can be abstracted by an enzymatic base, it has been
accepted by most enzymologists for almost 30 years. Moreover, the same mechanism was also assumed for the PAL
reaction,[28, 43] in which the existence of dehydroalanine had
been indicated by similar experiments to those described for
The E1cB mechanism for the PAL reaction is illustrated in
Scheme 3. In the first step the a ammonium group of phenylalanine must be deprotonated, which enables a nucleophilic
addition to the prosthetic electrophile to occur. In this way a
secondary ammonium group is formed. From this point three
pathways are possible.[43] By the E1cB mechanism, HSi
abstraction would lead to a benzylic carbanion (EC),
whereas by an E1 mechanism cleavage of the NCa bond
would result in an a carbocation (EC+). However, the
occurrence of both intermediates is unlikely because the
Angew. Chem. Int. Ed. 2005, 44, 3668 – 3688
in the presence of radiolabeled
urocanate a substantial portion of
the radioactivity was found in histidine, whereas [15N]ammonia was
not incorporated.[39] This observation indicated the involvement of a
relatively stable amino enzyme
intermediate, which could revert
to histidine in the presence of
seemed to support the formerly
accepted E1cB mechanism, as it
was believed that the amino group
was covalently bound to the enzymatic electrophile.[40, 43] A convincing alternative explanation for
Peterkofskys results is presented
in Section 4.2.
Although according to Hermes
et al.[43] both the concerted (E2) and
the carbonium-ion (E1) mechanisms are ruled out on the basis of
measurements, we also considered
this possibility. A positively
charged ammonium ion is known
to enhance the acidity of the neighboring hydrogen atoms.[44] Such an
effect may be significant in the
enzymatic elimination of ammonia
from aspartate and 3-methylaspartate. However, in these substrates
the b hydrogen atom is activated
additionally by the neighboring
carboxy group. It is therefore not
Scheme 3. Mechanism of the PAL reaction, as proposed by Hanson and Havir (1970)[28] and modisurprising that the corresponding
fied by Hermes et al. (1985).[43] The states EA, EA’, EC , EC+, EP, and E*+P are depicted according
enzymes aspartase and methylasto the proposal of Hermes et al.
partase do not have a prosthetic
electrophile, as revealed by the
determination of their crystal structures.[45–48] Moreover, these two ammonia-lyases belong to the
pKa value for the abstraction of a benzylic proton is more than
40, and the carbocation a to the carboxy group is also too high
enolase superfamily, as concluded on the basis of their
in energy. Although the following steps would be chemically
sequence similarity and structure, which suggests that in a
plausible, both reaction pathways are improbable because of
first step a proton is removed from position 3. Thus, the
the high-energy intermediates involved. As a third alternareaction is not concerted as was originally believed.
tive, a concerted reaction (E2 mechanism) that leads directly
In the last 10 years, no results that support the E1cB
from the EA’ state to the EP state was also considered.[43]
mechanism have been published. Although the E1cB mechanism is favored on the basis of modeling studies in a recent
In a thorough kinetic analysis of the PAL reaction,
publication that describes the first X-ray crystal structure of
Hermes et al. measured kinetic isotope effects for [15N]- and
PAL,[49a] a more recently published X-ray crystal structure of
[3-2H2]phenylalanine and concluded that the mechanism is
not concerted.
higher resolution[49b] supports another mechanism, which is
(For further details of the studies with
isotope-labeled substrates, see Section 6.2). The difficulty in
discussed in Section 4.2.
abstracting the nonacidic b proton is also discussed by
Hermes et al.,[43] but no solution to this chemical challenge
4.2. An Alternative Mechanistic Proposal: Friedel–Crafts Attack
is offered. Nevertheless, the kinetic 15N-isotope effect of
by the Prosthetic Electrophile
1 % is interpreted in favor of the addition of the amino
group to the prosthetic dehydroalanine residue.
A further mechanistically relevant observation was interAs discussed in the previous section, the main problem
preted in the same sense. In 1962, Peterkofsky reported that
with the “old” E1cB mechanism for ammonia-lyase-catalyzed
Angew. Chem. Int. Ed. 2005, 44, 3668 – 3688
2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
J. Rtey and L. Poppe
reactions is that it does not explain the abstraction of the
nonacidic b protons by the enzymatic base. The pKa value for
the abstraction of the benzylic protons is above 40,[50] so that
an extremely strong base would be required to abstract them
under harsh conditions, for example, at high temperature. At
the same time, the much more acidic ammonium hydrogen
atoms should remain untouched, as the positive charge is
required for the ammonium group to retain its leaving ability.
In the so-called Hofmann degradation of the second type, the
trimethylammonium group cannot lose its positive charge
even under harsh basic conditions.
The experimental basis for the postulation of an alternative mechanism emerged in an indirect way. It originated in
the search for the precursor of the prosthetic dehydroalanine
residue. It was known that dehydroalanine can be formed by
the dehydration of serine either in vitro or during the
biosynthesis of the lantibiotics.[51] After some controversy[22a, 52] the question was solved by specific mutations of serine
residues that are conserved in a number of HAL and PAL
sequences.[53] The mutation of Ser 143 in HAL from
P. putida[54] and Ser 202[*] in PAL from parsley (Petroselinum
crispum)[55] to alanine led to a decrease in the activity of the
enzyme by a factor of more than 1000 (Table 2).[55] The
Table 2: Comparison of the kinetic constants of HALs isolated from
Pseudomonas ATCC 11299b and P. putida, and mutant enzymes
expressed in E. coli BL21(DE3).[54]
HAL-derivative source
Km [mm]
Vmax [U mg1]
wild-type Pseudomonas ATCC 11299b
wild-type Pseudomonas putida
recombinant wild-type E. coli
recombinant mutant S112A
recombinant mutant S143A
recombinant mutant S393A
recombinant mutant S418A
mutation of other conserved serine residues had little or no
effect on the activity. This result identified the two serine
residues mentioned as the precursors of the catalytically
essential dehydroalanine residue. Interestingly, the mutation
of the essential serine residue to cysteine led to fully active
wild-type enzyme.[56] It can be concluded that the posttranslational modification can remove either water or SH2,
but in a site-specific manner.[56]
After obtaining these results, we turned to the question of
the role of dehydroalanine in the reaction mechanism. A
literature search helped to solve the problem. Klee et al.[57]
found that 5-nitrohistidine is also a substrate of HAL.
Moreover, whereas b-dideuterated histidine showed a kinetic
isotope effect of 1.5–2.0, no such effect was observed for bdideuterated 5-nitrohistidine.[57] The explanation for this
difference is illustrated in Scheme 4: The presence of the
nitro group leads to a decrease in the electron density of the
imidazole ring and thus to an increase in the acidity of the
[*] This residue is Ser 203 in the sequence of PAL from P. crispum
(P24481) in the SWISS-PROT database.
2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Scheme 4. Activation of the b hydrogen atoms by the nitro group of
5-nitro-l-histidine, a substrate of HAL.
b hydrogen atoms. As a consequence, the abstraction of the
b proton is no longer the rate-determining step, and no kinetic
isotope effect is observed.
This result suggested that the function of the prosthetic
dehydroalanine residue may be similar to that of the nitro
group, namely, the acidification of the b hydrogen atom. To
test this idea, the activity of 5-nitrohistidine was measured
both with wild-type HAL and with HAL derivatives mutated
at Ser 143 or inactivated by treatment with NaBH4.[58] The
reaction rate is practically independent of the HAL derivative
used. In other words, the presence of the nitro group makes
the prosthetic electrophile unnecessary. A further conclusion
is that the nitro group and the prosthetic electrophile do in
fact have the same function, namely, the acidification of the
b hydrogen atom.
This conclusion and the precedent for electrophilic attack
at the imidazole ring in the urocanase reaction[13] led to the
proposal of an alternative mechanism, which is depicted in
Scheme 5.[58] Attack at the imidazole ring of histidine by the
prosthetic electrophile creates a similar situation to the
presence of the nitro substituent and facilitates the abstraction of HRe[59] by the enzymatic base. The ammonium group is
then cleaved as ammonia, and, in a final step, the fragmentation of the adduct gives urocanate and regenerates the
prosthetic group.[58]
Such a mechanism is less plausible for the PAL reaction,
as the phenyl ring is not as electron rich as the imidazole ring.
Through the attack by the prosthetic electrophile, the
aromaticity would be lost at least transiently, which is
unfavorable energetically. This would be the first biological
Friedel–Crafts reaction.[60] However, a number of experiments supported such a mechanism for the PAL reaction.[61]
In analogous experiments to those described for HAL, the
PAL mutants Ser202Ala and Ser202Thr reacted much faster
with 4-nitrophenylalanine than with the unsubstituted substrate (kcat < 20).[61]
PALs from various sources also accept tyrosine as a
substrate, but react with it at a much lower rate (kcat < 50). If
the Friedel–Crafts-type attack plays a role in the mechanism,
then m-tyrosine should be a much better substrate, as is
indeed the case. Recombinant PAL from parsley even reacts
with m-tyrosine 10–20 % faster than with l-phenylalanine.
