A Change in the Internal Aldimine Lysine (K42) in O-Acetylserine Sulfhydrylase to
Alanine Indicates Its Importance in Transimination and as a General Base Catalyst†
Vaishali D. Rege,‡Nicholas M. Kredich,§Chia-Hui Tai,‡William E. Karsten,‡Klaus D. Schnackerz,|and
Paul F. Cook*,‡
Department of Biochemistry and Molecular Biology, UniVersity of North Texas Health Science Center at Fort Worth, 3500
Camp Bowie BouleVard, Fort Worth, Texas 76107-2699, Department of Medicine and Biochemistry, Duke UniVersity Medical
Center, Box 3100, Durham, North Carolina 27710-0001, and Theodor-BoVeri-Institut fu ¨r Biowissenschaften (Biozentrum) der
UniVersita ¨t Wu ¨rzburg, Physiologische Chemie I, Am Hubland, D-97074 Wu ¨rzburg, Germany
ReceiVed June 24, 1996; ReVised Manuscript ReceiVed July 29, 1996X
ABSTRACT: O-Acetylserine sulfhydrylase (OASS) is a pyridoxal 5′-phosphate dependent enzyme that
catalyzes a ?-replacement reaction forming L-cysteine and acetate from O-acetyl-L-serine (OAS) and sulfide.
The pyridoxal 5′-phosphate (PLP) is bound at the active site in Schiff base linkage with a lysine. In the
present study, the Schiff base lysine was identified as lysine 42, and its role in the OASS reaction was
determined by changing it to alanine using site-directed mutagenesis. K42A-OASS is isolated as an
external aldimine with methionine or leucine and shows no reaction with the natural substrates. Apo-
K42A-OASS can be reconstituted with PLP, suggesting that K42 is not necessary for cofactor binding
and formation of the external Schiff base. The apo-K42A-OASS, reconstituted with PLP, shows slow
formation of the external aldimine but does not form the R-aminoacrylate intermediate on addition of
OAS, suggesting that K42 is involved in the abstraction of the R-proton in the ?-elimination reaction.
The external aldimines formed upon addition of L-Ala or L-Ser are stable and represent a tautomer that
absorbs maximally at 420 nm, while L-Cys gives a tautomeric form of the external aldimine that absorbs
at 330 nm, and is also seen in the overall reaction after addition of primary amines to the assay system.
The use of a small primary amine such as ethylamine or bromoethylamine in the assay system leads to
the initial formation of an internal (γ-thialysine) or external (ethylamine) aldimine followed by the slow
formation of the R-aminoacrylate intermediate on addition of OAS. Activity could not be fully recovered,
and only a single turnover is observed. Data suggest a significant rate enhancement resulting from the
presence of K42 for transimination and general base catalysis.
Pyridoxal 5′-phosphate (PLP)1acts as a cofactor for many
enzymes catalyzing a wide variety of reactions in the
metabolism of amino acids, such as transamination, ?-elim-
ination, ?,γ-replacement, and racemization. In all PLP-
dependent enzymes, the carbonyl group of the coenzyme
binds to an ?-amino group of a lysine residue in the active
site, forming an internal aldimine. In the course of the
catalytic reaction, the lysine may be involved in one or more
functions including binding of PLP to the enzyme, formation
and stabilization of intermediates, and/or release of products.
The function of the lysine residue in Schiff base linkage has
been studied in many of the enzymes that fall in either the
R, ?, or γ families of PLP-dependent enzymes (Alexander
et al., 1994).
O-Acetylserine sulfhydrylase from Salmonella typhimu-
rium, a member of the ?-family of PLP-dependent enzymes
(Kredich & Tomkins, 1966), catalyzes a ?-replacement
reaction forming L-cysteine and acetate from OAS and sulfide
(Becker et al., 1969). A Bi-Bi ping-pong kinetic mechanism
has been proposed for OASS-A (Cook & Wedding, 1976;
Tai et al., 1993). A chemical mechanism has been postulated
in which the unprotonated R-amine of the amino acid
substrate attacks the C4′ carbon of the internal Schiff base
(Tai et al., 1995). A proton from the R-carbon of the
resulting external Schiff base is abstracted by the ?-amino
group of the active site lysine in the ?-elimination of acetate
from OAS. A protonated enzyme residue is also required
to assist in the elimination of the acetate leaving group to
form R-aminoacrylate, acting to hold the acetyl group out
of the plane of the PLP. After Michael addition of the
nucleophilic substrate, a proton is donated to the R-carbon
by the ?-amino group of the active site lysine to form the
external Schiff base with product. The final product is then
released, and the internal Schiff base with the active site
lysine is re-formed.
In the present study, the lysine forming the Schiff base to
PLP was identified using a combination of protein chemistry
and molecular biology. In an attempt to investigate the
functions of the Schiff base lysine, it was changed to an
†This work was supported by grants to P.F.C. from the National
Science Foundation (MCB 9405020) and the Robert A. Welch
Foundation (BK-1031), to NMK from the National Institutes of Health
(DK 12828), and to K.D.S. from the Deutsche Forschungsgemeinschaft
(Schn 139/11-3), and by Grant CRG. 900519 from the North Atlantic
Treaty Organisation Scientific Affairs Division to P.F.C. and K.D.S.
‡University of North Texas Health Science Center at Forth Worth.
§Duke University Medical Center.
|Theodor-Boreri-Institu ¨t fu ¨r Biowissenschaften (Biozentrum) der
Universita ¨t Wu ¨rzburg.
XAbstract published in AdVance ACS Abstracts, October 1, 1996.
1Abbreviations: PLP, pyridoxal 5′-phosphate; OAS, O-acetyl-L-
serine; OASS, O-acetylserine sulfhydrylase; SDS-PAGE, sodium
dodecyl sulfate-polyacrylamide gel electrophoresis; Hepes, N-(2-
hydroxyethyl)piperazine-N′-2-ethanesulfonic acid; HPLC, high-perfor-
mance liquid chromatography; WT, wild type; Tris, tris(hydroxymethyl)-
aminomethane; TNB, 5-thio-2-nitrobenzoate; R2?2-TS, tryptophan
synthase; AAT, aspartate aminotransferase.
