Enantioselective reductive amination of α-keto acids to α-amino acids by a pyridoxamine cofactor in a protein cavity

Article abstract of DOI:10.1021/ja954271z

Adipocyte lipid binding protein (ALBP) is a small 131 residue protein with a simple architecture that consists of two orthogonal planes of β-sheet secondary structure. This protein binds a variety of fatty acids in a large cavity formed between the two sheets such that the bound ligands are completely enclosed within the protein. In this paper, the synthesis of an ALBP conjugate (ALBP-PX) containing a pyridoxamine cofactor attached to a thiol within the protein interior is described. The conjugate was characterized by mass spectrometry, UV/vis spectroscopy, and gel filtration chromatography. ALBP-PX reductively aminates a number of alkyl, aryl, and side chain functionalized α-keto acids to α-amino acids with enantioselectivities as high as 94% ee.

Full text of DOI:10.1021/ja954271z

10702
J. Am. Chem. Soc. 1996, 118, 10702-10706
Enantioselective Reductive Amination of R-Keto Acids to
R-Amino Acids by a Pyridoxamine Cofactor in a Protein Cavity
Hao Kuang, Matthew L. Brown, Ronald R. Davies, Eva C. Young, and
Mark D. Distefano*
Contribution from the Department of Chemistry, UniVersity of Minnesota,
Minneapolis, Minnesota 55455
ReceiVed December 21, 1995. ReVised Manuscript ReceiVed August 9, 1996X
Abstract: Adipocyte lipid binding protein (ALBP) is a small 131 residue protein with a simple architecture that
consists of two orthogonal planes of â-sheet secondary structure. This protein binds a variety of fatty acids in a
large cavity formed between the two sheets such that the bound ligands are completely enclosed within the protein.
In this paper, the synthesis of an ALBP conjugate (ALBP-PX) containing a pyridoxamine cofactor attached to a
thiol within the protein interior is described. The conjugate was characterized by mass spectrometry, UV/vis
spectroscopy, and gel filtration chromatography. ALBP-PX reductively aminates a number of alkyl, aryl, and side
chain functionalized R-keto acids to R-amino acids with enantioselectivities as high as 94% ee.
Introduction
of achiral catalysts; additionally, ALBP (and related proteins)
has been the focus of numerous mutagenesis and protein folding
studies which suggests that the protein cavity is amenable to
modification.7
Catalytic enantioselective reactions are of great importance
in modern synthetic chemistry.1 One strategy that can be used
to design selective reaction catalysts is to incorporate nonspecific
achiral catalytic molecules into chiral cavities. A number of
cavities ranging from cyclodextrins to spherands have been
exploited for this purpose.2 Although successful in many cases,
an important limitation of such systems is the difficulty involved
in making incremental changes in the cavity size and shape.
Small changes in cavity structure frequently involve major
synthetic efforts. With the development of recombinant DNA
technology and the concomitant capabilities it affords for protein
manipulation, polypeptide based cavities offer tremendous
potential. Several investigators have studied proteins derivitized
with catalytic groups; however, most of the “semisynthetic
enzymes” constructed to date have involved modifications on
the surface or in clefts of the target proteins;3 such systems do
not allow the potential selectivity afforded by completely
encapsulating the catalyst to be examined.
Pyridoxal and pyridoxamine are nonphosphorylated deriva-
tives of cofactors that are involved in a number of important
biological reactions including transaminations, racemizations,
decarboxylations, and eliminations.8 Since the reactions of these
cofactors are well characterized and have been examined in a
number of proteinaceous and non-proteinaceous systems,9,10 we
decided to study their reactivity and selectivity in the ALBP
cavity. In this paper, the synthesis of an ALBP conjugate
containing a pyridoxamine cofactor, attached to a thiol within
the protein interior, and the reductive amination of seven R-keto
acids to amino acids by this construct are described. The study
here focuses on enantioselectivity since the attainment of such
selectivity remains one of the major challenges in modern
synthetic chemistry.