The proposed mechanism is illustrated in Scheme 6.[61]
More recently, the chemical feasibility of the electrophileassisted elimination of ammonia was demonstrated with a
chemical model.[62] One portion of the model compound
mimicked the essential parts of the substrate (phenylalanine),
and another portion mimicked the electrophilic Michael
acceptor (MIO) in a sterically appropriate position
(Scheme 7).
Angew. Chem. Int. Ed. 2005, 44, 3668 – 3688
Scheme 5. Alternative mechanism for the HAl reaction.[58]
The following measures were taken to facilitate the
reaction: a) The reaction was carried out in an intramolecular
fashion, b) a methoxy group was introduced to enhance the
nucleophilicity of the phenyl ring at the position of attack, and
c) the electrophilicity of the Michael acceptor was enhanced
by using an a,b-unsaturated aldehyde instead of the corresponding amide. Under Friedel–Crafts conditions (BF3·Et2O
or AlCl3) two diastereomeric tricyclic compounds were
isolated in excellent yields, but no trace of the product that
would arise from the elimination of dimethylamine was
detected. Evidently, the competing abstraction of a proton
from the ring, accompanied by rearomatization and completion of the substitution, suppressed the abstraction of the
benzylic proton. PAL prevents the abstraction of a ring
proton by excluding bases from the hydrophobic pocket
which the phenyl group occupies. Therefore, we had to take a
further measure to mimic this “negative catalysis”. When we
substituted the hydrogen atoms on the phenyl ring for methyl
groups (Scheme 7) the completion of the Friedel–Crafts
substitution was prevented, and we could isolate the desired
elimination product.[62]
Evidence for the active role of the phenyl ring in the
process was provided by deactivating the phenyl ring, either
by removing the methoxy group or by exchanging it for a nitro
group. In these cases no elimination products were
Angew. Chem. Int. Ed. 2005, 44, 3668 – 3688
Scheme 6. Mechanism of the PAL reaction with m-tyrosine as the substrate, assuming a Friedel–Crafts-like attack.
Scheme 7. Synthetic model system for modeling the PAL reaction.
4.3. Discovery of 5-Methylene-3,5-dihydroimidazol-4-one (MIO)
as the Prosthetic Electrophile of HAL and PAL
The prerequisite for site-directed mutation in HAL and
PAL was the cloning and heterologous expression of both
enzymes. Success was first achieved with HAL. The expres-
2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
J. Rtey and L. Poppe
sion of its gene from P. putida in E. coli by using the pT7-7
expression vector yielded about 100 mg of the pure enzyme
from 1 L of overnight culture of the recombinant bacteria.
The access to gram quantities of pure HAL facilitated its
crystallization.[58] However, the crystal structure could only be
solved after the mutation of a solvent-exposed cysteine
residue to alanine.[63b]
HAL from P. putida had been characterized previously by
biochemical methods.[52] The X-ray structure at a resolution of
2.1 confirmed that it is a homotetramer and also led to an
unexpected result, namely, that the prosthetic electrophile is
not dehydroalanine but MIO.[64] MIO can be regarded as a
modified dehydroalanine residue and is formed post-translationally by cyclization followed by the elimination of two
water molecules from the inner tripeptide Ala142-Ser143Gly144 (Scheme 8).
Figure 3. X-ray crystal structure of the homotetramer of HAL (a) and
one subunit (b).[64]
Scheme 8. Formation of the MIO group in HAL by post-translational
MIO is the only catalytically important protein-derived
cofactor with such a heterocyclic structure. An example of a
noncatalytic species is the fluorophore of the green fluorescent protein. This fluorophore also has an MIO structure, but
the exocyclic methylene group has a p-hydroxyphenyl substituent.[65]
As catalytically active HAL is also formed in an in vitro
translation system, the information for the biosynthesis of
MIO must be in the amino acid sequence and hence in the
folding of the polypeptide chain.[66] The homotetramer
exhibits D2 symmetry and consists mainly of a helices
(Figure 3). Each subunit can be subdivided into two domains.
The N-terminal domain is globular with eight helices and four
short b strands; the C-terminal domain consists of five long,
nearly parallel a helices surrounded by six other helices.[64]
In each active center (identified by the presence of the
MIO groups) there are amino acid residues from three
subunits (Figure 4). The importance of these active-site amino
acids for catalysis is supported by mutational analysis. The
much higher electrophilicity of MIO relative to dehydroalanine makes the Friedel–Crafts attack at the aromatic ring
2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Figure 4. Substrate-free active site of HAL.
In MIO, the delocalization of the lone pairs of electrons
on the nitrogen atoms into the Michael system is prevented.
In the case of the sp2-hybridized nitrogen atom the orbital of
the lone pair of electrons is orthogonal to the p orbitals of the
a,b-unsaturated carbonyl system, and in the case of the sp3hybridized nitrogen atom the sp3 !sp2 transition is prevented
by the fold of the polypeptide chain. However, the barrier to
this transition is not excessively high.[67] Moreover, the attack
of a nucleophile at the exocyclic double bond renders the
imidazole ring aromatic, which compensates for the transient
loss of aromaticity of the imidazolyl or phenyl group of the
The mechanism of the PAL reaction initiated by a
Friedel–Crafts attack is shown in Scheme 9. An example of
a lone pair of electrons on a nitrogen atom that is prevented
from undergoing delocalization into an adjacent carbonyl
group is provided by quinuclidone (Scheme 10), which does
not behave as a lactam, but as a ketoamine.
Angew. Chem. Int. Ed. 2005, 44, 3668 – 3688
Scheme 9. Mechanism of the PAL reaction involving a Friedel–Craftstype attack of the MIO group on the phenyl moiety of l-phenylalanine.
Figure 5. UV difference spectra of a) wild-type HAL and the mutant
Ser143Ala-HAL, and b) wild-type PAL and the mutant Ser202Ala-PAL.
Concentrations in mg mL1: ~ 1.6, * 0.8, + 0.4, & 0.2; wt = wild-type.
5. The Active Sites of HAL and PAL
5.1 Identification of Active-Site Amino Acids by Site-Directed
Scheme 10. Structure of quinuclidone.
In 1970, Klee reported a shoulder at about 315 nm in the
UV spectrum of HAL, but no explanation for its occurrence
could be offered at that time.[68] The idea emerged that the
cross-conjugated system of MIO could be responsible for the
shoulder. We recorded UV difference spectra of HAL and
PAL mutants that do not contain the MIO group and the
corresponding wild-type enzymes. In both cases a distinct
absorption maximum (lmax) was observed at 308 nm
(Figure 5).[69] This method has also been used by others to
show the presence of MIO.[70]
The very recently published X-ray crystal structures of
PAL from Rhodosporidium toruloides[49a] and parsley[49b]
further confirm the presence of MIO.
Angew. Chem. Int. Ed. 2005, 44, 3668 – 3688
The amino acids in the active site of HAL were identified
by X-ray crystal-structure analysis.[64, 71] In Scheme 11 the
relative positions of these amino acids are shown, together
with the substrate modeled into the active site.[72] All
mutations were carried out on the Cys273Ala mutant, which
had been used for the X-ray crystal-structure analysis. The kcat
value of this mutant was about 5 times lower than that of wildtype HAL, whereas its Km value was 4.5 times higher. The
kinetic constants of several mutants are listed in Table 3.[72]
The importance of Glu 414 was highlighted by the decrease in
the kcat value for the Glu414Ala mutant by a factor of 21 000.
The likely function of this residue is the abstraction of the HRe
proton from the substrate as the enzymatic base. As can be
seen from Scheme 11, Glu 414 is assisted in the abstraction of
HRe by Tyr 280 on a neighboring subunit. On the basis of the
structure of cysteine-inhibited HAL, Baedeker and Schulz
postulated a different mechanism, by which Tyr 280 abstracts
HRe with assistance from Glu 414.[67] There are arguments in
2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
J. Rtey and L. Poppe
Scheme 11. Model of the active site of HAL with the substrate arranged
according to the Friedel–Crafts-like mechanism.[74]
Table 3: Kinetic constants of active-site mutants of C273A HAL.[72]
in C273A HAL
Km [mm]
kcat [s1]
18 3
4.1 0.7
4.4 0.7
1.7 0.9
18.0 4
1.2 0.4
6.1 0.7
18 1
0.79 0.03
0.32 0.01
0.18 0.01
0.14 0.01
0.053 0.0025
0.018 0.001
0.011 0.001
0.0068 0.0004
0.001 0.0002
0.00086 0.00007
18 000
20 930
Table 4: Kinetic constants of active-site mutants of PAL.[74]
[a] kcat,C273A/kcat,wt = 0.21.
favor of both suggestions. The fact that the HAL mutant
Glu144Ala shows a greater loss of activity than Tyr280Phe
relative to the wild-type enzyme would support the former
view,[64, 72] whereas the observation that Gln 488 occupies a
position in PAL analogous to Glu 144 in HAL, but that
Tyr 351 is conserved, speaks for the second view.
