ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS

Vol. 291, No. 2, December, pp. 326-333, 1991

1 -Methyl-DL-tryptophan, P-(3-Benzofuranyl)-DL-alanine
(the Oxygen Analog of Tryptophan), and ,&[3-
Benzo(b)thienyl]-kalanine (the Sulfur Analog
of Tryptophan) Are Competitive Inhibitors
for lndoleamine 2,3-Dioxygenase’

Susan G. Cady2 and Masanori Sono

Department of Chemistry and Biochemistry, University of South Carolina, Columbia, South Carolina 29208

Received May 6, 1991, and in revised form August 26,199l

structural requirement for Trp to be metabolized by the

1-methyl-DL-Trp, B-(3-benzofuranyl)-DL-alanine (the enzyme. The inability of 1-methyl-Trp to serve as a sub-
oxygen analog of Trp), and 8-[3-benzo(b)thienyll-DL-al- strate for the dioxygenase supports a view that singlet
anine (the sulfur analog of Trp), each of which has a sub- oxygen is not the reactive oxygen species involved in the
stitution at the indole nitrogen atom, were found to be dioxygenation of Trp by the enzyme. Q 11~1 academic
the first examples of potent substrate analog competitive Pra, Inc.

inhibitors (Ki 7-70 PM) with respect to the substrates D-

Trp and I.-Trp for rabbit small intestinal indoleamine
2,3-dioxygenase. Binding studies using optical absorp-
tion and CD spectroscopy demonstrated that these three Indoleamine 2,3dioxygenase is a monomeric heme
inhibitors cause spectral changes upon binding to the na- protein that catalyzes the oxidative cleavage of the pyrrole
tive ferric, ferrous, ferrous-CO, and ferrous-NO en- ring of the indole nucleus of tryptophan (Tip)4 upon the
zymes. Such spectral effects of 1-methyl-DL-Trp on all insertion of two oxygen atoms of molecular oxygen to
of these enzyme derivatives were similar to those caused yield N’-formylkynurenine (1, 2). The spectroscopic
by L-Trp, while the sulfur and the oxygen analogs of Trp properties of various derivatives of this dioxygenase (3-
exhibit relatively small effects except for those observed 5) are very similar to those of analogous derivatives of
for the sulfur analog with CD spectroscopy. Each of these myoglobin (a molecular oxygen-binding heme protein)
three Trp analog inhibitors competes with L-Trp for the and horseradish peroxidase (a peroxide-utilizing heme
ferrous-CO enzyme, a model for the ferrous-O2 enzyme. protein) except for minor differences for their native ferric
The present findings demonstrate that, although substi- forms. The dioxygenase exhibits ligand binding properties
tution of a methyl group for the hydrogen atom on the similar to those of myoglobin but somewhat different from
indole nitrogen or of a more electron-inductive sulfur or those of the peroxidase (6). Several heme ligands including
oxygen atom for the indole nitrogen atom does not pre- cyanide, azide, carbon monoxide, 4-phenylimidazole, and
vent the binding of the resulting Trp analog to indole- norharman (7-9) have been known to inhibit the catalytic
amine 2,3-dioxygenase, the free form of the indole ni-
reaction of the dioxygenase by occupying the molecular
trogen base is an important physical and/or electronic
oxygen binding site, the heme iron. With respect to the
organic substrate Trp, 2,5-dihydro+phenyl-alanine is the
1 Supported by United States Public Health Service Grant GM 37696
only competitive inhibitor reported at present (10). No
(M.S.) substrate analog inhibitors have been known for the
* Portions of this work have been presented by S.G.C. in partial ful- dioxygenase. &Hydroxy-L-Trp has been found to inhibit
fillment of the Ph.D. requirements at the University of South Carolina. the metabolism of L-T&I by the mouse epididymal enzyme
Present address: Department of Chemistry, Arkansas State University,
State University, AR 72467-0419.
3 To whom correspondence should be addressed. ’ Abbreviation used: Trp, tryptophan.

