Phosphonosulfonates are potent, selective inhibitors of dehydrosqualene synthase and staphyloxanthin biosynthesis in staphylococcus aureus

Article abstract of DOI:10.1021/jm801023u

Staphylococcus aureus produces a golden carotenoid virulence factor called staphyloxanthin (STX), and we report here the inhibition of the enzyme, dehydrosqualene synthase (CrtM), responsible for the first committed step in STX biosynthesis. The most acti

Full text of DOI:10.1021/jm801023u

976
J. Med. Chem. 2009, 52, 976–988
Phosphonosulfonates Are Potent, Selective Inhibitors of Dehydrosqualene Synthase and
Staphyloxanthin Biosynthesis in Staphylococcus aureus
Yongcheng Song,† Fu-Yang Lin,¶ Fenglin Yin,¶ Mary Hensler,‡ Carlos A. Rodr´ıgues Poveda,§ Dushyant Mukkamala,¶
Rong Cao,¶ Hong Wang, Craig T. Morita, Dolores Gonza´lez Pacanowska,§ Victor Nizet,‡ and Eric Oldfield*,†,¶
Department of Chemistry, UniVersity of Illinois at UrbanasChampaign, 600 South Mathews AVenue, Urbana, Illinois 61801, Center for
Biophysics and Computational Biology, UniVersity of Illinois at UrbanasChampaign, 607 South Mathews AVenue, Urbana, Illinois 61801,
Department of Pediatrics and Skaggs School of Pharmacy and Pharmaceutical Sciences, UniVersity of California, San Diego, 9500 Gilman
DriVe, La Jolla, California 92093-0687, Instituto de Parasitologa´ y Biomedicina Lo´pez-Neyra, Consejo Superior de InVestigaciones
Cientı´ficas, Granada, Spain, and DiVision of Rheumatology, Department of Internal Medicine, EMRB 400F, UniVersity of Iowa CarVer
College of Medicine, Iowa City, Iowa 52242
ReceiVed August 14, 2008
Staphylococcus aureus produces a golden carotenoid virulence factor called staphyloxanthin (STX), and we
report here the inhibition of the enzyme, dehydrosqualene synthase (CrtM), responsible for the first committed
step in STX biosynthesis. The most active compounds are halogen-substituted phosphonosulfonates, with
Ki values as low as 5 nM against the enzyme and IC50 values for STX inhibition in S. aureus as low as 11
nM. There is, however, only a poor correlation (R2 ) 0.27) between enzyme and cell pIC50 () -log10 IC50)
values. The ability to predict cell from enzyme data improves considerably (to R2 ) 0.72) with addition of
two more descriptors. We also investigated the activity of these compounds against human squalene synthase
(SQS), as a counterscreen, finding several potent STX biosynthesis inhibitors with essentially no squalene
synthase activity. These results open up the way to developing potent and selective inhibitors of an important
virulence factor in S. aureus, a major human pathogen.
Introduction
CrtM, and involves the head-to-head condensation of two molecules
of farnesyl diphosphate (FPP) to produce the C30 species,
presqualene diphosphate, which is then converted to dehy-
drosqualene (Figure 1A).5 Since this condensation is remarkably
similar to the first step in mammalian cholesterol biosynthesis
(Figure 1B), we reasoned that known squalene synthase inhibitors,
developed in the context of cholesterol-lowering therapy, might
also inhibit dehydrosqualene synthase. This turns out to be the case,
and we recently reported that phosphonosulfonates such as 1 (BPH-
652 or rac-BMS-187745), developed by Bristol-Myers Squibb and
advanced through phase I/II human clinical trials,7,8 potently inhibit
S. aureus CrtM, as well as STX biosynthesis in the bacterium.6
Upon treatment with 1, the resulting nonpigmented S. aureus are
much more susceptible to killing by hydrogen peroxide and are
less able to survive in freshly isolated human whole blood than
are normally pigmented S. aureus. Moreover, in an in vivo systemic
S. aureus infection model, the bacterial counts in kidneys of mice
treated with 1 were reduced by 98%, compared to those of a control
group. These results show that 1 represents a novel lead compound
for virulence factor-based therapy of S. aureus infection. Here, we
report the synthesis and testing of a library of 38 phosphonosul-
fonates and related bisphosphonates against CrtM, against STX
biosynthesis in S. aureus, and as a counterscreen, against an
expressed human squalene synthase. We report both qualitative
and quantitative (QSAR, quantitative structure-activity relation-
ships) results for CrtM and STX biosynthesis inhibition, as well
as CrtM/SQS selectivity. In addition, we investigate how cell
activity can be predicted from enzyme (CrtM) inhibition results,
opening the way to the further development of potent and selective
inhibitors of S. aureus virulence.
Staphylococcus aureusis a major human pathogen, producing
a wide spectrum of clinically significant hospital- and community-
acquired infections. Methicillin-resistant strains of S. aureus
(MRSA) have now reached epidemic proportions and pose a
significant challenge to the public health. A recent CDC study has
shown that more people in the U.S. die from invasive MRSA each
year than do from HIV/AIDS.1,2 There is, therefore, an urgent need
to find new therapies. One unconventional approach to anti-
infective therapy involves blocking bacterial virulence factors,3 a
potential benefit of this strategy being that without the “life or
death” selective pressure exerted by classical antibiotics, bacteria
may be less prone to develop drug resistance.
An important virulence factor of S. aureus is the golden
carotenoid pigment, staphyloxanthin (STXa), whose numerous
double bonds can react with, and thus deactivate, the reactive
oxygen species (ROS) generated by neutrophils and macroph-
ages, making S. aureus resistant to innate immune clearance.4,5
STX has been shown to be essential for infectivity: bacteria
that lack staphyloxanthin are nonpigmented, are susceptible to
neutrophil killing, and fail to produce disease in mouse skin
and systemic infection models.4,6 STX biosynthesis is thus a
novel target for preventing or treating S. aureus infections. The
first committed step in STX biosynthesis is catalyzed by the enzyme
dehydrosqualene synthase, also called diapophytoene synthase or
* To whom correspondence should be addressed. Phone: 217-333-3374.
Fax: 217-244-0997. E-mail: eo@chad.scs.uiuc.edu.
†
Department of Chemistry, University of Illinois at UrbanasChampaign.
Center for Biophysics and Computational Biology, University of Illinois
¶
at UrbanasChampaign.
‡
University of California, San Diego.
Instituto de Parasitologa´ y Biomedicina Lo´pez-Neyra.
§
Results and Discussion
University of Iowa Carver College of Medicine.
a Abbreviations: CrtM, dehydrosqualene synthase; FPP, farnesyl diphos-
phate; QSAR, quantitative structure-activity relationship; SQS, squalene
synthase; STX, staphyloxanthin.
Qualitative Structure-Activity Relationships in CrtM
Inhibition. The inhibitor class we investigate here originated
in early research aimed at developing human squalene synthase
10.1021/jm801023u CCC: $40.75
2009 American Chemical Society
Published on Web 02/03/2009

Inhibition of Staphyloxanthin Biosynthesis
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 4 977
Figure 1. (A) Pathway for staphyloxanthin biosynthesis (in S. aureus). (B) Pathway for cholesterol biosynthesis. Each biosynthetic pathway is
thought to involve initial formation of presqualene diphosphate, catalyzed by CrtM (S. aureus) or squalene synthase. In the bacterium, the NADPH
reduction step is absent, resulting in production of dehydrosqualene.
Figure 2. Structures of phosphonosulfonate and bisphosphonate compounds and their activities (IC50 values, µM, in parentheses) against
dehydrosqualene synthase.
inhibitors as cholesterol lowering drugs. Early SQS inhibitors
were based on the structure of FPP, the substrate for SQS, with
the labile diphosphate group being replaced by a bisphos-
phonate. However, isoprenyl side chains were metabolized
in vivo. This could be overcome by incorporation of a
biphenyl isostere (e.g., 2),9 but these compounds potently
bound to bone, in addition to causing elevation in liver
enzymes. Phosphonosulfonates, on the other hand, bound only
weakly to bone, and the diphenyl ether phosphonosulfonates
(such as 1) were not broken down and did not cause liver
function abnormalities.10 Plus, in our previous work we found
that 1 had promising activity against CrtM as well as cellular

978 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 4
Song et al.
substituted fluoro and n-propyl groups, respectively) were
∼4× as active (IC50 ) 2.3 and 2.2 µM; Ki ≈ 5 nM) as 1.
Compound 7 (IC50 ) 7.2 µM), containing a p-trifluoromethyl
substituent, showed activity similar to that found with 1,
while 8, which contains an o-benzyl group, was ∼5× less
active than 1 (IC50 ) 37.2 µM). The phosphonosulfonates
containing biphenyl side chains (9-13) had, in general,
significantly reduced activities.
These results, together with the bacterial cell-based results
discussed below, are of interest because they show that the
biphenyl bisphosphonates are the most potent CrtM inhibitors,
with IC50 values of <1 µM. They do, however, have poor cell-
based activities, with IC50 > 1 µM, and consequently were not
selected for further development. The biphenyl phosphonosul-
fonates, on the other hand, had generally poor activity against
CrtM (average IC50 for the five compounds investigated of ∼50
µM), making them also less attractive candidates for develop-
ment. However, the diphenyl ether phosphonosulfonates had,
on average, an IC50 value of ∼11 µM (or a Ki of ∼30 nM), and
these compounds were very potent in cell based assays (i.e., in
inhibiting STX biosynthesis by S. aureus), as discussed in detail
below, and were thus selected for further development.
and in vivo activity.6 In the present study, we sought to
develop more potent CrtM inhibitors with improved activity
in S. aureus cells that, at the same time, have poor activity
against human SQS, reducing formation of the 1,10-dioic acid
FPP metabolite that is formed as a result of SQS inhibition.
We first synthesized a small library of five diphenyl ether
phosphonosulfonates, five biphenyl phosphonosulfonates, and
three biphenyl bisphosphonates, based in part on the types
of compound tested previously as SQS inhibitors, and
examined them for their activity against CrtM. The structures
and IC50 values (in parentheses) in CrtM inhibition of these
compounds are shown in Figure 2. Bisphosphonates (2 and
3) are the most potent CrtM inhibitors (IC50 ) 0.5 and 0.2
µM; Ki ≈ 1 and 0.5 nM), being an order of magnitude more
active than any of the phosphonosulfonates. However, these
compounds were found to have only modest activity in
inhibition of STX biosynthesis in S. aureus (see below), due
perhaps to poor cell uptake. Compound 4 (IC50 ) 5.4 µM),
an analogous phosphinophosphonate compound, was ∼10×
less active than was 2, showing similar activity in CrtM
inhibition as the phosphonosulfonates, which are likely to
have the same formal charge. For the phosphonosulfonates
with diphenyl ether side chains, 5 and 6 (which contain para-
To see if major improvements in CrtM activity might be
obtained by modifying the phosphonosulfonate side chain, we
next synthesized the eight analogues of 1, shown in Figure 3,
in which we modified the side chain heteroatom (O f NH, CH2,
and a carbazole), the position of the heteroatom (meta f para),
and the length of the alkyl side chain (N ) 1, 2, and 3 CH2
groups), and we synthesized the (R)- and (S)-enantiomers of 1,
since previously the S-form was found to be far more potent
than the (R)-form in inhibiting SQS.11 The optically active (S)-1
(IC50 ) 1.4 µM) was ∼30× more active than its (R)-enantiomer
in CrtM inhibition (Figure 3). Shorter linkers between the
aromatic ring and phosphonosulfonate headgroup, as found in
Figure 3. Structures of compounds and their activities (in parentheses) against dehydrosqualene synthase.