A large decrease in kcat was also observed for the mutant
His83Leu. His 83 is most likely involved in the binding of the
imidazole ring of the substrate. The distance (4.15 ) suggests
that the binding occurs through a water molecule or a
hydronium ion. Although in the X-ray crystal structure no
metal ion is observed, it has been found that Zn2+, Mn2+, and
similar metal ions increase the activity of HAL.[73] We suggest
that the species represented with an “X” in Scheme 11 could
be either a hydronium or a metal ion. The guanidino group of
Arg 283 is the counter ion of the carboxylate group of the
histidine substrate and can be substituted for Lys with a
relatively small loss of activity. However, its substitution by
Ile, a neutral amino acid, leads to a large decrease in the
kcat value. An even larger loss of activity is observed for the
mutation Tyr53!Phe. Tyr 53 interacts with the carboxylate
group of the substrate and may play a role in the abstraction
and binding of the a amino group. Tyr 53 may also be
responsible for the stability of the amino enzyme intermedi-
2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ate,[39] possibly with the assistance of Asn 195, whose substitution for Ala causes a 1000-fold decrease in kcat. Somewhat
controversial is the effect of the mutation Phe 329!Ala. One
research group[72] found a 100-fold decrease in activity,
whereas another[67] found a 2500-fold decrease. Phe 329 is
conserved in all members of the HAL/PAL family and in the
second model[67] interacts with the imidazolyl group of the
substrate. It is postulated that the phenyl group of Phe 329 is
required to induce MIO formation by steric compression. The
mutant Phe329Gly leads to an inactive HAL protein which
does not contain MIO,[71] whereas the mutant Phe329Ala
formed the MIO group but showed very low activity.
To elucidate the role of the active-site amino acids of PAL
it was necessary to resort to its high sequence homology with
HAL, which allowed modeling of the active site of PAL (see
also Section 5.2). On the basis of this model and the kinetic
constants of a number of PAL mutants (Table 4), functions of
the active-site residues could be proposed.[74] All except two
amino acids are found to occupy almost identical positions.
His 83 and Glu 414 in HAL are exchanged for Leu 138 and
Gln 488, respectively, in PAL.
In an attempt to transform PAL into HAL, the PAL
mutants Leu138His and Gln488Glu were prepared, as well as
the double mutant Leu138His/Gln488Glu.[74, 75] Whereas for
the single mutants the kcat values decreased only slightly, the
double mutant was 145 times less active than wild-type PAL
(Table 4). More dramatic was the change in the Km values for
in wt-PAL
Km [mm]
kcat [s1]
0.12 0.004
0.057 0.006
13.5 0.6
0.057 0.003
55 4.9
0.024 0.004
0.027 0.005
0.019 0.001
0.033 0.002
0.03 0.01
0.033 0.003
13.5 0.1
2.1 0.04
0.99 0.02
0.104 0.005
0.093 0.004
0.057 0.001
0.039 0.001
0.031 0.0001
0.022 0.002
0.0057 0.0004
0.005 0.001
75 000
the Leu138His mutants. These values are more than 100 times
higher than that for the wild-type enzyme, thus supporting the
view that Leu 138 is part of the hydrophobic binding pocket
that harbors the phenyl group of the substrate. Because of the
relatively strong adhesion of the phenyl group to this hydrophobic environment, the rate-limiting step in the PAL
reaction is the release of the product, (E)-cinnamate. Therefore, in contrast to the HAL reaction, no primary kinetic
deuterium-isotope effect is observed in the PAL reaction.[43]
Recently, it was shown that with the mutant Leu138His the
release of the product is faster and a kinetic deuteriumisotope effect (kH/kD = 2.3) is observed.[75] The assumption
that the PAL double mutant Leu138His/Gln488Glu would
Angew. Chem. Int. Ed. 2005, 44, 3668 – 3688
behave like HAL was only partially endorsed. Whereas the
Km value for histidine was similar to that with HAL, the
kcat value was 8000 times lower. It seems that amino acid
residues other than those in the active site also affect the
reaction rate.
The largest loss of activity was found with the PAL mutant
Tyr110Phe. Tyr 110 in PAL is thought to have the same
function as Tyr 53 in HAL, but it seems to be even more
important. Similarly, Asn 260 and Tyr 348 in PAL correspond
to Asn 195 and Tyr 280 in HAL and probably have the same
important functions. In this case the phenolate form of Tyr 348
is definitely the enzymatic base that abstracts the b HSi proton
of the substrate.[76] Arg 354 and Phe 400 in PAL correspond to
Arg 283 and Phe 329 in HAL, and their mutations affect the
kinetic constants of the two enzymes to a similar extent.[74]
Another interesting difference between HAL and PAL is
found with the mutants Ser143Thr and Ser203Thr. Although
practically inactive, the HAL mutant Ser143Thr shows a
maximum at 305 nm in the UV difference spectrum versus the
MIO-free mutant Ser143Ala, whereas no such absorption
maximum is observed in a similar experiment with the
Ser203Thr mutant of PAL. These results indicate that the
HAL mutant Ser143Thr still forms the MIO structure but that
the corresponding PAL mutant does not. The exchange of the
two other amino acid residues directly involved in MIO
formation to give HAL mutants Ala142Ser and Gly144Val led
to decreased Vmax values (by a factor of 6 and 44, respectively), but these effects are less dramatic.[77] Other mutations
at the same positions (Ala142Gly, Ala142Asp, and
Gly144Ala) resulted in moderate loss of activity. No activity
at all was reported for the HAL mutant Asp145Ala.[67]
5.2. Modeling the Active Sites of HAL and PAL
Molecular modeling was applied as a complementary tool
for the interpretation of the enzyme kinetic data from
reactions with HAL and PAL mutants modified at activesite residues. Thus, a histidine molecule was modeled into the
X-ray crystal structure of HAL[64] with its amino group at
binding distance from the exocyclic methylene carbon atom
of MIO and its carboxylate near the guanidino group of
Arg 283’.[67] Inspection of this arrangement revealed that there
is no strong base near the b hydrogen atom of the substrate
and no specific binding site for the imidazole ring of histidine,
thus suggesting that direct ammonia abstraction is not
compatible with the experimental structure.
However, the experimental HAL structure[64] was found
to be compatible with a Friedel–Crafts attack of the MIO
group at the aromatic ring of the substrate.[58, 61, 64] Two slightly
different substrate-binding models have been described for
HAL.[64, 71, 72] In the first model,[64, 72] the imidazole ring of
histidine is positioned close to Leu 146 and thus “below” the
MIO group. In the second, later, proposal[67] the imidazole
ring is positioned “above” MIO and interacts with Phe 329.
Both models are consistent with the functions of the amino
acid residues at the active site as concluded from mutational
experiments, and both rationalize the experimental finding
that urocanate is released before ammonia.[39]
Angew. Chem. Int. Ed. 2005, 44, 3668 – 3688
The electronic spectra of a truncated MIO model were
calculated at the PM3 level of theory and used to estimate the
degree of polarization of the MIO moiety in the substrate-free
state of HAL (Scheme 12).[72] The calculated absorption
Scheme 12. A model compound for nonprotonated MIO and its singly
protonated forms; the calculated UV absorption maxima are given.
maxima at l = 303 and 307 nm for the nonprotonated MIO
model were in good agreement with the experimentally
determined maximum at about l = 308 nm in the UV difference spectrum of HAL. Calculated UV spectra for MIO
structures fully protonated at the carbonyl oxygen atom, at
N1, or at N3 were significantly different, thus indicating that
MIO is not polarized significantly in substrate-free HAL.
To investigate the functions of the active-site residues of
PAL, the X-ray structure of HAL was used as a template for
homology modeling of the PAL structure (Figure 6).[74] In
agreement with biochemical data, PAL was modeled as
homotetramer (Figure 6, right). The model revealed that
catalytically important residues are located at highly isosteric
positions within two distinct regions of the protein chains in
both HAL and PAL.