326 0003.9861/91$3.00
Copyright 0 1991 by Academic Press, Inc.
All rights of reproduction in any form reserved.
 Trp ANALOG COMPETITIVE INHIBITORS
(ll), while this compound is a substrate for the rabbit
intestinal (12,13) and human placental dioxygenases (14).
On the other hand, many inhibitors have been known
for tryptophan 2,3dioxygenase, another heme-containing
dioxygenase (1516) that catalyzes the same reaction as,
but is distinct from, indoleamine 2,3-dioxygenase. In ad-
dition to the above-mentioned compounds (except for 4-
phenylimidazole that has not been examined), various
tryptophan analogs and derivatives such as &hydroxy-L-
Trp, tryptamine, serotonin (&hydroxytryptamine), 5 genase were performed at 25°C using 50 nM dioxygenase, either D- or
methyl-Trp, and 5-fluoro-Trp, all of which are substrates
for indoleamine 2,3dioxygenase (13, 17), have been re-
ported to inhibit tryptophan 2,3dioxygenase in either a
competitive or a noncompetitive [an inhibitor binds
equivalently to both the free (E) and the substrate-bound
(ES) enzyme] manner with respect to the substrate L-
Trp (15,18, 19).
Inhibitors for these two dioxygenases, especially sub-
strate analog competitive inhibitors, may serve as useful
probes for understanding what physical and electronic
structural factors are required for an organic compound
not only to be able to bind to the catalytic site of the
enzyme but also, more importantly, to serve as a substrate
for the enzyme. Such inhibitors might thus provide a clue
to the understanding of the mechanism of action of these
enzymes which has not yet been clarified.
In the course of a study on the effects of Trp modifi-
cations on the catalytic activity of indoleamine 2,3diox-
ygenase (17), we have discovered three Trp analogs which
serve as potent competitive inhibitors with respect to D-
Trp and L-Trp for the rabbit intestinal dioxygenase. These
compounds are 1-methyl-DL-Trp, @-(3-benzofuranyl)-DL-
alanine (the oxygen analog of Trp), and B- [3-
benzo(b)thienyl] -DL-alanine (the sulfur analog of Trp)
each of which has a substitution at the indole nitrogen
atom (see Fig. 1 for their structures). Enzyme kinetic
studies of the inhibition mechanism and spectroscopic
equilibrium binding studies of the interactions of these
Trp analogs with the dioxygenase will be reported in this
paper.
EXPERIMENTAL PROCEDURES
Materials. L-Trp was purchased from Sigma. D-Trp and l-methyl-
DL-T~~ were purchased from Aldrich. @-[3-Benzo(b)thienyl]-DL-alanine
hydrochloride, the sulfur analog of Trp, was a gift from Professor Tal-
mage R. Bosin at Indiana University. @-(3-Benzofuranyl)-DL-alanine,
the oxygen analog of Trp, was a gift from the National Institute of
Mental Health Chemical Synthesis Program (c/o Dr. J. Stephen Ken-
nedy and Dr. Albert A. Manian, NIMH). All of these Trp analogs were
an approximately 1:l mixture of the DL-forms as examined with reverse
phase high performance liquid chromatography (Cyclobond III chiral
column, ASTEC Inc., Whippang, New Jersey) (20). The sulfur and the
oxygen analogs were free of DL-Trp. Since the commercial samples of
1-methyl-DL-Trp contained -4% DL-Trp (17), it was recrystallized from
hot water to decrease the DL-Trp contamination to about 2% (1% for
each isomer). The DL mixtures of these three Trp analogs were used in
this study.
FOR INDOLEAMINE 2,3-DIOXYGENASE 327
Enzymes. Indoleamine 2,3dioxygenase was purified from rabbit
small intestine by the method of Shimixu et al. (13) except that the final
isoelectrofocusing step was omitted and instead step 6 (Sephadex G-
100 chromatography) was repeated 2-4 times. The purified native ferric
enzyme exhibited an AM/Am value of 1.7-1.8 in 20 mu potassium
phosphate buffer at pH 6.0 and 25’C and was >70% pure as judged by
sodium dodecyl sulfate gel electrophoresis (4). The amount of the enzyme
was expressed in terms of its heme content based on the absorbance at
405 nm [c = 159 mM-’ cm-’ at pH 6.0 and 25°C (S)]. Bovine liver catalase
was a product of Sigma.
Enzyme kinetic experiments. The assays of indoleamine 2,3dioxy-
L-Trp as a substrate, and an ascorbic acid-methylene blue (+catalase)
cofactor system as described previously for the standard assay (3) by
continuously following the absorbance increase at 321 nm [c = 3.75
mM-’ cm-’ (21)] due to the formation of N’-formylkynurenine in the
absence and presence of the Trp analog inhibitors. Enzyme kinetic data
[the initial rate of the product formation (V) vs substrate (Trp) con-
centration] were analyzed by employing three types of plots, the Line-
weaver-Burk (or the double-reciprocal) plot (l/Vvs l/[substrate]), the
Eadie-Hofstee plot (V vs V/[substrate]), and the Hanes-Woolf plot
([substrate]/Vvs [substrate]) (22).
Preparations of indoleamine $3-dioxygenuse derivatives and their ti-
trations with Trp analogs and heme &an&. Direct or indirect (com-
petitive) titration experiments for the binding of L-Trp, Trp analogs,
and the heme ligands azide and I-phenylimidasole to the ferric and the
ferrous dioxygenases were carried out in 0.1 M potassium phosphate
buffer at 25°C with optical absorption and CD spectroscopy as described
previously (3,9). Analysis of titration data has been described elsewhere
(3, 23). Due to the relatively low solubilities, the Trp analogs used in
this study were added from their 5-6 mM stock solutions in water or in
-5 mM HCl (cf. 50 mM D- and L-Trp stock solutions in water). Ferrous
derivatives (the unligated form and its complexes with CO and NO) of
the dioxygenase were prepared as described previously (3,4).
Spectroscopic measurements. Optical absorption spectra and CD
spectra were recorded on a Varian Cary 219 spectrophotometer and a
Jasco J-500A spectropolarimeter, respectively. Both instruments were
equipped with a circulator for temperature control (+l”C). All experi-
ments were carried out at 25’C with a O.l- or 0.2-cm cuvette. Except
for updated instrumentation (Jasco J-500A spectropolarimeter and its
data acquisition and storage systems and the use of IBM personal com-
puters), the basic methods and instrumental conditions used have been
described elsewhere (23).
RESULTS
Among the various Trp analogs examined, including
those having a substitution of a fluorine atom or a methyl
group for a single hydrogen atom on the indole ring, and
those having a heteroatom substitution of a sulfur or OX-
ygen atom for the indole nitrogen atom (17), the following
three compounds were found to markedly inhibit the cat-
alytic activity of indoleamine 2,3dioxygenase with D-Tip
and L-Trp as substrates: 1-methyl-DL-Trp, P-(3-benzo-
furanyl)-DL-alanine (the oxygen analog of Trp), and p-
[3-benzo(b)thienyl] -DL-alanine (the sulfur analog of Trp)
(see Fig. 1). None of these Trp analogs were found to
serve as substrates for the dioxygenase in this study.
Figure 2 shows a typical double-reciprocal plot for the
conversion of D-Trp to N’-formylkynurenine by the diox-
ygenase at pH 8.0 in the absence and presence of various
328 CADY AND SON0