Biochemistry 1996, 35, 13485-13493
S0006-2960(96)01517-6 CCC: $12.00© 1996 American Chemical Society
alanine using site-directed mutagenesis. Changes occurring
in binding of PLP, catalysis, and binding and release of
reactants, as well as possible structural differences at the
active site in the overall reaction of the mutant enzyme have
MATERIALS AND METHODS
Chemicals. 2-Bromoethylamine, ethylamine, and guani-
dinium chloride were purchased from ICN. O-Acetyl-L-
serine, PLP, L-alanine, L-cysteine, L-serine, Hepes, NaBH4,
pyridoxamine dihydrochloride, pyridoxal hydrochloride, R-N-
carbobenzoxy-L-lysine, L-glutamic acid, and CNBr were
purchased from Sigma. Endoproteinase Arg-C was from
Boehringer Mannheim Biochemicals. Sodium [3H]borohy-
dride (100 mCi, 9.5 Ci/mmol) was purchased from Amer-
Enzymes. Alkaline phosphatase and aspartate aminotrans-
ferase were obtained from Sigma.
Molecular Biology Reagents. Restriction enzymes were
purchased from Promega or USB. The DNA sequencing
was carried out using the kit from USB. For plasmid
purification, the Nucleobond AX kit (The Nest Group, Inc.)
was used.Oligonucleotides used for mutagenesis and
sequencing were prepared using a Biosearch oligonucleotide
Bacterial Strains and Plasmids. The bacterial strains used
in these experiments are Escherichia coli NM522 (hsd∆5,
∆(lac-pro), [F′, pro+, laclqZ∆M15]), Salmonella typhimu-
rium LB5000 (metA metE551 trpD2 leu hsdLT hsdA hsdB
and m+for all three modification systems), and Salmonella
typhimurium DW378 (trpC109 cysK1772 cysM1770). The
plasmid pRSM40 contains the S. typhimurium cysK gene on
a 1.48 ClaI-DraI fragment which was inserted in pT7T3
19U that had been digested with ClaI and SmaI (Monroe &
Kredich, 1990). The upstream end of cysK is nearest the
SmaI/DraI junction (Figure 1).
The plasmid, pRSM40, was linearized with EcoRI, and a
690 bp AVaI fragment containing the codon AAG for K42
was amplified using PCR. The oligonucleotide primers used
and primer 2 is complimentary to 5′-GCGCAGGCTTCAT-
CCCGGGCAACCT-3′ in the DNA sequence of cysK.
In oligonucleotide 1, AAG encoding for lysine was
changed as shown to GCG encoding for alanine. The
sequence CCCGAG is the recognition site for AVaI. In
oligonucleotide 2, the sequence CCCGGG is the recognition
site for AVaI.
The protocol for mutagenesis is shown in Figure 1B. The
ligated plasmids carrying the new GCG codon were trans-
formed into E. coli NM522 using the CaCl2 method
(Sambrook et al., 1989) and purified using the Nucleobond
AX kit. The complete gene was sequenced with a USB
sequencing kit to confirm the mutation and to check for PCR
errors. The mutant plasmids were transformed into S.
FIGURE 1: Procedure for site-directed mutagenesis. Panel A shows the map of the plasmid pRSM40 that carries the cysK gene. Panel B
shows the procedure for mutagenesis with the help of PCR.
13486 Biochemistry, Vol. 35, No. 41, 1996
Rege et al.
typhimurium LB5000 using a modified Hanahan method and
stored as phage P22 lysates (Monroe & Kredich, 1990). Since
LB5000 is r-m+for all three restriction-modification
systems, the DNA obtained from this strain is not degraded
in any other S. typhimurium strain.
Expression strains were constructed by transducing the
cysK-cysM-strain (DW378) of S. typhimurium with phase
P22 lysates containing mutant plasmids. Briefly, 2 µL of
the phage lysate was added to 100 µL of the DW378 culture
grown overnight at 37 °C with good aeration. The trans-
ductants were selected on LB plates with 100 µg/mL
ampicillin in an overnight incubation at 37 °C. The plasmid-
containing strain was then allowed to grow for 18 h at 37
°C and 250 rpm on Vogel-Bonner medium E (Vogel &
Bonner, 1956) supplemented with 0.5% glucose, 1% LB,
40 µM L-tryptophan, 500 µM reduced glutathione, and 100
µg/mL ampicillin. Reduced glutathione was the only sulfur
source in the medium and was added to derepress the cysteine
biosynthesis pathway (Kredich, 1971).
The mutant enzyme was purified by the method of Tai et
al. (1993). UV-visible absorption spectra were recorded
for each fraction of the Q-Sepharose and phenyl-Sepharose
columns. The fractions with an A280/424of 5.0 were pooled
and concentrated. The purity of the enzyme was further
tested via SDS-PAGE.
Preparation and Reconstitution of Apo-K42A-OASS. The
preparation of apo-K42A-OASS and its reconstitution were
performed using the method of Schnackerz and Cook (1995)
with the exception that OAS was not added prior to dialysis
against 5 M guanidinium chloride.
Modification of C43 with Bromoethylamine. To modify
K42-OASS with BEA, the holoenzyme was first incubated
with 5 M guanidinium chloride in 50 mM phosphate buffer,
pH 7.5, for 15 min, 40 mM 2-bromoethylamine was added
to the apo-K42A-OASS, and the preparation was incubated
for 14 h at room temperature. Reconstitution with PLP was
carried out as described (Schnackerz & Cook, 1995).
Synthesis of ?-N-Pyridoxyl-L-lysine. For the preparation
of ?-N-pyridoxyllysine, a mixture of 1.12 g of R-N-
carbobenzoxy-L-lysine (4 mmol) and 0.81 g of pyridoxal
hydrochloride (4 mmol) in 50 mL of CH3OH was incubated
with 0.8 g of sodium hydroxide (20 mmol) at 4 °C and stirred
under a nitrogen atmosphere until the reaction was complete
(30 min). Solid sodium borohydride was then added until
the yellow color of the solution disappeared. The mixture
was warmed to 25 °C, stirred for 1 h, brought to pH 6 with
concentrated HCl, and evaporated to dryness in Vacuo. The
protecting group on the R-nitrogen was removed by refluxing
with 6 N HCl for 2 h. After hydrolysis, the reaction mixture
was evaporated to dryness. The residue was dissolved in
H2O and applied to a cation exchange column (1.5 × 25
cm, Dowex 50 × 8, H+form, 200-400 mesh, Sigma), which
was washed with 200 mL of 1 N HCl and eluted with a 800
mL gradient of 1-5 N HCl. The fractions which absorb at
294 nm were combined and evaporated to dryness. Final
purification of the ?-pyridoxyllysine was accomplished using
a Beckman HPLC System Gold. Aliquots of the reaction
mixture were loaded onto an SP-5PW cation exchange
column (21.5 × 150 mm, Toso Haas) equilibrated with
solvent A (H2O, pH 2.5). The column was eluated by a 5
min linear gradient 0-7% solvent B (4 N NaCl, pH 2.5) at
a flow rate of 5 mL/min followed by a 60 min gradient of
7-10% solvent B at the same flow rate. The pyridoxyllysine
fractions were monitored at 294 nm, collected, and then
desalted using a size exclusion column (1.5 × 80 cm,
Sephadex G-10, Sigma).