In the region of the active site, the PAL model closely
resembles the X-ray crystal structure of HAL (Figure 7). All
the active-site residues in the PAL model occupy positions
that would be expected on the basis of comparison with the
HAL sequence and structure. Moreover, the model constructed for the PAL from Petroselinum crispum[86] (Pc-PAL;
Figure 6, right) shows excellent similarity with the recently
published X-ray crystal structures of the PAL from Rhodosporidium toruloides[49a] (Rt-PAL; Figure 6, middle column;
PDB codes for Rt-PAL: 1T6J and 1T6P) and Pc-PAL.[49b]
Interestingly, the residues His 137 and Gln 138 at the
active site of Rt-PAL are replaced by Phe 140 and Leu 141 in
the Pc-PAL model. This difference stems from differences in
the sequences of the two PALs in this region rather than
errors in the modeling of Pc-PAL. Moreover, the loop 105–
123, which contains the catalytically important Tyr 110
residue, is not present in the Rt-PAL structure.[49a] The
absence of this loop can lead to the ambiguous arrangement
of the substrate when modeling within this structure.
2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
J. Rtey and L. Poppe
Figure 6. Comparison of the X-ray crystal structures of HAL (from
Pseudomonas putida)[64] and PAL (from Rhodosporidium toruloides)[49a]
with the homology model of PAL (from Petroselinum crispum):[86] a) tetrameric structures; b) single subunits (the catalytically important residues are shown as stick models; moieties containing MIO prosthetic
groups are colored green).
Figure 8. Models of the cationic intermediate (a) and product-binding
states (b) at the PAL active site according to the Friedel–Crafts-like
plus ammonia (Figure 8 b) at the active site of Pc-PAL
provided more information about the catalytic functions of
the amino acid residues (Scheme 13).
A common feature of the tetrameric HAL and PAL
structures is that a tyrosine residue (Tyr 53 in HAL and
Figure 7. Model of the substrate-free active site of PAL (from
P. crispum)[86] superimposed on the active site of HAL (from
P. putida)[64] The lighter colors (yellow labels; light-blue, light-green,
and light-red ribbons) indicate the PAL model; the darker colors
(orange labels, blue, green, and red ribbons) indicate the HAL structure.
The modeling of a s-complex-like intermediate (Figure 8 a) and the product-binding state with (E)-cinnamate
2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Scheme 13. Proposed catalytic roles of the active-site residues on the
basis of the model of PAL.[86]
Angew. Chem. Int. Ed. 2005, 44, 3668 – 3688
Tyr 110 in PAL) is located at the edge of the channel through
which the substrate can enter or the product be released from
the active site. Mutations of Tyr to Phe at these positions lead
to a dramatic decrease in catalytic activity (Scheme 14).[74]
The existence of this narrow channel may account for the
experimental finding[39, 74] that after the elimination the
product (E)-cinnamate is released first (Scheme 14 e), followed by the ammonia in the form of an ammonium ion
(Scheme 14 f).
thiol groups of the two inhibitors are positioned at the correct
distance to attack MIO (relative to the 5-position of the
imidazole ring of histidine).
To isolate the chromophore that absorbs at l = 335–
340 nm, the inhibited HAL protein was denatured and
partially digested with pronases. After separation of the
peptides produced, those that contained the chromophore
were identified and characterized spectroscopically. Although
the oligopeptides described in the two reports differed in
length, they both contained the same chromophore structure
(1 and 2 in Scheme 15).[79] Viewed superficially, this finding
Scheme 15. Chromophore groups in the isolated oligopeptides after
treatment of the HAL protein with l-cysteine or after reaction with llysine followed by proteolytic digestion.
Scheme 14. Proposed mode of substrate binding (a–c), elimination
(d), and release of the product (e, f) in HAL and PAL, assuming a
Friedel–Crafts-like mechanism.
6. HAL and PAL Inhibitors and Alternative
6.1. Inhibition of HAL by Cysteine
Among the inhibitors of HAL, l-cysteine is the most
interesting, because its behavior provided clues as to the
reaction mechanism.[58] Inhibition by cysteine is reversible at
neutral pH and under anaerobic conditions, but irreversible
above pH 10.5 and in the presence of oxygen. An absorption
occurs at about l = 340 nm, the intensity of which is proportional to the degree of inhibition.[68b]
The first proposals for the structure of the chromophore
generated had been made before MIO was discovered as the
prosthetic group of HAL.[78] Soon after the discovery of the
MIO group[64] two research groups described the structure
and proposed mechanisms for its formation.[79] In both cases it
was assumed that the thiolate group of cysteine reacts as a
nucleophile with MIO, which is consistent with our previous
proposal.[58, 78b] The follwing arguments favor the attack of the
thiolate group: 1) l-Cysteine and l-homocysteine but no
other nonaromatic amino acids act as inhibitors, and 2) the
Angew. Chem. Int. Ed. 2005, 44, 3668 – 3688
supports the previously favored mechanism, in which the
a amino group of histidine is the nucleophile that reacts with
MIO.[28, 41–43] However, both research groups believe that the
isolated chromophore is the product of a rearrangement, as
illustrated in Scheme 16.[79] The initial addition of the thiolate
group of cysteine to MIO affords a highly nucleophilic
enolate, which, like in the photorespiration by the enzyme
ribulose-1,5-bisphosphate carboxylase (rubisco),[80] reacts
with dioxygen to give a peroxide anion. A vinylogous
thioester formed in an electrocyclic reaction coupled with a
1,3-H shift from a C atom to an O atom then undergoes
intramolecular aminolysis followed by disulfide formation
with a further cysteine molecule.
The postulated rearrangement was supported by the
observation of other proteolysis products. Merkel isolated
the adduct from lysine and MIO as the main product (3 in
Scheme 15).[79a,c] Evidently the initially formed vinylogous
thioester had undergone an intermolecular aminolysis with
the e amino group of lysine molecules present in the lysate as
the products of extensive proteolysis. This explanation is
consistent with the observation that the inhibited but intact
protein shows an absorption maximum at l = 338 nm,
whereas for the isolated cysteine adduct lmax = 335 nm was
found, and for the lysine adduct lmax = 332 nm.[79a,c]
The isolated pure peptide containing the chromophore
that absorbs at l = 335 nm was slightly unstable to chromatography on a reversed-phase column under acidic conditions,
as manifested by a smaller extra peak in the elution diagram.
This degradation product was also isolated and characterized
spectroscopically. From the spectral data (UV, NMR, MS) its
2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
J. Rtey and L. Poppe
Scheme 16. Proposed mode of addition of l-cysteine to MIO followed by oxidation and rearrangement.
chromophore was identified as an MIO group with a hydroxy
substituent on the exocyclic methylene group. The compound
forms two tautomers and is reductone-like, like vitamin C. Its
absorption maximum occurs at 310 nm, very close to that of
MIO (lmax = 308 nm). A possible mechanism for the formation of this degradation product involves a further rearrangement followed by ester hydrolysis (Scheme 17).[79a,c]
When l-[35S]cysteine was used as an inhibitor of HAL, it
was incorporated into the peptide containing the chromophore that absorbs at l = 335 nm.[79a,c] This finding seems to
contradict previous results.[68b, 78a] A plausible explanation for
this deviation could be that in the earlier studies the initially
formed chromophore had already been converted into
products devoid of cysteine. Recently, the X-ray crystal
structure of the cysteine-inhibited HAL was determined, but
the cysteine moiety was not observed despite very high
resolution.[67] As the crystals had been prepared several
months before they were submitted to X-ray crystallographic
examination, rearrangement of the chromophore might have
occurred, so that it was no longer possible to localize the
cysteine unit.
6.2. The Behavior of Substrate Analogues and Kinetic Isotope
Effects in the PAL Reaction
In recent years a number of research groups have
examined the above issues and interpreted the results in
2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Scheme 17. Degradation of the chromophore 1 upon hydrolysis.
Angew. Chem. Int. Ed. 2005, 44, 3668 – 3688
favor of the Friedel–Crafts-type mechanism.[81–83, 94] However,
before the new mechanism was proposed some relevant
publications had appeared,[43, 84] in which 2,5-dihydrophenylalanine (4, Scheme 18) was identified as a moderately good
Scheme 18. Partially hydrogenated phenylalanine derivatives for the
kinetic investigation of the PAL reaction.
substrate of PAL. In contrast to phenylalanine, the 3,3’dideutero derivative 5 exhibited a primary kinetic isotope
effect of about 2, thus indicating that the deprotonation is
partially rate limiting.[43] It seems that the activation of the
b HSi hydrogen atom is less efficient in this substrate
analogue, but still possible through the electrophilic attack
of MIO at the double bond next to the side chain. No
activation of the b hydrogen atoms is expected, however, in
1,4-dihydrophenylalanine (6). It was discovered that the Birch
reduction of l-phenylalanine affords, beside 2,5-dihydrophenylalanine as the main product, some 1,4-dihydro-l-phenylalanine.[81] As expected, this compound was not a substrate for
PAL but showed weak inhibitory properties.