coo- B
/

FIG. 1. Structures of the three Trp analog inhibitors used in this

study, l-methyl-DL-tryptophan (A), p-(3-benzofuranyl)-DL-ahmine (the
oxygen analog of Trp) (B, X = 0), and fi-[3-benxo(b)thienyl]-DL-ahmine
(the sulfur analog of Trp) (B, X = S).

concentrations (O-60 pM) of the oxygen analog of Trp, 8- 0 2. 5 5 7. 5 10

(3-benzofuranyl)-DL-alanine. A common intersection V/CO-Trpl (nmoles Product.min-'.mM-'1
point on the y-axis (= l/V,,) of the double-reciprocal
plot clearly indicates that this compound behaves as a
competitive inhibitor with respect to D-Trp. The replot
shown in the inset, where the apparent K,,, value (K”,Pp)
is plotted against the inhibitor concentration, yields the
Ki value of 9.9 PM. A competitive inhibition pattern in
double-reciprocal plots (not shown) was also observed for
1-methyl-DL-Trp and @-[3-benzo(b)thienyl]-DL-alanine
(the sulfur analog of Trp) (17).
The classification of the inhibition mechanism for these
Trp analogs was further analyzed by using the Eadie-
Hofstee plot (Fig. 3A) and the Hanes-Woolf plot (Fig.
3B). Consistent with the double-reciprocal plot results, a
competitive inhibition mechanism was confirmed from

FIG. 3. An Eadie-Hofstee plot (A) and a Hanes-Woolf plot (B) for

the conversion of D-Trp to the product (iV’-formylkynurenine) catalyzed
by indoleamine 2,3dioxygenase in the absence and presence of varying
concentrations of the inhibitor, o-(3-benzofuranyl)-DL-alanine (the ox-
ygen analog of Trp). The concentrations (O-60 PM) of the inhibitor used
for each plot are indicated in the figure. In the inset, a replot for the
apparent K,,, values (pmm), which were obtained from the slope (A) or
the r-axis intercept (B) of the original plot, versus the inhibitor con-
centration is shown for each case to determine the Ki values (Ki = 10.6
gM for A, and K; = 13.7 pM for B). The same kinetic data was used for
the plots shown in Fig. 2 and Fig. 3.

both the Eadie-Hofstee plot (a common intersection point

FIG. 2. A double-reciprocal plot of the rate of the product (N’-for-
mylkynurenine) formation catalyzed by indoleamine 2,3dioxygenase
versus the substrate ~-Tip concentration in the absence and presence
of varying concentrations of the inhibitor, j3-(3-benzofuranyl)-DL-alanine
(the oxygen analog of Trp). The concentrations (040 PM) of the inhibitor
used for each plot are indicated in the figure. In the inset, a replot for
the apparent K,,, values (ep), which were obtained from the x-axis
intercept of the original plot (-l/K”mm), versus the inhibitor concentra-
tion is shown to determine the Ki value (= 9.9 PM). The enzyme kinetics
assay was performed using 50 pmol of the dioxygenase at pH 8.0 and
25°C in a l-ml reaction mixture. See Experimental Procedures for further
details.
on the y-axis, Fig. 3A) and the Hanes-Woolf plot (parallel
lines, Fig. 3B) (22). The Ki values determined from these
two plots for the oxygen analog of Trp were 10.6 and 13.7
PM, respectively. The average Ki value (f standard error)
obtained from the three types of plots is 11 (+2) PM. The
average Ki values similarly obtained for 1-methyl-DL-Trp,
and the sulfur and the oxygen analogs of Trp using L-
Trp (pH 7.0) and D-Trp (pH 8.0) as substrates are sum-
marized in Table I. Comparable Ki values were obtained
at pH 7.0 with L-Trp and D-Trp (data not shown); these
substrates are expected to bind to the same catalytic site
of the dioxygenase (7,9). These three compounds are po-
tent inhibitors having the Ki values in the range of 7-70
 Trp ANALOG COMPETITIVE INHIBITORS FOR INDOLEAMINE 2,3-DIOXYGENASE 329