Results
Adipocyte lipid binding protein (ALBP) is a small 131 residue
protein with a simple architecture that consists of two orthogonal
planes of â-sheet secondary structure.4 A variety of fatty acids
ranging from palmitate to arachidonate bind in a large (600 Å3)
cavity formed between the two sheets.5 An interesting feature
of ALBP is that the bound ligands are completely enclosed
within the protein; a space filling representation of the ALBP
structure masks all atoms of the bound lipid.6 Thus, ALBP
possesses a true protein cavity and is a good system for
evaluating the ability of protein cavities to affect the selectivity
Design Considerations. Our design for the desired ALBP
conjugate originated with the observation that the thiol group
from Cys117 is located in the interior of the protein cavity; this
unique cysteine residue has previously been modified with a
variety of thiol-specific reagents.11 These studies indicate that
large, nonfatty acid molecules can be accomodated in the cavity
(7) (a) Sha, R. S.; Xu, Z.; Banaszak, L. J.; Bernlohr D. A. J. Biol. Chem.
1993, 268, 7885-7892. (b) Stump, D. G.; Lloyd, R. S.; Chytil, F. J. Biol.
Chem. 1991, 266, 4622-4630. (c) Jiang, N.; Frieden, C. Biochemistry 1993,
32, 11015-11021. (d) Kim, K.; Cistola, D. P.; Frieden, C. Biochemistry
1996, 35, 7553-7558. (e) Cistola, D. P.; Kim, K.; Rogl, H. Frieden, C.
Biochemistry 1996, 35, 7559-7565.
X Abstract published in AdVance ACS Abstracts, October 1, 1996.
(1) Seebach, D. Angew. Chem., Int. Ed. Engl. 1990, 29, 1320.
(2) (a) Dugas, H. Bioorganic Chemistry; Springer-Verlag New York,
Inc.: New York, 1996. (b) Breslow, R. Acc. Chem. Res. 1995, 28, 146-
153.
(3) (a) Kaiser, E. T.; Lawrence, D. S. Science 1984, 226, 505-511. (b)
Levine, H. L.; Nakagawa, Y.; Kaiser, E. T. Biochem. Biophys. Res. Comm.
1977, 76, 64-70. (c) Corey, D. R.; Schultz, P. G. Science 1987, 238, 1401-
1403.
(8) Martell, A. E. Acc. Chem. Res. 1989, 22, 115-124.
(9) (a) Breslow, R.; Hammond, M.; Lauer, M. J. Am. Chem. Soc. 1980,
102, 421-422. (b) Breslow, R.; Canary, J. W.; Varney, M.; Waddell, S.
T.; Yang, D. J. Am. Chem. Soc. 1990, 112, 5121-5219. (c) Cochran, A.
G.; Pham, T.; Sugasawara, R.; Schultz, P. G. J. Am. Chem. Soc. 1991, 113,
6672-6673. (d) Imperiali, B.; Roy, R. S. J. Am. Chem. Soc. 1994, 116,
12083-12084. (e) Kikuchi, J.-I.; Zhang, Z. Y.; Murakami, Y. J. Am. Chem.
Soc. 1995, 117, 5383-5384.
(4) Xu, Z.; Bernlohr, D. A.; Banaszak, L. J. Biochemistry 1992, 31,
3484-3492.
(5) LaLonde, J. M.; Bernlohr, D. A.; Banaszak, L. J. Biochemistry 1994,
33, 4885-4895.
(10) (a) Bruice, T. C.; Topping, R. M. J. Am. Chem. Soc. 1963, 85, 1480-
1487. (b) Bruice, T. C.; Topping, R. M. J. Am. Chem. Soc. 1963, 85, 1488-
1492. (c) Thanassi, J. W.; Butler, A. R.; Bruice, T. C Biochemistry 1965,
4, 1463-1470.
(6) Banaszak, L. J.; Winter, N.; Xu, Z.; Bernlohr, D. A.; Cowan, S.;
Jones, T. A. AdV. Protein Chem. 1994, 45, 89-151.
S0002-7863(95)04271-5 CCC: $12.00 © 1996 American Chemical Society

ReductiVe Amination of R-Keto Acids to R-Amino Acids
J. Am. Chem. Soc., Vol. 118, No. 44, 1996 10703
Figure 1. Stereoview of a model for the conjugate ALPB-PX showing pyridoxamine attached to a site in the protein interior. The overall fold of
ALBP is shown in a ribbon representation with Cys117 and the pyridoxamine shown in ball and stick format.
and that modification of Cys117 does not result in large
perturbations of the protein structure. Based on this knowlege,
a model for an ALBP pyridoxamine conjugate (Figure 1) was
constructed using the crystal structure of the ALBP/oleic acid
complex.
