Specific host-guest interactions in a protein-based artificial transaminase

Article abstract of DOI:10.1016/S0968-0896(01)00218-8

Artificial enzymes can be created by covalent attachment of a catalytic active group to a protein scaffold. Recently, we assembled an artificial transaminase by conjugation of intestinal fatty acid binding protein (IFABP) with a pyridoxamine derivative via a disulfide bond; the resulting constract catalyzed a transamination reaction 200-fold faster than free pyridoxamine. To identify the origin of this increased catalytic efficiency computer modeling was first used to identify two putative residues, Y14 and R126, that were in close proximity to the γ-carboxylate group of the substrate, α-ketoglutartate. These positions were mutated to phenylalanine and methionine, respectively, and used to prepare semisynthetic transaminases by conjugation to pyridoxamine (Px) or an N-methylated derivative (Px). Kinetic analysis of the resulting constructs showed that the R126 mutation reduced substrate affinity 3- to 6-fold while the additional Y14F mutation had a negligible effect. These results are consistent with a model for substrate recognition that involves an electrostatic interaction between the cationic guanidinium group of R126 and the anionic carboxylate from the substrate. Interestingly, one of the conjugates that contains an N-methylated pyridoxamine catalyzes a transamination reaction with a k_cat we have thus far obtained and is 34-fold greater than that for the free cofactor in the absence of the protein. Copyright

Full text of DOI:10.1016/S0968-0896(01)00218-8

Bioorganic & Medicinal Chemistry 9 (2001) 2461–2466
Specific Host–Guest Interactions in a Protein-Based
Artificial Transaminase
Dietmar Haring and Mark D. Distefano*
Department of Chemistry, University of Minnesota, Minneapolis, MN 55455, USA
Received 2 April 2001; accepted 5 June 2001
Dedicated to Professor Peter B. Dervan in recognition of his pioneering work in the fields of bioorganic chemistry
and molecular recognition and for his distinctive qualities as a scientific mentor
Abstract—Artificial enzymes can be created by covalent attachment of a catalytic active group to a protein scaffold. Recently, we
assembled an artificial transaminase by conjugation of intestinal fatty acid binding protein (IFABP) with a pyridoxamine derivative
via a disulfide bond; the resulting construct catalyzed a transamination reaction 200-fold faster than free pyridoxamine. To identify
the origin of this increased catalytic efficiency computer modeling was first used to identify two putative residues, Y14 and R126,
that were in close proximity to the g-carboxylate group of the substrate, a-ketoglutartate. These positions were mutated to phenyl-
alanine and methionine, respectively, and used to prepare semisynthetic transaminases by conjugation to pyridoxamine (Px) or an
N-methylated derivative (MPx). Kinetic analysis of the resulting constructs showed that the R126M mutation reduced substrate
affinity 3- to 6-fold while the additional Y14F mutation had a negligible effect. These results are consistent with a model for sub-
strate recognition that involves an electrostatic interaction between the cationic guanidinium group of R126 and the anionic car-
boxylate from the substrate. Interestingly, one of the conjugates that contains an N-methylated pyridoxamine catalyzes a
0
transamination reaction with a kcat value of 1.1 hꢀ1 which is the fastest value for kcat we have thus far obtained and is 34-fold
greater than that for the free cofactor in the absence of the protein. # 2001 Elsevier Science Ltd. All rights reserved.
Introduction
(Px) via a disulfide bond to Cys60 inside the cavity.6 A
MOLSCRIPT7 representation of this construct is shown
in Figure 1. That enzyme mimic, denoted here as
IFABP-Px, catalyzed the enantioselective transamina-
tion of a-ketoglutarate and various amino acids. A sig-
nificant increase (ca. 200-fold) in catalytic efficiency
(kcat/KM) was observed with IFABP-Px compared to the
free pyridoxamine. Kinetic analysis indicated that the
turnover number, kcat, of the protein conjugate was
approximately 4-fold greater than that for the free
cofactor whereas a 50-fold increase in substrate affinity
(KM) for the protein-based host–guest system was
found. Since the cofactor is tethered inside the cavity of
IFABP and the substrates are completely surrounded by
the protein during catalysis, the modulation of catalytic
efficiency we observed mediated by the protein host is
not surprising. To improve the efficiency of these sys-
tems, we are interested in introducing functional groups
via mutagenesis of the protein interior that will partici-
pate in both catalysis and substrate recognition.
