Constrained H-Phe-Phe-NH2 analogues with high affinity to the substance P 1-7 binding site and with improved metabolic stability and cell permeability

Article abstract of DOI:10.1021/jm400209h

We recently reported the discovery of H-Phe-Phe-NH₂ as a small and high affinity ligand for the substance P 1-7 (SP_1-7, H-Arg-Pro-Lys-Pro-Gln-Gln-Phe-OH) specific binding site and its intriguing ability to reduce neuropathic pain. With the overall aim to develop stable and orally bioavailable SP_1-7 mimetics, the dipeptide was chosen as a lead compound. Herein the structure-activity relationship (SAR) of a set of modified H-Phe-Phe-NH₂ analogues is presented together with their potential active uptake by PEPT1 transporter, intestinal permeability, and metabolic stability. Local constraints via peptide backbone methylation or preparation of cyclized analogues based on pyrrolidine were evaluated and were shown to significantly improve the in vitro pharmacokinetic properties. The SAR was rationalized by deriving a plausible binding pose for the high affinity ligands. Rigidification using a 3-phenylpyrrolidine moiety in the C-terminal of H-Phe-Phe-NH₂ resulted in high affinity and improved intrinsic clearance and intestinal epithelial permeability.

Full text of DOI:10.1021/jm400209h

Article
pubs.acs.org/jmc
Constrained H‑Phe-Phe-NH2 Analogues with High Affinity to the
Substance P 1−7 Binding Site and with Improved Metabolic Stability
and Cell Permeability
Rebecca Fransson,† Christian Skold,† Jadel M. Kratz,‡,∥ Richard Svensson,‡,⊥ Per Artursson,‡,⊥,#
̈
Fred Nyberg,§ Mathias Hallberg,§ and Anja Sandstrom*,†
̈
†Department of Medicinal Chemistry, Uppsala University, Box 574, SE-751 23 Uppsala, Sweden
‡Department of Pharmacy, Uppsala University, Box 580, SE-751 23 Uppsala, Sweden
§Department of Pharmaceutical Biosciences, Uppsala University, Box 591, SE-751 24 Uppsala, Sweden
∥Programa de Pos
Federal de Santa Catarina, 88.040-900, Florianop
́
-Graduaca
̧
o em Farmac
́
ia, Centro de Cien
̂
cias da Saude, Departamento de Cien
olis, SC, Brazil
̂
cias Farmaceu
̂
ticas, Universidade
́
̃
́
⊥The Uppsala University Drug Optimization and Pharmaceutical Profiling Platform (UDOPP), Chemical Biology Consortium
Sweden (CBCS), Uppsala University, Box 580, SE-751 23 Uppsala, Sweden
#Science for Life Laboratory, Uppsala University, SE-751 23 Uppsala, Sweden
ABSTRACT: We recently reported the discovery of H-Phe-
Phe-NH2 as a small and high affinity ligand for the substance P
1−7 (SP1−7, H-Arg-Pro-Lys-Pro-Gln-Gln-Phe-OH) specific
binding site and its intriguing ability to reduce neuropathic
pain. With the overall aim to develop stable and orally
bioavailable SP1−7 mimetics, the dipeptide was chosen as a lead
compound. Herein the structure−activity relationship (SAR)
of a set of modified H-Phe-Phe-NH2 analogues is presented
together with their potential active uptake by PEPT1
transporter, intestinal permeability, and metabolic stability.
Local constraints via peptide backbone methylation or
preparation of cyclized analogues based on pyrrolidine were
evaluated and were shown to significantly improve the in vitro pharmacokinetic properties. The SAR was rationalized by deriving
a plausible binding pose for the high affinity ligands. Rigidification using a 3-phenylpyrrolidine moiety in the C-terminal of H-
Phe-Phe-NH2 resulted in high affinity and improved intrinsic clearance and intestinal epithelial permeability.
activation of the naloxone-sensitive sigma receptor (σ1)
system.12 SP1−7 itself is, however, not an active ligand for the
σ1 receptor nor do the biological effects observed for SP1−7
appear to be mediated through any of the known tachykinin or
opioid receptors.12 Even though the inhibitory effect of SP1−7
on SP induced behavior can be reversed by the nonselective
opioid ligand naloxone, specific antagonists for the opioid
receptors (μ, δ, and κ) do not inhibit this effect.14,15 More
likely, the effects of SP1−7 are generated via an as yet undefined
macromolecular target at specific binding sites identified for this
heptapeptide in rat and mouse spinal cord and brai-
n.16−18Although most studies on SP1−7 have been done in
rodents, its relevance in humans is suggested from the presence
of SP and SP degrading enzymes in human and the presence of
this heptapeptide fragment in human cerebrospinal fluid
(CSF).19,20
INTRODUCTION
■
The undecapeptide substance P (SP, H-Arg-Pro-Lys-Pro-Gln-
Gln-Phe-Phe-Gly-Leu-Met-NH2)1 is a neurotransmitter and
neuromodulator in the central and peripheral nervous system
and is most well-known for its involvement in pain. It belongs
to the tachykinin family and constitutes the endogenous ligand
for the neurokinin-1 (NK-1) receptor.2,3 SP is degraded into
several bioactive fragments of which the N-terminal metabolite
SP 1−7 (SP1−7, H-Arg-Pro-Lys-Pro-Gln-Gln-Phe-OH) is the
most abundant.4−6
SP1−7 has been shown to reverse naloxone provoked opioid
tolerance and withdrawal, in contrast to SP and its C-terminal
fragments that amplify these effects.7−9 It has also been
demonstrated that SP1−7 attenuates the inflammatory10 and the
nociceptive11 effects exerted by SP. Moreover, it was recently
observed that SP1−7 induces antihyperalgesia in diabetic mice,
suggesting a possible influence on neuropathic pain.12 We find
this observation particularly interesting, since no satisfactory
treatment of neuropathic pain is available today.13 The effects
of SP1−7 are suggested to be mediated through an indirect
Received: February 8, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/jm400209h | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Table 1. Binding Affinity of H-Phe-Phe-NH2 Analogues to the SP1−7 Binding Site
a
Ki value determined at the same occasion as for 2−12. A Ki value of 1.5 nM was achieved in a previous run using other batches of membrane and
labeled ligand, as reported previously.23 Submitted sample contained compound 4 along with the ring closed form (i.e., the diketopiperazine) in a
92:8 relationship. Correspondence to each diastereomers undetermined. 8a (2:1 ratio of 8a and 8b), evaluated in triplicate at six concentrations in
b
c
d
each run.