TABLE I Since the above enzyme kinetic results suggested that

Inhibition Constants ( Kj) for Three Tryptophan Analogs for these three Trp analog inhibitors bind to the Trp binding
Their Inhibitory Effects on the Catalytic Activity of site of the dioxygenase, an equilibrium binding study using
Indoleamine 2,3-Dioxygenase’ optical absorption and CD spectroscopy was undertaken.
These compounds caused detectable spectral changes at
Ki’ (PM) -5 mM concentrations upon addition to the native ferric,
ferrous, ferrous-CO, and ferrous-NO enzymes. Figure 4
Inhibitorb pH 7d pH ae
shows the effects of these three Trp analogs on the optical
1-Me-DL-Trp -f 6.6 z!z0.6 absorption spectra of the native ferric (A) and the ferrous-
Oxygen analog 25 + 2 11 +2 NO enzymes (B) in comparison with the effects of L-Trp
Sulfur analog 70 t 4 30 +1 reported earlier (3,4). For both enzyme forms, l-methyl-
DL-T~~ (-5 mM) caused relatively large spectral changes
’ Ail values were obtained using 56 nM dioxygenase in 0.1 M potassium
phosphate buffer at 25°C. For the details for the assay, see Experimental which are similar to but somewhat smaller than those
Procedures. caused by L-Trp. However, the spectral effects of the sul-
’ 1-Me-DL-Trp, 1-methyl-DL-Trp; oxygen analog, j3-(3benzofuranyl)- fur and the oxygen analogs of Trp (-5 mM) on these two
DL-alanine; sulfur analog, &[3-benzo(b)thienyl]-DL-alanine. enzyme forms were, although detectable, quite small, es-
’ The values listed here are averages of Ki values (+ errors) determined
from three different plots described in the Experimental Procedures pecially for the native ferric enzyme. Except for the bind-
section. ing of 1-methyl-DL-Trp to the ferrous-CO enzyme, effects
d With L-Trp as substrate. of these Trp analogs for the unligated ferrous enzyme and
’ With D-Trp as substrate. its complex with CO were almost undetectable with optical
f Not determined.
absorption spectroscopy.
CD spectroscopy, on the other hand, provided clear ev-
PM. Among these Trp analogs, 1-methyl-DL-Trp is the idence for the binding of these Trp analogs to the various
most efficient inhibitor, followed by the oxygen analog, forms of the dioxygenase as displayed in Fig. 5 for the
and then by the sulfur analog. The Ki values appear to native ferric (A), ferrous (B), ferrous-CO (C), and ferrous-
be significantly (-2.5 times) lower at pH 8.0 than at NO (D) enzymes. First, the effects of 1-methyl-DL-Trp
pH 7.0. are very similar to, but consistently smaller than those

Wavelength (nm>

Wave 1ength (nm>

FIG. 4. Effects of the Trp analog inhibitors on the optical absorption spectra of native ferric indoleamine 2,3dioxygenase (IDO) (A) and of
the ferrous-NO dioxygenase (B). The spectra of the enzyme (70-75 PM) were recorded at pH 8.0 for both A and B in the absence (. * . ) and
presence of 1-methyl-DL-Trp (5.5 mM) (-), @-[3-benzo(b)-thienyll-DL-alanine (the sulfur analog of Trp) (4.1 mM) (- * -), and 6-(3benzofuranyl)-
DL-alanine (the oxygen analog of Trp) (5.1 mM) (0). The corresponding dioxygenase adducts with L-Trp (20 mM for the ferric enzyme and 2 mM
for the ferrous-NO enzyme) (- - -) are overplotted for comparison. The path length was 0.1 cm. In B, note that the scale of the 375-475 nm range
has been expanded to twice the scale of the 475-650 nm range.
330 CADY AND SON0

of L-Trp on the CD spectra of these enzyme forms (Fig. TABLE II

5B-5D). In all cases the CD spectra of the homogeneous Dissociation Constants (&) of the 1-Methyl-nL-TRP
l-methyl-IX-Tip complexes of the enzyme (i.e., >95% Complexes of Ferric and Ferrous-CO
saturated) (see below for &) are clearly distinguishable Indoleamine 2,3-Dioxygenase (IDO)”
from those of the L-Trp complexes (Fig. 5A-5D). This
indicates that the observed spectral changes caused by Kd for Kd for
PH ferric ID0 (mM) ferrous-CO ID0 (mM)
the 1-methyl-DL-Trp sample are not due to the presence
of small amounts of contaminating DL-T~~ (<l% for each 6.0 -5.6 0.31
isomer). Second, the effects of the sulfur analog (dot- 7.0 0.59 0.31b
dashed line) on the CD spectra are nearly as large as those 8.0 0.48' 0.33
of L-Trp for the ferrous (Fig. 5B) and the ferrous-NO
enzymes (Fig. 5D), while the sulfur analog-induced spec- ’ All values were determined in 0.1 M potassium phosphate buffer at
25°C using optical absorption spectroscopy.
tral changes are smaller for the native ferric (Fig. 5A) b Kd = 0.30 mM for L-Tip.
’ Kd = 0.36 mM for L-Trp.

