Structure Property Relationships of Carboxylic Acid Isosteres

Article abstract of DOI:10.1021/acs.jmedchem.5b01963

The replacement of a carboxylic acid with a surrogate structure, or (bio)-isostere, is a classical strategy in medicinal chemistry. The general underlying principle is that by maintaining the features of the carboxylic acid critical for biological activity, but appropriately modifying the physicochemical properties, improved analogs may result. In this context, a systematic assessment of the physicochemical properties of carboxylic acid isosteres would be desirable to enable more informed decisions of potential replacements to be used for analog design. Herein we report the structure-property relationships (SPR) of 35 phenylpropionic acid derivatives, in which the carboxylic acid moiety is replaced with a series of known isosteres. The data set generated provides an assessment of the relative impact on the physicochemical properties that these replacements may have compared to the carboxylic acid analog. As such, this study presents a framework for how to rationally apply isosteric replacements of the carboxylic acid functional group.

Full text of DOI:10.1021/acs.jmedchem.5b01963

Subscriber access provided by NEW YORK UNIV
Article
Structure Property Relationships of Carboxylic Acid Isosteres
Pierrik Lassalas, Bryant Gay, Caroline Lasfargeas, Michael J James, Van Tran, Krishna G Vijayendran, Kurt
R Brunden, Marisa C Kozlowski, Craig J Thomas, Amos B Smith III, Donna M. Huryn, and Carlo Ballatore
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.5b01963 • Publication Date (Web): 11 Mar 2016
Downloaded from http://pubs.acs.org on March 12, 2016
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Structure Property Relationships of Carboxylic Acid
Isosteres
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
Pierrik Lassalas, Bryant Gay, Caroline Lasfargeas, Michael J. James, Van Tran, Krishna G.
a
a
b
a
a
b
a
c
Vijayendran, Kurt R. Brunden, Marisa C. Kozlowski, Craig J. Thomas, Amos B. Smith III, Donna
a
a,*
M. Huryn and Carlo Ballatore
a,b,*
a
th
Department of Chemistry, School of Arts and Sciences, University of Pennsylvania, 231 South 34 St.,
b
Philadelphia, PA 19104-6323; Center for Neurodegenerative Diseases Research, Institute on Aging,
c
University of Pennsylvania, 3600 Spruce Street, Philadelphia, PA 19104-6323; National Center for
Advancing Translational Sciences, National Institutes of Health, Bethesda, MD 20850, USA
*
Corresponding Authors.
ABSTRACT: The replacement of a carboxylic acid with a surrogate structure, or (bio)-isostere, is a
classical strategy in medicinal chemistry. The general underlying principle is that by maintaining the
features of the carboxylic acid critical for biological activity, but appropriately modifying the
physicochemical properties, improved analogs may result. In this context, a systematic assessment of
the physicochemical properties of carboxylic acid isosteres would be desirable to enable more informed
decisions of potential replacements to be used for analog design. Herein we report the structure-property
relationships (SPR) of 35 phenylpropionic acid derivatives, in which the carboxylic acid moiety is
replaced with a series of known isosteres. The dataset generated provides an assessment of the relative
impact on the physicochemical properties that these replacements may have compared to the carboxylic
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 62
acid analog. As such, this study presents a framework for how to rationally apply isosteric replacements
of the carboxylic acid functional group.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
KEYWORDS: Carboxylic acid isostere; bioisostere; isosteric replacement; physicochemical properties;
structure property relationship; tetrazole.
Introduction. The replacement of an atom, or group of atoms, or even an entire scaffold of a
biologically active compound with a surrogate structure that exhibits broadly similar biological
properties is a fundamental strategy of medicinal chemistry, known as isosteric or bio-isosteric
1
-3
replacement. In general, for this approach to be successful, similarities must exist between at least
some of the properties of the isostere (we use the term ”isostere” in the broadest sense to include both
classical and non-classical isosteres, as well as bio-isosteres) and those of the fragment being replaced,
such that the new analogs retain the biological activities of the parent compound. At the same time,
however, the isosteric replacement must produce changes in the physicochemical properties or
susceptibility to metabolism compared to the parent compound in order to lead to improved derivatives.
The success of any isosteric replacement is invariably context dependent, however, and depends on the
particular molecular environment of the biological target (e.g., the ability of the target to accommodate
surrogate structures), and whether those surrogate structures will also provide the desired improvements
in properties. For this reason, a screening of a series of alternative structures is almost always
4
necessary. In this situation, the availability of experimental data detailing the structure-property
relationship (SPR) of the existing palette of isosteres is most desirable. However, when these data are
not available, the selection/prioritization of potential replacements is typically based on a variety of
factors, such as the historical success rate of specific isosteres, calculated physicochemical properties,
chemical/medicinal chemistry intuition, as well as synthetic accessibility.
ACS Paragon Plus Environment

