Targeting multiple chorismate-utilizing enzymes with a single inhibitor: Validation of a three-stage design

Article abstract of DOI:10.1021/jm100158v

Chorismate-utilizing enzymes are attractive antimicrobial drug targets due to their absence in humans and their central role in bacterial survival and virulence. The structural and mechanistic homology of a group of these inspired the goal of discovering inhibitors that target multiple enzymes. Previously, we discovered seven inhibitors of 4-amino-4-deoxychorismate synthase (ADCS) in an on-bead, fluorescent-based screen of a 2304-member one-bead-one-compound combinatorial library. The inhibitors comprise PAYLOAD and COMBI stages, which interact with active site and surface residues, respectively, and are linked by a SPACER stage. These seven compounds, and six derivatives thereof, also inhibit two other enzymes in this family, isochorismate synthase (IS) and anthranilate synthase (AS). The best binding compound inhibits ADCS, IS, and AS with K _i values of 720, 56, and 80 μM, respectively. Inhibitors with varying SPACER lengths show the original choice of lysine to be optimal. Lastly, inhibition data confirm the PAYLOAD stage directs the inhibitors to the ADCS active site.

Full text of DOI:10.1021/jm100158v

3718 J. Med. Chem. 2010, 53, 3718–3729
DOI: 10.1021/jm100158v
Targeting Multiple Chorismate-Utilizing Enzymes with a Single Inhibitor: Validation of a
Three-Stage Design
Kristin T. Ziebart, Seth M. Dixon, Belem Avila, Mohamed H. El-Badri, Kathryn G. Guggenheim, Mark J. Kurth,* and
Michael D. Toney*
Department of Chemistry, University of California, One Shields Avenue, Davis, California 95616
Received February 4, 2010
Chorismate-utilizing enzymes are attractive antimicrobial drug targets due to their absence in humans
and their central role in bacterial survival and virulence. The structural and mechanistic homology
of a group of these inspired the goal of discovering inhibitors that target multiple enzymes. Previously,
we discovered seven inhibitors of 4-amino-4-deoxychorismate synthase (ADCS) in an on-bead,
fluorescent-based screen of a 2304-member one-bead-one-compound combinatorial library. The
inhibitors comprise PAYLOAD and COMBI stages, which interact with active site and surface residues,
respectively, and are linked by a SPACER stage. These seven compounds, and six derivatives thereof,
also inhibit two other enzymes in this family, isochorismate synthase (IS) and anthranilate synthase
(AS). The best binding compound inhibits ADCS, IS, and AS with Ki values of 720, 56, and 80 μM,
respectively. Inhibitors with varying SPACER lengths show the original choice of lysine to be optimal.
Lastly, inhibition data confirm the PAYLOAD stage directs the inhibitors to the ADCS active site.
Antibiotic drug resistance, now a global health issue, has
increased dramatically over the past several decades.1 Rates of
methicillin-resistant Staphylococcus aureus (MRSAa) infec-
tions doubled between the years 2000 and 2005. They are now
the cause of more deaths in the United States thanAIDS.2 The
low profitability and short functional lifetime of antibiotics
has resulted in a steady decrease in research and development
by industry.1,3 As a result, only two new classes of antibiotics
have been introduced since 1967: oxazolidinones (2000) and
lipopeptides (2003).4
We have developed a streamlined approach for discovering
potential antimicrobial lead compounds, entailing massively
parallel, on-bead screens of one-bead-one-compound (OBOC),
mass-tag-encoded combinatorial libraries with fluorescently
labeled enzyme.5 A small (2304-member), proof-of-concept,
peptide-based library was designed to target a set of homologous
chorismate-utilizing enzymes: 4-amino-4-deoxychorismate syn-
thase (ADCS), isochorismate synthase (IS), anthranilate syn-
thase (AS), and salicylate synthase (SS). ADCS, IS, AS, and SS
catalyze the first committed steps in the formation of folate,
enterobactin (an iron-chelating compound, i.e. siderophore), try-
ptophan, and mycobactin (a siderophore from Mycobacterium
tuberculosis), respectively, in bacteria and apicomplexan para-
sites (Figure 1).6
Chorismate-utilizing enzymes are excellent antimicrobial
drug targets for several reasons: (1) they catalyze the first
committed step in formation of several metabolites critical for
survival and/or virulence (Figure 1), (2) they are structurally
and mechanistically homologous, suggesting the possibility
that one drug could inhibit multiple metabolic pathways, (3)
they are present only in bacteria, plants, fungi, and api-
complexan parasites (e.g., Plasmodium, Toxoplasma, and
Cryptosporidium)6 and absent in humans. Furthermore, these
enzymes and metabolic pathways have not been fully
exploited as drug targets. One example of successfully target-
ing these pathways is the sulfa drugs, which inhibit folate
biosynthesis through inhibition of dihydropteroate synthase
(DHPS), which occurs two steps downstream of ADCS.
The structural homology of ADCS, IS, AS, and SS, includ-
ing nearly identical active site environments, inspired our goal
to develop one compound that would inhibit more than one of
these enzymes. The paradigm of multitarget inhibition by a
single compound has been growing in popularity and accep-
tance in recent years.7 This strategy has been exploited to
discover treatments for conditions such as HIV-AIDS, cancer,
depression, and Mycobacterium tuberculosis infections.8-13 A
drug that inhibits multiple enzymes would be highly desirable
for its potentially increased potency, even at moderate binding
affinities, due to inhibition of multiple critical metabolic path-
ways which would combine to slow growth. Also, the increased
difficulty for development of primary resistance (i.e., muta-
tion of the target enzymes to overcome inhibition), which
would require multiple simultaneous and spontaneous muta-
tions, makes this concept attractive. Nevertheless, alternative
*Corresponding Authors. For M.D.T.: phone, 530-754-5282; fax,
530-752-8995; E-mail, mdtoney@ucdavis.edu. For M.J.K.: phone, 530-
752-8192; fax, 530-752-8995; E-mail, mjkurth@ucdavis.edu.
a Abbreviations: ACN, acetonitrile; ADC, 4-amino-4-deoxychorismate;
ADCL, 4-amino-4-deoxychorismate lyase; ADCS, 4-amino-4-deoxychoris-
mate synthase; AS, anthranilate synthase; DCM, dichloromethane; Dde,
1-(4,4-dimethyl-2,6-dioxocyclohexylidene)ethyl; DIC, diisopropylcarbodii-
mide; DHPS, dihydropteroate synthase; DIPEA, N,N-diisopropylethyla-
mine; DMF, dimethylformamide; DMSO, dimethylsulfoxide; DTT,
dithiothreitol; EDTA, ethylenediaminetetraacetate; ESI, electrospray ioniza-
tion; Fmoc, 9H-fluoren-9-ylmethoxycarbonyl; HOBt, n-hydroxybenzotria-
zole; HPLC, high performance liquid chromatography; HRMS, high
resolution mass spectrometry; IPTG, β-isopropylthiogalactoside; IS, iso-
chorismate synthase; LB, luria-bertani broth; LDH, lactate dehydro-
genase; MBHA, 4-methylbenzhydrylamine; MRSA, methicillin-resistant
Staphylococcus aureus; NADH, nicotinamide adenine dinucleotide, reduced
form; OBOC, one-bead one-compound; PLP, pyridoxal-50-phosphate; SDS-
PAGE, sodium dodecyl sulfate polyacrylamide gel electrophoresis; TEA,
triethanolamine; TFA, trifluoroacetic acid; TIS, triisopropylsilane
r
pubs.acs.org/jmc
Published on Web 04/01/2010
2010 American Chemical Society