and ferrous-CO enzymes (Fig. 5C) than those caused by

Ferrous-NO
FIG. 5. Effects of Trp analog inhibitors on the Soret region CD spectra
of indoleamine 2,3dioxygenase for its native ferric (A), ferrous (B),
ferrous-CO (C), and ferrous-NO (D) forms. Spectral acquisition was
performed at pH 8.0 for the ferric enzyme and at pH 7.0 for all the
ferrous enzyme derivatives using 70-80 pM dioxygenase and a O.l-cm
cuvette. Corresponding spectra of the L-Trp adducts are overplotted for
comparison. The spectra shown are: without Trp analogs ( * . o), with
1-methyl-DL-Trp (5.5 mM) (-), P-[3-benzo(b)thienyl]-DL-alanine (the
sulfur analog of Trp) (4.1 mM) (-*-), /3-(3-benzofuranyl)-DL-alanine
(the oxygen analog of Trp) (5.1 mM) (0), and L-Trp (20 mM for ferric,
0.4 mM for ferrous, 5 mM for ferrous-CO, and 2 mM for ferrous-NO
enzymes) (- - -). See Experimental Procedures for sample preparations.
L-Trp or 1-methyl-DL-Trp. Third, the oxygen analog (cir-
cle-dashed line) induces surprisingly little CD spectral
effects on these enzyme forms except for the case of the
ferric dioxygenase (Fig. 5A).
Taking advantage of the relatively large optical ab-
sorption spectral changes upon the binding of l-methyl-
DL-T~~ to the ferric (Fig. 4A), the ferrous-CO (not
shown), and the ferrous-NO dioxygenase (Fig. 4B), Kd
values for these complexes were spectrophotometrically
determined at pH values 6,7, and 8 and are summarized
in Table II for the first two enzyme forms. Interestingly,
the Kd values for the 1-methyl-DL-Trp complexes are very
similar to the corresponding values for L-Trp for each
enzyme form.5 The pH dependence of the Kd values is
also similar for 1-methyl-DL-Trp and L-Trp; the Kd value
is pH-independent for the ferrous-CO enzyme between
pH 6 and 8 while for the ferric enzyme, the Kd value de-
creases with the increase in pH. Note that, below pH 7.0
1-methyl-DL-Trp (Kd = -5.6 mM at pH 6.0) exhibits even
somewhat higher affinity for the native ferric enzyme than
L-Trp (Kd = 30 mM at pH 6.0) (3). The Kd value for the
ferrous-NO adduct with 1-methyl-DL-Trp was -70 PM
at pH 7.0.
The Kd values of the complex of the sulfur analog of
Trp with the unligated ferrous dioxygenase were deter-
mined by the CD spectroscopic titration method as shown
in Fig. 6. Analysis of the data using a double-reciprocal
plot (Fig. 6, inset) yields a Kd value of 0.43 mM at pH 7.0.
A Kd value of -0.1 mM was obtained for the ferrous en-
zyme complex with l-methyl-IX-Tip. The CD spectral
changes caused by the sulfur and the oxygen analogs were
’ Assuming that the L-isomer of 1-methyl-DL-Trp is the predominant
form that readily binds to the various forms of the dioxygenase as has
been demonstrated with the other Trp analogs and derivatives examined
(3, 17), 1-methyl-L-Trp appears to have a somewhat higher affinity for
the enzyme than L-Trp.
 Trp ANALOG COMPETITIVE INHIBITORS FOR INDOLEAMINE 2,3-DIOXYGENASE 331