On the assumption that the most difficult step in the PAL
reaction is the Friedel–Crafts-type attack with concomitant
dearomatization of the phenyl ring, it was anticipated that the
substitution of the phenyl group for the nonaromatic cyclooctatetraenyl group could facilitate the reaction. Racemic
cyclooctatetraenylalanine, however, showed a very low Vmax
value (0.6 % of the value for l-phenylalanine).[85] A possible
explanation for the low reactivity of this compound is that the
different size and nonplanar geometry of the cyclooctatetraene ring may make the electrophilic attack by MIO more
Another interesting aspect is the synergistic inhibition of
PAL by various phenols with glycine. Although glycine alone
does not inhibit PAL at all and the phenols are only very weak
inhibitors, together in 1:1 ratio they act as strong inhibitors.[83, 94] For example, the combination m-cresol/glycine
could be seen to correspond to m-tyrosine, which is an
excellent substrate of PAL. m-Cresol and glycine together
inhibit PAL competitively with a Ki value of 0.89 mm, which is
more than twenty times lower than that observed for glycine
or m-cresol alone.[94] Alunni et al.[83] analyzed this synergistic
inhibition systematically and found that the pair phenol/
glycine showed the strongest inhibition (Ki = 0.014 mm),
whereas the pair p-cresol/glycine hardly inhibited PAL at
Angew. Chem. Int. Ed. 2005, 44, 3668 – 3688
all. Except for the pair m-cresol/glycine, all other pairs
showed mixed inhibition.
The kinetic isotope effect was also studied with phenylalanine labeled with deuterium or tritium on the phenyl ring.
Gloge et al. found a conventional secondary kinetic isotope
effect of 1.09 0.01 on the overall reaction rate with
[2H5]phenylalanine.[94] Since a deuterated carbon atom
changes its hybridization from sp2 to sp3 in the first putative
chemical step, an inverse secondary isotope effect should be
expected. However, this effect may be overcompensated by
the influence of the four other deuterium atoms on the phenyl
Lewandowicz et al. confirmed the normal secondary
kinetic isotope effect with [2H5]phenylalanine through competitive methods, i.e. determination of the enrichment of the
deuterated material in the unconverted substrate.[82] The
obtained kH/kD value of 1.13 0.02 was similar to that
described by Gloge et al. (1.09 0.01).[94] Lewandowicz
et al. also measured secondary kinetic tritium-isotope effects
with phenylalanine labeled with tritium at the ortho position
of the phenyl group.[82] The isotope effect showed a strong
dependence on the reaction progress. For the first 5 %
conversion the kH/kT value was 0.85; it then reached the
normal value and was about 1.15 at 20 % conversion. The
authors interpreted these results in favor of the Friedel–
Crafts-type mechanism. Simple semiempirical calculations
supported this interpretation. It was concluded that with
[2H5]phenylalanine both the Hanson/Havir and the Friedel–
Crafts-type mechanism should give a conventional secondary
kinetic deuterium-isotope effect, but that the kH/kT value
should be higher for the latter.
6.3. N-Methylated l-Phenylalanines in the PAL Reaction
N-Methyl-l-phenylalanine, N-methyl-4-nitro-l-phenylalanine, and N,N-dimethyl-4-nitro-l-phenylalanine were
investigated as substrates or inhibitors of PAL from Petroselinum crispum.[86] Although N-methyl-l-phenylalanine was a
moderate substrate (Km = 6.6 mm, kcat = 0.22 s1), no reverse
reaction was observed with methylamine and (E)-cinnamate.
The Km value for ammonia in the reverse reaction with (E)cinnamate was determined to be 4.4 m at pH 8.8 and 2.6 m at
pH 10. N-Methyl-4-nitro-l-phenylalanine and N,N-dimethyl4-nitro-l-phenylalanine did not act as substrates but instead
showed strong inhibitory effects (Ki = 130 and 8 nm, respectively).[86] This finding can be better interpreted in terms of
the Friedel–Crafts-type mechanism than with the Hanson/
Havir mechanism.[28]
Molecular modeling of N-methyl- and N,N-dimethyl-4nitro-l-phenylalanine in the orientation that corresponds to
the Friedel–Crafts-type mechanism[86] into the active site of
PAL[74] demonstrated that the nitrated aromatic rings fit well
into the apolar binding pocket. On the other hand, the
increased gas-phase proton affinities of N-methyl- and N,Ndimethyl-l-phenylalanine relative to that of l-phenylalanine[87] indicate that the N-methylamino and N,N-dimethylamino moieties might be more strongly bound than the nonmethylated groups. Moreover, if the 4-nitro moiety on the
2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
J. Rtey and L. Poppe
aromatic ring hinders the addition to MIO, the N-methylated
group will remain anchored to Tyr 100, thus keeping the
b HSi hydrogen atom[76] far from the enzymatic base Tyr 348.
These effects may together account for the strong inhibition.
In contrast, if one considers that 4-nitro-l-phenylalanine
reacts with both wild-type PAL and the MIO-free mutant
Ser202Ala,[61] and that N-methyl-l-phenylalanine is a moderate substrate for wild-type PAL,[86] a faster reaction would be
expected by the E1cB route for 4-nitro-N-methyl-l-phenylalanine. An electron-withdrawing group at 4-position of the
phenyl ring would facilitate the deprotonation from the EA’
state (Scheme 3) and also stabilize the developing carbanion
in the EC state.
7. Use of the Reverse HAL
and PAL Reactions in Biocatalysis
Table 5: Synthesis of enantiomerically pure l-phenylalanine analogues by PAL catalysis.
The synthesis of enantiomerically pure natural amino acids and
their non-natural derivatives and
analogues is an important goal in
synthetic chemistry. For example,
the pharmacophores of protease
inhibitors, an extremely important
class of pharmaceuticals against
HIV, influenza, and human cytomegalovirus, have a phenylalanine2
like architecture.
As the ammonia-lyase reactions proceed by the stereodestruc3
[94, 95]
tive elimination of ammonia from
l-amino acids, the d-amino acids
could in principal be produced by
the enantioselective removal of the
corresponding l-amino acids from
the racemate. However, the stereoconstructive nature of the reverse
reaction (the enantioselective
addition of ammonia to a,b-unsa7
[85, 95]
turated acids) made this approach
more attractive for the preparation
of amino acids by biotransforma8
[85, 95]
The reversibility of the HAL
reaction was recognized about four
decades ago.[89] Under extreme
reaction conditions (4 m NH4OH,
pH 10) HAL can catalyze the reverse (amination) reaction in
vitro.[90] The synthetic utility of HAL is quite restricted,
because it reacts with only a limited number of analogues,
such as 5-nitrohistidine and 5- or 2-fluorohistidines.[57, 91]
However, it was recently reported that HAL can be used
for the production of a wider selection of aromatic l-amino
acids from the corresponding acrylic acid derivatives.[92]
More than 20 years ago it was recognized that PAL also
catalyzes the enantioselective addition of ammonia to aryl
acrylic acids when the ammonia concentration is increased to
5 m.[93] Furthermore, l-phenylalanine ammonia-lyase tolerates a broader, more structurally diverse range of substrates
than HAL, while strictly maintaining the enantioselectivity of
the ammonia addition. This broader substrate tolerance—in
addition to the demand for structures derived from phenylalanine in protease inhibitors—makes PAL a valuable tool for
the preparation of enantiomerically pure amino acids from
achiral acrylate precursors. Methods for PAL-catalyzed
biotransformations have been elaborated for the enantioselective synthesis of all pyridinylalanine isomers,[94] 5-pyrimidinylalanine and several fluoro- and chlorophenylalanines,[85, 95] and other substituted aryl alanines[85] (Table 5).
Various aryl alanines were prepared in enantiomerically pure
form (> 99 % ee) by PAL catalysis in moderate to almost
quantitative yields.
2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
[85, 95]
[85, 95]
[85, 95]
With the whole cells of Rhodotorula graminis as the PAL
source, no reaction was observed when 4-chloro-, 3,4dichloro-, 2,4,6-trichloro-, 4-bromo-, 4-methyl-, 4-formyl-, 4amino-, 4-nitro-, or 2-hydroxyphenylacrylate, naphth-2-ylacrylate, or furan-3-ylacrylate were tested as substrates.[95]
However, when the differences between Pc-PAL and RtPAL in the active site region are considered (see Section 5.2),
it seems reasonable that there may be slight differences in the
substrate tolerance of yeast and plant PAL enzymes.