We previously reported a structure−activity relationship
(SAR) study of the heptapeptide toward the specific binding
site that revealed that the first four N-terminal amino acids in
SP1−7 are not essential for binding, since each could be
substituted with an alanine or removed without significantly
affecting the affinity.21 Interestingly, amidation of the C-
terminal of SP1−7 resulted in a peptide with greater binding
affinity toward the SP1−7 binding site compared to the native
heptapeptide (Ki = 0.3 nM and Ki = 1.6 nM, respectively). In
analogy with the binding studies it was found that the amidated
analogue of SP1−7 was more potent than SP1−7 in reducing
abstinence symptoms in morphine dependent rats and in
inducing antihyperalgesia in diabetic mice.8,22
In parallel with the SAR investigation of SP1−7, similar studies
were performed on the endogenous μ-receptor agonist
endomorphin-2 (EM-2, H-Tyr-Pro-Phe-Phe-NH2).23 The
interest in this tetrapeptide originates from the observations
by Botros et al.17 that endomorphin-1 (EM-1, H-Tyr-Pro-Trp-
Phe-NH2) and EM-2 interact with the SP1−7 binding site.
However, a significant difference in binding affinity between the
two tetrapeptides was observed. EM-1 and EM-2 displayed a
1000-fold and a 10-fold lower affinity, respectively, compared to
the endogenous ligand SP1−7. From an alanine scan and
truncation study of EM-2 the same SAR as for SP1−7 was
observed. The C-terminal was shown to be crucial for binding
affinity to the SP1−7 binding site, and substitution of either of
B
dx.doi.org/10.1021/jm400209h | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a
the two amino acids in the N-terminal with alanine resulted in
only small differences in affinity. Truncation of EM-2 resulted
in identification of the dipeptide H-Phe-Phe-NH2 (1)
possessing a Ki of 1.5 nM.23 Notably, 1 was recently shown
to attenuate the signs of hyperalgesia, allodynia, and
nociception in diabetic mice after intrathecal administration,
in a similar manner compared to peptide SP1−7 and its
amidated analogue.24 Dipeptide (1) is therefore clearly a very
promising starting point for the development of future agents
for the treatment of neuropathic pain.
Given the interesting results obtained from 1, we felt
prompted to further optimize this lead compound into stable
and orally bioavailable SP1−7 mimetics. Herein we report the
synthesis and SAR of a set of constrained H-Phe-Phe-NH2
analogues evaluated in a binding assay displacing [3H]SP1−7. In
addition a plausible binding conformation of the high affinity
ligands was generated and the potential active uptake by the
PEPT1 transporter, intestinal permeability, and metabolic
stability of the dipeptide analogues were investigated.
Table 2. Metabolic Stability Data of Compounds 1−12
b
c
compd
Clint (μL min−1 mg−1
)
t1/2 (min)
12
597 40
1
121 39
2.7 1.5
64 11
4
2
3
22
15
4
1
4
92
0
5
38 15
40 16
6
175
28
5
7.9 0.2
7a
7b
3
9
3
50
6
16
103
6
d
8a
16
88 15
8b
9a
9b
10
11
12
7.6 1.9
187
36
98
32
14
13
5
39
16
44
99
4
7
6
5
3
4
5
1
1
107 12
a
b
Results are expressed as the mean SD. Clint = in vitro intrinsic
c
d
clearance. t1/2 = in vitro half-life. 8a (2:1 ratio of 8a and 8b).
RESULTS
■
Biological Evaluation. The dipeptides 1−12 in Table 1
were prepared using solid-phase peptide synthesis. The
compounds were evaluated in a binding assay using spinal
cord membrane from Sprague−Dawley rats and radioactive
[3H]SP1−7 as tracer.17 The binding affinities of compounds 1−
12 for the SP1−7 binding site are summarized in Table 1.
Methylation of the peptide backbone giving analogues 2, 3, 5,
and 6 resulted in lower affinity in general (Ki = 26−189 nM)
except for compound 4 , which maintained similar affinity as 1
(Ki of 9.4 and 8.4 nM, respectively). Rigidification of the N-
terminal phenylalanine using a pyrrolidine analogue resulted in
reduced binding affinities (7a and 7b, Ki = 33.9−50.5 nM).