Inhibition of Staphyloxanthin Biosynthesis
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 4 979
Table 2. CoMSIA Results for CrtM Inhibition
Table 1. IC50 Values of Diphenyl Ether Phosphonosulfonates in CrtM
Inhibition
CrtM enzyme
experimental
activity
CoMSIA pIC50 predictions
test setsb
IC50 pIC50 training
compda (µM) (M)
set
residual
1
2
3
4
5
3
2
0.18
0.48
1.7
2.1
2.1
2.2
2.2
2.2
2.3
2.5
4.6
5.4
5.8
6.3
6.9
7.2
7.9
6.74
6.32
5.76
5.68
5.67
5.66
5.66
5.65
5.63
5.6
5.34
5.27
5.24
5.2
5.16
5.14
5.1
5.04
5.01
4.98
4.87
4.69
4.66
4.61
4.58
4.55
4.43
4.36
4.27
4.14
4.09
3.91
3.77
3.77
3.3
6.61
0.13 6.54 6.53 6.53 6.49 6.58
6.55 -0.23 6.35 6.41 6.19 6.29 6.49
5.62 0.14 5.61 5.60 5.56 5.55 5.70
30
35
32
6
20
31
5
24
33
4
18
34
21
7
5.69 -0.01 5.69 5.41 5.51 5.58 5.76
5.69 -0.02 5.65 5.46 5.44 5.54 5.49
5.64
5.34
5.62
5.44
5.46
0.02 5.82 5.12 5.60 5.63 5.61
0.32 5.33 5.32 5.29 5.25 5.41
0.03 5.77 5.90 5.97 5.88 5.69
0.19 5.45 5.34 5.32 5.37 5.48
0.14 5.61 5.49 5.54 5.52 5.51
5.37 -0.03 5.35 5.38 5.36 5.28 5.44
5.21 0.06 5.09 5.13 4.87 5.06 5.23
5.34 -0.10 5.28 5.10 5.20 5.22 5.32
5.21 -0.01 5.24 5.47 5.54 5.40 5.23
5.18 -0.02 5.16 5.01 5.03 5.04 5.25
5.30 -0.16 5.59 5.33 5.51 5.47 5.32
5.15 -0.05 5.15 5.05 5.03 5.05 5.18
1
10
22
25
26
17
9
11
29
27
8
16
23
12
36
13
19
28
15
14
9.1
9.8
4.92
0.12 4.94 4.99 5.04 4.94 4.88
5.05 -0.04 5.23 5.28 5.32 5.43 5.05
5.10 -0.12 5.37 4.74 5.18 5.15 4.95
4.91 -0.04 4.86 5.03 4.67 4.69 4.83
4.72 -0.02 4.89 4.81 4.82 4.81 4.67
4.73 -0.07 4.79 4.79 4.85 4.77 4.68
4.75 -0.14 4.89 4.80 5.03 4.82 4.65
4.71 -0.13 4.81 4.44 4.66 4.57 4.61
4.55
4.43
4.36
4.23
10.5
13.5
20.4
21.9
24.5
26.3
28.2
37.2
43.7
53.7
72.4
81.3
123
170
170
500
0.00 4.55 4.55 4.55 4.55 4.55
0.00 4.35 4.31 4.27 4.34 4.68
0.00 4.36 4.36 4.80 4.36 4.36
0.04 4.28 4.37 4.32 4.33 4.24
4.15 -0.01 4.10 3.72 4.06 3.89 4.08
4.16 -0.07 4.28 4.48 4.46 4.50 4.15
3.86
0.05 3.91 4.16 4.08 4.12 3.84
3.79 -0.02 3.91 4.09 3.92 4.07 3.70
3.80 -0.03 3.90 3.85 3.96 4.20 3.82
3.33 -0.03 3.24 3.25 3.12 3.20 3.40
1700 2.77
2.65
0.12 2.53 2.65 2.50 2.56 2.71
r2
0.98
0.72
0.97 0.94 0.95 0.95 0.99
0.69 0.71 0.72 0.67 0.70
151.6 82.4 113.9 97.5 280.1
q2
Ftest 245.8
N
n
7
36
6
31
5
4
5
7
31 31
31 31
a Structures shown in Figures 2 and 3 and Table 1. b Bold values represent
predicted activities of compounds that were not included in the training
set.
14 and 15, led to essentially no CrtM activity (IC50 > 500 µM).
As for 16 (IC50 ) 43.7 µM) and 17 (IC50 ) 20.4 µM), replacing
the bridging -O- atoms with -NH- and -CH2-, respec-
tively, resulted in reduced CrtM inhibition activity. The 4-phe-
noxyphenyl analogue (18) was slightly more active than its
3-phenoxyphenyl counterpart (1) in CrtM inhibition but was
less active in the cell based assay, while compound 19, a fused
position significantly decrease activity. (4) Dihalogen substituted
compounds are more active than the parent compound 1. (5)
Substitution at the 6-position with F has little effect on activity,
but incorporation of an OMe group in this location abrogates
most activity. Clearly, the above represents a complex set of
empirical observations that require a more quantitative analysis.
Quantitative Structure-Activity Relationships in CrtM
Inhibition. To investigate the structure-activity relationships
outlined above, we carried out a comparative molecular similar-
ity indices analysis (CoMSIA)12 of the CrtM inhibition activities
of these phosphonosulfonate compounds. All molecules were
constructed and minimized using the MMFF94x force field in
the program MOE.13 Compounds were then aligned using the
flexible alignment module14 in MOE, which perceives common
features within the molecules (e.g., similar partial charge,
H-bond acceptor, aromaticity, hydrophobicity, etc.). We used
fully deprotonated phosphonosulfonate headgroups, based on
the observation that 1 binds two Mg2+ in the CrtM active site6
and our previous NMR and quantum chemical studies on
bisphosphonates, which indicate deprotonation when bisphos-
tricyclic analogue, had very poor activity against CrtM (IC50
)
170 µM). These results, together with those discussed above,
clearly indicated that investigating additional diphenyl ether
phosphonosulfonates would be of interest.
We therefore next synthesized 17 more substituted phospho-
nosulfonates whose structures and activities in CrtM inhibition
are shown in Table 1, together with the previously described
results for 1 and 5-7. The following structure-activity relation-
ship features can be seen from these results: (1) Diphenyl ether
phosphonosulfonates substituted with F or CF3 had diminished
activity when the substituent was in the 2′-position relative to
the corresponding 3′- or 4′-substituted analogues. (2) Halogen-
containing groups, including F, CF3, and Cl, generally enhance
activity. (3) Alkyl groups with various shapes and sizes located
at the 4′-position result in modest activity changes (when
compared to 1), while oxygen-containing groups at the same

980 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 4
Song et al.
Figure 4. Compound alignment and CoMSIA fields for CrtM inhibition: (A) flexible alignment of compounds, with only heavy atoms displayed
for clarity; (B) steric fields (green, favorable; red, disfavored); (C) hydrophobic fields (magenta, favorable; white, disfavored); (D) electrostatic and
H-bond donor fields (blue, positive-charge favorable; yellow, negative-charge favorable; cyan, H-bond donor disfavored).
Figure 5. (A) Correlation between staphyloxanthin (cell) activity and CrtM (enzyme) activity, showing R2 ) 0.27. (B) Improved correlation
between cell activity and enzyme activity with incorporation of a molecular descriptor, SlogP, with R2 ) 0.60. (C) Best cell activity prediction
obtained by using CrtM inhibition plus two additional descriptors; R2 ) 0.72 and Q2 ) 0.62. (D) correlation between CrtM activity and human SQS
activity, showing R2 ) 0.50.