Preparation and Characterization of ALBP-PX. To
prepare the desired ALBP conjugate, 1 was first synthesized in
five steps from pyridoxamine and purified by reversed-phase
HPLC. ALBP was purified from an overproducing E. coli
strain, JM105/pMON4 and was then reacted with compound 1
to generate the conjugate ALBP-PX (2). Thiol titration of the
protein before and after reaction with 1 revealed the loss of
Figure 2. Electrospray mass spectrum of ALBP-PX. The charges of
0.73 thiols per ALBP and indicated that the modification had
the peaks used to calculate the mass of ALBP-PX are noted on the
occurred at Cys117. Analysis of ALBP-PX by electrospray mass
spectrometry (Figure 2) indicated the presence of a new species
plot.
with molecular mass 14 722.14 ((2.94) that compares well with
the predicted value of 14 723.56 calculated for the conjugate.
Spectrophotometric analysis of ALBP-PX after chromatography
to remove unincorporated 1 revealed the presence of a new
absorption band at 320 nm characteristic of pyridoxamine
(spectrum A, Figure 3). While the extinction coefficient for
this absorption band appeared low in the native conjugate, it
increased significantly upon denaturation of ALBP-PX (spec-
trum B, Figure 3). Clearly, the microenvironment within the
ALBP cavity perturbs the pyridoxamine chromophore and/or
alters its protonation state. Comparison of the pyridoxamine
concentration in ALBP-PX under denaturing conditions (cal-
culated from a standard curve obtained for free pyridoxamine
Figure 3. UV/vis spectra of ALBP-PX under native and denaturing
conditions. (A) Spectrum of ALBP-PX (46 µM) under native conditions.
(B) Spectrum of ALBP-PX (46 µM) under denaturing conditions. (C)
Spectrum of 39 µM pyridoxamine (85% of 46 µM).
under the same conditions) with the total protein concentration
in a sample of ALBP-PX (determined by protein assay) indicated
that the modification reaction was 85% complete. This
compared favorably with the thiol titration data described above
and suggested that the desired conjugate had been prepared.
Gel filtration chromatography (Figure 4) of ALBP and ALBP-
PX under native conditions gave identical retention times
suggesting that ALBP-PX maintains the folded structure present
in the unmodified protein.
(11) Modification of this residue with bulky reagents blocks the binding
of fatty acids: (a) Buelt, M. K.; Bernlohr, D. A. Biochemistry 1990, 29,
7408-7413. (b) Buelt, M. K.; Xu, Z.; Banaszak, L. J.; Bernlohr, D. A.
Biochemistry 1992, 3l, 3493-3499. Alkylation of thiols positioned in the
cavity with fluorophores generates highly fluorescent derivatives: (c)
Frieden, C.; Jiang, N.; Cistola, D. P. Biochemistry 1995, 34, 2724-2730.
Both of these observations are consistent with modifications within the
cavity.
Reactions Performed by ALBP-PX. To examine the effect
of the protein cavity on reaction selectivity, the reductive
amination reactions shown below were studied. First, the initial

10704 J. Am. Chem. Soc., Vol. 118, No. 44, 1996
Kuang et al.
Figure 4. Gel filtration chromatography of ALBP-PX using a Superose
12 column. Inset: elution volume versus LN (molecular mass) of
standard proteins used for the molecular mass determination of ALBP-
PX. (l) Standards. (∆) ALBP. (∇) ALBP-PX.
Figure 6. HPLC analysis of a reaction of ALBP-PX with pyruvate.
Inset: calibration curve for the quantitation of D- and L-alanine showing
that the fluorescence intensities for the two diastereomeric derviatives
differ by 7.6%.