Sodium [3H]Borohydride Reduction of OASS-A. Twenty-
four milligrams of OASS-A was reduced by the addition of
0.1 M NaBH4containing 100 mCi of sodium [3H]borohy-
dride in 20 mM Hepes, pH 8, until the yellow color
disappeared. The reduced OASS-A was hydrolyzed in 6 N
HCl, 110 °C, for 24 h. The tritiated pyridoxyllysine was
purified as described above.
CNBr and Endo Arg-C CleaVage. Twenty-four milligrams
of NaB3H4-reduced OASS-A was dialyzed 5 times against
1 L of 20 mM Hepes in order to remove excess NaB3H4.
The reduced protein was digested in 70% trifluoroacetic acid
with 205 mg of CNBr for 24 h, and the reaction was
quenched by the addition of 10 volumes of H2O and the
mixture lyophilized. Lyophilized CNBr fragments were
further digested at pH 8 for 2 days in 0.1 M NH4HCO3and
0.1% SDS by endoproteinase Arg-C with a weight ratio of
20 to 1, OASS to Endo Arg-C.
Purification of the Pyridoxyl Peptide. The pyridoxyl
peptide was isolated and purified using a Beckman System
Gold HPLC with a Model 171 radioisotope detector. Mobile
phases were as follows: solvent A consisted of 0.1% TFA
in HPLC grade H2O (Bio-Rad), while solvent B consisted
of 0.1% TFA in HPLC grade CH3CN. Aliquots of the digest
were loaded onto a C18reversed phase column (4.6 × 250
mm, Beckman) equilibrated with solvent A. The column
was eluted with solvent A for 5 min followed by a 90 min
linear gradient of 0-60% solvent B at a flow rate of 1 mL/
min. Elution of peptide was simultaneously monitored using
radioactivity and the absorbance at 215 nm.
Degradation of Pyridoxyllysine to Pyridoxamine. The
pyridoxyllysine (20 mg) was dissolved in 0.5 mL of water
at 4 °C. Argon was passed through the solution which was
then treated with 0.06 mL of 1 N NaOH followed by 0.05
mL of 5% sodium hypochlorite. After 15 min, the solution
was added dropwise to a flask of boiling water (15 mL)
through which argon was bubbling. The solution was heated
for 10 min, then cooled to 4 °C, neutralized with 0.1 N HCl,
and evaporated to dryness in vacuo.
dissolved in water and passed over a cation exchange HPLC
column (21.5 × 150 mm, Toso Haas) using the same gradient
and flow rate as used for the purification of pyridoxyllysine.
The pyridoxamine was finally desalted using a size exclusion
column (1.5 × 80 cm, Sephadex G-10, Sigma).
Stereochemical Analysis of [3H]Pyridoxamine. In a typical
analysis, 0.5 mg of the tritiated pyridoxamine was added to
0.3 mL of 20 mM Tris buffer, pH 8.1, and combined with
10 mL of 3.4 mM R-ketoglutarate and 10 mL of 0.15 M
monosodium glutamate (Miles et al., 1982). The pH of the
solution was adjusted to 8 with 1 M potassium carbonate,
pH 8.7; 1.5 mg of apoasparate aminotransferase was added
in a closed vessel and incubated for 3 days at 37 °C, and
subsequently distilled to obtain the3H2O.
Spectral Studies. UV-visible spectra were measured on
a Hewlett-Packard, Model 8452A, photodiode array spec-
trophotometer. Spectra of apo-K42-OASS reconstituted with
PLP in 10 mM Hepes, pH 7.5, and different concentrations
of L-alanine, L-cysteine, O-acetyl-L-serine, or L-serine were
recorded at wavelengths from 250 to 600 nm using a 1 cm
light path. Both buffer and amino acid blanks were
The residue was
Function of the OASS Schiff Base Lysine
Biochemistry, Vol. 35, No. 41, 1996 13487
subtracted from the spectra. Data were plotted using Cricket
Fluorescence Studies. Fluorescence spectra of 2.5 µM
mutant enzyme were obtained on a Shimadzu RF5000U
spectrofluorometer in the absence and presence of amino
acids at 25 °C. Excitation was at 298 nm, and the excitation
slit widths were set at 5 nm. Emission was measured over
the wavelength range 300-550 nm.
Circular Dichroism Studies. CD spectra were collected
on an Aviv 62DS spectropolarimeter at 25 °C with a path
length of 0.2 cm. Enzyme concentrations of 100 µg/mL and
4 mg/mL were used for far-UV and visible CD spectra,
respectively. The buffer used for all spectra was 10 mM
phosphate, pH 7.0, and a buffer blank was subtracted from
Enzyme Assays. The activity of PLP-reconstituted apo-
K42A-OASS and the PLP-reconstituted bromoethylamine-
modified apo-K42A-OASS was monitored using the follow-
ing two assay methods. In the normal reaction, i.e.,
formation of L-cysteine from OAS and sulfide, the decrease
in sulfide concentration was observed using a computer-
assisted sulfide ion-selective electrode assay (Hara et al.,
1990). In addition, the formation of S-(3-carboxy-4-nitro-
phenyl)-L-cysteine (Tai et al., 1993) was measured by
monitoring the disappearance of TNB using a Gilford 2600
spectrophotometer at 412 nm (Ellman, 1959).
31P NMR Spectroscopy. Fourier transform
spectra were collected at 121.497 MHz on a Bruker AM300
SWB superconducting spectrometer using a 10-mm multi-
nuclear probehead with broadband1H decoupling. The NMR
tube spinning at 15-20 Hz contained the sample (2 mL)
and2H2O (0.2 mL) as field/frequency lock and was main-
tained at 20 ( 0.1 °C using a thermostated continuous air
flow. A spectral width of 2000 Hz was acquired in 8K data
points with a pulse angle of 60°. The exponential line
broadening used prior to Fourier transformation was 10 Hz.