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 9 3719
Figure 1. Chorismate is a central branch-point compound in bacterial, plant, fungal, and apicomplexan parasite metabolism. Chorismate-utilizing
enzymes are essential for survival and/or virulence in these organisms. The enzymes in bold (ADCS, AS, and IS) are the subject of this work.
resistance mechanisms (e.g., efflux pumps or covalent modi-
fications) would remain problematic.
Our inhibitor design, inspired by the popular cholesterol-
lowering statin drugs, incorporates three “stages”: PAYLOAD,
SPACER, and COMBI, as illustrated in Figure 2. The PAY-
LOAD is a chorismate mimic, designed to guide the compound
to the enzyme’s active site. The SPACER links PAYLOAD and
COMBI stages together and extends out of the active site to the
solvent-exposed surface region of the proteins. The COMBI
stage is where combinatorial diversity elements are introduced
to the library. Schreiber and co-workers first introduced the
concept of biased combinatorial libraries and used this techni-
que to discover ligands for the SH3 domain of phosphatidylino-
sitol 3-kinase14 and, later, for histone deacetylases.15
This strategy is similar to the fragment-based drug discov-
ery paradigm, where subcomponents of the final inhibitor are
identified and optimized individually, then connected in a
later stage. The fragment-based approach has the recognized
disadvantage that the process of connecting fragments can
be problematic.16 In contrast, the method employed here
addresses this major hurdle by connecting the fragments from
the outset. Importantly, it is enabled by the massively parallel
OBOC library screening and hit identification technology
utilized in our approach.
In previous work, OBOC library screening employing
ADCS in hit selection led to the identification and kinetic
Figure 2. (A) Inhibitor design strategy employs three “stages”: PAY-
characterization of seven ADCS inhibitors.5 The results pre-
LOAD, SPACER, and COMBI. X1 and X2 represent the two diversity
elements present in each member of the combinatorial library. (B)
sented therein established the efficacy of our general paradigm
for chorismate-utilizing enzyme inhibition based on combi-
PAYLOAD compound 1 is a chorismate mimic, designed to guide the
inhibitors to the active site. (C) 2was synthesized and assayed in order to
natorial chemistry and a staged inhibitor design. We further
demonstrate here that the COMBI stage interacts with surface
compare directly the inhibition results of the fully elaborated inhibitors
3-9 with the PAYLOAD (for description of X1 and X2, see Figure 6).

3720 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 9
Ziebart et al.
Table 1. Substrate Binding Constants for ADCS, IS, and AS
Scheme 1
ADCSa
IS
AS
K
m, chorismate (μM)
(saturating [Mg2þ])
13 ( 1
16 ( 1
3.9 ( 0.2
ordered,
chorismate
binds first
kinetic
mechanism
random
random
K
chorismate (μM)b
RKchorismate (μM)c
KMg2þ (μM)e
2200 ( 1100
N/Ad
3030 ( 540
29 ( 4
259 ( 46
2.5 ( 0.4
25.22 ( 0.02
14.87 ( 0.01
415.1 ( 0.5
244.7 ( 0.3
5 ( 2f
residues proximal to the active site and leads to an increased
binding affinity to the enzymes. In this work, our central
objective, inhibiting more than one enzyme with the same
compound, is demonstrated. The seven inhibitors discovered
in the ADCS screen, as well as six derivatives thereof, were
assayed against IS and AS; these and other results are
presented herein. Generally, the inhibitors previously char-
acterized with ADCS bind more tightly to IS and AS. Five
of the inhibitors display micromolar inhibition of ADCS, IS,
and AS, and the best overall binding compound inhibits
ADCS, IS, and AS with Ki values of 720, 56, and 80 μM,
respectively.
Additionally, the present work investigates two impor-
tant design features via synthesis and assay of four new
compounds against ADCS, IS, and AS. First, the optimal
length of the SPACER stage was tested by varying its
length. Second, the possibility of the COMBI stage being
principally responsible for the inhibition results was
probed by synthesis and assay of COMBI-only and COM-
BI-SPACER-only compounds. These were found to in-
hibit minimally and nonspecifically AS and ADCS, thus
verifying that all three stages are necessary for optimal
inhibition.
RKMg2þ (μM)g
N/Ad
a Data taken from ref 5. b Reflects the binding of chorismate to the
free enzyme. c Reflects the binding of chorismate to the [E Mg2þ
]
3
complex. In principle, this is the same as the chorismate KM in the pre-
sence of saturating Mg2þ d Not applicable when an ordered kinetic
.
mechanism is employed. e Reflects the binding of Mg2þ to the free
enzyme. f ADCS value reflects the binding of Mg2þ to the [E choris-
3
mate] complex. g Reflects the binding of Mg2þ to the [E chorismate]
3
complex.
The affinity of IS for chorismate is very low in the absence
of Mg2þ, as reflected by the high Kchorismate values relative to
RKchorismate. This contrasts with the 1.7-fold difference bet-
ween the AS values for Kchorismate and RKchorismate. Here, R is
a proportionality constant that defines the extent to which
binding of one substrate is enhanced if the other is already
bound. If R = 1, binding of one substrate has no effect on the
other. In IS reactions, R = 0.025; therefore, chorismate and
Mg2þ binding is highly cooperative. Conversely, AS has an R
value of 0.59, which indicates that binding of either substrate
to the free enzyme is much less dependent on the presence of
the other compared to IS and ADCS.
The random kinetic mechanism employed by AS and IS
contrasts with the ordered kinetic mechanism observed for
ADCS.5 In the case of ADCS, chorismate binds first, fol-
lowed by magnesium. Chorismate binding to ADCS elicits a
large conformational change. In support of this, substantial
differences exist between the “open” and “closed” forms of
ADCS X-ray crystal structures.17 Three glutamate residues,
Glu302, Glu436, and Glu439 (ADCS numbering used
throughout), located just above the putative chorismate
binding site, are strictly conserved among ADCS, AS, and
IS, and they are proposed to coordinate Mg2þ. The distance
between two of them, Glu302 and Glu439, reflects the extent
Lastly, the inhibition results are presented in the context of
the kinetic and structural properties of ADCS, IS, and AS,
which are Mg2þ-dependent enzymes. The kinetic mechanisms
employed by these enzymes explain and account for the
observed inhibition patterns with respect to chorismate and
the Mg2þ cofactor.
Results and Discussion
Kinetic Mechanisms of AS and IS. Initial rates of reactions
where chorismate and Mg2þ concentrations were varied
across wide ranges were measured. Double reciprocal plots
of the kinetic data for AS and IS are shown in Figures 3
and 4, respectively. Both the AS and IS data sets were
globally best-fit to eq 1, which describes a random kinetic
mechanism.
˚
of protein movement between open (25.3 A) and closed
(6.2 A) forms of ADCS (Figure 5). Because the Mg2þ binding
˚
site is not fully formed in the “open” structure, a random
kinetic mechanism is not possible for ADCS.
Conversely, AS and IS undergo a smaller conformational
change on chorismate binding. This is supported by crystallo-
graphic data.18-20 In the case of AS, one can compare the
structural differences between “closed” Serratia marces-
cens AS,18 solved with Mg2þ and benzoate present, and
“open” Salmonella typhimurium AS,19 which has no bound
ligands. (These two AS isozymes are nearly identical
proteins). The distance between Glu302 and Glu439 in
n
o.
vi ¼ Vmax½chorismateꢀ½Mg2þꢀ
n
K
chorismateKMg2þ þ RKchorismate½Mg2þꢀ
o
þ RKMg2þ ½chorismateꢀ þ ½chorismateꢀ½Mg2þꢀ
ð1Þ
˚
the open and closed forms changes from 10.7 to 6.8 A,
respectively. The change in distance between these ana-
logous glutamate residues in IS is of a similar magnitude:
Scheme 1 describes the equilibria and kinetic parameters
that correspond to eq 1. The binding constant values for
chorismate and Mg2þ were determined from nonlinear re-
gression analysis and are listed in Table 1. The ADCS kinetic
mechanism and substrate binding constant values were
described previously;5 those results are listed in Table 1 for
comparison with AS and IS.
˚
˚
9.9 A (open) versus 5.9 A (closed). Figure 5 depicts the
open and closed crystal structures of AS and IS. Here, the
MenF isochorismate synthase crystal structures were con-
sulted20 because structural data is not yet available for the
EntC IS protein that is the subject of kinetic analyses in
this work. MenF IS shares 24% sequence identity and 41%