Table 1. Conversion and Enantioselectivities of ALBP-PX
Reductive Amination Reactions
product
% conversiona major enantiomer
% ee
alanine (6a)
valine (6b)
leucine (6c)
norvaline (6d)
aminocaprylate (6e)
tyrosine (6f)
18 ( 3.6
28 ( 1.9
39 ( 1.2
21 ( 0.6
13 ( 1.3
42 ( 1.5
46 ( 5.6
D
L
42 ( 5.1
94 ( 0.2
L
b
54 ( 0.8
b
b
b
L
L
67 ( 7.3
84 ( 2.2
glutamate (6g)
a Conversion after 24 h. b No ee observed.
reaction is shown in Figure 6. Greater selectivity was obtained
in reactions using a bulkier substrate, R-keto isovalerate (4b),
which produced valine (6b). In this case, a 77% ee for L-valine
was obtained (24 h, 49% conversion). While performing the
experiments with 4b it was noticed that a small amount of
precipitate had formed suggesting that ALBP-PX was aggregat-
ing or denaturing under the reaction conditions. To eliminate
this problem, the ALBP-PX concentration was decreased from
156 to 50 µM, and NaCl was added (20 mM, final concentra-
tion). Under these new conditions 6b was produced from 4b
in 94% ee (24 h, 28% conversion). Keto acid 4c, a γ branched
substrate gave leucine (6c) in 54% ee (24 h, 39% conversion).
Unbranched keto acid substrates 4d (C5) and 4e (C8) gave
comparable levels of conversion to their respective amino acid
products, 6d and 6e, but without any selectivity. A keto acid
containing a phenolic side chain, 4f, produced L-tyrosine (6f)
in 67% ee (24 h, 42% conversion). A keto acid posessing a
carboxylic acid side chain, 4g, produced L-glutamate (6g) in
84% ee (24 h, 46% conversion). Conversions and selectivites
for all the substrates studied are sumarized in Table 1.
Figure 5. Initial rate data for the formation of pyridoxal from ALBP-
PX or PX in the presence of pyruvate. (O) Rate data for the formation
of pyridoxal from ALBP-PX. (0) Rate data for the formation of
pyridoxal from pyridoxamine.
rates of reactions between pyruvate (4a) and PX (3) or ALBP-
PX (2) were determined and are shown in Figure 5.12 The rate
of pyridoxal (5b) formation in the PX reaction was 1.0 × 10-7
M‚min-1, while the ALBP-PX reaction yielded the correspond-
ing aldehyde product (5a) at a rate of 5.7 × 10-8 M‚min-1
.
As a prelude to catalytic studies, the extent of product release
in ALBP-PX mediated reactions was examined. Reactions
between ALBP-PX and 4b were performed, subjected to
ultrafiltration, and analyzed by HPLC. The ultrafiltration step
allowed the concentrations of free 6b and 6b bound to ALBP-
PX to be determined. As sumarized in Table 2, no significant
amount of bound 6b was observed.
This 1.8-fold rate difference suggests that positioning the
pyridoxamine catalyst in the ALBP cavity does not substantially
affect the rate of the reaction.
The production of alanine (6a) in these two systems was
examined next. The enantiomeric excess of amino acid products
in the reactions was determined by forming diastereomeric
fluorescent derivatives and separating and quantitating these
derivatives by reversed-phase HPLC.13 Reactions performed
with PX produced no enantiomeric excess of D-alanine. In
contrast, ALBP-PX produced a 42% ee of D-alanine after
reacting for 24 h (18% conversion); a HPLC trace for this
Discussion
The experiments described here suggest that the ALBP cavity
can impart selectivity to reactions between pyridoxamine and
a number of R-keto acids with selectivities ranging from modest
to excellent. The low ee observed with alanine is likely due to
the small size of its side chain and its limited ability to serve as
(12) Banks, B. E. C.; Diamantis, A. A.; Vernon, C. A. J. Chem. Soc.
1961, 4135-4247.
(13) (a) Lindroth, P.; Mopper, K. Anal. Chem. 1979, 51, 1667-1674.