Inhibition of Staphyloxanthin Biosynthesis
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 4 981
phonates bind to Mg2+.15 All compounds were constructed with
the (S)-configuration, based on the observation that (S)-1 is far
more active than is its (R)-enantiomer in CrtM inhibition (as
also found in SQS inhibition10). The aligned compounds were
exported into Sybyl.16 Then, we used a partial least-squares
(PLS) method to regress the CrtM inhibitory activity and
CoMSIA field data. The 3D QSAR model yielded r2 ) 0.98,
q2 (number of components) ) 0.72 (7), Ftest ) 245.8, and a
pIC50 error of 0.12, as shown in Table 2. To further validate
the model, we performed five leave-five-out training/test sets,
obtaining on average a factor of 1.8× error between predicted
and experimental IC50 values (Table 2). The compound align-
ment and the CoMSIA fields are shown in Figure 4. There is a
relatively large hydrophobic contribution (46.6%) to the CoM-
SIA model, followed by steric (25.2%), electrostatic (19.9%),
and H-bond donor (8.3%) interactions. The CoMSIA results are
in good agreement with the more qualitative experimental
observations: steric and hydrophobic-favorable areas (in green
and magenta, respectively, in Figure 4B,C) at the 3′- and 4′-
positions correspond to generally increased activities of 1
Table 3. Staphyloxanthin Inhibition in S. aureus
compda staphyloxanthin IC50 (µM) compda staphyloxanthin IC50 (µM)
31
33
32
8
25
24
7
26
(S)-1
34
5
35
22
6
0.011
0.015
0.015
0.019
0.021
0.039
0.040
0.049
0.050
0.051
0.052
0.059
0.074
0.093
0.10
11
10
18
20
(R)-1
9
23
13
16
21
3
29
4
2
27
36
15
14
19
0.18
0.18
0.19
0.22
0.31
0.35
0.46
0.49
0.50
0.51
1.1
1.1
1.4
2.2
6.8
13.8
20.4
33.9
28
1
30
12
17
0.11
0.11
0.16
0.17
263
a Structures shown in Figures 2 and 3 and Table 1.
Table 4. CoMSIA Results for CrtM/hSQS Selectivity
CoMSIA predictionsd
test sets
experimental data
relative
hSQS IC50
(µM)
CrtM/hSQS
CrtM/hSQS
pIC50(CrtM) -
pIC50(hSQS) (M)
training
set
compda
selectivityb
selectivityc
residual
1
2
3
4
5
6
1.9
4.0
11.7
0.72
0.17
0.22
0.54
0.033
0.47
0.35
1.8
0.003
0.072
0.29
0.85
0.38
0.068
0.054
0.024
0.024
0.021
0.019
0.016
0.016
0.015
0.014
0.014
0.012
0.0098
0.0079
0.0072
0.0071
0.0071
0.0071
0.0065
0.0062
0.0036
0.0034
0.0033
0.0032
0.0029
0.0022
0.0019
0.0018
0.0018
0.0018
0.0011
0.0011
0.00073
7.1 × 10-5
295
132
24.0
18.6
8.3
8.3
7.1
6.6
5.8
5.6
5.1
4.8
4.7
4.1
3.4
2.8
2.5
2.5
2.5
2.5
2.2
2.1
1.3
1.2
1.1
1.1
1.0
0.8
0.6
0.6
0.6
0.6
0.4
0.4
0.3
0.02
-0.07
-0.42
-1.16
-1.27
-1.62
-1.62
-1.69
-1.72
-1.78
-1.79
-1.83
-1.86
-1.87
-1.93
-2.01
-2.10
-2.14
-2.15
-2.15
-2.15
-2.19
-2.21
-2.44
-2.47
-2.48
-2.49
-2.54
-2.66
-2.73
-2.74
-2.74
-2.74
-2.96
-2.96
-3.14
-4.15
-0.16
-0.38
-0.99
-1.29
-1.52
-1.52
-1.70
-2.12
-2.34
-1.98
-1.93
-2.00
-1.93
-1.64
-1.88
-2.15
-2.23
-2.29
-2.20
-2.31
-2.12
-2.08
-2.31
-2.63
-2.48
-2.43
-2.50
-2.50
-2.84
-2.65
-2.68
-2.70
-2.60
-2.92
-2.78
-4.16
0.09
-0.04
-0.17
0.02
-0.10
-0.10
0.01
0.40
0.56
0.19
0.10
0.14
0.06
-0.29
-0.13
0.05
0.09
0.14
0.05
0.16
-0.17
-0.28
-1.13
-1.41
-1.51
-1.60
-1.75
-2.22
-2.37
-2.01
-2.04
-1.84
-1.74
-1.62
-1.97
-2.12
-2.30
-2.32
-2.33
-2.31
-2.15
-2.08
-2.38
-2.54
-2.47
-2.47
-2.50
-2.56
-2.66
-2.62
-2.78
-2.74
-2.51
-2.91
-2.75
-4.17
-0.18
-0.26
-1.15
-1.40
-1.49
-1.45
-1.73
-2.21
-2.39
-1.81
-2.01
-1.80
-1.71
-1.39
-1.99
-2.12
-2.32
-2.33
-2.28
-2.31
-2.16
-2.06
-2.35
-2.47
-2.49
-2.40
-2.49
-2.58
-2.62
-2.62
-2.83
-2.72
-2.53
-2.96
-2.74
-4.16
-0.18
-0.31
-1.17
-1.43
-1.53
-1.60
-1.77
-2.38
-2.44
-1.90
-2.00
-1.82
-1.75
-1.50
-2.00
-2.18
-2.30
-2.42
-2.34
-2.29
-2.16
-2.16
-2.45
-2.47
-2.47
-2.34
-2.50
-2.63
-2.62
-2.67
-2.77
-2.76
-2.55
-2.96
-2.77
-4.16
-0.16
-0.43
-0.97
-1.29
-1.52
-1.51
-1.71
-2.09
-2.34
-2.00
-1.90
-1.99
-1.97
-1.63
-1.86
-2.09
-2.27
-2.29
-2.19
-2.47
-2.12
-1.99
-2.31
-2.55
-2.53
-2.41
-2.52
-2.51
-2.90
-2.66
-2.72
-2.68
-2.66
-2.93
-2.81
-4.18
-0.17
-0.27
-1.13
-1.42
-1.51
-1.89
-1.71
-2.22
-2.40
-2.02
-2.02
-1.88
-1.70
-1.64
-1.95
-2.13
-2.38
-2.34
-2.30
-2.30
-2.14
-2.05
-2.34
-2.56
-2.87
-2.40
-2.48
-2.52
-2.60
-2.61
-2.82
-2.74
-2.55
-2.89
-2.74
-4.18
25
19
26
7
10
29
30
27
9
13
3
4
11
36
24
18
5
33
22
28
31
20
17
23
34
1
0.79
0.020
0.042
0.017
0.032
0.069
1.1
0.014
0.008
0.069
0.18
0.020
0.023
3.7
0.001
0.004
0.079
0.13
-0.07
-0.13
-0.13
0.16
0.00
-0.06
-0.04
-0.16
0.11
-0.09
-0.06
-0.04
-0.36
-0.04
-0.36
0.01
14
2
35
16
12
32
15
21
8
0.002
0.55
0.005
0.003
r2
0.94
0.54
76.5
6
0.93
0.50
71.6
5
0.93
0.52
75.6
5
0.94
0.51
82.5
5
0.95
0.55
75.0
6
0.92
0.51
64.5
5
q2
Ftest
N
n
36
33
33
33
33
33
a Structures shown in Figures 2 and 3 and Table 1. b Selectivity ) IC50(hSQS)/IC50(CrtM). c Relative selectivity ) (selectivity of compound)/(selectivity
of 1). d Bold values represent predicted activities of compounds that were not included in the training set.

982 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 4
Song et al.
curves of representative compounds are shown in the Supporting
Information (Figure S1). Surprisingly, we found that STX
biosynthesis inhibition in S. aureus was poorly correlated with
CrtM (enzyme) inhibition, with R2 ) 0.27 for the CrtM/STX
inhibition pIC50 values, as shown in Figure 5A.
Presumably, this poor correlation is due to the fact that no
consideration of cell permeability or drug uptake is involved in
the in vitro enzyme inhibition data. To try to take this into
account, we therefore chose to use SlogP (the logarithm of the
octanol/water partition coefficient17) to describe this effect by
using the following equation:
pIC50(STX, cell) ) a[pIC50(CrtM)] + b(SlogP) + c (1)
where a, b, and c are regression coefficients from a linear
regression analysis. This yielded R2 ) 0.60 for the experimental-
versus-predicted pIC50 values (Figure 5B) or R2 ) 0.53 for a
leave-two-out (L2O) prediction test set, to be compared with
R2 ) 0.16 for a L2O test set of predictions using solely the
enzyme pIC50 results. Of course, there is no a priori reason to
use SlogP as the extra descriptor or to limit the method to use
of a single extra descriptor. So we next used a three descriptor
model:
pIC50(STX, cell) ) a[pIC50(CrtM)] + (b)(B) + (c)(C) + d
(2)
where B and C are all possible descriptor pairs available in MOE
that have non-Boolean values (i.e., the properties do not contain
0’s or 1’s). The top 10 “enzyme plus two descriptor” search
results are shown in the Supporting Information (Table S1),
rank-ordered by R2 value. Each of the top 10 results contains
pIC50 (CrtM) as the major descriptor, and the R2 value obtained
for the top solution, R2 ) 0.72, is clearly an improvement over
that obtained using CrtM, or CrtM and SlogP results (Figure
5C). We then used a leave-two-out method to produce a test
set result in which we recomputed all training sets minus the
two compounds of interest, then used the coefficients and
descriptors to predict the two omitted compounds. In this way,
the activity of each compound was predicted 37 times. The R2
in this leave-two-out test set was 0.62, a major improvement
over the R2 ) 0.16 using solely enzyme inhibition data (and
the same leave-two-out test set approach). To verify that this
predictivity did not occur by chance, we repeated the leave-
two-out process, using scrambled cell activity data. This process
was repeated 10 times, with the average R2 values (leave two
out, scrambled) being 0.10 (Supporting Information Table S1).
Clearly, then, phosphonosulfonates are potent inhibitors of
the CrtM enzyme, and their activity can be relatively well
predicted by using the combinatorial descriptor search
method, even when the cell/enzyme data are very poorly
correlated. Moreover, as might be expected on the basis of
the enzyme inhibition results (Figure 3), (S)-1 is far more
active than is (R)-1 in cells (Table 3).
Figure 6. CoMSIA fields for CrtM/SQS selectivity: (A) steric fields
(green, favorable; red, disfavored); (B) hydrophobic fields (magenta,
favorable; white, disfavored); (C) electrostatic fields (blue, positive-
charge favorable).
analogues with para- and meta-substituents on the distal phenyl
ring. Two large steric penalty areas (in red in Figure 4B) at the
4- and 6- positions of the proximal phenyl ring account for the
decreased activity of the biphenyl phosphonosulfonates (9-13
in Figure 2) and 36 (entry 21 in Table 1). In addition, a positive-
charge-favored region (in blue, shown in Figure 4D) within the
distal phenyl ring helps to account for the enhanced activity of
the (electron-withdrawing) halogen containing phosphonosul-
fonates (e.g., 5, 24, and 30). The negative-charge favored area
(in yellow, Figure 4D) obviously correlates with the activity of
the bisphosphonates (2 and 3) in CrtM inhibition, since the
sulfonate group is replaced by the more negatively charged
phosphonate group.
Staphyloxanthin Biosynthesis Inhibition. The results de-
scribed above are of interest since they indicate that dehy-
drosqualene (CrtM) can be potently inhibited by phosphono-
sulfonates, and by bisphosphonates. The question next arises:
how potent are these compounds in inhibiting STX biosynthesis
in S. aureus? We thus treated S. aureus bacteria with serially
diluted compounds at 37 °C for 3 days, after which the STX
pigment was extracted with methanol. Optical densities were
measured at 450 nm, and IC50 values for inhibition of pigment
formation for each compound were calculated using a standard
dose-response curve. The rank ordered IC50 values of the 38
compounds investigated are shown in Table 3, and dose-response
Selectivity of CrtM Inhibition. Since phosphonosulfonates,
such as 1, were originally developed as inhibitors of squalene
synthase in the context of cholesterol-lowering therapy,7,8,10 we
were also interested to see how they inhibit human SQS, since
SQS inhibition results in formation of a 1,10-dioic acid
metabolite (from unused FPP). We thus screened each of the
compounds described above for their activity against an
expressed human SQS enzyme. Results are shown in Table 4.
We find that CrtM inhibition is moderately correlated with
human SQS inhibition, R2 ) 0.50, as shown in Figure 5D,
consistent with the modest sequence homology of S. aureus
CrtM and human SQS (30% identity, 36% similarity).6 Of