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 9 3721
Figure 3. Double reciprocal plot of Mg2þ dependence indicates AS
employs an equilibrium random kinetic mechanism. The concen-
trations of chorismate are: 1 (O), 4 (b), 12 (0), 36 (9) μM.
Figure 4. Double reciprocal plot of Mg2þ dependence indicates an
equilibrium random kinetic mechanism for IS. The concentrations
of chorismate are: 33 (O), 80 (b), 120 (0), 200 (9) μM.
homology with EntC IS. Further, MenF has been reported
to employ a random kinetic mechanism.21
Because the magnesium binding site in AS and IS is more
fully formed in the absence of chorismate, it is possible for the
two substrates to bind independently of each other. Together,
the AS and IS structural and kinetic data support a mechanism
in which chorismate and Mg2þ bind proximal to one another
in a cooperative manner. The inhibition data for ADCS, AS,
and IS further support this conclusion (vide infra).
Inhibition Results. Inhibitors (3-9) identified from the
ADCS on-bead screen were resynthesized5 and assayed
against IS and AS (Figure 6). The results of those assays,
along with data for compounds 1 and 2, are presented in
Table 2. Remarkably, all of the inhibitors except 8 bind more
tightly to IS and AS than they do ADCS, even though ADCS
binding was employed in the initial bead screening bind-
ing assay. The best inhibitor of IS is 7 with a Ki value of
56 μM; the tightest binding inhibitor of AS is 5 with a Ki value
of 20 μM.
Figure 5. Overlay of open and closed X-ray structures for ADCS
(A), AS (B), and IS (C). Each protein’s C-R backbone is displayed in
stick form with the relevant glutamic acid residues labeled and side
chains rendered in sick form (Glu302, Glu436, Glu439; ADCS
numbering used throughout). The gray trace represents the closed
(i.e., substrate bound) structure; glutamic acid residues are colored
gray. The purple trace represents the open (i.e., apo) structure;
glutamic acid residues are colored purple. Chorismate (peach-
colored) and Mg2þ (magenta) were docked into the active site of
each structure.
where initial rates of reactions were measured across a
wide range of inhibitor and chorismate concentrations
with saturating magnesium. To assess the pattern of
inhibition with respect to magnesium, the same type of
analysis was performed for reactions in which choris-
mate was held constant at a saturating concentration
and magnesium concentrations were varied from 5 to
The Ki values in Table 2 are for inhibition against
chorismate. They were obtained from a global fit of data

3722 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 9
Ziebart et al.
Figure 6. Structures of the compounds identified from the ADCS bead screen assay. These inhibitors were subsequently assayed against AS
and IS.
The most appropriate comparison of inhibition by fully
staged inhibitors 3-9 to inhibition by the PAYLOAD alone
is with amide derivative 2 because the SPACER and PAY-
LOAD stages are connected by an amide bond. The inhibi-
tion of IS by fully elaborated 7 is 46-fold tighter than 2, and
compound 5 binds 38-fold more tightly to AS than 2.
Compared to ADCS, inhibitors 3-9 overall bind more
tightly to AS and IS; however, the degree to which the
COMBI stage improves inhibition is generally less for IS
and AS than ADCS. The Ki value of the best inhibitor of
ADCS, 6, is 360 μM, which is an 89-fold improvement
relative to 2. The greater impact of the COMBI stage on
ADCS binding may be related to the larger conformational
change that occurs on chorismate binding, relative to IS and
AS.
With AS, the lack of competitive inhibition versus chor-
ismate was unexpected. None of the compounds competi-
tively inhibits chorismate binding; 1-6, 8, and 9 display
mixed inhibition and 7 is a noncompetitive inhibitor. Given
the similarity between chorismate and 1 and 2, one would
expect these inhibitors to be competitive against chorismate.
The Ki0 values listed in Table 2 reflect the binding of inhibitor
Table 2. Inhibition of ADCS, IS, and AS by Library Compounds
Identified in an ADCS On-Bead Screen
Ki (μM)a
compd
ADCSb
IS
ASc
1
2
3
4
5
6
7
8
9
37000 ( 4000
32000 ( 2000
540 ( 60
3900 ( 500
770 ( 60
360 ( 30
720 ( 60
1100 ( 100
710 ( 80
competitive
1300 ( 100
2600 ( 210
260 ( 21
430 ( 64
210 ( 29
190 ( 19
56 ( 3
210 ( 18 550 ( 83
760 ( 6 2600 ( 300
28 ( 7
490 ( 73 1081 ( 117
20 ( 2 54 ( 4
250 ( 61 110 ( 7
80 ( 2
N/Ad
81 ( 20
1200 ( 99
170 ( 11
competitive
160 ( 16 389 ( 24
570 ( 44 1020 ( 59
mixed
inhibition
vs chorismate
inhibition
vs Mg2þ
competitive noncompetitive
competitive
a Inhibition constants reflect inhibition with respect to chorismate
c
binding. b Data taken from ref 5. Ki0 values, listed in the last column,
reflect the inhibitor binding to the [E chorismate] complex. d Not applicable
3
because compound 7 exhibited pure noncompetitive inhibition with respect to
chorismate.
500 μM. Representative plots of inhibition with respect to
each substrate for AS and IS are shown in Figures 7 and 8,
respectively.
to the [AS chorismate] complex. Except for compound 6, the
3
inhibitors bind more tightly to the free enzyme than to the
[AS chorismate] complex. The average of the Ki0/Ki ratios
3

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 9 3723
Figure 8. (A) IS inhibition by 7 with respect to chorismate. Con-
centrations of 7 are 0 (O), 250 (b), 500 (0) μM. (B) IS inhibition by 7
with respect to Mg2þ. Concentrations of 8 are 0 (O), 250 (b) μM.
Figure 7. (A) AS inhibition by 8 with respect to chorismate. Con-
centrations of 8 are 0 (O), 220 (b), 500 (0) μM. (B) AS inhibition by
8 with respect to Mg2þ. Concentrations of 8 are 0 (O), 200 (b), 375
(0) μM.
Scheme 2
among those inhibitors is 2.5. Scheme 2 depicts the set of
equilibria that describe noncompetitive and mixed inhibition
patterns; it additionally takes into account the random
kinetic mechanism of AS. For pure noncompetitive inhibi-
tors, Ki0 = Ki; for mixed inhibition, Ki0 ¼ Ki.
Figure 9. Dixon plot of AS inhibition by 1. Linear regression yields
an R value of >0.99 for each chorismate concentration, indicating
that the [AS chorismate 1] complex is catalytically inactive. The
concentrations of chorismate are: 2.1 (O), 4.1 (b), 8 (0), 10 (9), 20
(4) μM.
Additionally, the inhibition data for PAYLOAD 1 were
plotted as a Dixon plot (i.e., 1/vi vs [1]) for several choris-
mate concentrations (Figure 9). A linear Dixon plot rules
out the possibility of other inhibition scenarios, such as
partial inhibition and other types of mixed inhibition,
because these inhibition modes would result in a curved
Dixon plot. It also indicates that the [AS chorismate
3
3
and 9. Because of limited amounts of these compounds,
there is not a sufficient number of data points to construct
analogous Dixon plots for 2-5, 8, and 9. However, it is
assumed that these inhibitors follow the same inhibition
pattern as 1 based on the similar Ki0/Ki ratios observed for
compounds 1-5, 8, and 9.
3
3
inhibitor] complex is catalytically inactive.22 The linearity
of the data in Figure 9 confirms Scheme 2 as an accurate
description of AS inhibition by 1, and by extension, 2-5, 8,