(b) Nimura, N.; Kinoshita, T. J. Chromatogr. 1986, 352, 169-177. (c) Buck,
R. H.; Krummen, K. J. Chromatogr. 1987, 387, 255-265.

ReductiVe Amination of R-Keto Acids to R-Amino Acids
J. Am. Chem. Soc., Vol. 118, No. 44, 1996 10705
Table 2. Ultrafiltration Analysis of Valine Product from ALBP-PX Reactions
reaction 1
reaction 2
[valine] (µM)
volume (µL)
ee (%)
[valine] (µM)
volume (µL)
ee (%)
before filtration
after filtration, top layer
after filtration, bottom layer
12.2
14.0
14.6
93.8
94.5
94.3
18.7
16.9
16.3
96.1
96.2
95.9
20.1
60.6
17.6
64.9
a directing group while the selectivity for the D enantiomer
probably reflects a structural feature of the ketimine intermedi-
ate. If the largest substituent on the R carbon controls the
conformation about the imine linkage of the intermediate, then
the stereochemical results obtained here (D for alanine, L for
valine, leucine, tyrosine and glutamate) could result from
preferential protonation from the same face in all cases; the
results with alanine would simply reflect the unusual situation
in which the amino acid side chain is smaller than the carboxyl
group as shown below. The results observed with the alkyl
Experimental Section
Purification of ALBP. Large scale cultures of E. coli JM105/
pMON415 were grown in LB Media to an O.D.600 of 0.8 and induced
with nalidixic acid. The cells were then grown for an additional 4 h,
harvested by centrifugation, frozen in N2 (l), and stored at -80 °C.
For protein purification, 20 g of wet cell paste was resuspended in 50
mL of buffer A (25 mM imidazole-HCl, pH 7.0, 50 mM NaCl, 1 mM
2-mercaptoethanol, 5 mM EDTA, and 0.1 mM PMSF) and sonicated
at 0 °C for 5 min. The resulting cell extract was then clarified by
centrifugation, treated with 5 mL of 5% protamine sulfate in buffer A
(w/v), and centrifuged to remove the precipitated nucleic acids. The
supernatant was titrated to pH 5.0 with 2.0 M NaOAc, stirred overnight
at 4 °C, centrifuged to remove precipitated proteins, and concentrated
by ultrafiltration using a YM3 filter (Amicon). The solution of partially
purified protein was then fractionated by chromatography on a Sephadex
G-75 Superfine column (5 cm × 100 cm) equilibrated in buffer
containing 12.5 mM HEPES, pH 7.5, and 0.25 M NaCl. The fractions
were analyzed by SDS PAGE, and the pure fractions pooled and
concentrated by ultrafiltration. Approximately 75 mg of pure protein
was obtained. The purified protein was shown to be competent for
high affinity fatty acid binding in titration experiments with fluorescent
fatty acid analogs.7a
Synthesis of 5-(2-Pyridyldithio)pyridoxamine (1). Compound 1
was prepared by the reaction of 5-thiopyridoxamine dihydrobromide
(synthesized in four steps from pyridoxamine dihydrochloride)16 with
2,2′-dithiopyridine. A round bottom flask under a N2 atmosphere was
charged with 2,2′-dithiopyridine (661 mg, 3.0 mmol) dissolved in 15
mL of degassed MeOH. To this flask was added 5-thiopyridoxamine
dihydrobromide (100 mg, 300 µmol) dissolved in 5.7 mL of degassed
MeOH, dropwise over 1 h. The solvent was removed in Vacuo, and
the yellow solid was washed three times with 10 mL of CHCl3 and
then dissolved in 2 mL of H2O/CH3CN (1/1, v/v) containing 0.2% TFA.
The compound was purified by preparative reversed phase HPLC using
a Rainin Dynamax Microsorb C18 column (2.14 × 25 cm with a 5 cm
guard column). Chromatography was performed by injecting a 200
µL sample and eluting with a flow rate of 10 mL/min and an elution
profile consisting of isocratic elution with solvent A (H2O containing
5% CH3CN and 0.2% TFA) for 10 min followed by a 50 min linear
gradient from 0% to 40% solvent B (CH3CN containing 0.2% TFA).
Using this system, 1 eluted with a retention time of 32 min. Following
chromatography, the fractions containing 1 were pooled and concen-
trated in Vacuo to give 44 mg (48% yield) of a tan solid that darkened
upon storage. Compound 1 stored at -20 °C is stable for several weeks
but was found to be largely decomposed after 3 months. 1H NMR
(200 MHz, D2O) δ 8.25 (ddd, 1H, J ) 5.4, 1.8, 0.8), 7.93 (s, 1H), 7.87
(ddd, 1H, J ) 8.2, 7.6, 1.8), 7.61 (ddd, 1H, 8.2, 1.1, 1.1), 7.35 (ddd,
1H, J ) 7.4, 5.5, 1.1), 4.26 (s, 2H), 4.14 (s, 2H), 2.35 (s, 3H); 13C
NMR (50.32 MHz, D2O) δ 157.4, 156.1, 146.6, 146.5, 144.7, 138.6,
136.6, 135.3, 127.4, 126.4, 38.3, 37.1, 17.2; HR-CI MS calcd for
C13H16N3OS2 [M + H]+, 294.0734, found 294.0728.