Angew. Chem. Int. Ed. 2005, 44, 3668 – 3688
More recently, a PAL enzyme of plant origin was found to
be a useful reagent for the asymmetric synthesis of nonnatural amino acids.[96] PAL can also be catalytically active in
organic solvents, such as n-octanol.[97]
Enzymes can not only be used for synthetic processes as
single biocatalysts. Combinatorial biosynthetic approaches
can also be designed for the synthesis of natural products and
for the construction of libraries of non-natural analogues of
natural products. The usefulness of this strategy was demonstrated with synthesis of the chalcones naringenin and
pinocembrin from l-tyrosine and l-phenylalanine, respectively, by recombinant E. coli with an artificial gene cluster
assembled from the PAL gene of the yeast Rhodotorula rubra
and the genes that code for chalcone isomerase and chalcone
synthase of other organisms.[98]
Michael addition to a a,b-unsaturated acid derivative, and
therefore the easier part of the overall reaction.[70] Interestingly, the mutase also exhibits b-tyrosine racemase activity,
but a-tyrosine is not racemized. Thus, the readdition of
ammonia to the Michael system is reversible and not
enantiospecific. A mechanism for the l-tyrosine 2,3-aminomutase reaction is illustrated in Scheme 19.
8. Tyrosine 2,3-aminomutase: Another MIO Enzyme
A number of ammonia-lyases[88] and aminomutases use
different cofactors and mechanisms. Ethanolamine ammonialyase, for example, is dependent on coenzyme B12, as is blysine 5,6-aminomutase.[3] Both enzymes have radical intermediates. a-Lysine 2,3-aminomutase is a radical SAM (Sadenosyl-l-methionine) enzyme;[4] aspartate and methylaspartate ammonia-lyases do not require a cofactor.
Until recently HAL and PAL were the only enzymes
known to contain the MIO group. Christenson et al. discovered another enzyme, tyrosine 2,3-aminomutase,[70] which
contains the MIO group and uses a similar mechanism to that
of HAL and PAL. The enzyme is produced by Streptomyces
globisporus, and its product, (S)-b-tyrosine, is believed to be a
precursor of the antibiotic C-1027, which shows potent
anticancer and antimicrobial activity.[99] One portion of C1027 is (S)-3-chloro-4,5-dihydroxy-b-phenylalanine, which is
most likely formed from (S)-b-tyrosine, as 3,4-dihydroxyphenylalanine and 3-chlorophenylalanine are poor substrates of
the isolated mutase.[70] The gene that codes for the mutase was
isolated from S. globisporus, overexpressed in E. coli, and
shown to have a high sequence homology with the HAL/PAL
enzyme family.
Evidence for the presence of an MIO group in tyrosine
2,3-aminomutase has been provided in several ways: 1) The
mutation of serine to alanine in the inner-peptide sequence
Ala152Ser153Gly154 reduces the kcat/Km value by a factor of
640; 2) the UV difference spectrum between the wild-type
enzyme and its Ser153Ala mutant shows an absorption
maximum at ~ 310 nm; 3) nucleophiles such as NaBH4 or
KCN inhibit the mutase, but preincubation with l-tyrosine or
4-hydroxycinnamate protect the enzyme from inhibition;
4) the optimum pH value for the mutase reaction is about 8.8,
similar to those for the HAL and PAL reactions.
Further mechanistic studies revealed that 4-hydroxycinnamate is an intermediate in the mutase reaction and that it is
released slowly as a by-product of the reaction. In other
words, tyrosine 2,3-aminomutase also displays ammonia-lyase
activity. Normally, the mutase does not release the ammonia
formed at the active site, but adds it to the intermediate 4hydroxycinnamate at the b position. This second step is a
Angew. Chem. Int. Ed. 2005, 44, 3668 – 3688
Scheme 19. Proposed mechanism of action of l-tyrosine 2,3-aminomutase.[70]
The question arises as to whether there are still undiscovered MIO enzymes. Several mechanistic devices exist for
the activation of the inert b position of carboxylic acids.
Highly reactive radical species, such as 5’-deoxyadenosyl and
glycyl radicals, can abstract protons from inert positions to
form an activated substrate radical which can react further
under enzymatic control.[1] Electrophilic activation is an
alternative in all cases in which there is an aromatic ring at
the adjacent position. There are a few examples for which it
has still not been established definitively whether a radical or
an electrophilic activation mechanism operates.
2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
J. Rtey and L. Poppe
9. Conclusions and Outlook
Herein we have reviewed the elucidation of the mechanism by which the chemically difficult elimination of ammonia from histidine and phenylalanine is catalyzed by the
corresponding ammonia-lyases HAL and PAL. These
enzymes use the recently discovered superelectrophilic prosthetic group 5-methylene-3,5-dihydroimidazol-4-one to activate the nonacidic b hydrogen atoms of their substrates by a
Friedel–Crafts-type attack at the aromatic ring. In the scomplex generated, proton abstraction from the ring is
prevented by excluding any bases in the binding pocket of
the enzyme. Instead, the exocyclic proton is abstracted by an
appropriately positioned enzymatic base. The exocyclic
double bond formed in this way is the prerequisite for a
concomitant process that involves the elimination of ammonia, rearomatization, and fragmentation. The MIO group is
thereby regenerated, and the product, (E)-urocanate or (E)cinnamate, is formed.
The above mechanism is supported by the results of
several biochemical studies, as well as by the X-ray crystal
structures of HAL and PAL, and modeling studies. Before
these discoveries it was believed for thirty years that the
prosthetic electrophile was dehydroalanine and that the aNH2 group of the substrate added to this Michael acceptor.
This “old” mechanism was unsatisfactory for several reasons.
It did not explain how the nonacidic b proton (pKa > 40)
could be abstracted by an enzymatic base, or why most amino
acids do not inhibit HAL, but l-cysteine and l-homocysteine
do. Finally, dehydroalanine itself is not a good Michael
receptor, because the delocalization of the lone pair of
electrons on the nitrogen atom decreases its electrophilicity.
Since the publication of the “new” mechanism and in
particular of the X-ray crystal structure of HAL, a number of
research groups have reported results that support the
Friedel–Crafts-type mechanism, and no arguments against it
have been published.
Recently, a new bacterial MIO enzyme with tyrosine 2,3amino-mutase activity was discovered. We look forward to
the discovery of further MIO enzymes.
We thank Dr. Csaba Paizs for his help with the preparation of
the manuscript. J.R. thanks his former students for their
enthusiastic involvement in the MIO project: Dr. Andreas
Gloge, Dr. Birgid Langer (formerly Schuster), Dr. Martin
Langer, Dr. Dietrich Merkel, Dr. Gaby Morlock, Dr. Andrea
Pauling, Dr. Gunhild Reck, Dr. Dagmar Rther, Dr. Alexander
Skolaut, Dr. Sandra Viergutz, and Dr. Karl-Heinz Weber. J.R.
also thanks Professor Georg E. Schulz, Dr. Torsten F.
Schwede, and Dr. Mathias Baedeker for the fruitful collaboration on the X-ray crystal structure of HAL and for a plasmid
of the modified PAL gene.
We are grateful to Professor Nicholas Amrhein and Professor
Klaus Hahlbrock, who gave us the antibody against PAL and
the PAL gene from parsley, respectively. The excellent
technical assistance of Ingrid Merkler and Stefanie Vollmer is
also greatfully acknowledged.
The research at Karlsruhe was supported by the Deutsche
Forschungsgemeinschaft, the Fonds der Chemischen Industrie,
2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
and the Land Baden-Wrttemberg. L.P. thanks the Alexander
von Humboldt Foundation for a fellowship and OTKA (T048854) for financial support.
Received: July 21, 2004
Revised: October 29, 2004
Published online: May 20, 2005
[1] J. Rtey, Angew. Chem. 1990, 102, 373 – 379; Angew. Chem. Int.
Ed. Engl. 1990, 29, 355 – 361.
[2] a) P. A. Frey, M. K. Essenberg, R. H. Abeles, J. Biol. Chem.
1967, 242, 5369 – 5377; b) T. H. Finlay, J. Valinsky, K. Sato, R. H.
Abeles, J. Biol. Chem. 1972, 247, 4197 – 4207; c) Y. Zhao, P. Such,
J. Rtey, Angew. Chem. 1992, 104, 212 – 213; Angew. Chem. Int.
Ed. Engl. 1992, 31, 215 – 216; d) Y. Zhao, A. Abend, M. Kunz, P.
Such, J. Rtey, Eur. J. Biochem. 1994, 225, 891 – 896.
[3] For selected reviews, see: a) W. Buckel, B. T. Golding, Chem.
Soc. Rev. 1996, 25, 329 – 337; b) E. N. G. Marsh, C. L. Drennan,
Curr. Opin. Chem. Biol. 2001, 5, 499 – 505; c) R. Banerjee, Chem.
Rev. 2003, 103, 2083 – 2094; d) T. Toraya, Chem. Rev. 2003, 103,
2095 – 2127.
[4] P. A. Frey, O. T. Magnusson, Chem. Rev. 2003, 103, 2129 – 2148.
[5] P. Christen, P. K. Mehta, Chem. Rec. 2001, 1, 436 – 447.
[6] E. E. Snell, Methods Enzymol. 1986, 122, 128 – 135.