Inhibition of Staphyloxanthin Biosynthesis
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 4 983
Scheme 1. General Synthesis of Diphenyl Ether Phosphonosulfonatesa
a Reagents and conditions: (i) CuI, N,N-dimethylglycine, Cs2CO3, 1,4-dioxane, reflux; (ii) triethyl phosphonoacetate, NaH; (iii) Pd/C (5%) or Raney Ni,
H2; (iv) LiAlH4; (v) MsCl, NEt3, then NaI; (vi) cyclohexyl diethylphosphonomethylsulfonate, NaH; (vii) NH3, MeOH; (viii) Me3SiBr, then MeOH/KOH(aq).
Scheme 2a
a Reagents and conditions: (i) Pd(PPh3)4, K2CO3; (b) steps ii-v in Scheme 1; (c) NaH, tetraethyl methylenediphosphonate, then TMSBr; (d) NaH, triethyl
methylphosphinomethylphosphonate, then TMSBr; (e) steps vi-viii in Scheme 1.
Scheme 3a
a Reagents and conditions: (b) steps vi-viii in Scheme 1; (c) steps iv-viii in Scheme 1; (i) CuCl, DMF, 55 °C; (d) steps ii-viii in Scheme 1; (ii)
Pd(PPh3)4, K2CO3.
course, this modest correlation reflects something that is
potentially beneficial: that some good CrtM inhibitors are poor
hSQS inhibitors. Consider, for example, the 4′-n-propyl species
6 (Table 1). This compound has a 2.2 µM IC50 versus CrtM
and a 1.9 µM IC50 versus hSQS, Table 4, and is therefore a
much poorer hSQS inhibitor than is the parent compound 1.
For compound 1 (used previously in vivo6), the CrtM IC50 is
7.9 µM but 1 is a very potent hSQS inhibitor with an IC50 of
23 nM, Table 4. The results shown in Table 4 are rank-ordered
in terms of selectivity for CrtM over hSQS inhibition or
IC50(hSQS)/IC50(CrtM) such that a larger number means more
CrtM selectivity, and are also given in terms of selectivity
relative to 1, which is (selectivity of compound)/(selectivity of
1). The compound with the highest relative selectivity is thus
6, which is (1.9/2.2) ÷ (0.023/7.9) or ∼300× more selective a
CrtM inhibitor than is 1. Remarkably, both 1 and 6 have,
however, very similar IC50 values for STX biosynthesis (Table
3): 110 nM for 1 and 93 nM for 6. These results strongly suggest
that it is possible to make potent inhibitors of STX biosynthesis
(such as 6) that have much better relative selectivity in inhibiting
CrtM than does 1.
To put these observations on a more structural basis, we
carried out a QSAR/CoMSIA analysis, using the relative
selectivity results. Using the same compound alignment as
shown in Figure 4A, we obtained r2 ) 0.94, q2 ) 0.54, Ftest
)
76.5, and a pIC50 error of 0.20, as shown in Table 4. We also

984 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 4
Song et al.
validated this model by using five leave-three-out training/test
sets, finding on average a factor of 1.8× error between predicted
and experimental IC50 values (Table 4). There is a large
hydrophobic contribution (54.8%) to the CoMSIA field results,
followed by steric (29.5%) and electrostatic (15.7%) fields. The
CoMSIA fields are shown in Figure 6, and it is clear that
overlapping steric (green, Figure 6A) and hydrophobic (magenta,
Figure 6B) field features near the 4′-position are favored for
CrtM selectivity. Hydrophobic (white, Figure 6B) and steric (red,
Figure 6A) disfavored regions are located near the 2′- and 3′-
positions, respectively, and are responsible for the poor selectiv-
ity of compounds such as 8 and 21.
Compound 41 was reacted with the sodium salt of tetraethyl
methylenediphosphonate or triethyl methylphosphinometh-
ylphosphonate,19 followed by treatment with bromotrimethyl-
silane, to give bisphosphonates (2 and 3) or phosphinophos-
phonate (4), respectively. Following steps vi-viii in Scheme
1, biphenyl phosphonosulfonates (9-13) can be obtained from
the iodide 41.
Compounds (S)- and (R)-1 were synthesized according to a
published method,11 using trans-(1R, 2R)-N,N′-bismethylcyclo-
hexanediamine as a chiral auxiliary. The syntheses of other
compounds are illustrated in Scheme 3.
Compounds 14 and 15 were readily prepared from 3-phe-
noxybenzyl chloride and 3-phenoxyphenylacetic acid, respec-
tively, following the general method in Scheme 1. Compound
16 was prepared from aldehyde 42, which was obtained by the
coupling of nitrosobenzene and 3-formylphenylboronic acid, in
the presence of copper(I) chloride.20 Suzuki coupling of benzyl
bromide and 3-formylphenylboronic acid afforded 3-pheny-
laminobenzaldehyde 43. Compound 17 was then made from
aldehyde 43, following steps ii-viii in Scheme 1. Similarly,
compounds 18 and 19 were made following steps ii-viii in
Scheme 1, starting from available 4-phenoxybenzaldehyde and
9-ethyl-3-carbazolecarboxaldehyde, respectively.
Conclusions
The results described above are of interest for a number of
reasons. First, we made 38 phosphonosulfonate and bisphos-
phonate compounds and investigated their activity in inhibiting
S. aureus dehydrosqualene synthase (CrtM), the enzyme in-
volved in the first committed step in the biosynthesis of the
virulence factor, STX, in S. aureus. The most active compounds
were bisphosphonates, but they had poor activity in cells, while
many phosphonosulfonates were active in both enzyme and cell
assays. Second, we developed a QSAR model for CrtM
inhibition that had good predictivity. Third, we determined the
activity of each compound against STX biosynthesis, in S.
aureus. The pIC50 results were poorly correlated (R2 ) 0.27,
training; R2 ) 0.16, test set) with the enzyme inhibition pIC50
values. However, using a combinatorial descriptor search, we
obtained significant improvements in cell activity predictions
with R2 ) 0.72 (training) and R2 ) 0.62 (test set) results. Fourth,
we investigated the inhibition of human squalene synthase by
these compounds and used a QSAR method to enable good
predictions of CrtM/SQS selectivity, relative to that exhibited
by 1. Several CrtM inhibitors (e.g., 6 and 25) also showed very
potent activity in bacterial cell-based assays. Overall, these
results are of general interest, since they demonstrate that
diphenyl ether phosphonosulfonates are potent inhibitors of
dehydrosqualene synthase and of the STX biosynthesis in S.
aureus, and as such, they have considerable potential for further
development in a new approach to combating S. aureus
infections.
Experimental Section
All reagents used were purchased from Aldrich (Milwaukee, WI).
The purities of all compounds were routinely monitored by using
1H and 31P NMR spectroscopy at 400 or 500 MHz on Varian (Palo
Alto, CA) Unity spectrometers. Elemental analysis results are
provided in the Supporting Information (Table S2).
General Method A (Steps ii-viii in Scheme 1). Step ii. Triethyl
phosphonoacetate (3.3 mmol) was added dropwise to NaH (145
mg, 60% in oil, 3.6 mmol) suspended in dry THF (7 mL) at 0 °C.
To the resulting clear solution was added a benzaldehyde (3 mmol),
and after being stirred at room temperature for 0.5 h, the reaction
mixture was partitioned between diethyl ether (50 mL) and water
(50 mL). The organic layer was dried and evaporated.
Step iii. The residue oil was then hydrogenated in MeOH (15
mL), in the presence of 5% Pd/C (50 mg) or Raney Ni (500 mg)
when the aldehyde 37 was prepared using the CuI mediated reaction
(Scheme 1). The catalyst was filtered, and the filtrate was
concentrated and dried in vacuo.
Step iv. The resulting oil was dissolved in anhydrous THF (8
mL), and LiAlH4 (114 mg) was slowly added to the solution at
0 °C. After 1 h, the reaction was carefully quenched by adding a
few drops of water, and the reaction mixture was filtered.
Step v. The filtrate was evaporated to dryness and the alcohol
thus obtained redissolved in CH2Cl2 (10 mL) containing NEt3 (0.5
mL, 3.6 mmol). Methanesulfonyl chloride (230 µL, 3 mmol) was
added slowly at 0 °C. After 1 h of stirring at room temperature,
diethyl ether (50 mL) and water (50 mL) were added and the organic
layer was collected, washed with 1 N HCl and saturated NaHCO3,
dried, and evaporated to dryness. The oily residue was treated with
NaI (1.35 g, 9 mmol) in acetone (7 mL) at 60 °C for 1 h. The
reaction mixture was then partitioned between diethyl ether (50
mL) and water (50 mL) and the organic layer washed with 5%
Na2S2O3, dried, and evaporated to dryness to give an iodide, such
Chemistry: General Aspects
A general synthetic route to the diphenylether phosphono-
sulfonate compounds is shown in Scheme 1. If not commercially
available, a 3-phenoxybenzaldehyde 37 can be prepared with a
copper(I) iodide mediated coupling reaction18 from a substituted
halobenzene and a substituted hydroxybenzaldehyde, in a yield
of 70-90%. The aldehyde 37 was reacted with sodium
triethylphosphonoacetate in THF to give an R,ꢀ-unsaturated
carboxylate, which was hydrogenated, reduced to the alcohol
by treatment with LiAlH4, mesylated, then treated with NaI to
afford the iodide 38. This was then reacted with the sodium
salt of cyclohexyl diethylphosphonomethylsulfonate to give the
triester 39,10 typically in an overall yield of 40% from the
aldehyde 37. The triester 39 was deprotected by successive
treatments with ammonia in methanol, then bromotrimethylsi-
lane, followed by alkaline hydrolysis, affording the phospho-
nosulfonates as a tripotassium salt in ∼70% yield.
1
as 38. The iodide thus obtained was quite pure, according to H
and 13C NMR spectra, and may be used in the next step without
purification.
Step vi. Cyclohexyl diethylphosphonomethylsulfonate (470 mg,
1.5 mmol) was added to NaH (60 mg, 60% in oil, 1.5 mmol)
suspended in dry DMF (2 mL) at 0 °C. To the resulting clear
solution was added an iodide (1 mmol), and after being stirred at
room temperature for 3 h, the reaction mixture was partitioned
between diethyl ether (50 mL) and water (50 mL). The organic
layer was dried and evaporated and the residue subjected to a
Biphenyl bisphosphonates and phosphonosulfonates were
made similarly, as shown in Scheme 2. Iodide 41 was made
from a biphenylaldehyde 40, which is either commercially
available or prepared using a Suzuki coupling reaction from
4-bromobenzaldehyde and a substituted phenylboronic acid.