substituted R-keto acids follow a trend where the selectivity is
greatest when there is alkyl branching proximal to the develop-
ing chiral center. Valine (6b), with a â-methyl branch is
produced in high ee, whereas leucine (6c) with its γ-methyl
branch is generated in much lower ee; the nonbranched amino
acids 6d and 6e are produced with no enantioselectivity. The
ALBP-PX system appears to be quite general; a number of alkyl
and aryl R-keto acids as well as side chain functionalized R-keto
acids are accepted. This is in contrast to results obtained with
cyclodextrin based systems where only large substrates bind in
the cavity.9a,b ALBP-PX does not accelerate the rate of amino
acid production relative to free pyridoxamine in contrast to other
cyclodextrin- and protein-based systems where 20-200-fold rate
accelerations have been observed.9a,d This is probably due to
both the lack of specific interactions between the substrate and
conjugate, and the absence of rate-accelerating functional groups
within the cavity. The low affinity exhibited by ALBP-PX for
substrates appears to be mirrored in its product interactions.
Ultrafiltration experiments indicate that amino acid products are
predominantly free (unbound) and suggest that product release
is not a problem in this system and that catalytic turnover should
be possible. Additional experiments that examine the structural
basis for the observed enantioselectivity via mutagenesis and
crystallography and the ability to accomplish catalytic turnover
are in progress. We conclude that protein cavities can be useful
for the design of enantioselective catalysts and are using this
approach for the design of other catalytic systems.14
Synthesis of ALBP-PX. ALBP (13 mg, 920 nmol) in 32 mL of
200 mM HEPES, pH 7.5, was mixed with 1 (2.7 mg, 9.2 µmol)
dissolved in H20 (65 µL). The reaction was followed by monitoring
the production of 2-thiopyridone spectrophotometrically at 342 nm and
was complete after 30 min. It was then concentrated to 3 mL by
ultrafiltration with a Centriprep 10 cartridge, applied to a BioGel P6-
DG column (2.5 × 16 cm), and eluted with 200 mM HEPES, pH 7.5,
to remove unreacted 1 and 2-thiopyridone. The fractions containing
ALBP-PX were then concentrated by ultrafiltration and the protein
(15) Xu, Z.; Buelt, M. K.; Banaszak, L. J.; Bernlohr, D. A. J. Biol. Chem.
1991, 266, 14367-14370.
(16) Breslow, R.; Czarnik, A. W.; Lauer, M.; Leppkes, R.; Winkler, J.;
Zimmerman, S. J. Am. Chem. Soc. 1986, 108, 1969-1979.
(14) We have also prepared a conjugate of ALBP containing a Cu(II)
phenanthroline moiety within the cavity. This protein catalyzes the
enantioselective hydrolysis of amino acid esters.

10706 J. Am. Chem. Soc., Vol. 118, No. 44, 1996
Kuang et al.
concentration determined by the method of Bradford.17 The final yield
of the conjugate was 11 mg (84%). The extent of conjugation of ALBP
by 1 was determined by thiol titration with 5,5′-dithiobis(2-nitrobenzoic
acid) in 5.2 M guanidinium hydrochloride as described by Riddles et
al.18 and also by UV/vis spectrophotometry as described below.
Electrospray Mass Spectrometry. Samples for mass spectrometry
were obtained by extensive dialysis of solutions of ALBP-PX in 200
mM HEPES (1 mL, 1.1 mg/mL) against H2O/MeOH/HOAc (89/10/1,
v/v/v), (2 × 1 L). Spectra were obtained using a PE SCIEX API III
MS/MS instrument, and the mass of ALBP-PX was determined by
analysis of the intensity versus m/z data using HyperMass software.
Gel Filtration Chromatography. Gel filtration chromatography
was performed using an FPLC system (Pharmacia) and a Superose 12
HR 30 column. Samples (100 µL) were eluted with 25 mM HEPES,
pH 7.5, and 150 mM NaCl at a flow rate of 0.5 mL/min at 4 °C. The
columns were calibrated using bovine lung aprotinin (6500 g/mol), horse
heart cytochrome c (12 400 g/mol), bovine erythrocyte carbonic
anhydrase (29 000 g/mol) and bovine serum albumin (66 000 g/mol).