Page 3 of 62
Journal of Medicinal Chemistry
The importance of the carboxylic acid functional group in drug design is illustrated by the fact that
1
2
5
450 marketed drugs are carboxylic acid containing molecules. However, the presence of this
>
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
functional group in a drug or a drug candidate can be responsible for undesired consequences, such as
limited permeability across biological membranes, metabolic instability, and potential idiosyncratic
toxicities. To circumvent one or more of these shortcomings, medicinal chemists typically resort to
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
prodrug (e.g., ester prodrugs) or isosteric replacement strategies. As part of our continued interest in the
4, 6-9
area of isosteric replacements of the carboxylic acid functional group,
we set out to define the SPR
of a number of acidic moieties that are frequently used as replacements of the carboxylic acid in drug
design. In this particular context, the most important physicochemical parameters are arguably the
acidity and lipophilicity, as well as the effect that the isosteric replacements may have on compound
permeability. Since these three parameters are inter-related, even partial/incomplete information [e.g.,
just pK or distribution coefficient (logD ) value] can be valuable in anticipating the physiochemical
a
7.4
outcome of an isosteric replacement. However, a systematic SPR study enabling an accurate ranking of
specific properties of carboxylic acid isosteres would provide a rational basis on which design decisions
could be made. Somewhat surprisingly, thus far, such a study has never been reported. Although
comparison of experimentally determined physicochemical properties of a number of compounds
4
bearing carboxylic acid isosteres to the parent acid are available in the literature, these data are often
incomplete or fragmented. Furthermore, these data sets are not readily comparable given that different
isosteres are frequently found in completely unrelated molecules and often different methods are used to
define the experimental values. Thus, to enable a more rigorous side-by-side comparison, we have
assembled and characterized a library of 35 model compounds that are derivatives of phenylpropionic
acid. Each compound was then evaluated experimentally for water solubility, acidity, lipophilicity, and
passive permeability in a Parallel Artificial Membrane Permeability Assay (PAMPA). The effect of
each of the carboxylic acid replacements on plasma protein binding was also evaluated. The data
generated in these studies permit an assessment of the SPR of the most commonly incorporated
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 62
carboxylic acid isosteres, which may be useful in the selection and prioritization of potential
replacements of the carboxylic acid moiety to be used in analog design.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Results.
Library Design. The design of the library scaffold took multiple factors into consideration. To ensure
that the physicochemical properties of each member of the compound library were an accurate
representation as possible of the properties of the corresponding carboxylic acid isostere, it was
desirable that each model compound be only minimally functionalized. At the same time, given that the
assays employed to determine the physicochemical properties would rely upon either UV- or mass
spectrometry-based detection methods, the model compounds should both contain a chromophore and
exhibit sufficient molecular weight (i.e., MW ≥ 150 Da). Finally, a short aliphatic spacer separating the
chromophore from the acidic moiety was viewed as desirable to avoid possible interference caused by
conjugation that might affect the intrinsic physicochemical properties of the carboxylic acid surrogate.
Based on these considerations, phenylpropionic acid (1, Table 1) was selected as the reference
compound and each of the compounds in the library was designed as an analog thereof. In addition to
the technical reasons mentioned above, the phenylpropionic acid is also an attractive template as this
substructure is found in a wide range of biologically active compounds, thus making the compound
library directly relevant to medicinal chemistry.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The library of model compounds (36 entries) shown in Table 1 include examples from 24
different classes of carboxylic acid isosteres for which there is evidence of successful applications in
4
drug design. These examples were chosen to represent as wide a range of carboxylic acid surrogates as
possible with respect to structure and properties. They include non-carbon acyclic acids such as
phosphonic and sulfonic acids and sulfonamides, modified carbon-based acids such as hydroxamic acid
and acylurea, heterocycles like tetrazole and thiazolidinedione and carbacycles such as phenols and
squaric acid derivatives. For some of the classes of carboxylic acid isosteres that present multiple, non-
equivalent points of attachment or derivatization (e.g., sulfonamides, acyl sulfonamides), a minimal set
ACS Paragon Plus Environment

Page 5 of 62
Journal of Medicinal Chemistry
of representative congeners was prepared recognizing, however, that the physicochemical properties of
these particular classes of isosteres could be effectively modulated by varying the nature of the
substituents. Compounds 1, 6, 8, and 10 were commercially available, while all other analogs were
synthesized based on literature precedent (see Experimental section). In addition to the structures being
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1 13
established by standard methods ( H- and C-NMR, IR, HRMS), X-ray structures of compounds 1–3,
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
, 6, 8, 10–20, 22, 26, 28, 32, 35, 36 were also obtained (see Supporting Information).
Determination of Physicochemical Properties. The compound library was initially evaluated for kinetic
solubility in aqueous buffer (pH 7.4). These results, summarized in Table 1, reveal that the vast majority
of compounds exhibit aqueous solubility that is either comparable or greater than that of the parent acid
(1). The only exceptions were compounds 21 (which is significantly less soluble than 1) and 34, for
which a determination of aqueous solubility was not possible due to the fact that this compound was not
detected after the 24 h of incubation in aqueous buffer. With the knowledge that solubility would not
limit the subsequent property evaluation of the remaining members, each test compound was evaluated
for: (a) lipophilicity, by determining the distribution coefficient between n-octanol and aqueous buffer
at pH 7.4 (i.e., logD ) via shake-flask method; (b) acidity, by determining the pK values via capillary
7.4
a
electrophoresis; (c) permeability, using a PAMPA assay; and (d) plasma protein binding, by
determining the fraction unbound (fu) via equilibrium dialysis. Calculated logD , and pK values were
7
.4
a
1
also obtained for comparison employing ChemAxon. Table 1 assembles the accumulated data from
0
these studies.
To highlight how the properties of each carboxylic acid isostere change in comparison to acid 1,
Figures 1 and 2 illustrate the distribution of each of the measured values relative to 1. As seen from
Figure 1, the replacement of the carboxylic acid with these surrogate structures can lead to a
considerable variation of physicochemical properties relative to the reference compound 1: pK values
a
span over approximately 10 pK units (2–12), with most isosteres being either equally acidic or less
a
acidic than the parent acid. Only a handful of acidic moieties (e.g., non-carbon acids 6–9, as well as
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 62
squaric acid derivatives 30 and 31) had measured pK values below 4. The modulation of the pK of a
a
a
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
target molecule is often an important aspect in compound optimization. Influence of pK on compound
a
permeability (vide infra) as well as binding to the pharmacological target can be profound, therefore a
ranking of carboxylic acid isosteres based on their intrinsic acidity provides a starting point for
prioritizing which moieties are most likely to impart pK values within the desired range. In addition,
a
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
these data could be helpful in determining whether neighboring group substitution may be required to
modulate the pK values of specific isosteres of interest if the intrinsic acidity falls outside of the
a
optimal range.
In contrast to the wide range of pK values, lipophilicity (i.e., logD ) and permeability
a 7.4
coefficient (P ) values in the PAMPA assay appear to be more narrowly distributed within
app
approximately ~3 log units, with a near equal number of isosteres leading to a relative increase and a
decrease in lipophilicity and permeability. Examples of isosteres with logD7.4 most close to the acid
include tetrazole 16, oxazolidinedione 18, oxadiazol-5(4H)-thione 22, tetramic acid 24 and
cyclopentane-1,3-diones 25–27. Permeability coefficients (P ) were obtained for 33 of the 36 test
app
compounds. The data for the oxathiadiazole 21, and the fluorophenols derivatives 32 and 33 were
considered unreliable due to high non-specific binding (32) or compound precipitation (21 and 33) on
the donor plate during the assay (Table 1). Carboxylic acid isosteres with relatively high membrane
permeability in the PAMPA assay (i.e., log Papp > –5.8) included acylurea 15, sulfonamide 11,
thiazolidinedione 17, thiadiazol-5(4H)-one 20, cyclopentane-1,2-diones 28 and 29 and substituted
phenol 35.
Similar to the results of pK determination, the effect of isosteric replacements on plasma protein
a
binding was pronounced (Figure 2). Some derivatives displayed very high fraction unbound (fu) values
including the acylurea 15 (77%), hydroxamic esters 4 and 5 (68%, and 64%, respectively) and
ACS Paragon Plus Environment