Inhibition of Staphyloxanthin Biosynthesis
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 4 985
column chromatography (silica gel, ethyl acetate/hexane, 1:1) to
give a phosphonosulfonate triester, such as 39, as a colorless oil.
Step vii. The triester was treated with ammonium hydroxide (12
M, 1 mL) in MeOH (6 mL) at 60 °C for 3 h. The solvents were
evaporated, and the residue was subjected to ion exchange
chromatography (DOWEX 50WX8-200, H+ form, 3 mL) using
MeOH as eluent.
Step viii. The eluent was evaporated to dryness and the resulting
diethylphosphonosulfonic acid dissolved in anhydrous CH3CN
(3 mL) and treated with Me3SiBr (400 µL, 3 mmol) at 40 °C
overnight. The solution was evaporated to dryness and MeOH (5
mL) added to the residue. The solvent was removed in vacuo again,
and the residue was redissolved in MeOH (5 mL). Neutralization
with 2 M KOH to pH 8 gave a tripotassium salt of a phosphono-
sulfonate as a white powder.
2H, CH3CH2), 1.60-1.90 (m, 4H, -CH2CH2-), 2.40-2.50 (m,
4H, PhCH2 and PhCH2), 2.80-2.90 (m, 1H, CHSO3K), 6.70-7.20
(m, 8H, aromatic). 31P NMR (D2O): δ 14.4.
1-Phosphono-4-[3-(4-trifluoromethylphenoxy)phenyl]butyl-
sulfonic Acid Tripotassium Salt (7). Compound 7 was prepared
from 4-trifluoromethyl-iodobenzene (3 mmol) and 3-hydroxyben-
zaldehyde (4.5 mmol), following general methods B and A, as a
white powder (570 mg, 32% overall yield). Anal. (C17H15F3K3O7PS·
1
1.5 H2O) C, H. H NMR (400 MHz, D2O): δ 1.80-2.00 (m, 4H,
-CH2CH2-), 2.40-2.60 (m, 2H, PhCH2), 2.80-2.90 (m, 1H,
CHSO3K), 6.80-7.40 (m, 8H, aromatic). 31P NMR (D2O): δ 13.9.
1-Phosphono-4-[3-(2-benzylphenoxy)phenyl]butylsulfonic Acid
Tripotassium Salt (8). Compound 8 was prepared from 2-benzyl-
iodobenzene (3 mmol) and 3-hydroxybenzaldehyde (4.5 mmol),
following general methods B and then A, as a white powder (570
General Method B (Step i in Scheme 1).18 A mixture
containing a halobenzene (3 mmol), a 3-hydroxybenzaldehyde (4.5
mmol), CuI (58 mg, 0.3 mmol), N,N-dimethylglycine (93 mg, 0.9
mmol), and Cs2CO3 (2 g, 6 mmol) in 1,4-dioxane (8 mL) was
vigorously stirred at 90 °C for 18 h. The solvent was evaporated
and the residue partitioned between diethyl ether (50 mL) and water
(50 mL). The organic layer was successively washed with 5%
NaOH (2 × 20 mL), water (20 mL), and saturated NaCl (20 mL).
It was then dried and evaporated to give aldehyde 37 as a pale-
yellow oil, which is quite pure and may be used in the next step
directly. It may also be purified via a column chromatography.
1-Phosphono-4-(3-phenoxyphenyl)butylsulfonic Acid Tripo-
tassium Salt (1). Compound 1 was prepared from 3-phenoxyben-
zaldehyde (3 mmol), following general method A as a white powder
(680 mg, 36% overall yield). Anal. (C16H16K3O7PS·KBr·0.5H2O)
C, H. 1H NMR (400 MHz, D2O): δ 1.50-1.90 (m, 4H,
-CH2CH2-), 2.40-2.50 (m, 2H, PhCH2), 2.70-2.80 (m, 1H,
CHSO3K), 6.70-7.30 (m, 9H, aromatic). 31P NMR (D2O): δ 12.4.
4-(4-Biphenyl)butyldiphosphonic Acid Tetrapotassium Salt
(2). Compound 2 was prepared from 4-phenylbenzaldehyde (3
mmol), following steps ii-vi and then step viii of the general
method A, as a white powder (721 mg, 46% overall yield). Purity
was determined to be 87.3% by quantitative NMR spectroscopy.
1H NMR (400 MHz, D2O): δ 1.60-1.80 (m, 5H, -CH2CH2- and
CH), 2.56 (t, J ) 5.6 Hz, 2H, PhCH2), 7.20-7.60 (m, 9H, aromatic).
31P NMR (D2O): δ 21.2.
1
mg, 29% overall yield). Anal. (C23H22K3O7PS·4 H2O) C, H. H
NMR (400 MHz, D2O): δ 1.60-1.90 (m, 4H, -CH2CH2-),
2.40-2.50 (m, 2H, PhCH2), 2.80-2.90 (m, 1H, CHSO3K), 3.80
(s, 2H, Ph CH2Ph), 6.50-7.30 (m, 13H, aromatic). 31P NMR (D2O):
δ 12.7.
1-Phosphono-4-(4-biphenyl)butylsulfonic Acid Tripotassium
Salt (9). Compound 9 was prepared from 4-phenylbenzaldehyde
(3 mmol), following general method A, as a white powder (730
1
mg, 31% overall yield). Anal. (C16H16K3O6PS·2.5KBr) C, H. H
NMR (400 MHz, D2O): δ 1.70-1.90 (m, 4H, -CH2CH2-),
2.50-2.60 (m, 2H, PhCH2), 2.90-3.00 (m, 1H, CHSO3K),
7.20-7.60 (m, 9H, aromatic). 31P NMR (D2O): δ 14.4.
1-Phosphono-4-[4-(2,4-difluorophenyl)phenyl]butylsulfonic Acid
Tripotassium Salt (10). 4-(2,4-Difluorophenyl)benzaldehyde was
made by a Suzuki coupling reaction from 2,4-difluorophenylboronic
acid (3.6 mmol) and 4-bromobenzaldehyde (3 mmol), as described
previously.21 Compound 10 was prepared from the aldehyde thus
obtained, following general method A, as a white powder (470 mg,
1
28% overall yield). Anal. (C16H14F2K3O6PS) C, H. H NMR (400
MHz, D2O): δ 1.65-2.00 (m, 4H, -CH2CH2-), 2.45-2.60 (m,
2H, PhCH2), 2.80-2.85 (m, 1H, CHSO3K), 6.80-7.40 (m, 7H,
aromatic). 31P NMR (D2O): δ 12.5.
1-Phosphono-4-[4-(4-methylphenyl)phenyl]butylsulfonic Acid
Tripotassium Salt (11). Compound 11 was prepared from 4-(4-
methylphenyl)benzaldehyde (3 mmol), following general method
A, as a white powder (770 mg, 28% overall yield). Anal.
(C17H18K3O6PS·3.5KBr) C, H. 1H NMR (400 MHz, D2O): δ
1.70-1.90 (m, 4H, -CH2CH2-), 2.21 (s, 3H, Me), 2.40-2.50 (m,
2H, PhCH2), 2.90-3.00 (m, 1H, CHSO3K), 7.20-7.50 (m, 8H,
aromatic). 31P NMR (D2O): δ 14.5.
1-Phosphono-4-[4-(4-butylphenyl)phenyl]butylsulfonic Acid
Tripotassium Salt (12). 4-(4-Butylphenyl)benzaldehyde was made
by a Suzuki coupling reaction from 4-butylphenylboronic acid (3.6
mmol) and 4-bromobenzaldehyde (3 mmol).21 Compound 12 was
prepared from the aldehyde thus obtained, following general method
A, as a white powder (650 mg, 25% overall yield). Anal.
(C20H24K3O6PS·2.8KBr) C, H. 1H NMR (400 MHz, D2O): δ 0.74
(t, J ) 7.2 Hz, 3H, CH3), 1.10-1.20 (m, 2H, CH3CH2), 1.40-1.50
(m, 2H, CH3CH2CH2), 1.650-2.00 (m, 4H, -CH2CH2-), 2.45-2.60
(m, 4H, PhCH2 and PhCH2), 2.70-2.80 (m, 1H, CHSO3K),
7.20-7.50 (m, 8H, aromatic). 31P NMR (D2O): δ 12.4.
4-[4-(4-Trifluoromethylphenyl)phenyl)]butyldiphosphonic Acid
Dipotassium Salt (3). Compound 3was prepared from 4-trifluo-
romethylphenylbenzaldehyde (3 mmol), following steps ii-vi and
then step viii of general method A, as a white powder (671 mg,
1
42% overall yield). Anal. (C17H17F3K2O6P2) C, H. H NMR (400
MHz, D2O): δ 1.60-1.90 (m, 5H, -CH2CH2- and CH), 2.58
(t, J ) 5.6 Hz, 2H, PhCH2), 7.20-7.60 (m, 9H, aromatic). 31P NMR
(D2O): δ 21.5.
4-(4-Biphenyl)butyldiphosphonic Acid Dipotassium Salt (4).
Compound 4 was prepared from 4-phenylbenzaldehyde (3 mmol),
following steps ii-vi and then step viii of general method A, as a
white powder (470 mg, 21% overall yield). Anal. (C17H20K2O5P2)
C, H. 1H NMR (400 MHz, D2O): δ 1.19 (d, J ) 11.2 Hz, 3H, Me),
1.60-1.80 (m, 5H, -CH2CH2- and CH), 2.56 (t, J ) 5.6 Hz, 2H,
PhCH2), 7.20-7.60 (m, 9H, aromatic). 31P NMR (D2O): δ 18.3
(s, 1P), 49.6 (s, 1P).
1-Phosphono-4-[3-(4-fluorophenoxy)phenyl]butylsulfonic Acid
Tripotassium Salt (5). Compound 5 was prepared from 4-fluor-
oiodobenzene (3 mmol) and 3-hydroxybenzaldehyde (4.5 mmol),
following general methods B and A, as a white powder (425 mg,
26% overall yield). Anal. (C16H15FK3O7PS·1.5H2O) C, H. 1H NMR
(400 MHz, D2O): δ 1.60-1.90 (m, 4H, -CH2CH2-), 2.40-2.50
(m, 2H, PhCH2), 2.80-2.90 (m, 1H, CHSO3K), 6.60-7.10 (m, 8H,
aromatic). 31P NMR (D2O): δ 13.6.
1-Phosphono-4-[4-(3-propoxyphenyl)phenyl]butylsulfonic Acid
Tripotassium Salt (13). 4-(3-Propoxyphenyl)benzaldehyde was
made by a Suzuki coupling reaction from 3-propoxyphenylboronic
acid (3.6 mmol) and 4-bromobenzaldehyde (3 mmol).21 Compound
13 was prepared from the aldehyde thus obtained, following general
method A, as a white powder (539 mg, 30% overall yield). Anal.
1
(C19H22K3O6PS) C, H. H NMR (400 MHz, D2O): δ 0.74 (t, J )
7.2 Hz, 3H, CH3), 1.40-1.50 (m, 2H, CH3CH2), 1.65-2.00 (m,
4H, -CH2CH2-), 2.45-2.60 (m, 2H, PhCH2), 2.80-2.85 (m, 1H,
CHSO3K), 3.75 (t, J ) 7.2 Hz, 2H, OCH2), 6.80-7.40 (m, 7H,
aromatic). 31P NMR (D2O): δ 12.8.
1-Phosphono-4-[3-(4-propylphenoxy)phenyl]butylsulfonic Acid
Tripotassium Salt (6). Compound 6 was prepared from 4-propy-
lbromobenzene (3 mmol) and 3-hydroxybenzaldehyde (4.5 mmol),
following general methods B and A, as a white powder (670 mg,
27% overall yield). Anal. (C19H22K3O7PS·2.4 KBr) C, H. 1H NMR
(400 MHz, D2O): δ 0.73 (t, J ) 7.2 Hz, 3H, CH3), 1.40-1.50 (m,
(1S)-1-Phosphono-4-(3-phenoxyphenyl)butylsulfonic Acid Tri-
potassium Salt [(S)-1]. Compound (S)-1 was prepared from
methylphosphonic dichloride (3 mmol) and trans-(1R,2R)-N,N′-
bismethylcyclohexanediamine (3 mmol), following a published