Recently we introduced lysine residues into the IFABP
cavity at two positions and demonstrated that this new
functionality participated in covalent catalysis, altered
the rate versus pH profile and increased catalytic effi-
ciency as much as 4200-fold.8 However, it is also desir-
able to identify existing residues from the protein host
The search for rational designed biocatalysts with
desired catalytic activities and tailored substrate selec-
tivities has stimulated the development of exciting sys-
tems such as catalytic antibodies,1 coenzyme-amino acid
chimeras2 or semisynthetic enzymes.3 Protein-based
systems are particularly attractive scaffolds for the
assembly of new catalysts because their tertiary struc-
ture allows many interactions with substrates. Using
established recombinant DNA methodologies these
scaffolds can be designed at will. We have focused on a
class of proteins known as fatty acid binding proteins
(FABPs) that possess a large cavity for ligand bind-
ing.4,5 For example, intestinal fatty acid binding protein
(IFABP) forms a cavity of 600 A3, in which fatty acids
can be completely sequestered from solvent. Its well
established structure and easy overexpression of
mutants make this protein a versatile host for rational
design by molecular modeling and genetic engineering.
Recently, we assembled an artificial transaminase by
conjugation of IFABP with a pyridoxamine derivative
*Corresponding author. Tel.: +1-612-624-0544; fax: +1-612-626-
7541; e-mail: distefan@chem.umn.edu
0968-0896/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(01)00218-8

2462
D. Haring, M. D. Distefano /Bioorg. Med. Chem. 9 (2001) 2461–2466
¨
that form specific interactions with the reaction sub-
strates. If this could be determined, it would then allow
the specificity of these artificial transaminases to be
altered in a predictable manner.
molecule was bound to the aldehyde group of Px as a
ketimine and again optimized. A model for the overall
structure of this complex is shown in Figure 1. Interest-
ingly, the g-carboxylic acid group was in a position sui-
table for electrostatic interactions and hydrogen
bonding to Arg126 or hydrogen bonding to Tyr14,
respectively as shown in Figure 2 (left side).
As noted above, in the original conjugate IFABP-Px, a
significant increase in substrate affinity was observed;9
here, we examine the origin of this effect. Computer
modeling was first used to identify two putative resi-
dues, Y14 and R126, that were in close proximity to the
g-carboxylate group of the substrate, a-ketoglutartate.
We hypothesized that at least one of these amino acids
interacted with the substrate-derived carboxylate via
electrostatic or hydrogen bonding interactions. These
positions were mutated to phenylalanine and methio-
nine, respectively, and used to prepare semisynthetic
transaminases by conjugation to pyridoxamine (Px) or
an N-methylated derivative (MPx). Kinetic analysis of
the resulting constructs showed that the R126M muta-
tion reduced substrate affinity 3- to 6-fold while the
additional Y14F mutation had a negligible effect. These
results are consistent with a model for substrate recog-
nition that involves an electrostatic interaction between
the cationic guanidinium group of R126 and the anionic
carboxylate from the substrate.
When these residues were replaced by more hydro-
phobic isosteric amino acids (R126M and Y14F), the
overall orientation of a-ketoglutarate was unchanged
(Fig. 2, right side). However, these mutations yielded a
more hydrophobic environment and prevented favor-
able interactions with the g-carboxylic acid group. The
phenyl ring of Phe14 is oriented closer to the a-keto-
glutarate in comparison to the 4-hydroxy substituted
side chain of Tyr14. This suggests that the hydroxy
group prevented such an orientation because of a close
contact to the g-carboxylic acid group.
Recently, we developed a N-methylated pyridoxamine
conjugation reagent (MPx), which has a permanent
positive charge at the pyridine nitrogen.11 The resulting
electron-withdrawing effect should enhance the depro-
tonation of substrate-pyridoxal aldimine and ketimine
intermediates.12 The above described computational
modeling was repeated with this N-methylated cofactor.
Again, the g-carboxylic acid group of the ketoglutarate
bound to pyridoxal was in a position close the side
chain functional groups of Arg126 or Tyr14 as
shown in Figure 3 (left side). Interestingly, in this
case, the Y14F mutation resulted in a lengthening of
the distance between the g-carboxylic acid group of the
substrate and the phenyl ring from F14 (Fig. 3, right
side).
Results
Molecular modeling
The crystal structure of IFABP10 was used as a starting
point to generate computational models of IFABP-Px.