Page 7 of 62
Journal of Medicinal Chemistry
sulfonamide 10 (61%), while others [oxadiazol-5(4H)-thione 22, isoxazole 23, and substituted phenols
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
3
2 & 33] were virtually completely bound (i.e., fu < 1%).
Since many medicinal chemistry optimization projects rely on computational chemistry
programs to predict specific properties to prioritize compounds for synthesis, we employed programs in
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ChemAxon to calculate logD7.4 and pK values for all analogs. From the data depicted in Table 1, a
a
comparison of experimental and calculated logD7.4 or pK values reveals that the calculated values for
a
the carboxylic acid and most isosteres appear to correlate well with experimental results; however, there
are notable exceptions in which the discrepancy can be as large as 2 or 3 log unit (see Figure 4A/B and
Figure 5A/B). This variance is especially notable for the oxadiazol-5(4H)-thione (22) and conjugated
1,3-dicarbonyl systems giving rise to vinylogous acids (e.g., 24–27, 30 and 31). As such, caution should
be exercised when employing computational chemistry programs to predict properties of these structural
types.
While the specific value for each of the individual properties is useful, more helpful are the
correlations one can draw within the dataset. As stated above, most of the measured properties are inter-
related and one property certainly influences the other. Although no overt correlation was evident
between the fu and logD7.4 values (Figure 2), overall there appears to be a relationship linking the
ionization state of the acidic moiety with the fu. Thus, in most cases, the least acidic isosteres (i.e., pKa
>
8), such as the hydroxamic acids (i.e., 2 and 3) and esters (i.e., 4 and 5), sulfonamides (i.e., 10 and
1
1), and acyl-urea (i.e., 15) were found to lead to higher fu values (> 29%) relative to the carboxylic
acid and most other isosteres that are predominantly negatively charged at pH 7.4 (Table 1).
Evaluation of the data also revealed a reasonably good correlation between the lipophilicity and
2
permeability data (r =0.828; Figure 1). However, of note, some compounds having very similar acidity
and lipophilicity were found to exhibit significantly different permeability coefficients in the PAMPA
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 62
assay. For example, comparison of the amino squaric acid (31) with the phosphinic acid (7) derivative
suggests that the latter may be significantly less permeable than the former, in spite of comparable
acidity and lipophilicity (Table 1 and Figure 1). Relatively large differences in permeability coefficients
between compounds that are essentially isometric with respect to lipophilicity and acidity are also
observed when comparing carboxylic acid 1 with other isosteres, such as the cyclopentane-1,3-dione
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(25–27), or the oxadiazol-5(4H)-thione (20), or the tetrazole (16). In particular, the difference in
permeability between 1 and 16 was confirmed in three independent experiments (p < 0.01).
This observation indicated that, in addition to acidity and lipophilicity, other factors can influence
compound permeability in the PAMPA assay. One possibility is that the de-solvation energy may be
significantly different with different acidic moieties, even in those cases where acidity and lipophilicity
are closely comparable. Since the solvation/de-solvation energy is determined in large part by hydrogen-
bond (HB) interactions, we conducted further studies to compare specifically the HB capacity of
1
tetrazole 16 and carboxylic acid 1. For this endeavor, we employed a colorimetric assay in which the
1
H-bonding between a fluorescent pyrazinone sensor (HB acceptor) and the analyte HB donor (i.e., 1 or
1
6) was monitored by following the characteristic blue-shift of the maximum wavelength (λmax) of the
sensor in the UV spectrum (Figure 3). Given that the change in λmax of the sensor depends on the
strength of HB interaction with the analyte, stronger HB donors will cause comparatively larger shifts
than weaker HB donors. Interestingly, these titration experiments revealed that the tetrazole derivative
engages in significantly stronger HB interactions than the corresponding carboxylic acid (Figure 3).
These results suggest that if similar differences exist in the interaction with water, then a tighter
solvation and correspondingly higher de-solvation energy would be expected for 16 compared to 1,
leading to less permeability in a PAMPA assay. Further evaluation of this possibility will require,
however, additional studies to compare 16 and 1 in the deprotonated forms (i.e., tetrazolate and
carboxylate) as HB acceptors.
ACS Paragon Plus Environment

Page 9 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
Table 1.
Class
a
c
b
g
b
h
Cpd #
Structure
Aq. Solub.
(µM)
logD7.c
logD7.4 calc
PAMPA
% retention
pKa
pK calc.
a
PPB (% fu)
d
e
f
Pe (cm/s)
logPapp
*
Carboxylic acid
1
110.69 ±
3.04
–0.49 ±
0.19
–0.56
1.23
1.16
1.66E-06
± 3.48E-7
–6.8 ± 11
0.03
–5.79 ±
0.10
4.64
8.18
8.83
4.73
8.90
8.37
9.5 ± 0.4
29 ± 2
*
2
≥ 200
≥ 200
0.71
1.52
4.97E-06
4.53E-06
–5.30
–5.34
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
Hydroxamic
acids
*
3
1.1
37 ± 10
4
≥ 200
1.16
1.18
1.59
1.35
7.28E-06
4.60E-06
5.3
–5.14
–5.34
9.47
9.58
8.45
8.88
68 ± 3
64 ± 3
31 ± 5
Hydroxamic
esters
*
5
199.80 ±
0
–2.9
.49
*
6
152.36 ±
1.18
–1.14
–1.44
–1.45
–1.30
–1.54
–1.36
–1.17
–0.84
9.40E-08
1.70E-08
3.84E-08
–2.7
–2.1
–6.1
–6.8
–7.03
–7.77
–7.42
2.34
(8.49)
1.81
2.24
Phosphonic acid
Phosphinic acid
7
127.73 ±
1.96
1.98
<2.0
2.1
8.90 ± 0.06
0.31 ± 0.08
5.0 ± 0.7
*
Sulfonic acid
Sulfinic acid
8
≥ 200
–0.81
2.00
†
†
9
≥ 200
ND
ND
*
Sulfonamides
10
≥ 200
≥ 200
0.96
1.42
0.63
1.15
2.13E-06
1.05E-05
3.45E-07
4.2
8.1
1.4
–5.67
–4.98
–6.46
10.04
>12
11.38
12.06
4.08
61 ± 0.04
37 ± 0.06
12.8 ± 0.2
*
1
1
*
*
Acyl-
sulfonamides
12
13
199.70 ±
0.30
–1.02
–0.21
4.94
199.03 ±
1.24
0.17
–0.22
1.53E-06
2.2
–5.81
5.86
4.12
8.1 ± 0.2
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
*
*
*
Sulfonylureas
Acylurea
14
15
16
197.76 ±
.24
–1.23 ±
0.06
–0.87
0.57
0.10
2.61E-07
± 1.01E-7
–3.0 ± 5.39
–3.3
–6.61 ±
0.20
5.04
>12
5.09
4.14
11.77
5.08
31 ± 2
77 ± 1
2
≥ 200
≥ 200
1.42
1.63E-05
–4.79
Tetrazole
–0.25 ±
0
4.83E-07
± 1.48E-7
4.7 ± 2.8
–6.33 ±
0.15
1.1 ± 0.12
.10
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
*
*
*
Thiazolidine
diones
17
18
19
200.41 ±
0.41
1.07 ±
0.03
1.12
0.70
1.26
8.76E-06
± 1.06E-6
5.5 ± 1.7
–0.6
–5.06 ±
0.06
6.19
5.86
5.73
6.61
6.63
6.04
3.4 ± 0.11
14 ± 1
Oxazolidine
diones
≥ 200
–0.16
2.46E-06
–5.61
Oxadiazol-
5(4H)-one
≥ 200
0.32
1.22E-06
–4.0
–5.91
1.1 ± 0.12
*
Thiadiazol-
5
20
200.47 ±
0.47
1.66
ND
2.18
0.95
2.84
1.14E-05
–0.7
–1432
–3.4
–4.94
6.50
5.23
3.58
7.19
6.41
7.77
1.2 ± 0.45
ND
(4H)-one
‡
‡
Oxathiadiazole-
2-oxide
21
7.13 ± 0.74
1.10E-07
–6.96
*
Oxadiazol-
(4H)-thione
22
≥ 200
–0.25
3.27E-07
–6.49
0.65 ± 0.16
5
Isoxazole
23
24
≥ 200
≥ 200
0.46
1.34
1.34
4.65E-06
2.50E-06
–11
1.3
–5.33
–5.60
5.36
6.08
6.21
0.10 ± 0.10
ND
Tetramic acid
–0.35
10.54
2
5
194.93
±1.01
–0.70
–0.33
–0.60
2.32
2.71
2.16
2.12E-07
2.60E-07
1.54E-07
–3.0
–5.1
–8.8
–6.67
–6.58
–6.81
4.01
4.47
4.44
8.82
8.72
8.65
8.00 ± 0.35
11 ± 0.14
14 ± 0.6
Cyclopentane
1
*
26
≥ 200
,3-diones
2
7
199.04 ±
.76
0
ACS Paragon Plus Environment