986 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 4
Song et al.
method,11 as a white powder (640 mg, 29% overall yield). Anal.
(C16H16K3O7PS·1.2K2SO4 ·1.5H2O) C, H. Identical NMR spectra
as for 1.
(1R)-1-Phosphono-4-(3-phenoxyphenyl)butylsulfonic Acid Tri-
potassium Salt [(R)-1]. Compound (R)-1 was obtained as a minor
product during the synthesis of (S)-1, as described immediately
above (405 mg, 16% overall yield). Anal. (C16H16K3O7PS ·2K2SO4)
C, H. Identical NMR spectra as for 1.
1-Phosphono-2-(3-phenoxyphenyl)ethylsulfonic Acid Tripo-
tassium Salt (14). Compound 14 was prepared from 3-phenoxy-
benzyl chloride (1 mmol), following steps vi-viii of general method
A, as a white powder (285 mg, 55% overall yield). Anal.
(C14H12K3O7PS·C2H5OH) C, H. 1H NMR (400 MHz, D2O): δ
2.95-3.05 (m, 1H, CH2). 3.10-3.30 (m, 2H, CH and CH2),
6.70-7.30 (m, 9H, aromatic). 31P NMR (D2O): δ 13.8.
2.50-2.70 (m, 2H, PhCH2), 2.80-2.90 (m, 1H, CHSO3K),
6.70-7.20 (m, 8H, aromatic). 31P NMR (D2O): δ 13.8.
1-Phosphono-4-[3-(2-fluorophenoxy)phenyl]butylsulfonic Acid
Tripotassium Salt (21). Compound 21 was prepared from 2-flu-
oroiodobenzene (3 mmol) and 3-hydroxybenzaldehyde (4.5 mmol),
following general methods B and A, as a white powder (340 mg,
20% overall yield). Anal. (C16H15FK3O7PS ·2.5H2O) C, H. 1H NMR
(400 MHz, D2O): δ 1.70-1.90 (m, 4H, -CH2CH2-), 2.60-2.80
(m, 2H, PhCH2), 2.90-3.00 (m, 1H, CHSO3K), 6.60-7.20 (m, 8H,
aromatic). 31P NMR (D2O): δ 13.2.
1-Phosphono-4-[3-(3-trifluoromethylphenoxy)phenyl]butylsul-
fonic Acid Tripotassium Salt (22). Compound 22 was prepared
from 3-(3-trifluoromethylphenoxy)benzaldehyde (3 mmol), follow-
ing general method A, as a white powder (600 mg, 30% overall
1
yield). Anal. (C17H15F3K3O7PS) C, H. H NMR (400 MHz, D2O):
δ 1.80-2.00 (m, 4H, -CH2CH2-), 2.50-2.70 (m, 2H, PhCH2),
2.85-2.95 (m, 1H, CHSO3K), 6.80-7.40 (m, 8H, aromatic). 31P
NMR (D2O): δ 13.3.
1-Phosphono-3-(3-phenoxyphenyl)propylsulfonic Acid Tri-
potassium Salt (15). Compound 15 was prepared from 3-phenox-
yphenylacetic acid (3 mmol), following steps iv-viii of general
method A, as a white powder (380 mg, 25% overall yield). Anal.
(C15H14K3O7PS·0.25C2H5OH·0.5H2O) C, H. 1H NMR (400 MHz,
D2O): δ δ 2.00-2.10 (m, 2H, CH2), 2.65-2.80 (m, 2H, PhCH2),
2.90-3.00 (m, 1H, CHSO3K), 6.70-7.30 (m, 9H, aromatic). 31P
NMR (D2O): δ 14.1.
1-Phosphono-4-[3-(2-trifluoromethylphenoxy)phenyl]butylsul-
fonic Acid Tripotassium Salt (23). Compound 23 was prepared
from 2-trifluoromethyliodobenzene (3 mmol) and 3-hydroxyben-
zaldehyde (4.5 mmol), following general methods B and A, as a
white powder (205 mg, 12% overall yield). Purity was determined
1
to be 86.5% by quantitative NMR spectroscopy. H NMR (400
1-Phosphono-4-(3-phenylaminophenyl)butylsulfonic Acid Tri-
potassium Salt (16). A mixture of nitrosobenzene (3 mmol) and
CuCl (3 mmol) in anhydrous DMF (8 mL) was heated to 55 °C
for 0.5 h. Then 3-formylphenylboronic acid (3.3 mmol) was added
to the reaction mixture, which was then stirred for another 16 h.20
The product was then partitioned between diethyl ether (50 mL)
and water (50 mL) and the organic layer washed with saturated
NaHCO3, dried, then evaporated to dryness, giving 3- phenylami-
nobenzaldehyde as a pale-yellow oil, which may be used directly
without purification. Compound 16 was prepared from the aldehyde
thus obtained, following general method A, as a white powder (425
mg, 22% overall yield). Anal. (C16H17NK3O7PS·KBr·1.5H2O) C,
H, N. 1H NMR (400 MHz, D2O): δ 1.60-1.90 (m, 4H,
-CH2CH2-), 2.40-2.50 (m, 2H, PhCH2), 2.80-2.90 (m, 1H,
CHSO3K), 6.70-7.20 (m, 9H, aromatic). 31P NMR (D2O): δ 14.4.
1-Phosphono-4-(3-benzylphenyl)butylsulfonic Acid Tripotas-
sium Salt (17). 3-Benzylbenzaldehyde was prepared from benzyl
bromide (3 mmol) and 3-formylphenylboronic acid (3.3 mmol) by
a Suzuki coupling reaction.21 Compound 17 was prepared from the
aldehyde thus obtained, following general method A, as a white powder
(540 mg, 33% overall yield). Anal. (C17H20K3O7PS·0.25KBr·H2O)
C, H. 1H NMR (400 MHz, D2O): δ 1.50-1.90 (m, 4H,
-CH2CH2-), 2.40-2.50 (m, 2H, PhCH2), 2.80-2.90 (m, 1H,
CHSO3K), 3.88 (s, 2H, PhCH2Ph), 6.90-7.20 (m, 9H, aromatic).
31P NMR (D2O): δ 13.8.
1-Phosphono-4-(4-phenoxyphenyl)butylsulfonic Acid Tripo-
tassium Salt (18). Compound 18 was prepared from 4-phenoxy-
benzaldehyde (3 mmol), following general method A, as a white
powder (680 mg, 36% overall yield). Anal. (C16H16K3O7PS) C, H.
1H NMR (400 MHz, D2O): δ 1.60-1.90 (m, 4H, -CH2CH2-),
2.40-2.50 (m, 2H, PhCH2), 2.80-2.90 (m, 1H, CHSO3K),
6.80-7.30 (m, 9H, aromatic). 31P NMR (D2O): δ 13.7.
1-Phosphono-4-(9-ethylcarbazole-3-yl)butylsulfonic Acid Tri-
potassium Salt (19). Compound 19 was prepared from 9-ethyl-3-
carbazolecarboxaldehyde (3 mmol), following general method A,
as a white powder (600 mg, 36% overall yield). Anal. (C18H19N-
K3O6PS ·0.5C2H5OH·0.5H2O) C, H, N. 1H NMR (400 MHz, D2O):
δ 1.14 (t, J ) 7.2 Hz, 3H, CH3), 1.70-1.90 (m, 4H, -CH2CH2-),
2.60-2.70 (m, 2H, PhCH2), 2.80-2.90 (m, 1H, CHSO3K), 4.22
(q, J ) 7.2 Hz, 3H, NCH2), 7.05-7.40 (m, 5H, aromatic), 7.92
(s, 1H, aromatic), 8.00 (d, J ) 8.0 Hz, 1H, aromatic). 31P NMR
(D2O): δ 14.1.
MHz, D2O): δ 1.70-1.90 (m, 4H, -CH2CH2-), 2.50-2.65 (m,
2H, PhCH2), 2.80-2.90 (m, 1H, CHSO3K), 6.70-7.30 (m, 8H,
aromatic). 31P NMR (D2O): δ 13.7.
1-Phosphono-4-[3-(4-chlorophenoxy)phenyl]butylsulfonic Acid
Tripotassium Salt (24). Compound 24 was prepared from 3-(4-
chlorophenoxy)benzaldehyde (3 mmol), following general method
A, as a white powder (525 mg, 30% overall yield). Anal.
1
(C16H15ClK3O7PS·C2H5OH) C, H. H NMR (400 MHz, D2O): δ
1.60-1.90 (m, 4H, -CH2CH2-), 2.45-2.55 (m, 2H, PhCH2),
2.80-2.90 (m, 1H, CHSO3K), 6.70-7.10 (m, 8H, aromatic). 31P
NMR (D2O): δ 14.0.
1-Phosphono-4-[3-(4-tert-butylphenoxy)phenyl]butylsulfonic Acid
Tripotassium Salt (25). Compound 25 was prepared from 3-(4-
tert-butylphenoxy)benzaldehyde (3 mmol), following general method
A, as a white powder (610 mg, 35% overall yield). Anal.
(C20H24K3O7PS·1.5H2O) C, H. 1H NMR (400 MHz, D2O): δ 1.10
(s, 9H, CMe3), 1.60-1.85 (m, 4H, -CH2CH2-), 2.40-2.50 (m,
2H, PhCH2), 2.80-2.90 (m, 1H, CHSO3K), 6.60-7.40 (m, 8H,
aromatic). 31P NMR (D2O): δ 14.3.
1-Phosphono-4-[3-(4-benzylphenoxy)phenyl]butylsulfonic Acid
Tripotassium Salt (26). Compound 26 was prepared from 4-ben-
zyliodobenzene (3 mmol) and 3-hydroxybenzaldehyde (4.5 mmol),
following general methods B and A, as a white powder (510 mg,
1
28% overall yield). Anal. (C23H22K3O7PS·H2O) C, H. H NMR
(400 MHz, D2O): δ 1.60-1.90 (m, 4H, -CH2CH2-), 2.40-2.50
(m, 2H, PhCH2), 2.75-2.85 (m, 1H, CHSO3K), 3.79 (s, 2H,
PhCH2), 6.60-7.20 (m, 13H, aromatic). 31P NMR (D2O): δ 13.9.
1-Phosphono-4-[3-(4-hydroxyphenoxy)phenyl]butylsulfonic
Acid Tripotassium Salt (27). Compound 27 was prepared from
4-tert-butoxyiodobenzene (3 mmol) and 3-hydroxybenzaldehyde
(4.5 mmol), following general methods B and A, as a white powder
(690 mg, 22% overall yield). Anal. (C16H16K3O8PS·4 KBr ·3H2O)
C, H. 1H NMR (400 MHz, D2O): δ 1.60-1.90 (m, 4H,
-CH2CH2-), 2.40-2.50 (m, 2H, PhCH2), 2.90-3.00 (m, 1H,
CHSO3K), 6.60-7.20 (m, 8H, aromatic). 31P NMR (D2O): δ 14.5.
1-Phosphono-4-[3-(4-phenoxyphenoxy)phenyl]butylsulfonic
Acid Tripotassium Salt (28). Compound 28 was prepared from
4-phenoxyiodobenzene (3 mmol) and 3-hydroxybenzaldehyde (4.5
mmol), following general methods B and A, as a white powder
(610 mg, 30% overall yield). Anal. (C22H20K3O8PS · 0.5KBr ·
1
1.5H2O) C, H. H NMR (400 MHz, D2O): δ 1.60-1.90 (m, 4H,
-CH2CH2-), 2.40-2.50 (m, 2H, PhCH2), 2.80-2.90 (m, 1H,
CHSO3K), 6.60-7.25 (m, 13H, aromatic). 31P NMR (D2O): δ 13.7.
1-Phosphono-4-[3-[4-(furan-2-yl)phenoxy]phenyl]butylsul-
fonic Acid Tripotassium Salt (29). Compound 29 was prepared
from 4-(furan-2-yl)iodobenzene (3 mmol) and 3-hydroxybenzal-
dehyde (4.5 mmol), following general methods B and A, as a white
1-Phosphono-4-[3-(3-fluorophenoxy)phenyl]butylsulfonic Acid
Tripotassium Salt (20). Compound 20 was prepared from 3-flu-
oroiodobenzene (3 mmol) and 3-hydroxybenzaldehyde (4.5 mmol),
following general methods B and A, as a white powder (570 mg,
28% overall yield). Anal. (C16H15FK3O7PS·KBr·2.5H2O) C, H. 1H
NMR (400 MHz, D2O): δ 1.60-1.90 (m, 4H, -CH2CH2-),