Protein samples were dissolved in 50 mM Mes or Hepes
buffers containing 1 mM EDTA at the appropriate pH.
Changes in pH were performed by dialysis against the desired
buffer overnight. pH valves of the sample were determined
before and after the NMR measurement. Positive chemical
shifts in ppm are downfield changes with respect to 85%
ActiVe Site Sequence. The main radiolabeled peptide from
the elution profile of the Endo Arg-C/CNBr-cleaved OASS-A
containing the active site lysine was subjected to solid phase
sequencing using an Applied Biosystem gas phase protein
sequencer. The peptide is nine amino acids long and contains
>65% of the radioactivity in position 7. The sequence is
as follows: AsnProSerPheSerValLys(Pxy)CysArg.
above sequence exactly matches residues 36-44 of that
predicted based on the nucleotide sequence (Levy &
Danchin, 1988; Byrne et al., 1988), but is not the one
predicted based on the homology to other PLP enzymes
(Levy & Danchin, 1988). The cysteine immediately C-
terminal to the active lysine is the only one present in the
polypeptide. The identification of the pyridoxyl peptide and
the fact that the K42A mutation results in an inactive enzyme
indicate that lysine-42 is the enzyme group covalently bound
to the coenzyme PLP.
Stereochemical Analysis at C4′. To determine which face
of the internal Schiff base is reduced by sodium [3H]-
borohydride, the tritiated protein was hydrolyzed to [3H]-
pyridoxyllysine, which was subjected to hypochlorite oxi-
dation to produce [3H]pyridoxamine. Analysis of the tritium
distribution between the two diastereotopic hydrogens at C4′
of pyridoxamine made use of the stereospecificity of apoas-
partate aminotransferase for removal of the 4′-pro-S hydrogen
(Dunathan et al., 1968). The tritium from the 4′-pro-S
position is released into solution, while the tritium from the
4′-pro-R position is retained.
generated by reduction of the internal Schiff base with
sodium [3H]borohydride retained most of its tritium after
incubation with apoaspartate aminotransferase.
30 000 cpm added to the apoaspartate aminotransferase
incubation, 3280 cpm were released to solvent compared to
2245 for the minus apoenzyme control. These results agree
with the hypothesis put forth by Dunathan (Dunathan, 1971;
Dunathan & Voet, 1974) that a single surface (re face in
this case) of the active site PLP is accessible to solvent and
Growth and Yield of the Mutant Protein. When cells were
grown in shaker flasks overnight, the K42A mutant protein
was obtained in a yield of 20 mg/25 g of wet cell paste. The
mutant protein, as isolated, has λmaxvalues of 280 and 424
nm (Figure 2A) and a small amount of activity. Addition
of NaBH4 does not affect the absorbance at 424 nm, but
eliminates all of the activity. Preparation of larger quantities
of the K42A mutant was attempted by overnight growth in
a 12 L fermenter with forced aeration. After a lag of several
hours, growth was observed, and cells were harvested after
24 h. O-Acetylserine sulfhydrylase was isolated and found
to be wild type OASS-A.
Spectral Properties. WT OASS-A, as isolated, shows an
absorbance spectrum with two maxima, at 280 and 412 nm
FIGURE 2: Panel A shows UV/visible spectra of OASS-A (I) and
K42A-OASS (II) of equal concentration. Panel A shows UV/visible
spectra of OASS-A (I) and K42A-OASS (II) of equal concentration.
Panel B shows apo-K42A-OASS (I) and apo-K42A-OASS recon-
stituted with PLP (II).
13488 Biochemistry, Vol. 35, No. 41, 1996
Rege et al.
(Figure 2A, I), the latter corresponding to the internal Schiff
base (Cook et al., 1992). The ratio of A280/424is about 5.0
for the K42A mutant protein, compared to a value for A280/412
of 3.4 for WT OASS-A. Addition of sodium borohydride
alone has no effect on the 424 nm band of the K42A mutant
protein as isolated, but in the presence of 5 M guanidinium
chloride the absorbance at 424 nm disappears. The absorp-
tion spectrum of apo-K42A-OASS exhibits a single maxi-
mum at 280 nm (Figure 2B, I). Reconstitution of apo-K42A-
OASS with PLP produces additional absorption bands at 330
and 390 nm (Figure 2B, II). The ratio of A280/A(330+390)is
When excited at 298 nm, the fluorescence emission
spectrum of WT OASS-A shows two maxima at 337 and
498 nm; the ratio of A337/498is 8.0 (Figure 3, I). The band
at the shorter wavelength is a result of intrinsic tryptophan
fluorescence whereas the band at the longer wavelength is
due to delayed Schiff base fluorescence (McClure & Cook,
1994; Strambini et al., 1996). Excitation of the K42A mutant
protein at 298 nm shows an emission spectrum, with maxima
at 337 and 504 nm (Figure 3, II). The ratio for the K42A
mutant (A337/504) is around 4.5, which is a value between those
observed for the WT enzyme in the absence and presence
of acetate (Figure 3, III; McClure & Cook, 1994). Apo-
K42A-OASS shows no emission around 500 nm (data not
shown) as also seen for WT apo-OASS-A, and the emission
at 337 nm is increased in the apo-K42A mutant, when
compared to K42A-OASS.
The emission spectrum of apo-K42A-OASS reconstituted
with PLP is not identical to that of K42A-OASS, when
excited at 298 nm. The emission band at 500 nm has a lower
intensity than that of K42A-OASS as isolated. When excited
at 330 nm, the WT enzyme shows a major band with a λmax
of 484 nm and a weaker band with a λmax of 362 nm
(Strambini et al., 1996). Two bands are also observed in
the K42A mutant reconstituted with PLP, but λmax values
appear at 444 and 387 nm, respectively (data not shown).
Circular dichroism spectra for K42A-OASS as isolated are
identical to those of the WT enzyme in the far-UV (data not
shown), suggesting essentially no gross change in the
structure of the OASS as a result of the active site mutation.
In the visible region, however, K42A-OASS shows a
negative Cotton effect with λmax at 424 nm and a molar
ellipticity equal to that of the positive Cotton effect of the
WT enzyme at 412 nm (Figure 4). Apoenzyme shows no
ellipticity in the visible range whereas enzyme reconstituted
with PLP shows positive ellipticity centered around 330 and
390-400 nm (data not shown).
resonance at 5.3 ppm in the pH range 6.5-8.5 (data not
shown). This can be compared to a value of 5.2 ppm for
the WT enzyme (Cook et al., 1992).