Page 11 of 62
Journal of Medicinal Chemistry
1
*
Cyclopentane
,2-diones
28
195.03 ±
1.75
1.85
1.71
2.34
1.94
1.14E-05
1.70E-05
1.5
–4.94
–4.77
8.88
8.28
9.24
9.37
16 ± 4
ND
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
1
2
9
192.44 ±
4
–8.6
Cyclopentane
,2-diones
.06
1
†
†
30
31
165.79 ±
0.06
–0.84
–1.40
1.36
1.18
ND
–13
ND
<2.0
<2.0
6.56
7.97
8.20 ± 0.18
6.4 ± 0.34
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
Squaric acid
derivatives
≥ 200
2.24E-07
–6.6
–6.65
–5.15
*
#
3
2
109.17 ±
1
3.34
3.56
>3.78
2.16
1.76
3.88
3.85
3.91
2.55
2.31
7.05E-06
4.71E-06
ND
49
7.19
7.05
9.06
7.70
7.12
7.74
7.62
9.18
7.25
6.74
0.26 ± 0.08
0.52 ± 0.21
ND
.35
‡
3
3
148.70 ±
3
–32
ND
1.1
7.2
–5.33
ND
.91
§
Substituted
phenols
34
ND
*
3
3
5
6
≥ 200
1.23E-05
1.57E-05
–4.91
–4.80
12 ± 3
*
197.11 ±
0.29
1.60 ± 0.03
a
b
Kinetic solubility in aqueous phosphate buffer (pH 7.4) determined by LC/MS after 24 h of incubation (experiment run by Analyza);
10
Calculated values using ChemAxon ; Distribution coefficient between n-octanol and aqueous buffer (pH 7.4) determined by LC/MS (experiment
c
d
run by WuXi AppTech); Effective permeability (PAMPA assay run by Analyza); Membrane retention; Log of the apparent permeability
e
f
g
coefficient; Pk values determined by capillary electrophoresis, for diprotic compounds the second equivalence point is indicated in brackets
a
h
(experiment run by Analyza); Plasma protein binding, fraction unbound (fu) determined by equilibrium dialysis; X-ray crystal structure
*
†
revealing H-bond pattern is shown in the Supporting Information; Permeability value could not be calculated as the concentration of test
‡
compound in the acceptor plate was below limit of quantitation (<LOQ); Compound precipitated in donor plate; Compound appeared to exhibit
#
§
high non-specific binding; Compound appeared to be unstable during the 24h incubation time of the assay; ND = not determined.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 1. Plot showing lipophilicity (i.e., logD ), acidity (i.e., pK ), and permeability (i.e., logP ) of
a
7
.4
app
test compounds, relative to the carboxylic acid compound 1; logD and logP values for compounds
7
.4
app
1
, 16, 14 and 17 are the averages obtained from three independent experiments; (**) p < 0.01 by two-
tailed t-test confirming statistical significant difference in membrane permeability between 16 and 1; NS
indicates that the difference in logD7.4 values between 16 and 1 does not reach statistical significance.
ACS Paragon Plus Environment

Page 13 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. A plot of compound lipophilicity (i.e., logD ), plasma protein binding (i.e., fu), and acidity
7
.4
(i.e., pK ) relative to the carboxylic acid compound 1; fu values are the averages obtained from three
a
independent experiments; logD values for compounds 1, 16, 14 and 17 are the averages obtained from
7.4
three independent experiments.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 62
Figure 3. Titration curves from a colorimetric assay that monitors the blue-shift in λmax of the
pyrazinone sensor upon complexation with the HBD analyte (1 or 16).
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4. Comparison between experimental and calculated logD7.4 values; (A) the linear regression for
the entire data set; (B) the linear regression when major outliers (i.e., discrepancy > 1 log unit, shown in
red) are excluded.
ACS Paragon Plus Environment