UV/vis Spectroscopy. UV/vis spectra of ALBP-PX under native
conditions were obtained using 46 µM ALBP-PX as determined by a
Bradford protein assay in 200 mM HEPES, pH 7.5. Spectra of ALBP-
PX under denaturing conditions contained 46 µM ALBP-PX, as
determined above, and 4.8 M guanidinium hydrochloride in 200 mM
HEPES, pH 7.5. Spectra of pyridoxamine were obtained at 39 µM
(85% of 46 µM) in 200 mM HEPES, pH 7.5. The extinction
coefficients (ꢀ322) for pyridoxamine in 200 mM HEPES, pH 7.5, and
in the same buffer containing 4.8 M guanidinium hydrochloride were
determined to be 7000 and 6100 M-1 cm-1, respectively.
Pyridoxal Formation Kinetics. Reactions with the protein conju-
gate were performed using 157 µM ALBP-PX and 300 mM pyruvate
in 200 mM HEPES, pH 7.5, in a final volume of 250 µL at 23°C. At
1 h intervals, 45 µL samples were withdrawn from the reaction mixture
and quenched with 45 µL of 50% aqueous ethanolamine (v/v) and 10
µL of 10% SDS (w/v). The amount of pyridoxal formed in the reaction
was determined by spectrophotometric quantitation of its ethanolamine
Schiff base adduct (ꢀ362 ) 6740 M-1 cm-1) using a 100 µL ultrami-
crocuvette (Hellma) as described by Snell and Metzler.19 The addition
of SDS was used to denature the ALBP-PX and was necessary to obtain
an accurate absorbance value for the Schiff base adduct. Under the
conditions of the experiment, the SDS added altered the extinction
coefficient for the Schiff base adduct by less than 5%. Reactions with
pyridoxamine were performed using the concentrations described above
in a 10 mL total volume. At 1 h intervals, 400 µL samples were
withdrawn from the reaction mixture and quenched with 400 µL of
50% aqueous ethanolamine (v/v), and the pyridoxal product quantitated
as described above.
Reductive Amination Reactions. All reactions were performed at
37 °C, terminated by flash freezing in N2 (l), and stored at -80 °C
prior to HPLC analysis. Reactions between pyruvate, 4a, (0.3 M) and
ALBP-PX (160 µM) or PX were performed in 200 mM HEPES, pH
7.5, in a total volume of 70 µL. Reactions between R-keto isovalerate,
4b, (0.3 M) and ALBP-PX (160 µM) were first performed in 200 mM
HEPES, pH 7.5, in a total volume of 70 µL. After observing a
precipitate in the reactions containing 4b, the protein concentration was
reduced to 50 µM and NaCl was added to a final concentration of 20
mM. The concentrations of the other components were as described
above. Reactions between substrates 4c through 4g were performed
with keto acid (50 mM) and ALBP-PX (50 µM) in 200 mM HEPES
(pH 7.5) with 20 mM NaCl in a total volume of 100 µL.
Derivatization. Immediately before analysis, reaction samples were
thawed and derivatized with 10 µL of a reagent consisting of N-acetyl-
L-cysteine dissolved in “Incomplete o-Phthaldialdehyde Reagent Solu-
tion” (Sigma no. P7914) at a concentration of 5 mg/mL; these reagents
convert enantiomeric amines into diastereomeric isoindoles. The
derivatization reaction was allowed to proceed for 5 min at which time
the sample was immediately analyzed by HPLC.
and eluting with a flow rate of 0.75 mL/min using a gradient of solvent
A (80 mM sodium citrate and 20 mM sodium phosphate, pH 6.8) and
solvent B (MeOH). Detection of the isoindole derivatives was achieved
using a Beckman 6300A fluorescence detector (356 nm excitation filter
and 450 nm emission filter) and integration was performed with a HP
3393A integrator or System Gold software. Response factors for the
diastereomeric derivatives were obtained from calibration curves
generated by derivatization and analysis of standard solutions of racemic
amino acids. For alanine (6a), an elution profile consisting of isocratic
elution with 80% solvent A for 5 min followed by a 10 min linear
gradient from 20% to 40% solvent B and further elution with 60%
solvent A was employed. Using this system, the retention times for
the D- and L-alanine derivatives were 22.1 and 23.2 min, respectively.