However, corresponding rigidification of the C-terminal
phenylalanine gave 8a and 8b, of which one diastereomer
showed improved binding affinity (Ki = 2.2 nM) and one
diastereomer lower binding affinity (Ki = 93 nM). Because of a
difficult separation, the diastereomer 8a could not be
completely isolated from diastereomer 8b, which hence was
evaluated in a 2:1 ratio of 8a and 8b. However, from the
binding affinity difference one can still conclude that
diastereomer 8a has a significantly better binding affinity
compared to 8b. A methyl group on the β-carbon in the C-
terminal side chain as in compounds 9a and 9b reduced the
binding affinity 2−8 times. Elongation of the side chain with
one carbon gave compound 10, which showed similar affinity to
compound 1. Furthermore, small modifications on the aromatic
part of the C-terminal phenylalanine were well tolerated (cf. 1
with 11 and 12).
In the methylated series, analogues 2, 3, and 5 resulted in
more stable compounds with low or moderate risk for high first
pass metabolism and increased t1/2 up to 50 times (cf. 1 with
2). However, internal or C-terminal methylation of the
nitrogen (4 and 6, respectively) did not improve the metabolic
stability. In all the rigidified analogues the metabolic stability
increased and the half-life became 4−16 times longer (cf. 1 with
7a−8b). Furthermore, methylation at the β-carbon (compound
9a and 9b) or elongation of the phenylalanine (see compound
10) also improved the metabolic stability, while decoration of
the aromatic ring of the C-terminal phenylalanine did not
influence the stability (cf. 1 with 11 and 12).
In Vitro Permeability and Uptake Experiments. The
intestinal epithelial permeability, expressed as apparent
permeability coefficients (Papp), was determined from transport
rates across Caco-2 cell monolayers, as described previously.28
The experiments were run in both apical to basolateral (a−b)
and basolateral to apical (b−a) directions (Table 3). In general
a Papp value below 0.2 × 10−6 cm/s indicates low permeability, a
Papp value ranging from 0.2 × 10−6 to 1.6 × 10−6 cm/s
moderate permeability, and a Papp value above 1.6 × 10−6 cm/s
indicates high permeability.29 From Table 3 it is, however, clear
that significant efflux occurs for most compounds, presumably
by P-glycoprotein (PgP). However, despite the cell permeation
in the absorptive a−b direction being obscured by the efflux,
some compounds (2, 3, 4, 5, 7a,b, 8a,b, and 9b) show a
moderate to high a−b rate of transport. This indicates that for
these compounds the cell membrane and hence the cell
monolayer permeability is high to very high.
Some of the compounds showed a very low recovery, i.e.,
poor mass balance, especially in the a−b direction (see Table
3). This is most likely due to high activity of apical brush-
border peptidases. As seen in Table 3, there is a clear
distinction in the mass balance of unstable compounds when
comparing a−b and b−a. In the apical direction, the exposure
to the peptidolytic enzymes is immediate, whereas the exposure
is delayed when applying the compound initially on the
basolateral side. Thus, for the unstable compounds, the
calculations of the Papp values were corrected for the poor
mass balance by mathematically adjusting the apparent donor
concentration over the experimental time (the donor
concentration degradation was simulated by a single
In Vitro Metabolic Stability. The metabolic stability was
determined by incubating the compounds with pooled human
liver microsomes (HLM). In vitro half-life (t1/2) and in vitro
intrinsic clearance (Clint) were calculated using previously
published models, and the results are presented in Table 2.25,26
The Clint is used to rank compounds with regard to metabolic
stability. It can give a good indication of the in vivo hepatic
clearance when the overall clearance mechanism is metabolic
and when oxidative metabolism dominates (i.e., Clmetabolic
≫
Clrenal + Clbiliary + Clother). The following cutoff values were used
to classify the compounds with regard to metabolic stability:
Clint < 47 μL min−1 mg−1 indicates a low risk for high first pass
metabolism in vivo, 47 < Clint < 92 a moderate risk, and Clint
>
92 a high risk.27
C
dx.doi.org/10.1021/jm400209h | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a
Table 3. Uptake and Permeability Data of Compounds 1−12
b
uptake
Caco-2 permeability
c
(pmol/mg protein)/min
Papp (10−6 cm/s)
d
e
compd
CHO-PEPT1
CHO-K1
ratio PEPT1/K1
a−b
b−a
mass balance (%) (a−b)/(b−a)
ratio (b−a)/(a−b)
f
1
0.3 0.0
2.2 0.7
1.1 0.2
18.8 0.3
32.7 1.8
0.3 0.1
30.9 4.1
14.9 2.2
13.1 2.3
11.4 1.0
0.6 0.1
24.8 3.8
0.6 0.00
0.8 0.1
0.2 0.0
0.3 0.0
2.4 0.5
1.5 0.2
23.0 2.5
32.1 8.6
0.4 0.1
22.3 2.2
10.9 1.2
11.2 1.30
8.7 0.8
0.7 0.0
22.2 3.1
0.9 0.15
1.3 0.3
0.3 0.0
1.0
0.9
0.7
0.8
1.0
0.7
1.4
1.4
1.2
1.3
0.9
1.1
0.7
0.6
0.7
0.12 0.01
13.9 0.5
18.4 0.1
9.0 2.1
0.42 0.03
124 21
86.5 0.2
124 19
3:33
3.6
8.9
4.7
14
2
3
4
5
77:85
76:82
90:86
93:121
0.5:56
84:95
90:95
81:109
95:110
34:76
93:123
13:74
1:56
20.4 2.0
2.2 1.2
224
5
11
f
6
1.3 0.3
26.3 1.0
76.2 1.1
75.1 2.1
51.4 1.8
5.8 0.6
0.6
53
7a
7b
0.54 0.02
0.76 0.06
3.6 0.4
95
g
8a
21
8b
0.58 0.03
1.2 0.1
86
f
9a
5.0
39
9b
4.4 0.1
171
6
f
10
0.19 0.03
0.58 0.06
1.5 0.2
1.1 0.3
0.65 0.10
1.2 0.2
5.8
1.1
f
11
f
12
1:54
0.8
a
b
Results are expressed as the mean SD. All compounds were analyzed at 100 μM. See Experimental Section for experimental conditions. The
PEPT1 substrate [14C]GlySar was used as positive control. The use of the radioactive substrate allowed comparison with historical data and
consistency of performance over time of PEPT1 activity in this cell line. The ratio PEPT1/K1 for [14C]GlySar was equal to or larger than 5 under
c
d
e
f
the applied conditions. Papp = apparent permeability coefficient. a−b = apical to basolateral. b−a = basolateral to apical. Papp corrected for poor
mass balance using a single exponential decay function to account for compound degradation over the cause of the experiment. 8a (2:1 ratio of 8a
g
and 8b).