[7] G. Lukatela, N. Krauss, K. Theis, T. Selmer, V. Gieselmann, K.
von Figura, W. Saenger, Biochemistry 1998, 37, 3654 – 3664.
[8] a) J. E. Dove, J. P. Klinman, Adv. Protein Chem. 2001, 58, 141 –
174; b) S. J. Firbank, M. S. Rogers, C. M. Wilmot, D. M. Dooley,
M. A. Halcrow, P. F. Knowles, M. J. McPherson, S. E. Phillips,
Proc. Natl. Acad. Sci. USA 2001, 98, 12 932 – 12 937.
[9] G. Srinivasan, C. M. James, J. A. Krzycki, Science 2002, 296,
1459 – 1462.
[10] B. Hao, W. Gong, T. K. Ferguson, C. M. James, J. A. Krzycki,
M. K. Chan, Science 2002, 296, 1462 – 1466.
[11] S. K. Blight, R. C. Larue, A. Mahapatra, D. G. Longstaff, E.
Chang, G. Zhao, P. T. Kang, K. B. Green-Church, M. K. Chan,
J. A. Krzycki, Nature 2004, 431, 333 – 335.
[12] M. Thanbichler, A. Bck, Methods Enzymol. 2002, 347, 3 – 16.
[13] J. Klepp, A. Fallert-Mller, K. Grimm, W. E. Hull, J. Rtey, Eur.
J. Biochem. 1990, 192, 669 – 676.
[14] L. A. Chasin, B. Magasanik, J. Biol. Chem. 1968, 243, 5165 –
[15] P. Lund, B. Magasanik, J. Biol. Chem. 1965, 240, 4316 – 4319.
[16] H. K. Meiss, W. J. Brill, B. Magasanik, J. Biol. Chem. 1969, 244,
5382 – 5391.
[17] a) H. Tabor, A. H. Mehler, J. Biol. Chem. 1954, 210, 559 – 568;
b) T. G. Lessie, F. C. Neidhard, J. Bacteriol. 1967, 149, 1800 –
1810; c) B. E. Uhlim, A. J. Clark, J. Bacteriol. 1981, 163, 386 –
[18] R. G. Taylor, H. L. Levy, R. R. McInnes, Mol. Biol. Med. 1991, 8,
101 – 116.
[19] a) D. H. Hug, J. K. Hunter, J. Bacteriol. 1970, 152, 874 – 876;
b) A. R. Young, Phys. Med. Biol. 1997, 42, 789 – 802.
[20] a) T. Yoshida, K. Tada, Y. Honda, T. Arakawa, J. Tohoku, J. Exp.
Med. 1971, 104, 305 – 312; b) Z. Kalatatic, K. Lipovac, Z.
Jereniac, D. Juretic, M. Dumic, B. Zurga, L. Res, Metabolism
1980, 29, 1013 – 1019.
[21] D. Kessler, J. Rtey, G. E. Schulz, J. Mol. Biol. 2004, 342, 183 –
[22] a) M. W. Consevage, A. T. Phillips, Biochemistry 1985, 24, 301 –
308; b) M. Fessenmaier, R. Frank, C. Schubert, J. Rtey, FEBS
Lett. 1991, 286, 55 – 57.
[23] a) R. G. Taylor, J. Garcia-Heras, S. J. Sadler, R. G. Lafreniere,
H. F. Willrad, D. H. Ledbetter, R. R. McInnes, Cytogenet. Cell
Genet. 1991, 56, 178 – 181; b) R. G. Taylor, D. Grieco, G. A.
Clarke, R. R. McInnes, B. A. Taylor, Genomics 1993, 16, 231 –
Angew. Chem. Int. Ed. 2005, 44, 3668 – 3688
240; c) R. G. Taylor, R. R. McInnes, J. Biol. Chem. 1994, 269,
27 473 – 27 477.
a) “Regulation in the hut System”: B. Magasanik in The Operon
(Eds.: J. H. Miller, W. S. Reznikoff), Cold Spring Harbor
Laboratory Press, Cold Spring Harbor, 1987, pp. 373 – 387;
b) L. Hu, S. L. Allison, A. T. Phillips, J. Bacteriol. 1989, 171,
1489 – 1495; c) S. A. Boylan, L. J. Eades, Mol. Gen. Genet. 1984,
193, 92 – 97; d) A. J. Nieuwkoop, S. A. Baldauf, M. E. S. Hudspeth, R. A. Bender, J. Bacteriol. 1988, 170, 2240 – 2246; e) M.
Oda, A. Sugishita, K. Furukawa, J. Bacteriol. 1988, 170, 3199 –
a) K. Hahlbrock, D. Scheel, Annu. Rev. Plant Phys. Plant Mol.
Biol. 1989, 40, 347 – 369; b) C. Appert, E. Logemann, K.
Hahlbrock, J. Schmid, N. Amrhein, Eur. J. Biochem. 1994, 225,
491 – 499.
J. G. Anson, H. J. Gilbert, J. D. Oram, N. P. Minton, Gene 1987,
58, 189 – 199.
L. Xiang, B. S. Moore, J. Biol. Chem. 2002, 277, 32 505 – 32 509.
K. R. Hanson, E. A. Havir, Arch. Biochem. Biophys. 1970, 141,
1 – 17.
a) B. V. Charlwood, M. Pletsch, J. Herbs Spices Med. Plants 2002,
9, 139 – 151; b) T. Kusumi, Bio Ind. 1999, 16, 31 – 39; c) K. M.
Davies, S. J. Bloor, G. B. Spiller, Plant J. 1998, 13, 259 – 266.
J. Zoń, N. Amrhein, Liebigs Ann. Chem. 1992, 625 – 628.
a) G. M. Kishore, D. M. Shah, Annu. Rev. Biochem. 1988, 57,
627 – 663; b) E. Haslam, Shikimic Acid: Metabolism and Metabolites, Wiley, New York, 1993.
a) P. E. Flemming, U. Mocek, H. G. Floss, J. Am. Chem. Soc.
1993, 115, 805 – 807; b) K. Walker, H. G. Floss, J. Am. Chem. Soc.
1998, 120, 5333 – 5334.
W. W. Poon, B. N. Marbois, K. F. Faull, C. F. Clarke, Arch.
Biochem. Biophys. 1995, 320, 305 – 314.
A. Mitra, Y. Kitamura, M. J. Gasson, A. Narbad, A. J. Parr, J.
Payne, M. J. C. Rhodes, C. Sewter, N. J. Walton, Arch. Biochem.
Biophys. 1999, 365, 10 – 16.
PAL from Rhodotorula glutinis, Sigma Biochemicals and
Reagents, Cat. No. P1016.
J. A. Kyndt, T. E. Meyer, M. A. Cusanovich, J. J. Van Beumen,
FEBS Lett. 2002, 512, 240 – 244.
A. Hunter, J. Biol. Chem. 1912, 11, 537 – 545.
S. Edlbacher, Hoppe-Seylers Z. Physiol. Chem. 1926, 157, 106.
A. Peterkofsky, J. Biol. Chem. 1962, 237, 787 – 795.
T. A. Smith, F. H. Cordelle, R. H. Abeles, Arch. Biochem.
Biophys. 1967, 120, 724 – 725.
a) D. Hodgins, R. H. Abeles, J. Biol. Chem. 1967, 242, 5158 –
5159; b) D. Hodgins, J. Biol. Chem. 1971, 246, 2977 – 2985.
R. B. Wickner, J. Biol. Chem. 1969, 244, 6550 – 6552.
J. D. Hermes, P. M. Weiss, W. W. Cleland, Biochemistry 1985, 24,
2959 – 2967.
a) M. Varma, C. J. M. Stirling, J. Chem. Soc. Chem. Commun.
1981, 553 – 554; b) M. B. Smith, J. March, Advanced Organic
Chemistry, Wiley-VCH, Weinheim, 2001, p. 1321.
W. Shi, J. Dunbar, M. M. K. Jayasekera, R. E. Viola, G. K.
Farber, Biochemistry 1997, 36, 9136 – 9144.
M. M. K. Jayasekera, W. Shi, G. K. Farber, R. E. Viola, Biochemistry 1997, 36, 9145 – 9150.
a) C. V. Levy, P. A. Sedelnikova, Y. Kato, Y. Asano, D. W. Rice,
P. J. Baker, Struct. Fold. Des. 2002, 10, 105 – 113; b) Y. Asano, Y.
Kato, C. Levy, P. Baker, D. Rice, Biocatal. Biotransform. 2004,
22, 131 – 138.
M. Asuncion, W. Blankenfeldt, J. N. Barlow, D. Gani, J. H.
Naismith, J. Biol. Chem. 2002, 277, 8306 – 8311.
a) J. C. Calabrese, D. B. Jordan, A. Boodhoo, S. Sariaslani, T.