Formation of External Aldimine. Apo-K42A-OASS re-
constituted with PLP exhibits spectral changes upon addition
of amino acid reactants or analogs. Addition of 20 mM OAS
to the reconstituted enzyme leads to a time-dependent
decrease in the absorbance at 330 nm and a shift in the λmax
of the 390 nm band to 424 nm with a concomitant increase
in the absorbance of the latter (Figure 5, A). The reaction
is pH-independent. Formation of the external aldimine with
OAS is complete in about 90 min. L-Alanine and L-serine
produce similar changes over the same time period and show
saturation at 10 mM concentration. L-Cysteine at a concen-
tration of 15 mM produces an increase in the absorbance at
330 nm with a concomitant decrease at 390 nm (Figure 5,
B). A first-order plot of the absorbance change for the
formation of the L-Ala external Schiff base at 310 or 370
nm (wavelengths that produce maximum absorbance changes)
vs time gives a maximum rate constant of 0.048 min-1.
Similar data for OAS, serine, and cysteine give first-order
rate constants in the range of 0.035 or 0.04 min-1. The
31P NMR spectra of K42A-OASS show a single
FIGURE 3: Fluorescence spectra of OASS-A (I), K42A-OASS (II),
and addition of acetate to OASS-A (III). Each of these spectra
was collected at 25 °C in 10 mM Hepes, pH 7.5, with an enzyme
concentration of 2.5 µM for OASS-A and K42A-OASS.
FIGURE 4: Visible circular dichroism spectra of OASS-A (I) and
K42A-OASS (II). These spectra were taken at 25 °C with 10 mM
phosphate buffer, pH 7.5. The enzyme concentration used was 500
µg for both the spectra.
FIGURE 5: UV/visible difference spectra for external aldimine
formation. Each of these spectra was taken at 25 °C with 1 mg of
apo-K42A-OASS reconstituted with PLP in the absence and
presence of 10 mM OAS (A) or 10 mM L-cysteine (B). Spectra
are the difference of those in the presence and absence of amino
acid. The reactions were carried out in 10 mM Hepes, pH 7.5.
Function of the OASS Schiff Base Lysine
Biochemistry, Vol. 35, No. 41, 1996 13489
external aldimine species studied above are stable for at least
24 h. In the above experiments, no species absorbing at or
near 470 nm, corresponding to the R-aminoacrylate inter-
mediate, was formed (Cook et al., 1992), suggesting that the
?-amino group of lysine is important in the reaction.
Addition of any of the above amino acids to the recon-
stituted enzyme gives a decrease in the fluorescence at 337
nm and an increase in the fluorescence at 500 nm, with the
λmax at longer wavelength showing a blue shift with time
(data not shown). The rates of the changes are similar to
those obtained from UV/visible spectrophotometry. Forma-
tion of the external aldimine with OAS, L-Ala, or L-Cys was
also studied using circular dichroism. Addition of OAS and
L-Ala to the apoenzyme reconstituted with PLP shows a
decrease in the ellipticity around 390 nm. Addition of L-Cys
to the reconstituted apoenzyme, on the other hand, shows a
prominent decrease in the ellipticity at 330 nm.
Formation of R-Aminoacrylate. Addition of OAS to the
WT enzyme results in the formation of the R-aminoacrylate
intermediate accompanied by an absorbance decrease at 412
nm and increases at 330 and 470 nm (Cook et al., 1992).
The K42A mutant enzyme, when preincubated with 10 mM
ethylamine (Figure 6A), shows formation of the external
aldimine upon addition of OAS, followed by slow formation
of the R-aminoacrylate intermediate absorbing at 330 and
470 nm (Figure 6B). A maximum first-order rate constant
of 0.03 min-1is obtained from the time dependence of the
absorbance change at either 310 or 470 nm. Addition of
sulfide results in an increase in absorbance at 330 nm and a
decrease in the R-aminoacrylate species that absorbs at 470
nm (Figure 6C), giving a spectrum identical to that obtained
upon addition of L-Cys to the PLP-reconstituted apo-K42A-
OASS. Addition of an increasing concentration of ethyl-
amine gave no change in the species at 330 nm. The
resulting 330 nm species can be reduced by NaBH4, giving
an increase in the absorbance at 330 nm due to the presence
Circular dichroism experiments performed to study the
formation of the external aldimine made use of the apoen-
zyme reconstituted with PLP after addition of 10 mM
ethylamine. Addition of OAS to this enzyme species resulted
in CD maxima around 330 and 430 nm with negative
ellipticity.Addition of L-cysteine and L-alanine to the
enzyme treated with 10 mM ethylamine showed no change
in the spectrum.
Rates of L-Cysteine Formation. K42A-OASS as isolated
and PLP-reconstituted apo-K42A-OASS do not catalyze
either the overall reaction (formation of cysteine, with the
natural substrates, OAS, and sulfide) the deacetylase activity
(formation of pyruvate, ammonia, and acetate). As discussed
above, a small amount of contaminating WT activity was
detected in the K42A mutant enzyme as isolated, but it was
eliminated by borohydride reduction with no change in the
spectrum of the K42A mutant protein. Since the ?-amino
group of lysine is absent in the K42A mutant enzyme, two
different methods were attempted to restore activity, i.e.,
addition of ethylamine to the enzyme and chemical modi-
fication of C43 with bromoethylamine to form γ-thialysine
at position 43. Addition of 10 mM ethylamine to apo-K42A-
OASS reconstituted with PLP results in an absorbance shift
from 394 to 398-400 nm, suggesting the formation of an
external aldimine (Figure 6A). When apo-K42A-OASS
reconstituted with PLP is preincubated with ethylamine, the
rate of R-aminoacrylate formation from OAS (first half-
reaction) is slow compared to that with the wild type OASS-
A. Addition of sulfide to the R-aminoacrylate intermediate
causes a disappearance in the absorbance at 470 nm and at
400 nm with an increase in the absorbance at 330 nm
resulting from the cysteine external Schiff base (Figure 6C).