Page 15 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. Comparison between experimental and calculated pK values; (A) the linear regression for the
a
entire data set; (B) the linear regression when major outliers (i.e., discrepancy > 1 log unit, shown in
red) are excluded.
Discussion. The carboxylic acid is one of the most frequent functional groups in small molecules that
12
bind to protein targets, and is considered a privileged sub-structure. The acidity, combined with the
ability to establish relatively strong electrostatic interactions and hydrogen-bonds, as well as the ability
to participate in interactions with dipoles make this functional group highly versatile, and in turn able to
engage in molecular interactions with a wide array of complementary functional groups. These
properties also imply that the carboxylic acid moiety can impart relatively high water solubility, which
is an important attribute for a drug-like molecule. On occasion, these same properties can also
contribute to some of the deficits of carboxylic acids as drug candidates, including, for example, a
relatively poor permeability across biological membranes. In addition, other potential liabilities that
have been associated with the carboxylic acid functional group in drugs or drug candidates include a
1
relatively rapid metabolism mediated by uridine 5'-diphospho-glucuronosyl-transferase (UGTs). This
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 62
phase 2 metabolic process can be responsible for limited compound half-life and for the formation of
reactive acyl-glucuronides that can cause covalent modifications of proteins and potentially serious
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
4
adverse side effects. The replacement of the carboxylic acid with appropriate isosteres can be
especially useful to circumvent these possible drawbacks while maintaining the positive aspects of the
carboxylic acid functional group. Furthermore, in addition to these specific applications, carboxylic acid
isosteres can be employed more broadly in analog design to explore the effects of diversification of
structure and physicochemical properties of compounds of interest. The utility of this strategy is evident
from the fact that several carboxylic acid isosteres, such as tetrazole, thiazolidinedione, sulfonamides,
and acyl sulfonamides have ultimately led to the clinical development of entire classes of important
therapeutic agents (e.g., see Figure 6).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Based on the importance of acid isosteres to the field of medicinal chemistry and the critical role
that these fragments can play in tuning the properties of acidic, biologically active compounds, we
assembled a set of 35 carboxylic acid isosteres on the same scaffold and measured a focused set of key
physicochemical properties, such as acidity, lipophilicity, and permeability. Furthermore, given that
molecules possessing the carboxylic acid functionality are often found to exhibit high degree of plasma
1
5, 16
protein binding,
the effect of these isosteric replacements on the fraction unbound was also
investigated. Finally, using computational chemistry programs, we calculated pK and logD values to
a 7.4
compare to the experimental values. The dataset generated in this study provides an assessment of the
effect that such isosteres can have on the physicochemical properties of the corresponding carboxylic
acid. In addition, our results highlight some correlations between properties that can be used during the
selection/prioritization of isosteres for analog design.
ACS Paragon Plus Environment

Page 17 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 6. Representative examples of drugs/drug candidates containing carboxylic acid bioisosteres.
17
In agreement with previous studies, our results indicate that overall there is a clear correlation
between the experimental logD values and the apparent permeability coefficients in the PAMPA assay
7
.4
(Figure 1), such that, generally, more lipophilic, less acidic compounds exhibit higher rates of passive
diffusion. However, in selected cases the impact that isosteric replacements produce on the compound
permeability appears to deviate significantly from this relationship. These deviations are most
effectively illustrated by the comparison between tetrazole 16 and carboxylic acid 1. The tetrazole,
1
8, 19
which among the carboxylic acid isosteres is one that has been the focus of intense studies,
is
2
generally considered to be more lipophilic and thus potentially more permeable than the carboxylic
0
2
acid. However, our comparative studies with model compounds reveal that whereas the two acidic
1
moieties exhibit very similar pK and logD values, the tetrazole is in fact significantly less permeable
a 7.4
in the PAMPA assay. Interestingly, previous Caco2 bi-directional permeability studies with a series of
carboxylic acids and corresponding 1H-tetrazoles derivatives found that several tetrazoles were less
22, 23
permeable than the carboxylic acids due to active transport mechanisms.
However, in our studies,
given the artificial nature of the PAMPA assay, active transport systems, metabolism, and protein
binding are excluded from consideration. Therefore, any differences in the rate of transport should
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 62
ultimately result from specific differences in physicochemical properties between the acidic moieties.
Our studies found that the model compound bearing the tetrazole moiety can establish considerably
stronger HB interactions than the corresponding compound bearing the carboxylic acid. This result is
consistent with previous reports in which other 1H-tetrazoles were found to be relatively strong HB
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
donors, and suggests that in water the tetrazole derivative may be more tightly solvated than the
4
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
carboxylic acid leading to correspondingly higher de-solvation energies. As the de-solvation energy is
25, 26
known to affect compound permeability across biological membranes,
the different rates of
diffusion that have been observed for 16 and 1 in the PAMPA assay may be related, at least in part, to
the difference in HB strength between the tetrazole and the carboxylic acid. In this regard, it should be
noted that the lack of equivalence between the hydrogen-bonding and the acidity scale is not
2
4, 27
unexpected, as this phenomenon has been documented.
In addition to the differences in the permeability rates in the PAMPA assay, 16 and 1 were also
found to exhibit significant differences in the plasma protein-binding assay with the tetrazole derivative
being more tightly bound than the corresponding carboxylic acid compound. A similar observation was
reported when comparing the plasma protein-binding of other match paired tetrazoles and carboxylic
2
acid analogs. Also consistent with previous reports,
8
15, 16
which illustrated the importance of the
ionization state of molecules on plasma protein binding, a broader examination of the plasma protein-
binding data from all carboxylic acid isosteres evaluated in this study suggests that derivatives that are
predominantly neutral at pH 7.4 generally exhibit larger fu values. However, there seems to be no clear
correlation between the plasma protein binding and the lipophilicity of test compounds. Although this
outcome appears to be in disagreement with previous studies that reported a relatively good correlation
between the calculated logD7.4 values and plasma protein binding of a series of carboxylic acid
29
molecules, it is possible that such a correlation can be established only within specific families of
closely related compounds and not between different classes of carboxylic acid isosteres. Indeed when
we compared the fu values of 1, 25–27 with the corresponding, but more lipophilic congeners that were
ACS Paragon Plus Environment