The pyridoxal derivative was linked to Cys60 and its
conformation energy-minimized. An a-ketoglutarate
Figure 1. Stereoview of a computational model of a pyridoxamine moiety conjugated to Cys60 of intestinal fatty acid binding protein (IFABP). a-
Ketoglutarate is bound to the cofactor with its g-carboxylic group pointing to Tyr14 and Arg126. The entrance to the protein is in the upper left
corner. The model is based on the crystal structure of IFABP and was generated using MOLSCRIPT. The left and central images provide a con-
vergent stereoview while the central and right images give a divergent stereoview.

D. Haring, M. D. Distefano /Bioorg. Med. Chem. 9 (2001) 2461–2466
¨
2463
Conjugate preparation and characterization
from 6 L cultures, 35 mg (V60C/R126M) or 52 mg
(V60C/R126M/Y14F), respectively, of homogeneous
proteins were isolated. The cofactor was tethered to
Cys60 by addition of 10-fold excess 5-(2-pyridy-
ltdithio)pyridoxamine (1a) in 20 mM HEPES buffer (pH
7.5) at room temperature as shown in Figure 4. The
excess of reagent was removed by gel filtration chroma-
tography and the resulting conjugates IFABP-Px126
and IFABP-Px126/14 showed UV bands of both the
protein (280 nm) and the pyridoxamine (326 nm); UV
spectra of these two conjugates are shown in Figure 5.
In order to test the computational models the IFABP-
mutants V60C/R126M and V60C/R126M/Y14F were
prepared. Site-directed mutations were introduced in
the expression vector of pMON-IFABP recombinant
plasmid and the proteins overexpressed in Escherichia
coli JM105. Both IFABP-mutant proteins were found in
the supernatant after cell lysis and purified by ammo-
nium sulfate precipitation, ion exchange chromato-
graphy and gel filtration chromatography.6a,13 Starting
Using a similar approach as described above, the con-
jugation reagent 1-methyl-5-(2-pyridyltdithio)pyridox-
amine (1b, Fig. 4) was used to tether the cofactor via a
disulfide bond to Cys60 of IFABP-V60C/R126M and
IFABP-V60C/R126M/Y14F. The resulting conjugates,
IFABP-MPx126 and IFABP-MPx126/14, were pre-
pared and isolated by gel filtration on a 20 mg scale. UV
spectra, shown in Figure 6, of these MPx-containing
constructs again displayed bands from both the protein
and pyridoxamine but in this case, the pyridoxamine
absorbence was red-shifted to a lmax of 338 nm. As
previously reported, this shift is characteristic of the N-
methylated cofactor.
Kinetic analysis of artificial transaminases containing the
R126M and Y14F mutations
The newly prepared Px and MPx conjugates catalyzed
the transamination of a-ketoglutarate and phenylala-
nine yielding glutamic acid as shown in Figure 7. All of
the conjugates functioned catalytically, performing
multiple turnovers in 24 h incubations. Furthermore,
the observed catalysis occurred in an enantioselective
fashion yielding l-glutamate with an enantiomeric pur-
ity of 40–86% ee. In general, these enantioselectivities
are less than those observed for the parent systems. For
example IFABP-Px produces l-glutamate in 95% ee
while the mutant IFABP-Px126 yields this product in
86% ee. The presence of the second mutation in
IFABP-Px126/14 further reduces the enantiomeric pur-
ity to 67% ee; a similar trend was observed with the
MPx-based constructs. The kinetic parameters for the
transamination reaction catalyzed by the different con-
jugates were determined by varying the concentration of
the keto acid at a constant, saturating, concentration of
phenylalanine;14 these values are tabulated in Table 1
(IFABP-Px conjugates) and Table 2 (IFABP-MPx con-
Figure 2. Computer models of the IFABP-Px active site with a-keto-
glutarate bound in the ketimine form. The g-carboxylic group of the
substrate interacts with Arg 126 and Tyr 14 (left). The mutants
Arg126Met or Arg126Met/Tyr14Phe (black), respectively, are shown
on the right side for comparison.
0
jugates). The substrate affinity KM of IFABP-Px was
lowered by the mutation Arg126Met by a factor of
three. The double mutation Arg126Met/Tyr14Phe
0
resulted in an only slightly lower affinity. The kcat
Figure 3. Computer models of the IFABP-MPx active site with a-
ketoglutarate bound in the ketimine form. The g-carboxylic group of
the substrate interacts with Arg 126 and Tyr 14 (left). The mutants
Arg126Met or Arg126Met/Tyr14Phe (black), respectively, are shown
on the right side for comparison.