methyl-DL-Trp and L-Trp compete for the catalytic site

Wavelength 6-d
FIG. 6. A CD spectral titration of ferrous indoleamine 2,3dioxygenase
(IDO) with @-[3-benzo(b)thienyl]-DL-alanine (the sulfur analog Trp).
The titration experiments were performed with a O.l-cm cuvette at pH
7.0 containing -75 pM dioxygenase under nitrogen in the presence of
a slight excess of sodium dithionite (AsI < 2). The arrows show the
direction of the CD spectral changes upon the addition of the sulfur
analog of Trp. Shown in the inset is a double-reciprocal plot of the
negative CD intensity increase at 420 nm (A(Ae,,)) versus the concen-
tration of the sulfur analog of Trp.
not large enough for Kd value determinations for the other
enzyme derivatives.
Next, competitive binding experiments for L-Trp with
these three Trp analogs for indoleamine 2,3dioxygenase
were performed. For this purpose effects of the three Trp
analogs (1-methyl-DL-Trp, and the sulfur and the oxygen
analogs of Trp) on the L-Trp affinity of the ferrous-CO
enzyme (a model for the catalytically active species, fer-
rous-O2 enzyme) were examined with optical absorption
spectroscopy. 1-Methyl-DL-Trp (0, 0.3, 0.6, and 0.9 mM)
caused a linear increase in the pzp value for the L-Trp-
ferrous-CO enzyme complex, and the x-axis intercept
value (-0.27 mM) (plot not shown) reasonably coincides
with the Kd value (0.31 mM) that was independently de-
termined for the 1-methyl-DL-Trp complex of the ferrous-
CO enzyme (Table II). This clearly demonstrates that l-
of the enzyme (see Appendix in Ref. 23 for the analysis).
The sulfur analog (0,0.6,1.2, and 1.8 mM) and the oxygen
analog (0, 2.0, and 4.0 mM) also yielded a straight line
with the respective x-intercept values of -1.25 (mM) and
-3.5 (mM). These values (with positive sign) most likely
represent the Kd value of their complexes with the ferrous-
CO dioxygenase, although direct. determination of the Kd
values was not practically feasible due to the relatively
low solubilities of these Trp analogs.
Finally, how the binding of the Trp analog inhibitors
affects the affinity of heme ligands for the ferric and the
ferrous enzymes was investigated. Azide and 4-phenylim-
idazole were chosen as heme ligands for the ferric and
ferrous enzymes, respectively, since the affinities of these
ligands for the enzyme have previously been shown to be
sensitive to the binding of substrates to the enzyme (6).
At a concentration of 2 mM and at pH 7.0, all of these
Trp analogs exhibited positive cooperativity with azide
for the binding to the ferric enzyme and negative coop-
erativity with 4-phenylimidazole for the ferrous enzyme
as summarized in Table III. This is indicated by the de-
crease in the apparent Kd value for the azide affinity (pos-
itive cooperativity) and by the increase in the apparent
Kd value for the 4-phenylimidazole affinity (negative
cooperativity). Among the Trp analogs and L-Trp, l-
methyl-DL-Trp had the strongest positive cooperativity
with azide for the ferric enzyme, followed by L-Trp. The
sulfur and the oxygen analogs had relatively weak, com-
parable, but significant effects. For the ferrous enzyme-
4-phenylimidazole complexes, the negatively cooperative
effect of L-Trp was the strongest, followed by comparable
influences by 1-methyl-DL-Trp and the sulfur analog. The
oxygen analog only slightly affected the Kd for the 4-
phenylimidazole-ferrous enzyme complex. The observed
cooperative effects of these Trp analogs provide additional
support for the binding of these compounds to the active
site of the dioxygenase.

TABLE III
Effects of Indole 1-N-Substituted TRP Analogs on the Dissociation Constant (I&) of the Ferric Indoleamine 2,3-Dioxygenase
(IDO)-Azide Complex and the Ferrous IDO-4-Phenylimidazole (4-PheImid) Complex”

Trp analog Concn. of Trp Kd of ferric IDO-azide Concn. of Trp Kd of ferrous IDO-4-PheImid
added* analog (mM) complex (mM) analog (mM) complex (mM)

None 5.3 0.40

L-Trp 1 3.3 0.2 1.17
2 2.8
l-Me-DL-Trp 2 1.1 0.5 1.0
Sulfur analog 2 4.1 0.5 0.80
Oxygen analog 2 4.1 0.5 0.44