exponential decay function using the initial (c0) concentration
and the final (cfinal) concentration).
having a high uptake into CHO cells also have a higher
permeability in Caco-2 cells.
The permeability data in the a−b direction showed that the
methylated analogues 2−6 had high permeability (ranging from
2.2 × 10−6 to 20 × 10−6 cm/s). Thus, backbone methylation
significantly increased the cell permeability in comparison to
the reference compound H-Phe-Phe-NH2 (1, 0.12 × 10−6 cm/
s). The cyclized analogues 7a, 7b, and 8b together with
compound 9a were classified as having moderate permeability
(ranging from 0.5 × 10−6 to 1.2 × 10−6 cm/s) and the
analogues 8a and 9b to have high permeability (3.6 × 10−6 and
4.4 × 10−6 cm/s, respectively). The C-terminal modified
analogue with a one-carbon elongation in the side chain,
compound 10 (Papp = 0.19 × 10−6 cm/s), showed similar low
permeability as seen for compound 1. On the other hand, the
methylated and the fluorinated C-terminal modified H-Phe-
Phe-NH2 analogues (11 and 12) were accompanied by a slight
improvement in permeability compared to compound 1, with
moderate Papp values of 0.58 × 10−6 and 1.5 × 10−6 cm/s,
respectively. Moreover, compounds 11 and 12, together with
compound 6, were the only analogues that did not show
significant efflux. The rest of the compounds displayed a 5- to
95-fold higher transport rate in the b−a direction than in the
a−b direction (Table 3).
Uptake studies with CHO-K1 and CHO-PEPT1 cells
(control and PEPT1 stably transfected cells, respectively)
were performed, and the results are expressed as (pmol/mg of
protein)/min (Table 3). If the compounds are substrates for
the peptide transporter PEPT1 and actively transported into
the cell, the PEPT1/K1 ratio should be greater than 1. The
PEPT1/K1 ratios for the synthesized compounds ranged from
0.6 (11) to 1.4 (7a and 7b), which indicates that these
compounds are not actively transported. The rate of compound
uptake into CHO cells is not directly comparable to compound
permeability across the Caco-2 cell monolayers. However, a
qualitative comparison is possible. Thus, in general, compounds
Computational Analysis. The pharmacophore hypothesis
generation program Phase30 was used to investigate if the low-
energy conformations of the high affinity ligands could position
important structural features in common regions of space not
accessible to the ligands with lower affinity. The dipeptides with
Ki values of 10 nM or lower and known stereochemistry were
defined as strong binders (1, 4, 10, and 12). No compounds in
the series were devoid of affinity, but the three dipeptides with
lowest affinity (Ki = 70−190 nM) and known stereochemistry
were defined as weak binders (2, 3, and 6). With the
requirement that all four strong binders should match a
derived pharmacophore hypothesis the maximum number of
pharmacophore features was found to be 6 in order to obtain
pharmacophore hypotheses. Further evaluation of these
hypotheses identified two sets of pharmacophore features.
One set resulting in 242 hypotheses consisted of two hydrogen
bond acceptors, one donor, one positively charged group, and
two aromatic rings. The other set was very similar with the
difference that two hydrogen bond donors were included, of
which one replaced the positively charged group. This set of
pharmacophore features resulted in 2145 hypotheses. The
hypotheses were scored by matching the strong binders to the
pharmacophore hypotheses. With the default filter and
complete clustering 39 hypotheses with scores of 3.42−3.65
remained after scoring. A final score for each hypothesis was
derived by subtracting the matching scores of the weak binders
from the score obtained when matching the strong binders,
resulting in scores of 0.77−1.57. The 10 pharmacophore
hypotheses with the highest final scores were visually inspected,
and the compounds excluded from the set of strong binders
were aligned to the hypotheses, with the requirement that they
match all six pharmacophore features. The top ranked
pharmacophore hypothesis represented a common binding
pose for the strong binders that the weak binders had
D
dx.doi.org/10.1021/jm400209h | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
difficulties attaining. Binding poses from alignment to the
remaining pharmacophore hypotheses either could not explain
the affinity of the rigidified analogues (e.g., neither 8a nor 8b
could match the hypothesis, although one of them has the
highest affinity in the series) or resulted in high fitness scores
for some of the weak binders. The top ranked pharmacophore
hypothesis is shown together with 1 in Figure 1, and the fitness
rationalized. Both 2 and 6 could attain a binding pose that
matched all but one pharmacophore feature in the
pharmacophore hypothesis, resulting in a fairly low penalty to
their scores. The third weak binder 3 had a significantly lower
score than the strong binders and was considered to be
correctly classified. A quantitative model validation was not
performed, since all defined strong binders and weak binders
were used to derive the binding pose model and the Ki values of
the diastereomers 7−9 were not assigned within the pairs.