Vannelli, Biochemistry 2004, 43, 11 403 – 11 416; b) H. Ritter,
G. E. Schulz, Plant Cell 2004, 16, 3426 – 3436.
F. G. Bordwell, Y. Zhao, J. Org. Chem. 1995, 60, 3932 – 3933.
Angew. Chem. Int. Ed. 2005, 44, 3668 – 3688
[51] H. G. Sahl, R. W. Jack, G. Bierbaum, Eur. J. Biochem. 1995, 230,
827 – 853.
[52] J. Rtey, Biochim. Biophys. Acta 2003, 1647, 179 – 184.
[53] “Novel Cofactors”: B. Langer, M. Langer, J. Rtey, Adv. Protein
Chem. 2002, 55, 175 – 214.
[54] M. Langer, G. Reck, J. Reed, J. Rtey, Biochemistry 1994, 33,
6462 – 6467.
[55] B. Schuster, J. Rtey, FEBS Lett. 1994, 349, 252 – 254.
[56] a) B. Langer, A. Lieber, J. Rtey, Biochemistry 1994, 33, 14 034 –
14 039; b) B. Langer, D. Rther, J. Rtey, Biochemistry 1997, 36,
10 867 – 10 871.
[57] C. B. Klee, K. L. Kirk, L. A. Cohen, Biochem. Biophys. Res.
Commun. 1979, 87, 343 – 348.
[58] M. Langer, A. Pauling, J. Rtey, Angew. Chem. 1995, 107, 1585 –
1587; Angew. Chem. Int. Ed. Engl. Angew. Chem. Int. Ed. 1995,
34, 1464 – 1465.
[59] J. Rtey, H. Fierz, W. P. Zeylemaker, FEBS Lett. 1970, 6, 203 –
[60] J. Rtey, Naturwissenschaften 1996, 83, 439 – 447.
[61] B. Schuster, J. Rtey, Proc. Natl. Acad. Sci. USA 1995, 92, 8433 –
[62] M. Rettig, A. Sigrist, J. Rtey, Helv. Chim. Acta 2000, 83, 2246 –
[63] a) D. Hernandez, J. G. Stroh, A. T. Phillips, Arch. Biochem.
Biophys. 1993, 307, 126 – 132; b) T. F. Schwede, M. Bdeker, M.
Langer, J. Rtey, G. E. Schulz, Protein Eng. 1999, 12, 151 – 153.
[64] T. F. Schwede, J. Rtey, G. E. Schulz, Biochemistry 1999, 38,
5355 – 5361.
[65] M. Orm, A. B. Cubitt, K. Kallio, L. A. Gross, R. Y. Tsien, S. J.
Remington, Science 1996, 273, 1392 – 1395.
[66] D. Merkel, J. Rtey, unpublished results.
[67] M. Baedeker, G. E. Schulz, Eur. J. Biochem. 2002, 269, 1790 –
[68] a) C. B. Klee, J. Biol. Chem. 1970, 245, 3143 – 3152; b) C. B. Klee,
Biochemistry 1974, 22, 4501 – 4507.
[69] D. Rther, D. Merkel, J. Rtey, Angew. Chem. 2000, 112, 2592 –
2594; Angew. Chem. Int. Ed. 2000, 39, 2462 – 2464.
[70] a) S. D. Christenson, W. Liu, M. D. Toney, B. Shen, J. Am. Chem.
Soc. 2003, 125, 6062 – 6063; b) S. D. Christenson, W. Wu, M. A.
Spies, B. Shen, M. D. Toney, Biochemistry 2003, 42, 12 708 –
12 718.
[71] M. Baedeker, G. E. Schulz, Structure 2002, 10, 61 – 67.
[72] D. Rther, L. Poppe, S. Viergutz, B. Langer, J. Rtey, Eur. J.
Biochem. 2001, 268, 6011 – 6019.
[73] C. B. Klee, J. Biol. Chem. 1972, 247, 1398 – 1406.
[74] D. Rther, L. Poppe, G. Morlock, S. Viergutz, J. Rtey, Eur. J.
Biochem. 2002, 269, 3065 – 3075.
[75] S. Viergutz, J. Rtey, Chem. Biodiversity 2004, 1, 296 – 302.
[76] a) K. R. Hanson, R. H. Wightman, J. Staunton, A. R. Battersby,
J. Chem. Soc. Chem. Commun. 1971, 185 – 186; b) G. W. Kirby, J.
Michael, J. Chem. Soc. Chem. Commun. 1971, 187.
[77] S. Viergutz, J. Rtey, unpublished results.
[78] a) D. Hernandez, J. G. Stroh, A. T. Phillips, Arch. Biochem.
Biophys. 1993, 307, 126 – 132; b) K. Weber, J. Rtey, Bioorg.
Med. Chem. 1996, 4, 1001 – 1006.
[79] a) D. Merkel, PhD thesis, Universitt Karlsruhe, 1999; b) D.
Galpin, B. E. Ellis, M. E. Tanner, J. Am. Chem. Soc. 1999, 121,
10 840 – 10 841; c) D. Merkel, J. Rtey, Helv. Chim. Acta 2000, 83,
1151 – 1160.
[80] T. Andrews, G. Lorimer, N. Tolbert, Biochemistry 1973, 12, 11 –
[81] A. Skolaut, J. Rtey, Arch. Biochem. Biophys. 2001, 393, 187 –
[82] A. Lewandowicz, J. Jemielity, M. Kańska, J. Zoń, P. Paneth,
Arch. Biochem. Biophys. 1999, 370, 216 – 221.
[83] S. Alunni, A. Cipiciani, G. Fioroni, L. Ottavi, Arch. Biochem.
Biophys. 2003, 412, 170 – 175.
2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
J. Rtey and L. Poppe
[84] K. R. Hanson, E. A. Harry, C. Ressler, Biochemistry 1979, 18,
1431 – 1439.
[85] A. Gloge, J. Zoń, . Kvri, L. Poppe, J. Rtey, Chem. Eur. J.
2000, 6, 3386 – 3390.
[86] S. Viergutz, L. Poppe, A. Tomin, J. Rtey, Helv. Chim. Acta 2003,
86, 3601 – 3612.
[87] S. Campbel1, 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 – 5264.
[88] a) L. Poppe, J. Rtey, Curr. Org. Chem. 2003, 7, 1297 – 1315;
b) H. Kamachi, H. Aoki, Bio Ind. 2003, 20, 12 – 20.
[89] V. R. Williams, J. M. Hiroms, Biochim. Biophys. Acta 1967, 139,
214 – 216.
[90] R. L. Fuchs, J. F. Kane, J. Bacteriol. 1985, 167, 98 – 101.
[91] C. B. Klee, K. L. Kirk, L. A. Cohen, P. McPhie, J. Biol. Chem.
1975, 250, 5033 – 5040.
[92] H. Aoki, H. Kamachi (Showa Denko K. K., Japan), Jpn. Kokai
Tokkyo Koho (2004), JP 2004041107 A2 [Chem. Abstr. 2004, 140,
2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
[93] a) S. Yamada, K. Nabe, N. Izuo, K. Nakamichi, I. Chibata, Appl.
Environ. Microbiol. 1981, 42, 773 – 778; b) C. T. Evans, K.
Hanna, C. Payne, D. Conrad, M. Misawa, Enzyme Microb.
Technol. 1987, 9, 417 – 421; c) M. Yanaka, D. Ura, A. Takahashi,
N. Fukuhara (Mitsui Toatsu Chemicals), JP 06,113,870 (1994)
[Chem. Abstr. 1994, 121, 155941y].
[94] A. Gloge, B. Langer, L. Poppe, J. Rtey, Arch. Biochem.
Biophys. 1998, 359, 1 – 7.
[95] W. Liu, (Great Lakes Chemical Co.) USP 5,981,239 (1999)
[Chem. Abstr. 1999, 131, 321632].
[96] H. Aoki, H. Kamachi (Showa Denko K. K., Japan), PCT Int.
Appl. (2003), WO 2003000915 A1 [Chem. Abstr. 2003, 138,
[97] D. G. Rees, D. H. Jones, Biochim. Biophys. Acta 1997, 1338,
121 – 126.
[98] K. Masafumi, H. E. Il, O. Yasuo, H. Sueharu, J. Ind. Microbiol.
Biotechnol. 2003, 30, 456 – 461.
[99] J. L. Hu, Y. C. Xie, M. Y. Xie, R. Zhang, T. Otani, Y. Mihami, Y.
Yamada, T. Marunaka, J. Antibiot. 1988, 41, 1575 – 1579.
Angew. Chem. Int. Ed. 2005, 44, 3668 – 3688
Без категории
Размер файла
1 291 Кб
elimination, ammonia, enzymatic, mechanism, phenylalanine, typed, friedelцcrafts, histidine
Пожаловаться на содержимое документа