Inhibition of Staphyloxanthin Biosynthesis
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 4 987
powder (690 mg, 22% overall yield). Anal. (C20H18K3O8PS·2
the Gasteiger-Marsili method.22 CoMSIA indices were calculated
on a rectangular grid containing each of the sets of aligned
molecules using steric, electrostatic, hydrophobic, H-donor, and
H-acceptor fields, using default grid spacing and probe atoms. We
then used a partial least-squares (PLS) method to correlate the 3D
structural features with activity. The optimal number of components
was determined with the SAMPLS cross-validation method.23 Each
of the QSAR models was then validated by performing five training/
test set predictions.
For cell activity predictions, a set of 117 molecular descriptors,
including SlogP,17 were computed within MOE and exported to
MATLAB24 for a combinatorial descriptor search. The entire
descriptor space was searched exhaustively to find the combinations
of descriptors that gave the best regression coefficient (highest R2)
for the equation
1
KBr·1.5H2O) C, H. H NMR (400 MHz, D2O): δ 1.60-1.90 (m,
4H, -CH2CH2-), 2.40-2.50 (m, 2H, PhCH2), 2.70-2.80 (m, 1H,
CHSO3K), 6.40-7.60 (m, 11H, aromatic). 31P NMR (D2O): δ 12.4.
1-Phosphono-4-[3-(3,4-difluorophenoxy)phenyl]butyl-
sulfonic Acid Tripotassium Salt (30). Compound 30 was prepared
from 3,4-difluoroiodobenzene (3 mmol) and 3-hydroxybenzaldehyde
(4.5 mmol), following general methods B and A, as a white powder
(525 mg, 29% overall yield). Anal. (C16H14F2K3O7PS·0.5C2H5OH·
1
2.5H2O) C, H. H NMR (400 MHz, D2O): δ 1.60-2.00 (m, 4H,
-CH2CH2-), 2.40-2.50 (m, 2H, PhCH2), 2.80-2.90 (m, 1H,
CHSO3K), 6.50-7.30 (m, 7H, aromatic). 31P NMR (D2O): δ 12.8.
1-Phosphono-4-[3-(3,4-dichlorophenoxy)phenyl]butyl-
sulfonic Acid Tripotassium Salt (31). Compound 31 was prepared
from 3-(3,4-dichlorophenoxy)benzaldehyde (3 mmol), following
general method A, as a white powder (540 mg, 28% overall yield).
Anal. (C16H14Cl2K3O7PS·0.5C2H5OH·3H2O) C, H. 1H NMR (400
MHz, D2O): δ 1.60-2.00 (m, 4H, -CH2CH2-), 2.40-2.50 (m,
2H, PhCH2), 2.80-2.90 (m, 1H, CHSO3K), 6.80-7.40 (m, 7H,
aromatic). 31P NMR (D2O): δ 12.4.
pIC50(pigment) ) a[pIC50(enzyme)] + (b)(B) + (c)(C) + d
(3)
where B and C are MOE descriptors and a-d are coefficients. A
leave-two-out cross-validation was performed to test the predictivity
of the model, where all combinations of two compounds were
excluded from the data set and the descriptor combinations
reevaluated for the remaining (training set) compounds. The
regression equation obtained from each run was then used to
calculate the cell activity of the left out compounds (the test set),
and the R2 from all such leave-two-out predictions is reported
in the text. Finally, a scrambling analysis was performed in which
the cell activity values for all the compounds were scrambled and
the leave-two-out cross-validation was performed using the scrambled
data set.
CrtM Enzyme Inhibition Assay. The expression and purifica-
tion of the S. aureus CrtM as well as the inhibition assays were
carried out by using our previous methods.6 In brief, CrtM with a
histidine tag was overexpressed in E. coli BL21 (DE3) cells. After
an overnight growth, an initial 50 mL of culture was transferred
into 1 L of LB medium supplemented with 100 µg/mL ampicillin.
Induction was carried out with 1 mM IPTG for 4 h, when the cell
culture reached an OD of 0.6 at 600 nm. Protein was purified with
a Ni-NTA column, using a 100 mL linear gradient of 0-0.5 M
imidazole in 50 mM Tris-HCl buffer at pH 7.4. Enzyme inhibition
assays were carried out, in duplicate, in 96-well plates, with a total
of 200 µL of reaction mixture in each well. The reaction was
monitored by using a continuous spectrophotometric assay for
phosphate releasing enzymes.25 The reaction buffer contained 50
mM Tris-HCl, 1 mM MgCl2, 450 µM FPP, pH 7.4. The compounds
investigated were preincubated with 2 µg of CrtM for 30 min at
20 °C. The IC50 values were obtained by fitting the inhibition data
to a normal dose-response curve, using GraphPad PRISM, version
4.0, software for windows (GraphPad Software Inc., San Diego,
CA, www.graphpad.com). Ki was calculated on the basis of the
IC50 value and the reported Km of CrtM.26
1-Phosphono-4-[3-(benzofuran-5-yloxy)phenyl]butylsulfon-
ic Acid Tripotassium Salt (32). Compound 32 was prepared from
5-bromobenzofuran (3 mmol) and 3-hydroxybenzaldehyde (4.5
mmol), following general methods B and A, as a white powder
(380 mg, 22% overall yield). Anal. (C18H16K3O8PS·2H2O) C, H.
1H NMR (400 MHz, D2O): δ 1.60-1.90 (m, 4H, -CH2CH2-),
2.40-2.50 (m, 2H, PhCH2), 2.80-2.90 (m, 1H, CHSO3K),
6.60-7.60 (m, 9H, aromatic). 31P NMR (D2O): δ 14.1.
1-Phosphono-4-[3-(3,5-difluorophenoxy)phenyl]butyl-
sulfonic Acid Tripotassium Salt (33). Compound 33 was prepared
from 3,5-difluoroiodobenzene (3 mmol) and 3-hydroxybenzaldehyde
(4.5 mmol), following general methods B and A, as a white powder
(415 mg, 25% overall yield). Anal. (C16H14F2K3O7PS·H2O) C, H.
1H NMR (400 MHz, D2O): δ 1.60-1.90 (m, 4H, -CH2CH2-),
2.40-2.50 (m, 2H, PhCH2), 2.80-2.90 (m, 1H, CHSO3K),
6.40-7.20 (m, 7H, aromatic). 31P NMR (D2O): δ 14.2.
1-Phosphono-4-[3-(3,5-dichlorophenoxy)phenyl]butyl-
sulfonic Acid Tripotassium Salt (34). Compound 34 was prepared
from 3-(3,5-dichlorophenoxy)benzaldehyde (3 mmol), following
general method A, as a white powder (570 mg, 26% overall yield).
Anal. (C16H14Cl2K3O7PS ·C2H5OH·KBr) C, H. 1H NMR (400 MHz,
D2O): δ 1.60-1.90 (m, 4H, -CH2CH2-), 2.45-2.55 (m, 2H,
PhCH2), 2.85-2.95 (m, 1H, CHSO3K), 6.80-7.40 (m, 7H,
aromatic). 31P NMR (D2O): δ 12.8.
1-Phosphono-4-[3-(4-fluorophenoxy)-6-fluoro-phenyl]butyl-
sulfonic Acid Tripotassium Salt (35). Compound 35 was prepared
from 4-fluorophenol (4.5 mmol) and 3-bromo-6-fluorobenzaldehyde
(3 mmol), following general methods B and A, as a white powder
(525 mg, 30% overall yield). Anal. (C16H14F2K3O7PS·0.25KBr·
H2O) C, H. 1H NMR (400 MHz, D2O): δ 1.60-1.90 (m, 4H,
-CH2CH2-), 2.50-2.70 (m, 2H, PhCH2), 2.75-2.85 (m, 1H,
CHSO3K), 6.60-7.00 (m, 7H, aromatic). 31P NMR (D2O): δ 13.6.
1-Phosphono-4-[3-(4-fluorophenoxy)-6-methoxy-phenyl]butyl-
sulfonic Acid Tripotassium Salt (36). Compound 36 was prepared
from 4-fluorophenol (4.5 mmol) and 3-bromo-6-methoxybenzal-
dehyde (3 mmol), following general methods B and A, as a white
powder(470mg, 25%overallyield). Anal. (C17H17FK3O8PS·0.4KBr ·
Staphyloxanthin Biosynthesis Inhibition Assay. The S. aureus
strain used was the WT clinical isolate (Pig1).4 S. aureus was
cultured in THB (1 mL) in the presence of inhibitor compounds
for 72 h, in duplicate. Cells were then centrifuged and washed twice
with PBS. STX was extracted with MeOH, and the OD was
determined at 450 nm using a Perkin-Elmer MBA 2000 (Norwalk,
CT) spectrophotometer. The IC50 values were obtained by fitting
the OD data to a normal dose-response curve, using GraphPad
PRISM.
1
2H2O) C, H. H NMR (400 MHz, D2O): δ 1.50-1.90 (m, 4H,
-CH2CH2-), 2.30-2.50 (m, 2H, PhCH2), 2.60-2.70 (m, 1H,
CHSO3K), 3.62 (s, 3H, OMe), 6.65-7.00 (m, 7H, aromatic). 31P
NMR (D2O): δ 13.6.
Computational Methods. CoMSIA analysis was performed with
default settings in Sybyl16 (version 7.3). All compounds were
geometrically optimized, using the MMFF94x force field, then
aligned in the program MOE,13 utilizing the flexible alignment
module.14 The alignment was carried out by performing up to 1000
flexible refinement iterations using a gradient test of 0.01-1.0 with
hydrophobe, log P, and partial charge similarity features, as well
as the default options (H-bond donor, acceptor, aromaticity, polar
hydrogens, and volume). The alignments were exported into the
Sybyl program, where atomic charges were determined by using
Human SQS Enzyme Expression, Purification, and Inhibi-
tion Assay. A DNA sequence encoding a double truncated protein
lacking residue 31 at the N-terminus and residue 46 at the
C-terminus was amplified using the following primers: 5′-
CATATGGACCAGGACTCGCTCAGCAGC and 3′-GGATCCT-
CAATTCTGCGTCCGGATGGT. The corresponding amplified
insert was initially cloned in the vector pGEMT (Promega). Plasmid
was digested with the endonucleases NdeI and BamHI, and the
resulting fragment was cloned into the bacterial expression vector
pET-28a to give pET28a-HsSQS, which was used to transform E.