However, on the basis of the qualitative analysis, we considered
the model to be suitable for the SAR analysis and to support
the assignment of Ki values within the pairs of diastereomers.
Furthermore, since the model provides a rational binding mode
for compounds with high affinity to the SP1−7 binding site, the
model may also be used as a template in future virtual screening
campaigns to search for novel ligands to this target.
DISCUSSION
■
Figure 1. Binding pose of 1 derived from alignment to the top score
pharmacophore hypothesis. The pharmacophore hypothesis features
are orange torus = aromatic feature, blue sphere = hydrogen bond
donor feature, and red sphere = hydrogen bond acceptor feature.
A potential specific receptor for the N-terminal partial
sequences of SP in mouse spinal cord was already discussed
in the beginning of the 1980s,31 but it was not until 1990, when
it was found that the binding was specific, saturable, and
reversible, that the hypothesis of a specific SP1−7 receptor was
proposed.16 During the 1990s the biological roles of the N-
terminal fragment of SP were investigated by several
groups.7,14,32,33 More recently the SARs of SP1−7 and other
ligands (e.g., EM-2) have been investigated by our group,
leading to the discovery of dipeptides with high affinity to the
SP1−7 binding site and with interesting pharmacological
properties such as antinociceptive effects in animal models of
neuropathic pain.21,23,24
scores for the compounds attaining this pharmacophore
hypothesis are presented in Table 4. In general the defined
Table 4. Fitness Score for the Alignment to the Top
Pharmacophore Hypothesis for Compounds 1−12 and Their
Set Membership in the Analysis
a
b
compd
fitness
set
1
3
SB
2
2.19
1.7
WB
WB
SB
In the present study we are addressing the challenging task of
transforming these dipeptides into compounds with more
druglike features. One way to achieve this is to introduce local
constraints in the peptide backbone by incorporation of N-
methylamino acids.34 The advantage of incorporating a single
methyl group is that it provides a balance between decreasing
conformational entropy sufficient to change the binding affinity
while still allowing sufficient conformational flexibility for the
ligand to adopt the conformation required for molecular
recognition to the target. Furthermore, the incorporation of a
methyl group often provides metabolic stability because of
resistance to proteolytic degradation.35,36
In the dipeptide series presented herein we introduced a
methyl group in each position along the backbone of the
dipeptide H-Phe-Phe-NH2. Highest influence on binding
affinity is seen when substituting the N-terminal amine or the
C-terminal amide with a methyl group (2 and 6, respectively),
resulting in reduced binding affinities by 17−23 times. A
possible explanation for this can be seen when looking at the
proposed binding pose for the series (Figure 2). The binding
3
4
2.58
2.27
2.07
1.38
0.38
2.65
1.13
2.7
5
NA
WB
NA
NA
NA
NA
NA
NA
SB
6
7a
7b
8a
8b
9a
9b
10
11
12
0.69
2.14
2.92
2.96
NA
SB
a
b
The sum of the vector, site, and volume scores (range 0−3). SB =
strong binder, compounds used for generating pharmacophore
hypotheses (correspond to actives in phase). WB = weak binders,
compounds used for filtering out hypotheses that match the lowest
affinity ligands (correspond to inactives in phase). NA = not assigned,
compounds not used in the phase analysis.
strong binders in the series have the highest fitness scores. Only
10 (Ki = 6.2 nM) was not among the very top scored
compounds. Compounds 5 and 11 were not included in the
sets used to derive the common binding pose, since they are
outside the limits for a set membership in this investigation.
However, especially 11 may be considered as a strong binder
with Ki = 11.5 nM and was also shown to be able to attain the
binding pose with a fitness score of 2.92, as good as the strong
binder set in the series. Although the top scoring
pharmacophore hypothesis also suffered from a rather high
fitness score for the weak binders 2 and 6, their scores could be
Figure 2. Binding poses of 2 and 6 when aligned to the top ranked
pharmacophore hypothesis.
E
dx.doi.org/10.1021/jm400209h | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
pose presented includes the spatial arrangement of six
important interactions features that all high affinity compounds
can present to the target protein, namely, two hydrogen bond
acceptors, two hydrogen bond donors, and two aromatic rings.
In 2 the methyl group on the N-terminus is positioned in the
direction of the hydrogen bond donor vector from the amine to
the protein in the pharmacophore hypothesis, suggesting a
nonoptimal interaction in this part of the molecule with the
target protein.
A β-methylation of the phenylalanine side chain can have an
effect on how the side chain interacts with the binding site.37
When a methyl group was introduced on the β-carbon in the C-
terminal phenylalanine in 1, a 4-fold difference between the
(S,S)- and the (R,R)-configuration could be observed (cf. 9a
with 9b). Inspecting the fitness scores of the two diastereomers
when matched to the pharmacophore supports 9a with the
(S,S)-configuration of the β-methylphenylalanine as the
analogue with highest binding affinity. Furthermore, a β-
methylation reduces the conformational flexibility of the
phenylalanine side chain by imposing a steric rotational
constraint, which can be advantageous by reducing entropic
penalty upon binding. However, this was not advantageous for
the binding, since 1 possessed 2 times higher binding affinity
compared to 9a.