Page 19 of 62
Journal of Medicinal Chemistry
brominated in the para position, we found that within each of the matching pairs the fu was
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
considerably lower for the brominated derivative (see Supporting Information).
Finally, the discrepancies between the calculated and the experimental logD7.4 and pK values
a
that are evident for all vinylogous acids tested likely arise from these or similar vinylogous systems
being absent or underrepresented in the training set used by the software. As such, the predictive ability
of the software could be readily improved by simply updating the training set with the data provided in
this study. Our results in general underscore the importance of experimental values and suggest that
caution should be exercised when calculated physicochemical properties are used to prioritize
carboxylic acid isosteres for analog design.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Conclusions. Taken together, the data generated in this study permit an assessment of the relative
differences in physicochemical properties of carboxylic acid isosteres, compared to the corresponding
carboxylic acid. This data set comprises a benchmark that may be useful to compare the
physicochemical properties of carboxylic acid isosteres and ultimately permit more informed and
effective prioritizations of isosteric replacements in medicinal chemistry for analog design.
Experimental Section
3
The synthesis of test compounds is highlighted in Schemes 1–11. The hydroxamic acids 2 and 3
0
31
32
(Scheme 1) were prepared starting respectively from carboxylic acid 1 and O-benzylhydroxylamine.
3
Hydroxamic ester derivative 4 was obtained via 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide
3
3
4
(EDCI) mediated coupling of 1 with O-methyl hydroxylamine, whereas reverse hydroxamic ester 5
was obtained by O-alkylating acetohydroxamic acid, following procedures described by Wrigglesworth
3
5
and co-workers.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 1. Synthesis of hydroxamic acid and ester derivatives. Reagents and Reaction conditions: (a)
CDI, hydroxylamine, CH CN, r.t., 14 h, (70%); (b) N-methylmorpholine, EDCI•HCl, O-methyl
3
hydroxylamine, r.t., 16 h, (83%); (c) (i) K CO , Na CO , MeOH/H O, 50 ˚C, 10 min; (ii) (2-
2
3
2
3
2
bromoethyl)benzene, 50 ˚C, 16 h, (32%); (d) Boc O, THF/H O, NEt , r.t., 2 h, (45%); (e) (i) NaH,
2
2
3
DMF, r.t., 30 min; (ii) (2-bromoethyl)benzene, r.t., 16 h, (53%); (f) (i) TFA/CH Cl , r.t., 18 h, (85%);
2
2
(
ii) Ac O, DMAP, pyridine, CH Cl , r.t., 2 h, (79%); (g) H (1 atm), Pd/C, MeOH, r.t., 3 h, (77%).
2 2 2 2
3
6
37
With respect to phosphinic (7 ) and sulfinic acid (9 ) derivatives (Scheme 2), the former was obtained
by reacting triethylphosphite with phenethylmagnesium bromide, while the latter was prepared in two
steps starting from 2-phenylethanethiol, via an oxidation-reduction sequence.
ACS Paragon Plus Environment

Page 21 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 2. Synthesis of phosphinic and sulfinic acid derivatives. Reagents and Reaction conditions:
(a) Phenethylmagnesium bromide, Et O, 5 ˚C to r.t., 16 h, (32%); (b) NCS, CH Cl /H O, r.t., 3.5 h,
2
2
2
2
(
90%); (c) Na SO , NaHCO , CH CN/H O, r.t., 24 h, (75%).
2
2
3
3
3
3
8
39
The model compounds bearing the sulfonamide (11 ) and acylurea (15 ) functional groups were
synthesized starting respectively from 3-phenylpropan-1-amine and benzyl urea as previously described
Scheme 3). The acylsulfonamide derivatives 12 and 13 (Scheme 3) were obtained by coupling 1 with
(
the appropriate sulfonamides, whereas sulfonylurea 14 was obtained by reacting methylsulfonamide
with benzyl-carbamoylimidazole, which can be readily prepared from benzylamine hydrochloride and
4
0
1,1’-carbonyldiimidazole (CDI).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 62
Scheme 3. Synthesis of sulfonamide, acylsulfonamide, sulfonylurea, and acylurea derivatives. Reagents
and Reaction conditions: (a) MsCl, NEt , CH Cl , 4 ˚C, 24 h, (88%); (b) methanesulfonamide,
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
3
2
2
EDCI•HCl, DMAP, CH Cl , 45 ˚C, 8 h, then r.t., 48 h, (70%); (c) (i) 4 M HCl in dioxane, r.t., 5 min; (ii)
2
2
CDI, DMF, CH CN, r.t., 2 h; (iii) NaH, methanesulfonamide, DMF, r.t., 16 h, (11%); (d) (i) TBTU,
3
DIEA, CH Cl , r.t., 25 min; (ii) N,N-dimethylsulfamide, r.t., 14 h, (81%); (e) Ac O, CH Cl , 50˚C, 24 h,
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
2
2
2
then r.t., 48 h, (75%).
4
1
42
The tetrazole derivative 16 (Scheme 4) was prepared from nitrile 41 and sodium azide, or
alternatively, from the cyanoalkylamide 42 and TMS-azide, under Mitsunobu conditions, followed by
4
3
elimination of the acrylonitrile.
Scheme 4. Synthesis of tetrazole derivative. Reagents and Reaction conditions: (a) EDCI•HCl, HOBt,
DMF, DIEA, 3-aminopropionitrile, r.t., 15 h, (95%); (b) (i) PPh , DIAD, TMSN , 0 ˚C, 30 min, then r.t.,
3
3
2
h, then 50 ˚C, 17 h; (ii) DBU, CH Cl , r.t., 6 h, (63%); (c) NH OH•HCl, DMSO, 90 ˚C, 3.5 h, (61%);
2 2 2
(d) NaN , ZnBr , H O, 150 ˚C, 24 h, (20%).
3
2
2
4
In the case of thiazolidine-2,4-dione (17 ) and oxazolidine-2,4-dione (18) derivatives, these model
4
compounds were obtained by reacting the appropriately lithiated heterocycle with benzyl bromide
(Scheme 5).
ACS Paragon Plus Environment

Page 23 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 5. Synthesis of thiazolidine-2,4-dione and oxazolidine-2,4-dione derivatives. Reagents and
Reaction conditions: (a) (i) n-BuLi, THF, –78 ˚C, 10 min (ii) benzyl bromide, –78 ˚C to r.t, 1.5 h,
(53%); (b) (i) t-BuOK, diethyl carbonate, MeOH, 70 ˚C (microwave irradiation), 19 h, then r.t., 24 h;
(ii) n-BuLi, THF, –78 °C to 0 ˚C, 30 min; (iii) benzyl bromide, –78 ˚C to r.t., 2 h, (5%).
For the synthesis of the oxadiazole derivatives (19 and 22) as well as related thiadiazole (20) and
oxathiadiazole (21) compounds (Scheme 6), common precursor N'-hydroxy-3-phenylpropanimidamide
(43) was reacted with the appropriate acylating agent according to conditions reported by Kohara and
45
co-workers.
Scheme 6. Synthesis of 5-oxo-1,2,4-oxadiazole and related thiadiazole derivatives. Reagents and
Reaction conditions: (a) NH OH, H O, EtOH, 75 ˚C, 5.5 h, (92%); (b) (i) isobutyl chloroformate,
2
2
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 62
pyridine, DMF, 0 ˚C to r.t., 16 h; (ii) toluene, 120 ˚C, 2 h, then 140 ˚C, 24 h, (59%); (c) (i) 1,1'-
thiocarbonyldiimidazole, THF, r.t., 40 min; (ii) BF •OEt , THF, r.t., 3 h, (28%); (d) SOCl , THF,
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
3
2
2
CH Cl , 0 ˚C, 1 h, (<10%); (e) DBU, 1,1'-thiocarbonyldiimidazole, CH CN, r.t., 24 h, (43%).
2
2
3
4
The synthesis of known 3-hydroxyisoxazol 23 was conducted starting from Meldrum’s acid as
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
highlighted in Scheme 7.
Scheme 7. Synthesis of isoxazol-3-ol derivative. Reagents and Reaction conditions: (a) (i) pyridine,
CH Cl , 0 ˚C, 10 min; (ii) phenylacetyl chloride, 0 ˚C, 4 h, (91%); (b) N,O-di-Boc-hydroxylamine,
2
2
toluene, 65 ˚C, 16 h, (quant.); (c) 4 M HCl, MeOH, r.t., 16 h, (10%).
4
7
The synthesis of a tetramic acid derivative (24 ) was accomplished according to known procedures in
three steps, starting from N-Boc protected phenylalanine and Meldrum’s acid, while structurally related
cyclopentane-1,3-diones (25 and 26) were obtained via chemoselective benzylation of the appropriate
isobutyl-protected cyclopentane-1,3-dione with benzyl bromide under conditions reported by Koreeda
4
and co-workers, followed by hydrolysis of the vinylogous ester. The 1,3-dione derivative 27 was
8
prepared in one step from cyclopentane-1,3-dione and benzaldehyde under reductive alkylation
49
conditions.
ACS Paragon Plus Environment