3724 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 9
Ziebart et al.
Figure 10. Compounds 10, 5, 11, and 12 comprise different SPACER elements: L-ornithine, L-lysine, S-propyl-L-cysteine, and L-R-
homolysine, respectively. These compounds are composed of identical PAYLOAD and COMBI moieties.
The competitive inhibition with respect to Mg2þ reveals that
these compounds are binding near the AS active site. (This
result also confirms that the [AS chorismate inhibitor] com-
to the distance between the amide carbonyl carbon of the
PAYLOAD and the R-carbon of the amino acid (e.g.,
ornithine, lysine, cysteine) with its side chain in the extended
conformation. The Ki values reflect inhibition against chor-
ismate. Determination of the inhibition mechanism (i.e.,
competitive versus noncompetitive) followed a process simi-
lar to that described for inhibitors 3-9.
AS and ADCS are most sensitive to SPACER length. A
shorter SPACER strongly decreases binding to ADCS, and a
longer SPACER is most detrimental to AS binding. How-
ever, it is difficult to interpret the effect of SPACER length
on AS inhibition because these compounds are not compe-
titive inhibitors of chorismate (vide supra). IS was the least
sensitive to the SPACER length; the difference in inhibition
by 5, 10, and 11 was minimal. Despite a small improvement
in inhibition of IS by a longer SPACER (11), this benefit does
not counterbalance the cost paid with ADCS for the longer
SPACER. The results in Table 3 support the deduction from
the crystal structure that lysine provides the optimal
SPACER length for ADCS inhibitors.
Validation of PAYLOAD-Directed Active Site Binding.
The inhibitor design strategy is based on two principles: (1)
the PAYLOAD directs these compounds to the active site,
and (2) synergistic PAYLOAD and COMBI interactions
strengthen enzyme binding. PAYLOAD-deleted compounds
14 and 15 (Figure 12) were synthesized and assayed against
ADCS, IS, and AS to test the possibility that inhibition
of these two enzymes is mainly due to COMBI interactions
with surface residues. Ki values are listed in Table 4. The
results show that inhibition is best when all three stages
(PAYLOAD, SPACER, COMBI) are present. Furthermore,
inhibition of ADCS and AS by 14 and 15 is noncompetitive
against both substrates, indicating that the two substrates
and the inhibitor can bind simultaneously to the enzyme and
that the inhibitor is not binding to the active site. One must
interpret the AS inhibition data with caution. For AS, the
results in Table 4 indicate that the PAYLOAD directs
3
2þ 3
plex is catalytically inactive because Mg is required for
catalysis). The X-ray crystal structure of S. marcescens
AS, a nearly identical homologue of the S. typhimurium
AS under study here, was solved in the presence of magne-
sium.18 In this structure, magnesium is positioned in the
active site immediately proximal to the C1 carboxylate
group of the chorismate analogue benzoic acid. Addi-
tionally, three glutamate residues surround and coordi-
nate Mg2þ. The inhibitors likely bind near the analogous
glutamate residues to competitively inhibit Mg2þ from
binding. To propose the precise inhibitor binding loca-
tion to AS would be speculative. Therefore, one must
evaluate all subsequent inhibition data in light of this
uncertainty.
Conversely, in the IS inhibition studies, each inhibitor
competitively inhibited chorismate and noncompetitively
inhibited magnesium. The different inhibition patterns with
respect to each substrate are a consequence of the random
kinetic mechanism employed by IS and AS. Separate,
adjacent binding sites exist for chorismate and Mg2þ, and
the inhibitors bind to only one of these locations, allowing
the other substrate to bind.
Optimal SPACER Length. Lysine was originally chosen as
the SPACER element for our discovery library after docking
studies with the ADCS crystal structure17 indicated its
4-carbon side chain is an optimal length. To test this deduc-
tion, compounds 10 and 11 were synthesized and assayed
against ADCS, IS, and AS (Figure 10). Analogue 10 was
prepared with L-ornithine, such that the SPACER contains
one fewer carbon relative to lysine. Because of the high cost
and limited availability of a L-R-homolysine precursor, 11 (an
isosteric analogue of 12) was instead synthesized (Figure 11).
The inhibition constants obtained from kinetic assay of 5,
10, and 11 are listed in Table 3. SPACER lengths correspond

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 9 3725
Figure 11. Synthesis of compound 11. Reagents and conditions: (i) Boc2O, Et3N, DCM, reflux, overnight; (ii) I2, PPh3, 1H-imidazole, DCM,
rt, 3.5 h; (iii) Cs2CO3, DMF, methylester of L-cysteine hydrochloride, rt, overnight; (iv) LiOH, MeOH, rt, 2 h; (v) Fmoc-OSu, THF, Na2CO3,
0 °C, 1 h; (vi) TFA:DCM (3:7), rt, 30 min; (vii) DIPEA, H2O/ACN, rt, 7 h; (viii) HOBt, DIC; (ix) 20% piperidine in DMF, 2 ꢁ 10 min;
(x) HOBt, DIC; (xi) TFA/H2O/TIS (95, 2.5, 2.5, v/v/v).
Table 3. The Effect of SPACER Length on Inhibition of ADCS, IS,
and AS
Ki (μM)a
ADCSb
ISc
ASd
˚
SPACER length (A)
inhibitor
10
5
4.85
6.30
7.81
>1000e
770 ( 60
2500 ( 120
360 ( 34
210 ( 29
150 ( 3
41 ( 6
20 ( 2
96 ( 14
11
a Ki values reflect the inhibition with respect to chorismate. b Each
inhibitor competitively inhibited chorismate and Mg2þ. c Each inhibitor
displayed competitive inhibition versus chorismate and noncompetitive
Figure 12. COMBIþSPACER (14) and COMBI-only (15) com-
pounds were synthesized and assayed against ADCS, IS, and AS to
verify that the PAYLOAD directs the inhibitors to the active site
and that all three stages are necessary for optimal inhibition.
inhibition versus Mg2þ
.
d Each inhibitor displayed mixed inhibition
e No inhibition
versus chorismate and competitively inhibited Mg2þ
.
was observed up to 1000 μM; assaying above 1000 μM was not possible
due to instrument and assay limitations.
inhibitors 3-9 to the active site, only insofar as one can
consider magnesium’s binding site the active site.
Table 4. Validation of PAYLOAD-Directed Active Site Binding to
ADCS, IS, and AS
Two possibilities can explain the competitive inhibition of
IS by PAYLOAD-deleted 14 and 15. In the first possibility,
the COMBI stages of 14 and 15 only bind to IS when
chorismate is absent, in the same manner as the fully staged
inhibitors (e.g., 3-9), and prevent chorismate binding. IS
undergoes a much smaller conformational change on chor-
ismate binding and, by extension, PAYLOAD binding
(Figure 5), compared to ADCS. Therefore, the surface
residues of IS are more similar in orientation in the unbound
and PAYLOAD-bound forms. This would favor COMBI-
SPACER-only 14 and COMBI-only 15 binding to the same
region of IS as do the COMBI elements of fully staged
inhibitors. In the second possibility, the COMBI stage of
14 and 15 interact with IS in a fundamentally different
manner than in the fully staged inhibitors. If the first is
correct, then the optimal binding mode for the COMBI stage
ADCS
Ki (μM)
770 ( 60
IS
mode Ki (μM) mode Ki (μM) mode
Ca 20 ( 2 N-Cc,d
AS
inhibitor
5
210 ( 29 Ca,b
14
15
5700 ( 1000 N-Cb,c 790 ( 88 Ca,b 110 ( 6 N-Cb,c
6800 ( 2000 N-Cb,c 500 ( 72 Ca,b 200 ( 11 N-Cb,c
a C denotes competitive inhibition with respect to chorismate. b Non-
competitive inhibition with respect to Mg2þ was exhibited. c N-C
denotes noncompetitive inhibition with respect to chorismate. d Compe-
titive inhibition with respect to Mg2þ was observed.
alone (i.e., 14 and 15) is that observed for the fully staged
inhibitor. If the second is correct, then the PAYLOAD stage
enforces a specific binding mode for the COMBI stage that is
suboptimal but nevertheless enhances overall binding.
The present work is a major step toward realization of our
goal to find a single, potent compound that inhibits multiple