Elongation of the C-terminal phenylalanine side chain was
well tolerated (cf. 1 with 10), which indicates that there is room
for modification in this part of the molecule. Interestingly, this
is in contrast to the previous SAR investigation that resulted in
identification of H-Phe-Phe-NH2. Then the replacement of the
C-terminal phenylalanine for a phenylglycine resulted in a
completely inactive compound and even replacing the phenyl
ring with thiophene resulted in significantly lower affinity.23
Moreover, results herein suggest that the binding pocket seems
to tolerate substituents on the aromatic moiety, since both
compound 11 and compound 12 showed excellent binding
affinities, especially compound 12 (Ki = 3.3 nM) in which a
hydrogen in the meta position has been exchanged by a fluorine
atom. This is somewhat contradictory to the result of the
previous SAR investigation where the same C-terminal amino
acids as in 11 and 12 in combination with O-methylated tyrosin
and cyclohexylalanine as the N-terminal amino acids,
respectively, were devoid of affinity to the SP1−7 binding site.
However, in the latter case the main problem might be that the
change in both terminal parts was done concurrently, which
gave a negative synergistic effect upon binding. Altogether, it
seems that minor C-terminal changes are allowed if the N-
terminal is preserved.
The in vitro pharmacokinetic properties of the compounds
were also studied, i.e., active uptake via the dipeptide
transporter PEPT1 in CHO cells, permeability and efflux in
Caco-2 cells, and metabolic stability in human liver microsomes
(Tables 2 and 3). The metabolic stability study was
concentrated to the liver, since this organ constitutes the
major site of metabolism for most drugs.25 By use of the in vitro
parameter intrinsic clearance (Clint), an estimate of the rate of
liver metabolism (typically oxidative) of the compounds under
in vivo conditions is obtained, which gives an indication of
which compounds to select for further preclinical develop-
ment.26 Methyl groups incorporated into the peptide backbone
can increase the metabolic stability. Indeed, the stability
increased for all the methylated analogues of H-Phe-Phe-NH2
except for the internal (4) and the C-terminal (6) N-
methylated analogues. However, the apparent instability of
the internal N-methylated compound is partly explained by the
fact that the opened analogue to some extent might ring-close
into the corresponding diketopiperazine form as mentioned
earlier. To our satisfaction, the rigidified analogues of H-Phe-
Phe-NH2 resulted not only in metabolically more stable
compounds (7a−8b) but also in the most potent binder 8a.
In fact, 8a turned out to be 7 times more stable than H-Phe-
Phe-NH2, which makes this analogue a good candidate for
replacing 1 in further animal studies. Modifications of the C-
The hydrogen bond donor feature assigned to the amine
would also be compatible with a positively charged feature in
the series, and this was also found in other pharmacophore
hypotheses. However, if that was the case, the directional
property of the feature would be lost and compound 2 would
have a higher fitness to the pharmacophore hypothesis. The
reduced affinity observed for 6, in which the primary amide was
methylated to give a secondary amide, is also accounted for in
the pharmacophore hypothesis because the hydrogen bond
pattern for the C-terminal amide is optimal for a primary
amide, which is present in all high affinity compounds in the
series. This can be seen in Figure 2 where 6 is aligned to the
pharmacophore, unable to match this hydrogen bond donor
feature. When 1 was methylated on the internal nitrogen, no
influence on the affinity potency could be observed (see
compound 4). The internal methylated dipeptide showed a
tendency to ring-close to the corresponding diketopiperazine to
some extent. Although the initial concentration of diketopiper-
azine was only ∼10% in the test sample, further formation in
the assay cannot be excluded, making the results hard to
interpret in this case. However, internal methylation of the
dipeptide, either in the opened or in the diketopiperazine form,
seems to be allowed in terms of binding affinity.
To investigate the spatial arrangement of the phenylalanine
side chains and the effect of more strongly rigidified H-Phe-
Phe-NH2 analogues, the compounds with side chain constraints
7 and 8 were synthesized. This modification was shown to
display specific steric requirements as might be expected in a
defined receptor pocket. When the cis-3-phenylpyrrolidine
moiety was introduced at the N-terminal (7a and 7b), the
affinity dropped 4−6 times. Both these diastereomers show low
fitness scores to the proposed pharmacophore, indicating that
the compounds cannot adopt an optimal binding conformation.
Interestingly, however, when the cis-3-phenylpyrrolidine moiety
was incorporated in the C-terminal, the phenyl group seems to
be well positioned in the active binding site in one of the
diastereomers, resulting in almost 4 times higher binding
affinity. From a comparison of 8a and 8b (Figure 3), a clear
difference in fitness to the pharmacophore hypothesis is seen,
supporting the (S,S)-configuration of the cis-3-phenylpyrroli-
dine in the high-affinity analogue 8a.
Figure 3. Binding poses of 8a and 8b when aligned to the
pharmacophore hypothesis.
F
dx.doi.org/10.1021/jm400209h | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
terminal phenylalanine in the form of elongation as in
compound 10 or β-methylation also improved the stability.
However, o-methyl and m-fluoro substitutions on the aromatic
part of the phenylalanine do not seem to have any effect on the
metabolism compared to H-Phe-Phe-NH2 (1) itself.