For valine (6b), an elution profile consisting of isocratic elution with
60% solvent A for 5 min followed by a 10 min linear gradient from
40% to 45% solvent B and further isocratic elution with 55% solvent
A gave retention times for the L- and D-valine derivatives of 18.9 and
21.8 min, respectively. For tyrosine (6f), an elution profile consisting
of isocratic elution with 70% solvent A for 5 min followed by a 5 min
linear gradient from 30% to 35% solvent B and further isocratic elution
with 65% solvent A giving retention times for the L- and D-tyrosine
derivatives of 21.2 and 25.1 min, respectively. The amino acid
derivatives discussed below were separated with a modified set of
solvents consisting of solvent A (80 mM sodium citrate and 20 mM
sodium phosphate, pH 6.8, containing MeOH, 10:1, v/v) and solvent
B (MeOH). For leucine (6c), an elution profile consisting of isocratic
elution with 60% solvent A for 5 min followed by a 5 min linear
gradient from 40% to 50% solvent B and further isocratic elution with
50% solvent A gave retention times for the L- and D-leucine derivatives
of 22.3 and 23.6 min, respectively. For norvaline (6d), an elution
profile consisting of isocratic elution with 70% solvent A for 5 min
followed by a 5 min linear gradient from 30% to 45% solvent B and
further isocratic elution with 55% solvent A gave retention times for
the norvaline derivatives of 25.2 and 26.1 min; respectively. For
aminocaprylate (6e), an elution profile consisting of isocratic elution
with 70% solvent A for 5 min followed by a 5 min linear gradient
from 30% to 40% solvent B and further isocratic elution with 60%
solvent A gave retention times for the aminocaprylate derivatives of
29.2 and 30.5 min, respectively. For glutamate (6g), the samples were
eluted isocratically with 100% solvent A. Retention times for the L-
and D-glutamate derivatives were 15.6 and 17.1 min, respectively.
Ultrafiltration Experiments. Two 100 µL solutions containing
ALBP-PX (50 µM), R-ketoisovaleric acid, 4b, (0.3 M) and NaCl (20
mM) in HEPES (0.2 M, pH 7.5) were allowed to react for 24 h.
Samples (10 µL) were withdrawn from each reaction and analyzed to
determine the total valine concentration. The remaining sample from
each reaction was then subjected to ultrafiltration with a Ultrafree-MC
PLGC (Millipore) filter unit (10 000 nominal molecular mass limit for
passage). The samples were centrifuged at 5000 × g for 10 min, and
the volume of liquid in each layer was measured. Samples (10 µL)
from the liquid above (top layer) and below (bottom layer) the filter
along with the original 10 µL samples were then analyzed by HPLC
(as described above) to determine the valine concentration. Similar
experiments were performed with Ultrafree-MC PLCC (Millipore) and
Centricon 3 (Amicon) filtration units that have 5000 and 3000 nominal
molecular mass limits for passage, respectively.
Acknowledgment. The authors thank M. Silberglitt for
modeling ALBP-PX, L. Banaszak for providing the X-ray
coordinates of ALBP, and D. Bernlohr for the plasmid pMON4.
This work was supported by grants from the National Science
Foundation (NSF-CHE-9506793) and the Petroleum Research
Fund (PRF#28140-G4). M.L.B. was supported by the National
Science Foundation REU program (NSF-CHE 9424203) and
R.R.D. was supported by a National Institutes of Health Training
Grant (NIH 2T32GM08347-06).
HPLC Analysis. All reaction mixtures were separated by reversed
phase HPLC using a Rainin Microsorb-MV C18 column (5 µm, 4.6 ×
250 mm). Chromatography was performed by injecting a 10 µL sample
Supporting Information Available: 1H NMR, 13C NMR
and HR-CI MS data for compound 1 (3 pages). See any current
masthead page for ordering and Internet access instructions.
(17) Bradford, M. M. Anal. Biochem. 1976, 72, 248-254.
(18) Riddles, P. W.; Blakeley, R. L.; Zerner, B. Methods Enzymol. 1983,
91, 49-59.
(19) Metzler, D. E.; Snell, E. E. J. Am. Chem. Soc. 1952, 74, 979-980.
JA954271Z