988 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 4
Song et al.
(5) Clauditz, A.; Resch, A.; Wieland, K. P.; Peschel, A.; Gotz, F.
Staphyloxanthin plays a role in the fitness of Staphylococcus aureus
and its ability to cope with oxidative stress. Infect. Immun. 2006, 74,
4950–4953.
(6) Liu, C. I.; Liu, G. Y.; Song, Y.; Yin, F.; Hensler, M. E.; Jeng, W. Y.;
Nizet, V.; Wang, A. H.; Oldfield, E. A cholesterol biosynthesis
inhibitor blocks Staphylococcus aureus virulence. Science 2008, 319,
1391–1394.
(7) Sharma, A.; Slugg, P. H.; Hammett, J. L.; Jusko, W. J. Clinical
pharmacokinetics and pharmacodynamics of a new squalene synthase
inhibitor, BMS-188494, in healthy volunteers. J. Clin. Pharmacol.
1998, 38, 1116–1121.
(8) Sharma, A.; Slugg, P. H.; Hammett, J. L.; Jusko, W. J. Estimation of
oral bioavailability of a long half-life drug in healthy subjects. Pharm.
Res. 1998, 15, 1782–1786.
(9) Ciosek, C. P., Jr.; Magnin, D. R.; Harrity, T. W.; Logan, J. V.; Dickson,
J. K., Jr.; Gordon, E. M.; Hamilton, K. A.; Jolibois, K. G.; Kunselman,
L. K.; Lawrence, R. M.; et al. Lipophilic 1,1-bisphosphonates are
potent squalene synthase inhibitors and orally active cholesterol
lowering agents in vivo. J. Biol. Chem. 1993, 268, 24832–24837.
(10) Magnin, D. R.; Biller, S. A.; Chen, Y.; Dickson, J. K., Jr.; Fryszman,
O. M.; Lawrence, R. M.; Logan, J. V.; Sieber-McMaster, E. S.; Sulsky,
R. B.; Traeger, S. C.; Hsieh, D. C.; Lan, S. J.; Rinehart, J. K.; Harrity,
T. W.; Jolibois, K. G.; Kunselman, L. K.; Rich, L. C.; Slusarchyk,
D. A.; Ciosek, C. P., Jr. R-Phosphonosulfonic acids: potent and
selective inhibitors of squalene synthase. J. Med. Chem. 1996, 39,
657–660.
coli BL21 (DE3)RP strain (Novagen) for overexpression. This
cloning procedure resulted in the addition of a six-histidine tag to
the N-terminus of double truncated HsSQS.
Bacteria expressing the constructs were cultured in Luria-Bertani
medium supplemented with kanamycin (30 µg/mL) and chloram-
phenicol (34 µg/mL) at 37 °C, until the cells reached an OD of 0.4
at 600 nm, and were then induced at 37 °C for 4 h by incubation
with 1 mM isopropyl-1-thio-ꢀ-D-galactopyranoside. Cells were
harvested by centrifugation (10 min, 4000 rpm) and resuspended
in 10 mL of lysis/elution buffer (20 mM NaH2PO3, pH 7.4, 10
mM CHAPS, 2 mM MgCl2, 10% glycerol, 10 mM mercaptoethanol,
500 mM NaCl, 10 mM imidazole, and a protease inhibitor cocktail
(Roche), disrupted by sonication, and centrifuged at 16 000 rpm
for 30 min. The supernatant (40 mL) was then applied to a HiTrap
nickel-chelating HP column (Amersham Biosciences). Enzyme
purification was performed according to the manufacturer’s instruc-
tions using a Pharmacia FPLC system. Unbound protein was washed
off with 50 mM imidazole. Then the His6-HsSQS was eluted with
1 M imidazole. Purity was confirmed by SDS-PAGE electro-
phoresis. Fractions containing the pure enzyme were pooled and
dialyzed against buffer A (25 mM sodium phosphate, pH 7.4, 20
mM NaCl, 2 mM dithiothreitol, 1 mM EDTA, 10% glycerol, 10%
methanol), concentrated, then stored at -80 °C.
SQS activity was based on measuring the conversion of [3H]FPP
to [3H]squalene. Final assay concentrations were 50 mM MOPS
(pH 7.4), 20 mM MgCl2, 5 mM CHAPS, 1% Tween-80, 10 mM
DTT, 0.025 mg/mL BSA, 0.25 mM NADPH, and 7.5 ng of purified
recombinant human SQS. The reaction was started with the addition
of substrate (3HFPP, 0.1 nmol, 2.22 × 106 dpm), and the final
volume of the reaction was 200 µL. After incubation at 37 °C for
5 min, an amount of 40 µL of 10 M NaOH was added to stop the
reaction, followed by 10 µL of a (100:1) mixture of 98% EtOH
and squalene. The resulting mixtures were mixed vigorously by
vortexing. Then 10 µL aliquots were applied to 2.5 × 10 cm
channels of a silica gel thin layer chromatogram, and newly formed
squalene was separated from unreacted substrate by chromatography
in toluene-EtOAc (9:1). The region of the squalene band was
scraped and immersed in Hydrofluor liquid scintillation fluid and
assayed for radioactivity. IC50 values were calculated from the
hyperbolic plot of percent of inhibition versus inhibitor concentra-
tion, using GraphPad PRISM.
(11) Lawrence, R. M.; Biller, S. A.; Dickson, J. K.; Logan, J. V. H.; Magnin,
D. R.; Sulsky, R. B.; DiMarco, J. D.; Gougoutas, J. Z.; Beyer, B. D.;
Taylor, S. C.; Lan, S. J.; Ciosek, C. P.; Harrity, T. W.; Jolibois, K. G.;
Kunselman, L. K.; Slusarchyk, D. A. Enantioselective synthesis of
R-phosphono sulfonate squalene synthase inhibitors: chiral recognition
in the interactions of an R-phosphono sulfonate inhibitor with squalene
synthase. J. Am. Chem. Soc. 1996, 118, 11668–11669.
(12) Klebe, G.; Abraham, U.; Mietzner, T. Molecular similarity indices in
a comparative analysis (CoMSIA) of drug molecules to correlate and
predict their biological activity. J. Med. Chem. 1994, 37, 4130–4146.
(13) Molecular Operating EnVironment (MOE); Chemical Computing
Group, Inc.: Montreal, Quebec, 2006.
(14) Labute, P.; Williams, C.; Feher, M.; Sourial, E.; Schmidt, J. M. Flexible
alignment of small molecules. J. Med. Chem. 2001, 44, 1483–1490.
(15) Mao, J.; Mukherjee, S.; Zhang, Y.; Cao, R.; Sanders, J. M.; Song, Y.;
Zhang, Y.; Meints, G. A.; Gao, Y. G.; Mukkamala, D.; Hudock, M. P.;
Oldfield, E. Solid-state NMR, crystallographic, and computational
investigation of bisphosphonates and farnesyl diphosphate synthase-
bisphosphonate complexes. J. Am. Chem. Soc. 2006, 128, 14485–
14497.
(16) Sybyl; Tripos: St. Louis, MO, 2007.
(17) Wildman, S. A.; Crippen, G. M. Prediction of physicochemical
parameters by atomic contributions. J. Chem. Inf. Comput. Sci. 1999,
39, 868–873.
(18) Ma, D.; Cai, Q. N,N-Dimethyl glycine-promoted Ullmann coupling
reaction of phenols and aryl halides. Org. Lett. 2003, 5, 3799–3802.
(19) Teulade, M.-P.; Savignac, P.; Aboujaoude, E. E.; Collignon, N.
Carbanions phosphonates [alpha]-lithium: synthe`se, basicite´ compare´e
et stabilite´ a` l’autocondensation. J. Organomet. Chem. 1986, 312, 283–
295.
Acknowledgment. This work was supported by the U.S.
Public Health Service (Grant HD051796 to V.N., Grants
AI074233, GM073216, and GM65307 to E.O., and Grants
AR045504, CA113874, and AI057160 to C.T.M.). Y.S. was
supported by a Leukemia and Lymphoma Society Special
Fellowship.
(20) Yu, Y.; Srogl, J.; Liebeskind, L. S. Cu(I)-mediated reductive amination
of boronic acids with nitroso aromatics. Org. Lett. 2004, 6, 2631–
2634.
(21) Sanders, J. M.; Song, Y.; Chan, J. M.; Zhang, Y.; Jennings, S.;
Kosztowski, T.; Odeh, S.; Flessner, R.; Schwerdtfeger, C.; Kotsikorou,
E.; Meints, G. A.; Gomez, A. O.; Gonzalez-Pacanowska, D.; Raker,
A. M.; Wang, H.; van Beek, E. R.; Papapoulos, S. E.; Morita, C. T.;
Oldfield, E. Pyridinium-1-yl bisphosphonates are potent inhibitors of
farnesyl diphosphate synthase and bone resorption. J. Med. Chem.
2005, 48, 2957–2963.
(22) Gasteiger, J.; Marsili, M. Iterative partial equalization of orbital
electronegativity, a rapid access to atomic charges. Tetrahedron 1980,
36, 3219–3228.
(23) Bush, B. L.; Nachbar, R. B., Jr. Sample-distance partial least squares:
PLS optimized for many variables, with application to CoMFA.
J. Comput.-Aided Mol. Des. 1993, 7, 587–619.
(24) MATLAB, version 7.4; The Mathworks, Inc.: Natick, MA, 2007.
(25) Rieger, C. E.; Lee, J.; Turnbull, J. L. A continuous spectrophotometric
assay for aspartate transcarbamylase and ATPases. Anal. Biochem.
1997, 246, 86–95.
(26) Ku, B.; Jeong, J. C.; Mijts, B. N.; Schmidt-Dannert, C.; Dordick, J. S.
Preparation, characterization, and optimization of an in vitro C30
carotenoid pathway. Appl. EnViron. Microbiol. 2005, 71, 6578–6583.
Supporting Information Available: Details of computational
cell activity predictions, activity of selected compounds on human
cell lines, elemental analysis results, and dose-responsive curves
of representative compounds in STX inhibition. This material is
available free of charge via the Internet at http://pubs.acs.org.
References
(1) Bancroft, E. A. Antimicrobial resistance: It’s not just for hospitals.
J. Am. Med. Assoc. 2007, 298, 1803–1804.
(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 methicillinresistant Staphylococcus
aureus infections in the United States. J. Am. Med. Assoc. 2007, 298,
1763–1771.
(3) National Research Council. Treating Infectious Diseases in a Microbial
World: Report of Two Workshops on NoVel Antimicrobial Therapeu-
tics; National Academies Press: Washington, DC, 2006; p 21.
(4) Liu, G. Y.; Essex, A.; Buchanan, J. T.; Datta, V.; Hoffman, H. M.;
Bastian, J. F.; Fierer, J.; Nizet, V. Staphylococcus aureus golden
pigment impairs neutrophil killing and promotes virulence through
its antioxidant activity. J. Exp. Med. 2005, 202, 209–215.
JM801023U