Furthermore, the intestinal membrane permeability and the
transport mechanism of the analogues in this series were
investigated. The dipeptide/tripeptide transporter PEPT1 is
expressed in the intestine and ensures efficient absorption of
small peptides from digestion of dietary proteins. PEPT1 is also
important for absorption of a variety of peptidomimetic drugs,
such as β-lactam antibiotics and prodrugs of nucleoside
analogues, such as valaciclovir.38,39 In order to achieve efficient
binding and activation of the transporter, some important
structural features of dipeptides need to be fulfilled: (1) no
substitution at the N-terminal amine, (2) a negatively charged
C-terminus, (3) an amide carbonyl oxygen, and (4) the
interaction of the N-and C-terminal amino acids side chain with
a proposed hydrophobic pocket, preferably aromatic
groups.40−42 Thus, an efficient transport with PEPT1 was not
expected for the synthesized dipeptides described in this paper
which all contain primary amides in the C-terminal. However,
the aromaticity, the dipeptide character, and the fact that some
dipeptides with C-terminal amides have been shown to be
PEPT1 substrates motivated us to investigate if the dipeptides
in this series could be actively transported over the membrane.
In general, the uptake studies using CHO cells transfected with
the transporter PEPT1 indicated that the compounds are not
good substrates for this transport system (see PEPT1/K1 ratio
in Table 3, all around 1). A very weak increase of PEPT1
activation could be seen for the dipeptides incorporating the cis-
3-phenylpyrrolidine moiety (7a−8b), and this might be
explained by the study of Vig et al.43 where they demonstrated
that dipeptides with a proline at the C- or N-terminal with the
proper attendant amino acid exhibits both high affinity and
activation of the PEPT1 transporter.
Although the compounds were not actively transported, the
permeability data in the a−b direction generally indicated that
small modifications of a peptide can have a significant effect on
cell permeability in Caco-2 cells, in particular backbone
modifications (see Table 3). Thus, the lead compound H-
Phe-Phe-NH2 (1) has poor cell-permeability of 0.12 × 10−6
cm/s, whereas the backbone methylated compounds 2−6, a
proline-based analogue (8a), and a β-methylated compound
(9b) are highly permeable compounds (Papp values of (2.2−18)
× 10−6 cm/s). However, one should be aware that some of the
peptide compounds also were degraded over the Caco-2 cell
membrane, as reflected in poor mass balances. This was the
case for the nonbackbone modified peptides 1, 6, 9a, and 10−
12. The degradation is most probably a result of protease
activity in the cell membrane in the apical region. It is well-
known that brush border proteases are expressed in Caco-2-
cells.44,45 The set of sensitive peptides, all with a normal
peptide backbone, are well in line with the structural preference
for proteases. A majority of the peptides with a low recovery
were also accompanied by a high clearance in the human liver
microsomes (Table 2), indicating also a contribution of
protease mediated degradation here besides the normally
dominating oxidative metabolism. To investigate this, we
performed a separate experiment with HLM supplemented
with or without NADPH (cofactor for CYP450) (Figure 4).
From Figure 4 it is clear that both oxidative and proteolytic
activity is important for the more unstable compound (6, (−)
Figure 4. Metabolic stability of compounds 3 (red) and 6 (blue) in
human liver microsomes with (+) and without (−) the CYP450
cofactor NADPH.
t
1/2 = 28 min, (+) t1/2 = 10 min), and for the stable compound
only oxidative metabolism contributes to the degradation (3,
(−) t1/2 = no metabolism detected, (+) t1/2 = 60 min).
Satisfyingly, compound 8a with highest binding affinity (Ki =
2.2 nM) seems to possess both high stability (Clint = 16 μL
min−1 mg−1 and a high stability in Caco-2-cells) and high
permeability (3.6 × 10−6 cm/s).
The efflux transporter PgP has an important role in limiting
entry of various drugs into the central nervous system, which
makes it an important factor to be considered when working
with CNS active compounds.46 PgP is also expressed in the
human intestine, but its effect on limiting absorption of average
dosed drugs is less pronounced than in the CNS. In drug
discovery, an efflux ratio of 2 or larger is considered
significant.47 Several of the analogues studied in here displayed
a large efflux (the ba/ab ratio ranging from 0.6 to 95), which
might be explained by interactions with PgP. Nevertheless, in
the Caco-2 cells other efflux transporters are also present and
could be involved in the secretion of the compounds. To be
noticed though is that the efflux was substantially lower for the
methylated analogues (2−6) compared to the rigidified ones
(7a−8b). Also, the stereochemistry seemed to influence the
efflux, since there were big differences in the ba/ab ratios
between all the diastereomers (cf. 7a and 7b, 8a and 8b, 9a and
9b).
CONCLUSION
■
Constrained and modified dipeptides analogues of H-Phe-Phe-
NH2 have been synthesized and evaluated regarding their
binding affinity, uptake, permeability, and metabolic stability. In
general, local constraints such as methylation of the peptide
backbone or incorporation of pyrrolidines into the peptide
backbone delivered compounds with increased metabolic
stability and improved intestinal permeability. Unfortunately,
the propensity for efflux was not reduced. C-Terminal
rigidification of H-Phe-Phe-NH2 using a cis-3-phenylpyrrolidine
moiety with (S,S)-configuration (compound 8a) was the only
modification resulting in retained affinity to the SP1−7 binding
site, besides improved stability and cell permeability. Moreover,
in this study we have shown that it is possible to make changes
in the C-terminal, e.g., substitution with o-methyl or m-fluoro in
the aromatic part or elongation of the phenylalanine side chain
without any loss in binding affinity. The results herein will be of
significant value for further design of bioavailable research tools
useful in forthcoming animal studies to clarify the physiological
role of SP1−7 in pain of neuropathic origin.