’ All values were determined in 0.1 M potassium phosphate buffer, pH 7.0, and at 25’C with optical absorption spectroscopy.
* See footnote b to Table I for the abbreviation and chemical names of the compounds. L-Trp is included for comparison.
332 CADY AND
DISCUSSION
The three Trp analog inhibitors for indoleamine 2,3-
dioxygenase that were discovered in the present study, l-
methyl-DL-Trp, @-(3-benzofuranyl)-DL-alanine (the ox-
ygen analog of Trp), and j3-[3-benzo(b)thienyl]-DL-alanine
(the sulfur analog of Trp), are the first examples of potent
Trp analog competitive inhibitors for this dioxygenase.
Each of these Trp analogs involves a substitution at the
indole nitrogen atom. Judged by the discriminatory bind-
ing affinity of the enzyme for the L-isomer of Trp (3, 7)
(see also footnote 5), the L-form for each DL-T~~ analog
enantiomer pair is the likely predominant inhibitor. At-
tempts to examine the effects of each separate D- and L-
form of these compounds were not made in this study.
Based on the similar spectroscopic effects and the
binding affinities for 1-methyl-L&Tip and L-Trp, l-
methyl-DL-Trp appears to be capable of binding to the
catalytic site of the dioxygenase in a manner very similar,
if not identical, to that of the substrate L-Trp. This, for
the first time, indicates that the replacement of the hy-
drogen atom on the indole nitrogen of Trp by a bulkier
and electron-donating methyl group does not hinder the
resulting l-methyl-Tip from binding to the enzyme. A
similar study on tryptophan 2,3-dioxygenase with l-
methyl-DL-Trp has not been reported.
The inability of 1-methyl-DL-Trp to serve as a substrate
for the dioxygenase, therefore, strongly suggests that the
presence of the methyl substituent on the indole nitrogen
is interfering with the oxygenation of Trp by the enzyme.
Relevant to the present finding, the free indole nitrogen
has been shown in past model studies to be essential for
the catalytic oxygenation of some indole derivatives (24,
25). In these model studies, manganese phthalocyanine
or cobalt-tetraphenylporphine was used as the catalyst to
oxidatively cleave several 2,3-alkyl-substituted indoles in
the presence of OZ. In contrast, in the singlet oxygen-
mediated (photosensitized) oxygenation of 3-alkyl indoles,
methylation of the indole nitrogen does not prevent the
oxygenation of the compounds (26). The present result
with 1-methyl-DL-Trp suggests that singlet oxygen is not
the reactive oxygen species involved in the dioxygenation
of Trp by this enzyme. Although this point has been spec-
ulated from chemical and theoretical model studies (27),
the present study provides further experimental support
for this view using the dioxygenase enzyme system.
The quite efficient inhibitory effects (Ki = lo-70 PM)
of the sulfur and oxygen analogs of Trp on the metabolism
of Trp by the dioxygenase are noteworthy. It was dem-
onstrated that these two Trp analogs compete with L-Trp
for the ferrous-CO enzyme (Fig. 5). The estimated Kd
values of the complexes of the ferrous-CO enzyme with
these two Trp analogs (Kd = ~10~~ M) are somewhat
larger than the Ki values (on the order of lop5 to LOP4M).
However, this is not unreasonable since a similar corre-
SON0
lation was observed for L-Trp: Kd = -0.3 X 10e3 M for
its complex with the ferrous-CO enzyme vs K,,, = -2.5
X 10d5 M at pH 7.0 (3). The inability of the sulfur and
oxygen analogs of Trp to serve as substrates can be at-
tributed either (a) to their improper binding to the cat-
alytic site of the dioxygenase which may be unsuitable
for the oxygenation reaction to take place or (b) to the
difference in the intrinsic reactivity at the 3-C position
of indole, benzothiophene, and benzofuran; the latter two
are the sulfur and oxygen analogs of indole, respectively.
Possibility (a) is mentioned because the adducts of these
Trp analogs with the various enzyme derivatives exhibit
spectroscopic properties which are significantly different
from those of the corresponding L-Trp and 1-methyl-DL-
Trp complexes (Figs. 4 and 5). However, the latter pos-
sibility (b) is more likely the major reason.
It has been generally assumed that the first step in the
enzymatic oxygenation of Trp is the formation of 3-hy-
droperoxyindolenine. This assumption is based on the
chemical behavior of this peroxy compound which was
initially demonstrated by Witkop’s pioneering work in
the early 1950s (28-30). Although the exact nature of the
“reactive dioxygen” species involved in the enzymatic
dioxygenation of Trp has not been understood, the elec-
trophilic addition of this dioxygen species to the indole
3-C position is most likely the first step in the reaction.
Compared with indole compounds, the reactivity at the
3-C position (of the heterocycle) of both benzofuran and
benzothiophene compounds toward electrophilic attack