Figure 4. Reaction used for the preparation of IFABP conjugates
containing Px or MPx cofactors. Prot-SH is the thiol group from the
unique cysteine residue, Cys60, present in IFABP-V60C, used for the
preparation of all the constructs described here.

2464
D. Haring, M. D. Distefano /Bioorg. Med. Chem. 9 (2001) 2461–2466
¨
Figure 5. UV spectra of IFABP-Px126 (left) and IFABP-Px126/14 (right).
Figure 6. UV spectra of IFABP-MPx126 (left) and IFABP-MPx126/14 (right).
Although it is not possible to unambiguously determine
the reason for this rate increase without additional
experimental data, it is interesting to note that this effect
occurs upon removal of the positive charge at position
126 but only with the MPx conjugate. Comparison of
the modeling results shown in Figure 2 for the Px con-
jugates and Figure 3 for the MPx conjugates shows that
there are subtle differences in the position and con-
formation of the pyridoxamine moiety. In the case of
the Px conjugates, there is a compression of the distance
between the protein (such as position 14) and the sub-
strate-derived g-carboxylate group while with the MPx
conjugates, this distance expands. Such changes may
alter the nature of the rate determining step either by
affecting the conformation of the substrate–pyridox-
amine complex (which must be planar for efficient cata-
lysis) or by perturbing the electrostatics inside the
cavity. Additional structural and mechanistic experi-
ments will be necessary to reveal the underlying reasons
for these significant rate differences.
Figure 7. Transamination reaction catalyzed by IFABP-Px con-
jugates.
values were also influenced by these mutations and
found to be down by 40% compared to the original
IFABP-Px. The N-methylated pyridoxamine conjugates
revealed a similar behavior for the substrate affinity.
0
The mutations in position 126 and 14 increased the KM
values up to seven-fold compared to the original IFABP-
MPx. Interestingly, the turnover numbers increased dis-
tinctively with IFABP-MPx126 having a nearly 5-fold
0
higher kcat than IFABP-MPx. This increased turnover is
0
particularly significant since the observed value for kcat
0
of 1.1 hꢀ1 is the fastest value for kcat we have thus far
obtained in any of our semisynthetic enzymes based on
IFABP and is 34-fold greater than that for the free
cofactor in the absence of the protein.
Finally, while the data described here clearly supports
the idea of a specific interaction between the substrate-
derived g-carboxylate group and Arg126, it is note-
worthy that the catalytic efficiency of mutant conjugates
lacking Arg126 is still 34- to 68-fold greater that that of
the free cofactor. Thus additional interactions between
the protein and the pyridoxamine must occur that pro-
mote increased catalytic activity. Inspection of the
region surrounding the pyridoxamine moiety in the
IFABP cavity, modeled in Figure 8, indicates that a
Table 1. Kinetic parameters for the transamination catalyzed by
IFABP-Px mutants
IFABP-Px IFABP-Px126 IFABP-Px126/14
0
KM (mM)
)
1.8ꢁ0.5
0.29ꢁ0.02
0.16
5.5ꢁ1.1
0.18ꢁ0.01
0.03
6.0ꢁ0.9
0.11ꢁ0.004
0.02
0
kcat (hꢀ1
kcat0/KM (hꢀ1mMꢀ1
)
0

D. Haring, M. D. Distefano /Bioorg. Med. Chem. 9 (2001) 2461–2466
¨
2465
Table 2. Kinetic parameters for the transamination catalyzed by
IFABP-MPx mutants
the enormous flexibility of this ‘chemogenetic’
approach.
IFABP-MPx IFABP-MPx126 IFABP-MPx126/14
0
KM (mM)
)
6.8ꢁ1.3
0.23ꢁ0.01
0.03
40ꢁ6
1.10ꢁ0.06
0.03
48ꢁ7
0.78ꢁ0.05
0.02
Experimental
Materials
0
kcat (hꢀ1
kcat0/KM
(hꢀ1 mMꢀ1
0
)
Pyridoxamine dihydrochloride was converted into 5-(2-
pyridyltdithio)pyridoxamine (1a) as described pre-
viously by Kuang et al.5a Similarly, 1-methyl-5-(2-pyr-
idyltdithio)pyridoxamine (1b) was prepared using a
procedure previously reported.11 The mutagenic primers
were obtained from the Microchemical Facility of the
Institute of Human Genetics (University of Minnesota).