3726 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 9
Ziebart et al.
chorismate-utilizing enzymes. The inhibition results pre-
sented in Tables 2-4 validate the design approach: lysine
was the correct choice as SPACER; the inhibitors are
directed to the active site of ADCS by the PAYLOAD;
all three stages are necessary for optimal inhibition. Five
compounds (3, 5-7, 9) display low- to midmicromolar
inhibition of three enzymes. Considering that these com-
pounds were discovered in an ADCS on-bead screen, the fact
that better inhibition is observed against IS and AS is
remarkable.
The AS inhibition results presented in Tables 2-4 are
more difficult to interpret due to the noncompetitive inhibi-
tion of chorismate binding by 1-9. It will be necessary to
optimize the PAYLOAD such that it competitively inhibits
chorismate binding before going forward with future library
syntheses. Concurrent with our original report of this pep-
tide-based library, others reported low micromolar inhibi-
tion of AS by 1-carboxyethoxy-containing aromatic
acids.23,24 Among the reported chorismate analogues, the
very simple analogue 1 was chosen to serve as the PAY-
LOAD in the present work because it was best suited for
solid phase synthesis (e.g., compound stability, lack of
stereocenters, pliant protecting group requirements, syn-
thetic accessibility, etc.) in this small, proof-of-concept li-
brary. The next phase of this project is to move to larger,
nonpeptide based libraries with more drug-like properties.
Future efforts will employ the more potent 1-carboxyethoxy-
containing PAYLOAD. If COMBI elements in future non-
peptide based libraries can improve binding by a modest
factor of 40 (as demonstrated for 5 against IS) and if the
PAYLOAD has a low micromolar Ki,23,24 discovery of
fully staged, multienzyme inhibitors with low- to midnano-
molar inhibition is a realistic goal with unique therapeutic
potential.
voltage, and Xterra MS C18 column (2.1 mm ꢁ 50 mm ꢁ 3.5
μm). High-resolution mass spectral (HRMS) data was acquired
on a Thermo Fisher LTQ Orbitrap mass spectrometer (San Jose,
CA) by flow-injection analysis in the positive ion mode using
the IonMax electrospray source with 0.1% formic acid and
methanol (MeOH) as the mobile phases. The source voltage was
5.5 kV, sheath gas setting of 8, and capillary temperature of
250 °C. Compounds synthesized here were confirmed to be
95þ% pure by LC-MS under the conditions described above.
Enzyme Preparation. The plasmid construct bearing the genes
for the partial complex of anthranilate synthase (TrpE2TrpG2)
was a generous gift from Professor Ronald Bauerle, University
of Virginia. Partial complex anthranilate synthase (AS) from
Salmonella typhimurium consists of two TrpE:TrpG heterodi-
mers. Plasmid pSTS23 contains two tandem stop codons en-
gineered into the trpD gene, such that only the region encoding
amidotransferase activity is expressed; the polypeptide trans-
lated from this truncated gene is named TrpG.30
AS was prepared in the following manner: Escherichia coli
cells (strain CB694) harboring the pSTS23 construct were grown
in luria-bertani broth (LB) media at 37 °C to an OD600 of 2.5-3.
Cells were harvested by centrifugation and resuspended in lysis
buffer (10 mM TEA, pH 7.8, 10 mM mercaptoethanol, 5 mM
MgCl2, 1 mM EDTA, 0.5 mg/mL lysozyme, 0.2 units/mL
DNase I) prior to disruption by sonication. Cell debris was
pelleted by centrifugation at 14000 rpm. Ammonium sulfate
(23%) was added to the soluble extract; the resulting precipitate
was resuspended in a minimum volume of start buffer (10 mM
TEA, pH 7.8, 10 mM mercaptoethanol, 5 mM MgCl2, 1 mM
EDTA) and loaded onto Q-Sepharose Fast Flow (Pharmacia)
resin. Protein was eluted with a linear gradient of 500 mL of
0-300 mM KCl in starting buffer. The purest fractions, as
judged by sodium dodecyl sulfate polyacrylamide gel electro-
phoresis (SDS-PAGE) analysis, were concentrated by ultrafil-
tration and dialyzed against 20 mM KPi pH 7.5, 50 mM KCl,
1 mM DTT. Purified enzyme was flash-frozen and stored at
-80 °C. Protein concentration was measured with the Lowry
protein assay kit (Bio-Rad) using IgG as standard. Yield was
∼142 mg purified protein/10 g cell paste.
Experimental Procedures
The entB gene encoding isochorismatase (EntB) was ampli-
fied from E. coli K12 genomic DNA. The PCR primers were
designed to introduce NdeI and XhoI restriction sites into the
PCR products. The sequences of the primers used were: 50-
GGAATTCCATATGGCTATTCCAAAATTACA-30 (forward)
and 50-CCCCTCGAGTCATTATTTCACCTCGCGGGAG-
AG-30 (reverse). PCR products of Taq DNA polymerase were
cloned into the pCR2.1-TOPO vector using the TOPO TA
Cloning kit (Invitrogen). Clones containing entB inserts were
identified and isolated by blue-white screening and were
subsequently subcloned into the pET28a vector (Novagen)
using NdeI and XhoI restriction sites for expression as a
6xHis-tagged fusion protein.
BL21(DE3) E. coli cells harboring the pET28a-entB construct
were grown in LB media at 37 °C to an OD600 of 0.6. Over-
expression proceeded for six hours at 37 °C following induction
by 0.5 mM β-isopropylthiogalactoside (IPTG). Cells were har-
vested by centrifugation and resuspended in lysis buffer (20 mM
Na2HPO4, 500 mM NaCl, 10 mM imidazole, pH 7.4, 0.5 mg/mL
lysozyme, 0.2 units/mL DNase I) prior to disruption by sonica-
tion. Cell debris was pelleted by centrifugation at 14000 rpm for
45 min. The supernatant was incubated at 4 °C for 45 min with
Chelating Sepharose Fast Flow resin (Pharmacia) that had been
charged with Ni2þ. The resin was loaded into a column and
washed with 10 column volumes of starting buffer (20 mM
Na2HPO4, 500 mM NaCl, 10 mM imidazole, pH 7.4). Protein
was eluted with a linear gradient of 500 mL of 10-300 mM
imidazole in starting buffer. The purest fractions, as judged by
SDS-PAGE analysis, were concentrated by ultrafiltration
and dialyzed against 20 mM KPi pH 7.5, 50 mM KCl, 1 mM
DTT. Purified enzyme was flash-frozen and stored at -80 °C.
Materials. All reagents and solvent that were purchased from
commercial suppliers were used without further purification.
Triethanolamine (TEA), bicine, KCl, HCl, NaOH, (NH4)2SO4,
KH2PO4, K2HPO4, MgCl2, glacial acetic acid (AcOH), aceto-
nitrile (ACN), ethylenediaminetetraacetate (EDTA), and
L-glutamine, were purchased from Fisher. Pyridoxal 50-phos-
phate (PLP), dithiothreitol (DTT), nicotinamide adenine dinu-
cleotide, reduced form (NADH), lysozyme, anthranilic acid,
and 2-mercaptoethanol were purchased from Sigma-Aldrich.
The enzymes lactate dehydrogenase (LDH) and DNase I were
purchased from Roche. Chorismate was prepared according to
a literature procedure.25 Fmoc rink amide-MBHA resin
(capacity, 0.59 mmol/g; MBHA, 4-methylbenzhydrylamine;
Fmoc, 9H-fluoren-9-ylmethoxycarbonyl) was purchased from
Hecheng Science and Technology Co., Ltd., and all calculations
for the synthesis were based on a substitution of 0.59 mmol/g.
Fmoc-3-chloro-L-phenylalanine was purchased from Chem-
Impex International, Inc. Fmoc-Ala-OH, Fmoc-Lys(Dde)-
OH, Fmoc-Orn(Dde)-OH, and Fmoc-OSu were purchased
from Novabiochem. The coupling reagents, HOBt and DIC,
were purchased from AK Scientific. All other synthetic reagents
and solvents were purchased from Sigma-Aldrich. Compound
13 was prepared according to literature procedures.26-29 All
infrared spectra were determined on a Genesis II Mattson FT-
IR spectrometer. 1H and 13C NMR were measured in DMSO-d6
at 600 MHz. Preparative high performance liquid chromato-
graphy (HPLC) was performed on a Waters system (2487 dual
wavelength absorbance detector, 600 controller, and a 2767
sample manager) with the following specifications: electrospray
(þ) ionization (ESI), mass range 150-1500 Da, 20 V cone