Page 25 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 8. Synthesis of tetramic acid and cyclopentane-1,3-dione derivatives. Reagents and Reaction
conditions: (a) (i) Meldrum’s acid, DMAP, CH Cl , EDCI•HCl, 0 ˚C to r.t., 15 h; (ii) EtOAc, reflux, 30
2
2
min; (iii) TFA/CH Cl , r.t., 15 min, (91%); (b) (i) LDA, THF, –78 ˚C, 20 min; (ii) benzyl bromide, –78
2
2
˚C to r.t., 3 h, (27–45%); (iii) acetone, HCl (2 M), r.t., 16 h, (42–67%); (c) Hanztsch ester, L-proline,
benzaldehyde, CH Cl , r.t., 16 h, (60%).
2
2
5
The synthesis of cyclopentane-1,2-dione derivative 28 (Scheme 9) was accomplished in two steps by
0
5
reacting known di-potassium salt 45 with benzyl bromide to give intermediate 46, followed by a de-
1
52
benzylation/de-carboxylation sequence of the latter to obtain the desired 1,2-dione. The 1,2-dione 29
was synthesized via magnesium methoxide induced cyclization of α,β-unsaturated dione 47.
Scheme 9. Synthesis of cyclopentane-1,2-dione derivatives. Reagents and Reaction conditions: (a)
benzyl bromide, DMF, 120 ˚C, 2 h, (47%); (b) AcOH, HCl (37%), 130 ˚C, 3 h, (71%); (c) (i)
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 62
benzaldehyde, NaOH, MeOH, r.t., 24 h; (ii) TsOH•H O, acetone, r.t., 38 h, (73%); (d) Mg(OMe) ,
2
2
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
MeOH, reflux, 90 min, (53%).
The cyclobutane-1,2,3-trione derivative 30 (Scheme 10) was obtained by reacting diethyl squarate with
benzylmagnesium bromide, followed by acid-mediated hydrolysis. Squaric acid monoamide derivative
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
3
3
1
(Scheme 10) was prepared by direct condensation of benzylamine with squaric acid under
microwave assisted conditions.
Scheme 10. Synthesis of squaric acid derivatives. Reagents and Reaction conditions: (a)
benzylmagnesium bromide, THF, 0 ˚C to r.t., 40 min, (82%); (b) 3 M HCl, acetone, r.t., 3 h, (66%); (c)
benzylamine, 200 ˚C (microwave irradiation), 20 min, (22%).
The synthesis of fluorophenol derivative 32 (Scheme 11) was completed in 4 steps starting from 2,6-
difluorophenol. Thus, after bromination and protection of the phenol as benzyl-ether, Negishi cross-
coupling conditions were employed to obtain 51, which was finally deprotected by hydrogenolysis to
give model compound 32. The corresponding meta-benzylated isomer 33 was obtained via selective
ortho-borylation of protected phenol 52, followed by a Suzuki-Miyaura cross-coupling reaction with
54
55
benzyl bromide, and final deprotection. Finally, the synthesis of substituted phenols 34–36 was
accomplished starting from 2-hydroxythiophenol via benzylation reaction (34), followed by S-oxidation
to the corresponding sulfoxide (35) and sulfone (36).
ACS Paragon Plus Environment