Using 9 µM apo-K42A-OASS reconstituted with PLP, a
single turnover is observed with the catalytic cycle stopping
at the L-cysteine external Schiff base. The maximum first-
order rate of formation of the cysteine external Schiff base
is (3.0 ( 0.4) × 10-3s-1, 105-fold lower than V/Etfor the
WT enzyme (Tai et al., 1995). Apparent K values for OAS
and sulfide in the formation of the cysteine external Schiff
base are 10 mM and 10 µM, respectively, compared to Km
values of 1 mM and 6 µM, respectively, measured for the
rate of formation of cysteine by the WT enzyme. In the
second case, C43, located next to the active site lysine, was
FIGURE 6: Panel A shows the UV/visible spectra of apo-K42A-
OASS reconstituted with PLP (I) and formation of external aldimine
in the presence of 10 mM ethylamine (II). The enzyme concentra-
tion used was 500 µg. Panels B and C show the difference between
the spectrum in the presence of amino acid and that of free enzyme
as a function of time. The reaction is that catalyzed by apo-K42A-
OASS reconstituted with PLP in the presence of 10 mM ethylamine.
Panel B shows, first, the formation of the external aldimine (I),
followed by the formation of R-aminoacrylate (II) external Schiff
base. Times are 5, 15, 30, 60, and 90 min from top to bottom at
400 nm. Panel C shows the formation of L-Cys after addition of
sulfide to the reaction in panel B. Times are 5, 15, 30, and 60 min
from top to bottom at 400 nm. The reaction was carried out at 25
°C with 10 mM OAS and 10 µM sulfide in 10 mM Hepes, pH 7.5.
The enzyme concentration used for the reactions was 1 mg.
13490 Biochemistry, Vol. 35, No. 41, 1996
Rege et al.
modified with bromoethylamine, giving γ-thialysine which
forms an internal Schiff base with an absorbance at 398 nm
similar to that of K42A-OASS after addition of ethylamine.
Reduction of the γ-thialysine OASS with sodium borohy-
dride results in a bleaching of the 400 nm absorbance and
an increase in the absorbance at 320 nm. Denaturation with
4 M guanidinium chloride and dialysis produce no change
in the 320 nm absorbance. Again, a single turnover appears
to occur, and rate data are essentially identical to those
obtained with ethylamine. The alternative substrate, TNB,
gave no discernible rate with the reconstituted mutant
Location of the Schiff Base Lysine. Based on the data
obtained, a mixture of peptides was subjected to sequencing.
The sequence of one of the peptides was readily acquired
due to the constant yield per cycle. When compared to the
amino acid sequence derived from the gene sequence, the
radioactively labeled peptide is identical to positions 36-
44 of the N-terminal sequence of OASS-A (Byrne et al.,
1988) (Figure 7). Most of the radioactivity was found at
position 42, indicating that this is the location of the Schiff
K42A Mutant Enzyme. (A) Growth. Surprisingly, growth
under conditions of forced aeration produced the wild type
enzyme. However, the background strain, DW378, is cysK-,
cysM-as a result of point mutations, eliminating the activity
of OASS-A and OASS-B (Dreyfuss & Monty, 1962). In
addition, DW378 is a recombinant positive strain. As a result
of the selective pressure of growth in the absence of cysteine,
homologous recombination likely occurs between the cysK
gene on the plasmid-carrying mutation K42A and the
chromosomal cysK gene to produce the WT enzyme. As
stated under Results, K42A mutant protein could be obtained
in good yield only by overnight growth in shaker culture.
(B) Spectral Properties of K42A-OASS in the Absence of
Amino Acids. The K42A mutant enzyme, as isolated, shows
absorbance maxima at 280 and 424 nm, the latter corre-
sponding to the external aldimine of PLP with a mixture of
free methionine and leucine based on an amino acid analysis
of isolated K42A that was heat-denatured and dialyzed.
Similar spectra have been observed for the external Schiff
bases of OASS with L-cysteine3or L-serine (Schnackerz et
al., 1995). The external aldimine of K42A-OASS as isolated
cannot be reduced with sodium borohydride under conditions
normally used for the reduction of internal and/or external
aldimines of PLP. Only in the presence of 5 M guanidinium
chloride is the reduction with borohydride feasible, indicating
that the external aldimine of K42A-OASS as isolated may
be in a closed conformation. It is very difficult to remove
PLP from K42A-OASS, even though the mutant enzyme
lacks Lys-42, which in WT OASS-A binds PLP covalently
via an internal aldimine linkage. The presence of 5 M
guanidinium chloride is necessary to open the PLP binding
site to facilitate resolution of the cofactor to produce apo-
K42A-OASS. Reconstitution of the apoenzyme with PLP
results in two absorption bands at 330 and 390 nm. Free
PLP absorbs at 388 nm with a shoulder at 325 nm at neutral
pH (Peterson & Sober, 1954). The spectrum of reconstituted
apo K42A-OASS is very similar to that of free PLP at neutral
pH, confirming the presence of the free aldehyde at the active
The fluorescence emission spectrum of WT OASS exhibits
maxima at 337 and 500 nm when excited at 298 nm
(McClure & Cook, 1994; Strambini et al., 1996). The 337
nm band results largely from intrinsic tryptophan fluores-
cence, while the 500 nm band results from delayed fluores-
cence of the internal Schiff base (Strambini et al., 1996).
Addition of L-cysteine to WT OASS to form the external
Schiff base3(or acetate which binds to the R-carboxylate
subsite, mimicking cysteine) results in a significant enhance-
ment of the long-wavelength band a blue shift in its λmaxto
490 nm (McClure & Cook, 1994; Schnackerz et al., 1995;
Strambini et al., 1996). The fluorescence spectrum of K42A-
OASS, as isolated, also shows maxima at 337 and 504 nm,
when excited at 298 nm. The ratio of A337/504 for K42A-
OASS as isolated is about halfway between that of the WT
OASS and the cysteine external Schiff base, suggesting the
presence of about 50% occupancy of K42A with an external
aldimine between PLP and a free amino acid (methionine
and leucine as indicated under Results). Apo-K42A-OASS,
on the other hand, shows fluorescence emission only at 337
nm, in agreement with results on WT apo-OASS (McClure
& Cook, 1994). In apo-K42A-OASS, the fluorescence
emission at 337 nm is increased when compared to K42A-
OASS, likely the result of quenching upon binding of PLP
to the apoenzyme, as is also observed for WT OASS and its
apoenzyme (McClure & Cook, 1994; Strambini et al., 1996).