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 9 3727
regression analysis.
Protein concentration was measured with the Lowry protein
assay kit (Bio-Rad) using IgG as standard. Yield was ∼350 mg
purified protein/28 g cell paste.
Vmax½chorismateꢀ
ꢀ
ꢁ
IS was expressed from a pET28a-entC construct31 and pur-
ified from BL21 (DE3) E. coli cells harboring the pET28a-entC
plasmid construct according to a literature procedure,31 which is
similar to that described above for EntB. Protein yield was 213
mg/42 g cell paste.
vi ¼
ð2Þ
ð3Þ
ð4Þ
½Iꢀ
KM
þ ½chorismateꢀ
Ki
Vmax½chorismateꢀ
ꢀ
ꢁ
ꢀ
ꢁ
vi ¼
vi ¼
½Iꢀ
½Iꢀ
Preparation of the plasmid constructs bearing the genes for
PabA, PabB, and ADCL was described previously.32 Protein
expression and purification of these three enzymes followed
literature procedures.32 Protein yields were as follows: 83 mg
PabA/15 g cell paste; 330 mg ADCS/20 g cell paste; 466 mg
ADCL/37 g cell paste.
KM 1 þ
þ ½chorismateꢀ 1 þ
Ki
Ki
Vmax½chorismateꢀ
ꢀ
ꢁ
ꢀ
ꢁ
½Iꢀ
½Iꢀ
KM 1 þ
þ ½chorismateꢀ 1 þ
Ki
RKi
Spectrofluorometric AS Activity Assay. All kinetic assays
were performed on
a PerkinElmer LS 50B lumine-
Inhibitor Synthesis and Characterization. Synthesis of 2-(3-(-
(S)-5-((S)-1-Amino-3-(3-chlorophenyl)-1-oxopropan-2-ylamino)-
4-(3-hydroxy-4-methyl-2-nitrobenzamido)-5-oxopentylcarbamoyl)-
phenoxy)acetic Acid (10). Rink amide MBHA (1 g, 0.59 mmol/g
loading capacity) resin was weighed into a plastic column and
swollen in dimethylformamide (DMF) for 2 h, after which the
DMF was drained, 20% piperidine in DMF was added to the
resin, and the column was placed on a shaker for 2 ꢁ 10 min.
After the resin was washed (DMF 2ꢁ, MeOH 2ꢁ, dichloro-
methane (DCM) 2ꢁ, MeOH 2ꢁ, and DMF 2ꢁ), Fmoc-3-
chloro-L-phenylalanine (3.0 equiv) and HOBt (5.0 equiv) were
premixed in dry DMF (10 mL) and added to the resin, followed
by addition of diisopropylcarbodiimide (DIC) (5.0 equiv). The
column was placed on a shaker until the Kaiser test was negative
(∼4 h). The resin was washed according to the procedure out-
lined above and the Fmoc group was removed. Following the
standard wash, Fmoc-L-Orn(Dde)-OH (3.0 equiv) and n-hydro-
xybenzotriazole (HOBt) (5.0 equiv) were dissolved in DMF and
added to the resin, with subsequent addition of DIC (5.0 equiv).
The column was placed on a shaker until the Kaiser test was
negative. The resin was washed, followed by Fmoc-deprotection
standard conditions. After the resin was washed, a solution of
3-hydroxy-4-methyl-2-nitrobenzoic acid (3.0 equiv) and HOBt
(5.0 equiv) in DMF were added to the resin followed by
addition of DIC (5.0 equiv). The column was placed on a shaker
until the Kaiser test was negative. After the wash, the 1-(4,4-
dimethyl-2,6-dioxocyclohexylidene)ethyl (Dde) group was re-
moved by addition of hydrazine (2% in DMF, 2 ꢁ 10 min). The
resin was washed, and a DMF solution of 3-(2-tert-butoxy-2-
oxoethoxy)benzoic acid (3.0 equiv), and HOBt (5.0 equiv) was
added, followed by addition of DIC (5.0 equiv). The resin was
washed and a cleavage mixture of trifluoroacetic acid (TFA), H2O,
and triisopropylsilane (TIS) (95:2.5:2.5, v/v/v) was added. The
column was placed on a shaker for 2 h, and the filtrate was drained
and collected. The TFA was evaporated under a constant stream of
nitrogen, and the resulting residue was precipitated from ether and
stored in the refrigerator overnight. The ether-crude mixture was
centrifuged and the ether decanted. The crude mixture was purified
by reversed-phase HPLC (630 mg, 63% yield, purity >99%); mp
154-156 °C. IR (thin film, selected peaks): ν 3266, 1637, 1124,
1324, 1229, 1020, 685 cm-1. HRMS (ESI) m/z [M þ H]þ calculated
for C31H33ClN5O10: 670.1916; found 670.1909. 1H NMR (DMSO,
600 MHz): δ 13.01 (br s, 1 H), 10.06 (s, 1 H), 8.64 (d, 1 H, J = 7.8
Hz), 8.43 (t, 1 H, J = 5.4 Hz), 7.94 (d, 1 H, J = 8.4 Hz), 7.45-7.05
(m, 12 H), 4.72 (s, 2 H), 4.45 (m, 1 H), 4.32 (m, 1 H), 3.24 (m, 2 H),
3.01 (dd, 1 H, J = 13.8, 5.4 Hz), 2.83 (dd, 1 H, J = 13.8, 9.0 Hz),
2.27 (s, 3 H), 1.70-1.46 (m, 4 H). 13CNMR(DMSO-d6, 150 MHz):
δ 172.4, 171.1, 170.1, 165.7, 164.3, 157.6, 147.1, 140.3, 139.5, 136.0,
132.7, 131.8, 131.3, 129.8, 129.4, 129.0, 128.0, 127.3, 126.3, 119.9,
119.0, 117.3, 113.1, 64.5, 53.4, 53.2, 38.9, 37.3, 29.2, 25.8, 16.5.
Synthesis of 2-(3-(3-((R)-3-((S)-1-Amino-3-(3-chlorophenyl)-
1-oxopropan-2-ylamino)-2-(3-hydroxy-4-methyl-2-nitrobenza-
mido)-3-oxopropylthio) propylcarbamoyl)phenoxy)acetic Acid
(11). A procedure identical to that described for 10 was fol-
lowed, with the exception that compound 13 was used instead of
scence spectrophotometer at 25 °C. The activity assay for
AS was based on a previously described method.30 The assay
directly monitors anthranilate emission at 390 nm using
excitation at 313 nm. Reactions were initiated by addition
of anthranilate synthase following a 5 min incubation at
25 °C. Each 1 mL reaction consisted of 100 mM phos-
phate buffer, pH 7.0, 20 mM L-glutamine, 5 mM MgCl2,
and 2.2 nM AS.
Spectrophotometric IS and ADCS Activity Assays. Kinetic
assays were performed on either a Kontron Uvikon 930 or a
Shimadzu UV-2450 UV-vis spectrophotometer at 25 °C. The
activity assay for IS was described previously.31 Formation of
isochorismate was monitored at 278 nm (Δεisochorismate-chorismate
=
10211 M-1 cm-1).33 Each 1 mL reaction consisted of 50 mM
triethanolamine, pH 7.8, 1 mM MgCl2, and 50 nM IS. Reac-
tion mixtures were incubated for 5 min at 25 °C prior to initiation
with IS.
ADCS activity was monitored by following the decrease in
absorbance at 340 nm, due to oxidation of NADH in a coupled
assay with excess LDH and ADCL. Each 0.5 mL reaction
contained 0.1 M bicine, pH 8.5, 0.02 M L-glutamine, 5 mM
MgCl2, 20 μM PLP, 100 μM NADH, 8 μM ADCL, 2 units
LDH, 0.5 μM PabA, and 0.5 μM PabB. Reaction mixtures were
incubated for 5 min at 25 °C prior to initiation with an
equimolar solution of PabA and PabB. The rate of 4-amino-4-
deoxychorismate (ADC) formation was confirmed to be linearly
dependent on PabB concentration in the coupled assay.
Inhibition Assays and Ki Determination. Because of strong
absorbance at 278 nm by the inhibitors, the IS activity assay
described above was incompatible as an inhibition assay. There-
fore, inhibition data was acquired by measuring initial rates in
a coupled assay with excess EntB and LDH enzymes. The
decrease in absorbance at 340 nm due to oxidation of NADH
was monitored. Each 0.9 mL reaction contained 50 mM trietha-
nolamine, pH 7.8, 1 mM MgCl2, 100 μM NADH, 2 μM EntB,
2 units LDH, and 75 nM IS. The rate of isochorismate forma-
tion was confirmed to be linearly dependent on IS concentration
in this coupled assay. The AS inhibition assay is the same as that
described above; ADCS inhibition data was collected spectro-
photometrically using the ADCL-LDH coupled assay described
above.
Inhibitors were dissolved in water and brought to pH 8.5
with 25 mM NaOH. Inhibitor concentrations were varied from
10 μM to 8 mM, chorismate concentration was varied from
0.5 to 5 KM, and magnesium concentration was saturating (e.g.,
1 mM for IS reactions, 5 mM for ADCS and AS reactions). To
determine the inhibition pattern with respect to magnesium,
Mg2þ concentration was varied from 5 to 500 μM, and chor-
ismate concentration was saturating (e.g., 200 μM for IS and
ADCS reactions, 40 μM for AS reactions). Ki values were
determined by globally fitting the data to competitive (eq 2),
noncompetitive (eq 3), and mixed (eq 4) kinetic equations via the
software program GraFit, Version 4.0.21 (Erithacus). Reported
Ki values are from the best global fit, as determined by nonlinear