The site-directed mutagenesis kit (‘Quick-Change’) was
obtained from Stratagene. The expression vector
pMON-IFABP was used previously for the preparation
of IFABP-V60C.15 Protein purification was carried out
at 4 ꢂC unless otherwise noted.
number of residues are in close proximity to the cofac-
tor and substrate. Although IFABP-Px conjugates are
still quite primitive catalysts compared to highly evolved
natural transaminases, they share important char-
acteristics including an active site architecture that
involves multiple interactions between catalyst and
substrates that act in concert to increase catalytic effi-
ciency.
Conclusions
Modeling
Molecular modeling experiments with IFABP-Px iden-
tified Arg126 and Tyr14 as possible residues for binding
the g-carboxylic acid group of a-ketoglutarate via
hydrogen bonds or electrostatic interactions. Mutant
conjugates were produced that lacked these residues.
Kinetic analysis suggested that Arg126 but not Try14
was important for binding the substrate. These results
demonstrate that molecular modeling can be used to
accurately predict the existence of specific interactions
in these semisynthetic systems. This should be useful for
introducing new residues that can alter the substrate
specificity or catalytic efficiency of these artificial
enzymes. Finally, it is interesting to note that the con-
Molecular modeling experiments were performed using
Insight II 97.0/Discover 3.00 (Molecular Simulations).
The potential energy was calculated on basis of a con-
sistent valence forcefield and minimized sequentially by
steepest descents and conjugate gradients. The IFABP
backbone and all side chains except Cys60, Arg126 and
Tyr14 were fixed.
Mutagenesis and protein purification
Site-directed mutations were introduced in the expression
vector of pMON-IFABP recombinant plasmid with the
QuickChange mutagenesis kit. The sequence for the
mutagenic oligomers was 50-GACCGGAATCA-
GAACTTCGAAAAGTTCATGGAG-30 for Y14F and
50-GAAGGAGTGGAGGCCAAGATGATCTTTAAG-
AAGGAATAG-30 for R126M. The mutant proteins were
expressed in E. coli JM105. Cultures were grown in LB
struct IFABP-MPx126 which has
a significantly
0
increased kcat value is the product of a strategy that
involves both chemical and genetic methods; the protein
host was manipulated by mutagenesis while the catalytic
properties of the pyridoxamine cofactor were enhanced
via chemical synthesis (N-methylation). This highlights
Figure 8. Active site of IFABP-Px in the ketimine form with a-ketoglutarate. The residues Leu72, Tyr14, Tyr117, Arg126, Leu36, Gln115, Arg106
and Leu38 (clockwise, starting form the upper left corner) are shown. Pyridoxamine is bound to Cys60 of IFABP via a disulfide linkage. The left and
central images provide a convergent stereoview while the central and right images give a divergent stereoview.

2466
D. Haring, M. D. Distefano /Bioorg. Med. Chem. 9 (2001) 2461–2466
¨
media (6ꢃ1 L) with shaking at 37 ꢂC and induced with
nalidixic acid for 3–4 h prior to cell harvesting. Upon
cell lysis, both mutant proteins were found in the solu-
ble protein fraction. The mutant proteins were purified
by ammonium sulfate precipitation followed by ion-
exchange chromatography (QAE-Sephadex) and gel fil-
tration chromatography (Sephacryl S-100) as previously
described.6a,13
References and Notes
1. (a) Tramontano, A.; Janda, K. D.; Lerner, R. A. Science
1986, 234, 1566. (b) Pollack, S. J.; Jacobs, J. W.; Schultz, P. G.
Science 1986, 234, 1570.
2. Imperiali, B.; Roy, R. S. J. Am. Chem. Soc. 1994, 116,
12083.
3. Levine, H. L.; Nakagawa, Y.; Kaiser, E. T. Biochem. Bio-
phys. Res. Comm. 1977, 76, 64.
4. Banaszak, L.; Winter, N.; Xu, Z.; Bernlohr, D. A.; Cowan,
S.; Jones, T. A. Adv. Protein Chem. 1994, 45, 89.