Similar results are found for D-serine dehydratase (Schnack-
erz et al., 1973; Federiuk & Shafer, 1983), tryptophanase
(Tokushige et al., 1980), and tryptophan synthase (Strambini
et al., 1992). The emission band of PLP-reconstituted apo-
K42A-OASS at 500 nm has a lower intensity than K42A-
OASS, as isolated, when excited at 298 nm. Reconstituted
apo-K42A-OASS exhibits two emission bands at 444 and
387 nm, respectively, when excited at 330 nm, qualitatively
similar to WT OASS-A which shows maxima at 484 and
362 nm, respectively. For the WT OASS, which exists as
an internal aldimine, the major band centered at 484 nm is
characteristic of a ketoenamine tautomer, whereas the weaker
band around 362 nm is typical for an enolimine tautomer
(Strambini et al., 1996). In the case of the PLP-reconstituted
2The 418 nm tautomer observed for wild-type enzyme is actually
that of S-(3-L-alanyl)-L-cysteine with the active site PLP, resulting from
attack of the ?-thiol of L-cysteine on the R-aminoacrylate intermediate
(Woehl et al., 1996).
3The addition of L-cysteine actually results in a Schiff base between
S-(3-L-alanyl)-L-cysteine (Woehl et al., 1996).
FIGURE 7: Deduced amino acid sequence of O-acetylserine sulf-
hydrylase A from Salmonella typhimurium (Bryne et al., 1988; Levy
& Danchin, 1988).
Function of the OASS Schiff Base Lysine
Biochemistry, Vol. 35, No. 41, 1996 13491
apo-K42A-OASS, the 444 nm band may represent the
ketoenol form of free PLP, while the weaker 387 nm band
may represent the enolaldehyde form of free PLP. The
fluorescence data support the presence of an external
aldimine in K42A-OASS as isolated.
The CD spectrum of K42A-OASS, as isolated, shows in
the visible region a negative Cotton band with a maximum
at 424 nm centered on the visible absorbance band. This
can be compared to a visible CD band at 412 nm for the
WT enzyme, also centered on its visible absorbance band
(Schnackerz et al., 1995). Addition of L-cysteine to the WT
enzyme causes a red shift in both the visible absorbance and
CD bands. The visible CD band has a positive sign and is
of equal intensity to that observed in the absence of L-cysteine
(Schnackerz et al., 1995). The K42A-OASS as isolated
exists as a mixture of methionine and leucine Schiff bases.
Since these external Schiff bases have a visible Cotton band
opposite in sign to that of the WT cysteine external Schiff
base, there are two possible explanations. The difference
in chemical structure of Met, Leu, and Cys is responsible
for the opposite sign, or the opposite face of the external
Schiff base interacts with the protein surface in the cysteine
external Schiff base compared to the Met/Leu external Schiff
bases. Since the K42A-OASS as isolated cannot be reduced,
these possibilities cannot be distinguished at this time.4In
the PLP-reconstituted apo-K42A-OASS, CD bands with
positive ellipticity around 330 and 390-400 nm (centered
on the visible absorbance bands) are observed. The31P
signal of the internal aldimine of the WT OASS is 5.2 ppm,
independent of pH (Cook et al., 1992), while that of the
cysteine external Schiff base is shifted slightly downfield to
5.3 ppm (Schnackerz et al., 1995). The31P NMR spectra
of K42A-OASS, as isolated, show a single resonance at 5.3
ppm, consistent with the occurrence of an external aldimine
in K42A-OASS as isolated.
The Schiff base lysine has been replaced in three other
PLP-dependent enzymes: K258A in aspartate aminotrans-
ferase (Toney & Kirsch, 1993), K145A in D-amino acid
aminotransferase (Nishimura et al., 1991), and K87T in R2?2-
tryptophan synthase (Miles et al., 1989). The most closely
related is K87T in R2?2-tryptophan synthase which is isolated
as an external aldimine with free L-serine. The spectral
properties of the apo-K87T-tryptophan synthase reconstituted
with PLP are identical to those of the apo-K42A-OASS
reconstituted with PLP. The Schiff base lysine mutants of
the aminotransferases, however, are isolated with the free
aldehyde of PLP bound.
(C) Spectral Properties of K42A-OASS in the Presence of
Amino Acids. The external Schiff base forms slowly upon
addition of OAS, L-Cys, L-Ser, or L-Ala to the PLP-
reconstituted mutant enzyme. The final equilibrium mixture
of species is similar to that observed for the WT OASS in
the case of L-Ser, that is a mixture of tautomers with λmax
values at 330 and 418 nm, but differs in the case of L-Cys
with the 330 nm tautomer observed for K42A-OASS
compared to a 418 nm tautomer observed for the wild type
enzyme (Schnackerz et al., 1995). The slow formation of
the external Schiff base was also observed for K87T R2?2-
TS (Lu et al., 1993) and K258A-AAT (Toney & Kirsch,
(D) Regeneration of ActiVity. All amino acids tested,
OAS, L-cysteine, L-serine, and L-alanine, with PLP-recon-
stituted apo-K42A-OASS form an external Schiff base slowly
(0.04 min-1) compared to WT OASS (700-1000 s-1; Woehl
et al., 1996). Cordes and Jencks (1962) demonstrated that
the rate constants for reactions in solution of imines of PLP
with semicarbazide are greater than those for the parent
aldehyde alone with semicarbazide. The present results are
consistent with the suggestion of Cordes and Jencks that the
rate constant for formation of the extenral aldimine from the
WT OASS internal aldimine should be greater than that for
enzyme-bound free aldehyde of PLP. Thus, significant rate
enhancement is realized as a result of K42 allowing tran-
simination to occur in the catalytic cycle.
Addition of 10 mM ethylamine to reconstituted apo-K42A-
OASS in the presence of OAS shows the formation of
external aldimine with λmaxat 398 nm, followed by the slow
formation of the R-aminoacrylate intermediate. Addition of
sulfide results in an increase in the absorbance at 330 nm
and a decrease in the R-aminoacrylate intermediate, giving
a spectrum identical to that obtained upon addition of
cysteine to reconstituted apo-K42A-OASS. Similar spectral
changes in the presence of OAS were observed for K42A-
OASS modified at position 43 to a γ-thialysine.5
restored activity is 105-fold lower than that of WT enzyme,
and only one turnover is carried out in both half-reactions.