Page 27 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 11. Synthesis of substituted phenol derivatives. Reagents and Reaction conditions: (a) NBS,
DMF, 0 ˚C to r.t., 40 h, (86%); (b) (i) K CO , acetone, 70 ˚C, 1 h; (ii) benzyl bromide, 70 ˚C, 3 h then
2
3
®
r.t., 14 h, (99%); (c) ZnBr , benzylmagnesium chloride, PEPPSI-IPr , THF, r.t., 16 h, (74%); (d) H (1
2
2
atm), Pd/C, MeOH or EtOH, r.t., 1 h, (97–99%); (e) (i) n-BuLi, THF, –50 ˚C, 20 min; (ii) B(O-i-Pr) , –
3
5
0 ˚C to r.t., 1 h; (f) (PPh ) Pd(Br)(Succ), benzyl bromide, Na CO , THF/H O, 60 ˚C, 14 h, (42%); (g)
3 2 2 3 2
(i) KHCO , DMF, r.t., 5 min; (ii) benzyl bromide, r.t., 16 h, (93%); (h) AcOH, UHP, r.t., 20 h, (98%);
3
(l) N-methylmorpholine N-oxide, K OsO •2H O, CH Cl , r.t., 20 h, (95%).
2
4
2
2
2
Materials and methods. All solvents were reagent grade. All reagents were purchased from Aldrich
or Fisher Scientific and used as received. Thin layer chromatography (TLC) was performed with 0.25
mm E. Merck pre-coated silica gel plates. Flash chromatography was performed with silica gel 60
(particle size 0.040–0.062 mm) supplied by Silicycle and Sorbent Technologies. TLC spots were
detected by viewing under a UV light, or using KMnO or ceric ammonium molybdate stains. Melting
4
points (mp) were acquired on a Thomas-Hoover apparatus and are uncorrected. Infrared (IR) spectra
1
13
were recorded on a Jasco Model FT/IR-480 Plus spectrometer. Proton ( H) and carbon ( C) NMR
spectra were recorded on a Bruker AMX-500 spectrometer. Chemical shifts were reported relative to the
residual solvent’s peak. High-resolution mass spectra were measured at the University of Pennsylvania
Mass Spectrometry Center on either a VG Micromass 70/70H or VG ZAB-E spectrometer. Single-
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 62
crystal X-ray structure determinations were performed at the University of Pennsylvania with an Enraf
Nonius CAD-4 automated diffractometer. Analytical reverse-phased (Sunfire™ C18; 4.6x50 mm, 5 mL)
high-performance liquid chromatography (HPLC) was performed with a Waters binary gradient module
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
525 equipped with Waters 2996 PDA and Waters micromass ZQ. All samples were analyzed
employing a linear gradient from 10% to 90% of CH CN in H O over 8 min and flow rate of 1 mL/min,
3
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
and unless otherwise stated, the purity level was >95%. Preparative reverse-phase HPLC purifications
were performed on a Gilson instrument (i.e., Gilson 333 pumps, a 215 liquid handler, 845Z injection
module, and PDA detector) employing Waters SunFire™ preparative C OBD™ columns (5 ꢀm 19x50
1
8
or 19x100 mm). Unless otherwise stated, HPLC purifications were carried out employing a linear
gradient from 10% to 90% of CH CN in H O for 15 min with a flow rate of 20 mL/min. Unless
3
2
otherwise stated, all final compounds were found to be >95% as determined by HPLC/MS and NMR.
3-phenylpropionic acid (1): Commercially available. X-Ray quality crystals were obtained by slow
evaporation from a CH Cl solution (see Supporting Information): mp (CH Cl ) 49–51 °C.
2
2
2
2
N-(Hydroxy)-3-phenylpropanamide (2): To a stirred solution of 1 (1.00 g, 6.66 mmol, 1.00 equiv)
in anhydrous CH CN (22 mL) at r.t. under N , CDI (1.30 g, 7.99 mmol, 1.20 equiv) was added in one
2
3
portion, and the resulting solution was stirred for 35 min. A solution of hydroxylamine (50% wt in H O,
2
0
.450 mL, 7.32 mmol, 1.10 equiv) was then added dropwise, and the solution was stirred for 14 h at r.t..
The reaction was quenched by careful addition of 1 M KHSO (until pH 1–2), then extracted with
4
EtOAc (×3). The combined organic extracts were washed with phosphate buffer (pH 7), H O, brine,
2
then dried over Na SO , filtered, and concentrated in vacuo. Recrystallization from EtOAc/hexanes
2
4
provided the title product as a colorless to off-white solid (0.774 g, 4.69 mmol, 70%). X-Ray quality
crystals were obtained by layer diffusion of cyclohexane into a CH Cl solution (see Supporting
2
2
1
Information): mp (CH Cl /cyclohexane) 73–74 °C; H NMR (500 MHz, CDCl ) δ 9.08 (br s, 2H), 7.27–
2
2
3
1
3
7
.14 (m, 5H), 2.90 (t, J = 7.8 Hz, 2H), 2.40 (t, J = 7.8 Hz, 2H) ppm; C NMR (125 MHz, CDCl ) δ
3
ACS Paragon Plus Environment

Page 29 of 62
Journal of Medicinal Chemistry
1
1
71.3, 140.3, 128.8, 128.5, 126.6, 34.9, 31.5 ppm; IR (KBr) ν 3204, 3059, 3025, 2925, 2861, 1642,
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
–
496, 1452, 1382 cm ; HRMS (ES ) calculated for C H NNaO [M+Na] 188.0687, found 188.0688.
1
+
+
9
11
2
N-Hydroxy-N-phenethylacetamide (3): To a solution of 39 (0.183 g, 0.679 mmol, 1.00 equiv) in
MeOH (3.40 mL), Pd/C (10% wt (wet), 0.018 g) was added. The suspension was stirred under an H2
atmosphere (balloon) for 3 h. The reaction mixture was then filtered through a pad of Celite, washed
with EtOAc, and concentrated in vacuo. Purification by a short silica gel column chromatography (20%
EtOAc in hexanes) afforded the title compound (0.083 g, 0.46 mmol, 68%). X-Ray quality crystals were
obtained by slow evaporation from a CH Cl /hexanes solution (see Supporting Information): mp
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
2
1
(CH Cl /hexanes) 49–51 °C; H NMR (500 MHz, CDCl ) mixture of E/Z isomers δ 7.44–7.08 (m, 3H),
2
2
3
1
.97–3.71 (m, 2H), 2.98 (dt, J = 32.7, 6.7 Hz, 1H), 2.10 (s, 1H), 1.61 (s, 1H) ppm; C NMR (125 MHz,
3
3
CDCl ) mixture of E/Z isomers δ 172.3, 165.2, 138.7, 138.0, 129.1, 128.9, 128.6, 127.1, 126.5, 51.6,
3
–1
+
4
9.7, 33.5, 33.1, 20.5, 18.4 ppm; IR (KBr) ν 3172, 2922, 2848, 1614 cm ; HRMS (ES ) calculated for
+
C H NO [M+H] 180.1025, found 180.1020.
10 14
2
N-Methoxy-3-phenylpropanamide (4): To a solution of 1 (0.210 g, 1.40 mmol, 1.00 equiv) in
CH Cl at r.t. under N , N-methylmorpholine (0.190 mL, 1.70 mmol, 1.20 equiv) was added dropwise,
2
2
2
and the reaction mixture was stirred for 15 min. EDCI•HCl (0.328 g, 1.70 mmol, 1.20 equiv) was added,
and the suspension was stirred until a clear solution was obtained. O-methylhydroxylamine
hydrochloride (0.146 g, 1.70 mmol, 1.20 equiv) was then added and the solution was stirred for 16 h at
r.t.. The reaction was quenched with satd. aq. NH Cl, and extracted with CH Cl (×3). The combined
4
2
2
organic extracts were washed with H O, brine, then dried over Na SO , filtered, and concentrated in
4
2
2
vacuo. Purification by silica gel column chromatography (2–10% MeOH in CH Cl ) afforded the title
2
2
1
compound (0.210 g, 1.17 mmol, 83%) as a clear oil: H NMR (500 MHz, CDCl ) δ 10.05 (s, 1H), 7.28–
3
13
.21 (m, 5H), 3.66 (s, 3H), 2.98 (t, J = 7.8 Hz, 2H), 2.45 (app t, 2H) ppm; C NMR (125 MHz, CDCl )
3
7
δ 170.3, 140.5, 128.6, 128.4, 126.4, 64.0, 34.8, 31.5 ppm; IR (KBr) ν 3182, 3000, 2974, 2936, 1659,
–
1
+
+
1
497, 1453 cm ; HRMS (ES ) calculated for C H NNaO [M+Na] 202.0844, found 202.0838.
10 13 2
ACS Paragon Plus Environment