3728 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 9
Ziebart et al.
Fmoc-L-Orn(Dde)-OH. Therefore, removal of the Dde group was
not necessary, nor was addition of 3-(2-tert-butoxy-2-oxoethoxy)-
benzoic acid. The crude mixture was purified by reversed-phase
HPLC (27 mg, 31% yield, purity >99%); mp 168-170 °C. IR
(neat, selected peaks): 3266, 16.38, 1580, 1543, 1532 cm-1. HRMS
(ESI) m/z [M þ H]þ calcd for C32H34ClN5O10S: 716.1793; found
(2) Klevens, R. M.; Morrison, M. A.; Nadle, J.; Petit, S.; Gershman,
K.; Ray, S.; Harrison, L. H.; Lynfield, R.; Dumyati, G.; Townes,
J. M.; Craig, A. S.; Zell, E. R.; Fosheim, G. E.; McDougal,
L. K.; Carey, R. B.; Fridkin, S. K. Invasive methicillin-resistant
Staphylococcus aureus infections in the United States. JAMA,
J. Am. Med. Assoc. 2007, 298, 1763–1771.
(3) Jarvis, L. M. An Uphill Battle. Chem. Eng. News 2008, 86, 15–20.
(4) Fischbach, M. A.; Walsh, C. T. Antibiotics for Emerging Patho-
gens. Science 2009, 325, 1089–1093.
(5) Dixon, S.; Ziebart, K. T.; He, Z.; Jeddeloh, M.; Yoo, C. L.; Wang,
X.; Lehman, A.; Lam, K. S.; Toney, M. D.; Kurth, M. J. Amino-
deoxychorismate synthase inhibitors from one-bead one-com-
pound combinatorial libraries: “staged” inhibitor design. J. Med.
Chem. 2006, 49, 7413–7426.
(6) Roberts, F.; Roberts, C. W.; Johnson, J. J.; Kyle, D. E.; Krell, T.;
Coggins, J. R.; Coombs, G. H.; Milhous, W. K.; Tzipori, S.;
Ferguson, D. J.; Chakrabarti, D.; McLeod, R. Evidence for the
shikimate pathway in apicomplexan parasites. Nature 1998, 393,
801–805.
1
716.1864. H NMR (DMSO-d6): δ 13.05 (s, 1H), 10.09 (s, 1H),
8.72 (dd, J = 8.2, 1H), 8.47 (s, 1H), 8.14 (d, J = 8.2, 1H),
7.44-7.36 (m, 5H), 7.28 (s, 1H), 7.23 (d, J = 5.4, 2H), 7.21-7.12
(m, 3H), 7.06 (d, J = 8.2, 1H), 4.72 (s, 2H), 4.48 (m, 2H), 3.31 (s,
2H), 3.03 (m, 1H), 2.89-2.77 (m, 2H), 2.77-2.64 (m, 1H),
2.64-2.54 (m, 2H), 2.27 (s, 3H), 1.75 (m, 2H). 13C NMR
(DMSO-d6): δ 172.9, 170.7, 170.4, 166.4, 164.6, 158.3, 147.6,
140.9, 140.3, 136.4, 133.3, 132.4, 132.0, 130.0, 129.7, 129.6,
128.7, 127.4, 127.0, 120.6, 119.7, 117.9, 113.3, 65.1, 54.3, 53.6,
38.9, 37.8, 33.7, 29.7, 29.4, 17.2.
Synthesis of N-((S)-6-Amino-1-((S)-1-amino-3-(3-chloro-
phenyl)-1-oxopropan-2-ylamino)-1-oxohexan-2-yl)-3-hydroxy-
4-methyl-2-nitrobenzamide (14). A procedure identical to that
described for 10 was followed, with the exception that Fmoc-
L-Lys(Dde)-OH was used instead of Fmoc-L-Orn(Dde)-OH and
the addition of a DMF solution of 3-(2-tert-butoxy-2-oxoethoxy)-
benzoic acid (3.0 equiv), and HOBt (5.0 equiv), followed by
addition of DIC (5.0 equiv) was not performed after the Dde
deprotection step. The crude mixture was purified by reversed-
phase HPLC (212 mg, 71% yield, purity >99%); mp 134-135 °C.
(7) Morphy, R.; Rankovic, Z. Designed multiple ligands. An emerging
drug discovery paradigm. J. Med. Chem. 2005, 48, 6523–6543.
(8) Morphy, R.; Kay, C.; Rankovic, Z. From magic bullets to designed
multiple ligands. Drug Discovery Today 2004, 9, 641–651.
(9) Weber, G.; Prajda, N.; Lui, M. S.; Denton, J. E.; Aoki, T.; Sebolt,
J.; Zhen, Y. S.; Burt, M. E.; Faderan, M. A.; Reardon, M. A. Multi-
enzyme-targeted chemotherapy by acivicin and actinomycin. Adv.
Enzyme Regul. 1982, 20, 75–96.
(10) Hibasami, H.; Maekawa, S.; Murata, T.; Nakashima, K. Antitumor
effect of a new multienzyme inhibitor of polyamine synthetic path-
way, methylglyoxal-bis(cyclopentylamidinohydrazone), against hu-
man and mouse leukemia cells. Cancer Res. 1989, 49, 2065–2068.
(11) Zhan, P.; Liu, X. Y. Designed Multiple Ligands: An Emerging
Anti-HIV Drug Discovery Paradigm. Curr. Pharm. Des. 2009, 15,
1893–1917.
(12) Arora, P.; Goyal, A.; Natarajan, V. T.; Rajakumara, E.; Verma, P.;
Gupta, R.; Yousuf, M.; Trivedi, O. A.; Mohanty, D.; Tyagi, A.;
Sankaranarayanan, R.; Gokhale, R. S. Mechanistic and functional
insights into fatty acid activation in Mycobacterium tuberculosis.
Nat. Chem. Biol. 2009, 5, 166–173.
(13) Barkan, D.; Liu, Z.; Sacchettini, J. C.; Glickman, M. S. Mycolic
Acid Cyclopropanation is Essential for Viability, Drug Resistance,
and Cell Wall Integrity of Mycobacterium tuberculosis. Chem. Biol.
2009, 16, 499–509.
IR (neat, selected peaks): 2358, 2314, 1683, 1647, 1635, 1544 cm-1
.
HRMS (ESI) m/z [M þ H]þ calculated for C23H28ClN5O6:
506.1806; found 506.1832. 1H NMR (DMSO-d6): δ 10.14 (s,
1H), 8.62 (d, J = 8.1, 1H), 7.94 (d, J = 8.2, 1H), 7.72 (s, 3H),
7.45 (s, 1H), 7.39 (d, J = 8.5, 1H), 7.28 (d, J = 7.2, 1H), 7.27-7.23
(m, 2H), 7.19 (d, J = 6.8, 1H), 7.14 (s, 1H), 4.51-4.41 (m, 1H),
4.28 (d, J= 5.2, 1H), 3.04 (dd, J= 4.9, 13.8, 1H), 2.84 (dd, J=9.1,
13.8, 1H), 2.75-2.74 (m, 2H), 2.29 (s, 3H), 1.69-1.44 (m, 4H), 1.26
(s, 2H). 13C NMR (DMSO-d6): δ 172.9, 171.5, 165.1, 148.0, 140.9,
140.0, 133.4, 129.7, 128.7, 128.4, 126.8, 126.2, 119.9, 119.8, 105.0,
54.0, 40.5, 37.8, 31.9, 27.2, 22.9, 17.0, 16.9.
(14) Chen, J. K.; Lane, W. S.; Brauer, A. W.; Tanaka, A.; Schreiber, S.
L. Biased combinatorial libraries: Novel ligands for the SH3
domain of phosphatidylinositol 3-kinase. J. Am. Chem. Soc.
1993, 115, 12591–12592.
(15) Sternson, S. M.; Wong, J. C.; Grozinger, C. M.; Schreiber, S. L.
Synthesis of 7200 small molecules based on a substructural analysis
of the histone deacetylase inhibitors trichostatin and trapoxin. Org.
Lett. 2001, 3, 4239–4242.
Synthesis of N-((S)-1-((S)-1-Amino-3-(3-chlorophenyl)-1-oxo-
propan-2-ylamino)-1-oxopropan-2-yl)-3-hydroxy-4-methyl-2-
nitrobenzamide (15). A procedure identical to that described
for 14 was followed, with the exception that Fmoc-L-Ala-OH
was used instead of Fmoc-L-Lys(Dde)-OH; therefore, the
Dde deprotection was not performed. The crude mixture
was purified by reversed-phase HPLC (38 mg, 30% yield,
purity >99%); mp 240-242 °C. IR (neat, selected peaks):
3274, 2969, 2358, 1642, 1545, 1418, 1478 cm-1. HRMS (ESI)
m/z [M þ H]þ calculated for C20H21ClN4O6: 449.1228; found
449.1267. 1H NMR (DMSO-d6): δ 10.04 (s, 1H), 8.63 (d, J =
7.3, 1H), 7.89 (d, J = 8.2, 1H), 7.35 (d, J = 7.8, 2H),
7.20-7.25 (m, 3H), 7.11-7.16 (m, 3H), 4.38 (dd, J = 8.6,
13.4, 1H), 4.28 (t, J = 7.2, 1H), 2.99 (dd, J = 4.9, 13.8, 1H),
2.81 (dd, J = 8.8, 13.8, 1H), 2.25 (s, 3H), 1.19 (d, J = 7.1, 1H).
13C NMR (DMSO-d6): δ 173.1, 172.3, 164.7, 147.8, 141.1,
140.2, 133.3, 132.5, 132.0, 130.5, 129.7, 128.7, 127.7, 126.9,
119.8, 54.1, 49.7, 37.7, 18.3, 17.2.
(16) Erlanson, D. A. Fragment-based lead discovery: a chemical up-
date. Curr. Opin. Biotechnol. 2006, 17, 643–652.
(17) Parsons, J. F.; Jensen, P. Y.; Pachikara, A. S.; Howard, A. J.;
Eisenstein, E.; Ladner, J. E. Structure of Escherichia coli amino-
deoxychorismate synthase: architectural conservation and diversity
in chorismate-utilizing enzymes. Biochemistry 2002, 41, 2198–2208.
(18) Spraggon, G.; Kim, C.; Nguyen-Huu, X.; Yee, M. C.; Yanofsky,
C.; Mills, S. E. The structures of anthranilate synthase of Serratia
marcescens crystallized in the presence of (i) its substrates, choris-
mate and glutamine, and a product, glutamate, and (ii) its end-
product inhibitor, L-tryptophan. Proc. Natl. Acad. Sci. U.S.A 2001,
98, 6021–6026.
(19) Morollo, A. A.; Eck, M. J. Structure of the cooperative allosteric
anthranilate synthase from Salmonella typhimurium. Nat. Struct.
Biol. 2001, 8, 243–247.
(20) Parsons, J. F.; Shi, K. M.; Ladner, J. E. Structure of isochorismate
synthase in complex with magnesium. Acta Crystallogr., Sect. D:
Biol. Crystallogr. 2008, 64, 607–610.
(21) Kolappan, S.; Zwahlen, J.; Zhou, R.; Truglio, J. J.; Tonge, P. J.;
Kisker, C. Lysine 190 is the catalytic base in MenF, the menaqui-
none-specific isochorismate synthase from Escherichia coli: impli-
cations for an enzyme family. Biochemistry 2007, 46, 946–953.
(22) Segel, I. H. Enzyme Kinetics: Behavior and Analysis of Rapid
Equilibrium and Steady State Enzyme Systems. Wiley: New York,
1975; pp xxii, 957.
The synthesis of compounds 1-9 has been described pre-
viously.5
Acknowledgment. M.J.K. acknowledges support of this
work from the National Science Foundation (CHE-0910870).
Supporting Information Available: Proton and carbon NMR
spectra for compounds 10, 11, 14, and 15. This material is
available free of charge via the Internet at http://pubs.acs.org.
(23) Payne, R. J.; Bulloch, E. M.; Abell, A. D.; Abell, C. Design and
synthesis of aromatic inhibitors of anthranilate synthase. Org.
Biomol. Chem. 2005, 3, 3629–3635.
(24) Payne, R. J.; Kerbarh, O.; Miguel, R. N.; Abell, A. D.; Abell, C.
Inhibition studies on salicylate synthase. Org. Biomol. Chem. 2005,
3, 1825–1827.
References
(1) Coates, A.; Hu, Y.; Bax, R.; Page, C. The future challenges facing
the development of new antimicrobial drugs. Nat. Rev. Drug
Discovery 2002, 1, 895–910.