The spectral changes observed suggest that the reaction is
terminated with the formation of the cysteine external Schiff
base and the cysteine cannot be displaced, a prerequisite to
start a new cycle. Data are consistent with the above
conclusion that K42 is important for transimination, not only
to form the external Schiff base, for example, with OAS,
but also to form the internal Schiff base and release the amino
acid product, for example, cysteine. The lack of ability of
ethylamine to displace cysteine may be entropic, since
ethylamine lacks the advantage of being locked into place,6
or may be geometric resulting from ethylamine being bound
at the active site in such a way that it is unable to carry out
the displacement effectively. Planas and Kirsch (1991) have
suggested that the decreased basicity of the γ-thialysine group
in K258γ-thialysine-AAT is considered to be principally
responsible for the 14-fold lower Vmax value compared to
WT AAT.The 105-fold lower activity of the C43-γ-
thialysine derivative of K42A is likely a result of geometric
considerations, since C43 is one amino acid removed from
the normal Schiff base lysine position.5
In conclusion, lysine-42 of OASS-A facilitates the forma-
tion and dissociation of the OAS and L-cysteine external
Schiff bases, allowing more facile transimination. It also
participates as a general base in the first half-reaction,
abstracting the R-proton of OAS, and as a general acid in
the second half-reaction, donating a proton to the R-carbon.
4The three-dimensional structure of OASS-A has recently been
solved (P. Burkhard, E. Hohenester, G. S. J. Rao, P. F. Cook, and J.
N. Jansonius, unpublished results), and crystallization trials are now
in progress on the K42A-OASS as isolated.
5The observed restoration of activity upon bromoethylamine treat-
ment may also result from the presence of bromoethylamine acting in
the same manner as ethylamine. The similarity in kinetic data obtained
for apo-K42A-OASS reconstituted with PLP in the presence of
ethylamine or after bromoethylamine treatment is consistent with this
alternate interpretation. In addition, the crystal structure of WT OASS
suggests C43 is behind the active site, inaccessible to bromoethylamine.
6We thank the reviewer for this suggestion.
13492 Biochemistry, Vol. 35, No. 41, 1996
Rege et al.
REFERENCES Download full-text
Alexander, F. W., Sandmeier, E., Mehta, P. K., & Christen, P.
(1994) Eur. J. Biochem. 219, 953-960.
Becker, M. A., Kredich, N. M., & Tomkins, G. M. (1969) J. Biol.
Chem. 244, 2418-2427.
Byrne, C. R., Monroe, R. S., Ward, K. A., & Kredich, N. M. (1988)
J. Bacteriol. 170, 3150-3157.
Cook, P. F., & Wedding, R. T. (1976) J. Biol. Chem. 251, 2033-
Cook, P. F., Hara, S., Nalabolu, S. R., & Schnackerz, K. D. (1992)
Biochemistry 31, 2298-2303.
Cordes, E. H., & Jencks, W. P. (1962) J. Am. Chem. Soc. 84, 832-
Dreyfuss, J., & Monty, K. J. (1962) J. Biol. Chem. 238, 1019-
Dunathan, H. C. (1971) AdV. Enzymol. Relat. Areas Mol. Biol. 35,
Dunathan, H. C., & Voet, J. G. (1974) Proc. Natl. Acad. Sci. U.S.A.
Dunathan, H. C., Davis, L., Kury, P. G., & Kaplan, M. (1968)
Biochemistry 7, 4532-4537.
Ellman, G. L. (1959) Arch. Biochem. Biophys. 82, 70-77.
Federiuk, C. S., & Shafer, J. A. (1983) J. Biol. Chem. 258(9), 5372-
Hara, S., Payne, M. A. Schnackerz, K. D., & Cook, P. F. (1990)
Protein Expression Purif. 1, 70-76.
Kredich, N. M. (1971) J. Biol. Chem. 246, 3474-3484.
Kredich, N. M., & Tomkins, G. M. (1966) J. Biol. Chem. 241,
Levy, S., & Danchin, A. (1988) Mol. Microbiol. 2, 777-783.
Lu, Z., Nagata, S., McPhie, P., & Miles, E. W. (1993) J. Biol. Chem.
McClure, G. D., Jr., & Cook, P. F. (1994) Biochemistry 33, 1674-
Miles, E. W., Hauch, D. R., & Floss, H. G. (1982) J. Biol. Chem.
Miles, E. W., Kawasaki, H., Ahmed, S. A., Morita, H., Morita, H.,
& Nagata, S. (1989) J. Biol. Chem. 264, 6280-6287.
Monroe, R. S., & Kredich, N. M. (1990) J. Bacteriol. 172(12),
Nishimura, K., Tanazawa, K., Yoshimura, T., Esaki, N., Futaki,
S., Manning, J. M., & Soda, K. (1991) Biochemistry 30, 4072-
Peterson, E. A., & Sober, H. A. (1954) J. Am. Chem. Soc. 76, 169.
Planas, A., & Kirsch, J. F. (1991) Biochemistry 30(33), 8268-
Sambrook, J., Fritsch, E. F., & Maniatis, T. (1989) Molecular
Cloning: A Laboratory Manual, pp 1.82-1.84, Cold Spring
Harbor Laboratory Press, Cold Spring Harbor, NY.
Schnackerz, K. D., & Cook, P. F. (1995) Arch. Biochem. Biophys.
Schnackerz, K. D., Erlich, J. H., Wipf, A., & Wirths, G. G. (1973)
Z. Physiol. Chem. 354, 1241.
Schnackerz, K. D., Tai, C.-H., Simmons, J. W., III, Jacobson, T.
M., Rao, G. S. J., & Cook, P. F. (1995) Biochemistry 34, 12151-
Strambini, G., Cioni, P., Peracchi, A., & Mozzarelli, A. (1992)
Biochemistry 31, 7527-7534.
Strambini, G., Cioni, P., & Cook, P. F. (1996) Biochemistry 35,
Tai, C.-H., Nalabolu, S. N., Jacobson, T. M., Minter, D. E., & Cook,
P. F. (1993) Biochemistry 32, 6433-6442.
Tai, C.-H., Nalabolu, S. R., Simmons, J. W., Jacobson, T. M., &
Cook, P. F. (1995) Biochemistry 34, 12311-12325.
Tokushige, M., Iinuma, K., Yamamoto, M., & Nishijima, Y. (1980)
Biochem. Biophys. Res. Commun. 96, 863-869.
Toney, M. D., & Kirsch, J. F. (1993) Biochemistry 32, 1471-1479.
Vogel, H. J., & Bonner, D. M. (1956), J. Biol. Chem. 218, 97-
Woehl, E. U., Tai, C.-H., Dunn, M. F., & Cook, P. F. (1996)
Biochemistry 35, 4776-4783.
Function of the OASS Schiff Base Lysine
Biochemistry, Vol. 35, No. 41, 1996 13493