5. (a) Kuang, H.; Brown, M. L.; Davies, R. R.; Young, E. C.;
Distefano, M. D. J. Am. Chem. Soc. 1996, 118, 10702. (b)
Distefano, M. D.; Kuang, H.; Qi, D.; Mazhary, A. Curt. Opin.
Struct. Biol. 1998, 8, 459.
Conjugation to pyridoxamine cofactor
Prior to conjugation, the IFABP-mutant proteins were
desalted by gel filtration (Biorad P6-DG) and the con-
centration of free thiol determined by titration with
DTNB.
The
conjugation
reagent
5-(2-pyr-
6. (a) Kuang, H.; Davies, R. R.; Distefano, M. D. Bioorg.
Med. Chem. Lett. 1997, 7, 2055. (b) Kuang, H.; Distefano,
M. D. J. Am. Chem. Soc. 1998, 120, 1072. (c) Qi, D.; Kuang,
H.; Distefano, M. D. Bioorg. Med. Chem. Lett. 1998, 8, 875.
7. Kraulis, P. J. J. Appl. Crystallogr. 1991, 24, 946.
8. Haring, D.; Distefano, M. D. Bioconjugate Chem. 2001, 12,
385.
9. Transamination reactions catalyzed by free pyridoxamine
or protein conjugates thereof follow saturation kinetics and
can be fit to a Michaelis–Menton kinetic model with KM and
kcat terms. In the case of free pyridoxamine, KM is probably
equal to the dissociation constant for the ketimine, Schiff base
complex. Given that we have been unable to detect substrate
binding to FABPs in the absence of the pyridoxamine moiety,
we believe that the KM values obtained with the protein con-
jugates also reflect the stabilities of the Schiff base complexes.
Thus, comparison of these KM values provides a measure of
the increase in substrate affinity due to the presence of the
protein component of the construct.
idyldithio)pyridoxamine 1 was added in 10-fold excess
and the reaction was allowed to proceed for 18 h at
20 ꢂC in 20 mM HEPES (pH 7.5). The conjugated pro-
tein was purified by gel filtration (Biorad P6-DG) and
the concentration of pyridoxamine determined by its
UV absorbence (e326=6140 Mꢀ1 cmꢀ1 in 6 M guanidine,
20 mM HEPES, pH 7.5). Conjugation with 1-methyl-5-
(2-pyridyltdithio)pyridoxamine (1b) was performed in
an analogous manner. UV spectra of the purified con-
jugates were recorded at a protein concentration of
10 mM in 20 mM HEPES pH 7.5.
Kinetic analysis
Reactions were performed with 50 mM catalyst, 5 mM
phenylalanine (in the case of PMP 5 mM tyrosine
was used) and varying concentrations of a-ketogluta-
rate in 0.2 M HEPES (pH 7.5) at 37 ꢂC. The formation
of glutamate was monitored by reversed-phase
HPLC.5a,16
10. Sacchettini, J. C.; Gordon, J. I.; Banaszak, L. J. J. Mol.
Biol. 1989, 208, 327.
11. Kuang, H.; Haring, D; Qi, D.; Mazhary, A.; Distefano,
M. D. Bioorg. Med. Chem. Lett. 2000, 10, 2091.
12. (a) Maley, J. R.; Bruice, T. C. J. Am. Chem. Soc. 1968, 90,
2843. (b) Maley, J. R.; Bruice, T. C. Arch. Biochem. Biophys.
1970, 136, 187.
13. Sacchettini, J. C.; Banaszak, L. J.; Gordon, J. I. Mol. Cell.
Biochem. 1990, 98, 81.
Acknowledgements
We thank L. J. Banaszak and M. R. Lees for their sup-
port with the site-directed mutagenesis and D. P. Cis-
tola for providing pMON-IFABPV60C. The Minnesota
Supercomputing Institute provided computer time for
the modeling. This work was supported by the National
Science Foundation (CHE9807495) and a fellowship of
the Deutsche Akademische Austauschdienst (NATO-
Programm).
14. Although kinetic experiments were performed in the pre-
sence of saturating levels of the amino acid substrate, the
kinetic parameters obtained cannot be rigorously denoted as
kcat and KM since they were not obtained from extrapolations
to infinite amino acid concentration. For this reason, they are
0
0
designated kcat and KM
15. Jiang, N.; Frieden, C. Biochemistry 1993, 32, 11015.
.
16. Buck, R. H.; Krummen, K. J. Chromatogr. 1987, 387, 255.