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 9 3729
(25) Grisostomi, C.; Kast, P.; Pulido, R.; Huynh, J.; Hilvert, D.
Efficient in Vivo Synthesis and Rapid Purification of Chorismic
Acid Using an Engineered Escherichia coli Strain. Bioorg. Chem.
1997, 25, 297–305.
(26) Ensch, C.; Hesse, M. Total syntheses of the spermine alkaloids
(-)-(R,R)-hopromine and (()-homaline. Helv. Chim. Acta 2002,
85, 1659–1673.
(27) Bolognese, A.; Fierro, O.; Guarino, D.; Longobardo, L.; Caputo,
R. Chiral aminoalkyl cation equivalents. 1. One-pot synthesis of
orthogonally protected enantiopure S-(aminoalkyl)cysteine deri-
vatives. Eur. J. Org. Chem. 2005, 169–173.
(28) Lopez, P. E.; Isidro-Llobet, A.; Gracia, C.; Cruz, L. J.; Garcia-
Granados, A.; Parra, A.; Alvarez, M.; Albericio, F. Use of
p-nitrobenzyloxycarbonyl (pNZ) as a permanent protecting group
in the synthesis of Kahalalide F analogs. Tetrahedron Lett. 2005,
46, 7737–7741.
(29) Tietze, L. F.; Panknin, O.; Major, F.; Krewer, B. Synthesis of a
novel pentagastrin-drug conjugate for a targeted tumor therapy.
Chem.;Eur. J. 2008, 14, 2811–2818.
(30) Bauerle, R.; Hess, J.; French, S. Anthranilate synthase-
anthranilate phosphoribosyltransferase complex and subunits of
Salmonella typhimurium. Methods Enzymol. 1987, 142, 366–386.
(31) He, Z.; Toney, M. D. Direct detection and kinetic analysis of covalent
intermediate formation in the 4-amino-4-deoxychorismate synthase
catalyzed reaction. Biochemistry 2006, 45, 5019–5028.
(32) He, Z.; Stigers Lavoie, K. D.; Bartlett, P. A.; Toney, M. D.
Conservation of mechanism in three chorismate-utilizing enzymes.
J. Am. Chem. Soc. 2004, 126, 2378–2385.
(33) Morollo, A. A.; Bauerle, R. Characterization of composite amino-
deoxyisochorismate synthase and aminodeoxyisochorismate lyase
activities of anthranilate synthase. Proc. Natl. Acad. Sci. U.S.A
1993, 90, 9983–9987.