(substitution) is known to be much lower (31).
It should be mentioned that the sulfur analog of Trp
was reported to be an effective antagonist for Trp in mi-
crobiological systems (32). In relation to the present find-
ing, an in uiuo metabolism study by Bosin et al. (33)
showed that 48 h after the intraperitoneal administration
of the sulfur analog of Trp to rats, no significant hetero-
cycle (thiophene ring)-cleaved metabolites of this com-
pound such as N’-formylkynurenine and kynurenine an-
alogs were detected in the urine or feces or in rat tissues.
Only alanine side chain-modified metabolites were found.
ACKNOWLEDGMENTS
We thank Professor Talmage R. Bosin at the Pharmacy Department
at the School of Medicine, Indiana University, Bloomington, Indiana,
for his generous gift of &[3-benzo(b)thienyl]-DL-alanine hydrochloride,
the National Institute of Mental Health Chemical Synthesis Program
(c/o Drs. J. Stephen Kennedy and Albert A. Manian, Rockville, Mary-
land) for the generous gift of @-(3-benzofuranyl)-DL-alanine, Professor
John H. Dawson for the use of his laboratory facilities at the University
of South Carolina, for his interest in this research, and for his helpful
suggestions for this study, and Professor Robert Phillips at the Chemistry
Department of University of Georgia at Athens, Georgia, for his helpful
suggestions and information about the physical and chemical properties
of several tryptophan analogs.
REFERENCES
1. Hayaishi, O., Hiram, F., Fujiwara, M., Ohnishi, T., and Nukiwa, T.
(1975) FEBS Proc. Meet. 10. 131-144.
 Trp ANALOG COMPETITIVE INHIBITORS
2. Hayaishi, 0. (1976) J. B&hem. (Tokyo) 79.13-21.
3. Sono, M., Taniguchi, T., Watanabe, Y., and Hayaishi, 0. (1980) J.
Biol. Chem. 255,1339-1345.
4. Sono, M., and Dawson, J. H. (1984) Biochim. Biophys. Acta 789,
170-187.
5. Sono, M. (1986) Biochemistry 25,6089-6097.
6. Sono, M. (1990) Bioc~mistry 29,1451-1460.
7. Yamamoto, S., and Hayaishi, 0. (1967) J. Biol. C&m. 242,5260-
5266.
8. Ohnishi, T., Hirata, F., and Hayaishi, 0. (1977) J. Biol. Chem. 252,
4643-4647.
9. Sono, M., and Cady, S. G. (1989) Biochemistry 28,5392-5399.
10. Watanabe, Y., Fujiwara, M., Hayaishi, O., Takei, T., and Umezawa,
H. (1978) Biochem. Biophys. Res. Commun. 85,273-279.
11. Ozaki, Y., Nichol, C. A., and Duch, D. S. (1987) Arch. Biochem.
Biophys. 257, 207-216.
12. Hirata, F., and Hayaishi, 0. (1972) Biochem. Biophys. Res. Common.
47,1112-1119.
13. Shimizu, T., Nomiyama, S.,. Hirata, F., and Hayaishi, 0. (1978) J.
Bid Chem. 253,4700-4706.
14. Takikawa, O., Kuroiwa, T., Yamazaki, F., and Kido, R. (1988) J.
Bid. Chem. 263,2041-2048.
15. Feigelson, P., and Brady, F. 0. (1974) in Molecular Mechanisms of
Oxygen Activation (Hayaishi, O., Ed.), pp. 87-133, Academic Press,
New York.
16. Ishimura, Y., Makino, R., Ueno, R., Sakaguchi, K., Brady, F. O.,
Feigelson, P., Aisen, P., and Hayaishi, 0. (1980) J. Biol. Chem. 255,
3835-3837.
17. Cady, S. G. (1989) Effects of Tryptophan Analogs on the Catalytic
Activity and Binding Activity of Indoleamine 2,3-Dioxygenase, Ph.D.
Dissertation, University of South Carolina.
18. Frieden, E., Westmark, G. W., and Schor, J. M. (1961) Arch.
Biochem. Biophys. 92, 176-182.
FOR INDOLEAMINE 2,3-DIOXYGENASE 333
19. Eguchi, N., Watanabe, Y., Kawanishi, K., Hashimoto, Y., and Ha-
yaishi, 0. (1984) Arch. Biochem. Biophys. 32, 602-609.
20. Armstrong, 0. W., Yang, K., Han, S. M., and Menges, R. A. (1987)
Anal. Chem. 59.2594-2596.
21. Dawson, R. M. C., Elliott, D. C., Elliott, W. H., and Jones, K. M.
(1986) Data for Biochemical Research, 3rd ed., p. 23, Oxford Uni-
versity Press, New York.
22. Wood, W. B., Wilson, J. H., Benbow, R. M., and Hood, L. E. (1981)
Biochemistry: A Problems Approach, 2nd ed., pp. 144-172, Benjamin
and Cummings, Menlo Park, California.
23. Sono, M., Andersson, L. A., and Dawson, J. H. (1982) J. Bid. Chem.
257,8308-8320.
24. Uchida, K., Soma, M., Naito, S., Onishi, T., and Tamaru, K. &cm.
L&t. 1978,471-474.
25. Dufour, M. N., Crumbliss, A. L., Johnston, G., and Gaudemer, A.
(1980) J. Mol. Catal. 7, 277-287.
26. Saito, I., Matsuura, T., Nakagawa, M., and Hino, T. (1977) Act.
Chem. Res. 10,346352.
27. Hamilton, G. A. (1974) in Molecular Mechanisms of Oxygen Acti-
vation (Hayaishi, O., Ed.), pp. 405-451, Academic Press, New York.
28. Witkop, B., and Patrick, J. B. (1951) J. Am. Chem. Sot. 3, 2196-
2200.
29. Ek, A., Kissma, H., Patrick, J. B., and Witkop, B. (1952) Experientiu
8,36-40.
30. Nakagawa, M., Watanabe, H., Kodato, S., Okajima, H., Hino, T.,
Flippen, J. L., and Witkop, B. (1977) Proc. Natl. Acad. Sci. USA
74,4730-4733.
31. Streitwieser, A. Jr., and Heathcock, C. H. (1981) Introduction to
Organic Chemistry, 2nd ed., pp. 1077-1082, Macmillan, New York.
32. Avakian, S., Moss, J., and Martin, G. J. (1948) J. Am. Chem. Sot.
70,3075-3076.
33. Bosin, T. R., Bickers, R. G., and Dinner, A. (1978) Drug Metab.
Dispos. 6,31-37.