G
dx.doi.org/10.1021/jm400209h | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(dd, J = 5.0, 10.4 Hz, 1H), 7.18−7.36 (m, 10H). 13C NMR (CD3OD)
δ 32.1, 37.9, 39.3, 55.5, 64.1, 127.9, 129.0, 129.5, 130.2, 130.5, 130.51,
135.0, 138.4, 167.7, 174.8. HRMS (M + H+): 326.1870, C19H23N3O2
requires 326.1869.
EXPERIMENTAL SECTION
■
General Methods. Preparative RP-HPLC was performed on a
system equipped with a Zorbax SB-C8 column (150 mm × 21.2 mm)
or a 10 μm Vydac C18 column (250 mm × 22 mm), in both cases
with UV detection at 230 nm. Analytical RP-HPLC−MS was
performed on a Gilson-Finnigan ThermoQuest AQA system (Onyx
monolithic C18 column, 50 mm × 4.6 mm; MeCN/H2O gradient
with 0.05% HCOOH) in ESI mode, using UV (214 and 254 nm) and
MS detection. The purity of each of the peptides was determined by
RP-HPLC using the columns ACE 5 C18 (50 mm × 4.6 mm) and
ACE 5 phenyl (50 mm × 4.6 mm) or Thermo Hypersil Fluophase RP
(50 mm × 4.6 mm) with a H2O/MeCN gradient with 0.1% TFA and
UV detection at 220 nm. All peptides showed purity above 95%.
NMR spectra were recorded on a Varian Mercury plus spectrometer
Peptide 3 (α-Me-Phe-Phe-NH2). The analogue was prepared
according to the general procedure using Rink amide MBHA resin
(157 mg, 100 μmol), Fmoc-Phe-OH (155 mg, 400 μmol), Fmoc-α-
Me-Phe-OH (161 mg, 400 μmol), HATU (152 mg, 400 μmol), and
DIEA (139 μL, 800 μmol). Reaction time was 20 h. The crude peptide
was purified by RP-HPLC to give 3 as a white solid in 3.6 mg (11.2%
yield). HPLC purity: C18 column, >98%; Fluophase, >99%. 1H NMR
(CD3OD) δ 1.38 (s, 3H), 2.98 (dd, J = 10.1, 13.9 Hz, 1H), 3.30 (d, J =
14.4 Hz, 1 H), 3.23 (dd, J = 5.2, 13.9 Hz, 1H), 3.31 (d, J = 14.4 Hz,
1H), 4.78 (dd, J = 5.2, 10.1 Hz, 1H), 7.2−7.36 (m, 10H). 13C NMR
(CD3OD) δ 22.8, 39.2, 43.7, 55.7, 61.8, 127.9, 129.1, 129.5, 130.0,
130.4, 131.4, 134.2, 138.4, 171.5, 175.5. HRMS (M + H+): 326.1866,
C19H23N3O2 requires 326.1869.
(1H at 399.8 MHz and 13C at 100.5 MHz or 1H at 399.9 MHz and 13
C
at 100.6 MHz) at ambient temperature. Chemical shifts (δ) are
reported in ppm referenced indirectly to TMS via the solvent residual
signal. Exact molecular masses were determined on a Micromass Q-
Tof2 mass spectrometer equipped with an electrospray ion source at
the Department of Pharmaceutical Biosciences, Uppsala University,
Sweden. All other chemicals and solvents were of analytical grade from
commercial sources.
General Synthesis of Peptides 1−12. Coupling. The peptides
1−12 were synthesized manually from Rink amide MBHA resin (0.66
mmol/g) or methylindole AM resin (0.63 mmol/g) in 2 mL
disposable syringes fitted with porous polyethylene filter. Standard
Fmoc conditions were used. The Fmoc protecting group was removed
by treatment with 20% piperidine in DMF (2 × 1.5 mL, 2 + 10 min),
and the polymer was washed with DMF (6 × 1.5 mL, 6 × 1 min).
Coupling of the appropriate amino acid Fmoc-AA-OH (1.5 or 4 equiv)
was performed in DMF (1.5 mL) using N-[(1H-benzotriazole-1-yl)-
(dimethylamino)methylene]-N-methylmethanaminium hexafluoro-
phosphate N-oxide (HBTU, 1.5 or 4 equiv) in the presence of
DIEA (3 or 8 equiv). The resin was washed with DMF (5 × 1.5 mL, 5
× 1 min) and subsequently deprotected and washed as described
above. After completion of the coupling cycle the resin was also
washed with several portions of DMF, CH2Cl2, and MeOH before it
was dried in a vacuum overnight.
Cleavage. The final peptide was cleaved from the resin by treatment
with triethylsilane (100 μL) and 95% aqueous trifluoroacetic acid
(TFA, 1.5 mL) followed by agitation for 2 h at room temperature. The
resin was filtered off and washed with TFA (2 × (0.3−0.5) mL). The
filtrate was collected in a centrifuge tube and concentrated in a stream
of nitrogen. Cold diethyl ether (≤7 mL) was used to precipitate the
product, which was collected by centrifugation, washed with cold
diethyl ether (≤3 × 7 mL), and dried.
Purification. The crude peptide was dissolved in MeCN/0.1%
aqueous TFA, filtered through a 0.45 μm nylon membrane, and
purified in one to two runs by RP-HPLC. Selected fractions were
analyzed by RP-HPLC and RP-HPLC−MS, and those containing pure
product were pooled and lyophilized. The reported yields are based on
the loading of the starting resin.
Peptide 4 (Phe-N-Me-Phe-NH2). The analogue was prepared
according to the general procedure using Rink amide MBHA resin
(157 mg, 100 μmol), Fmoc-N-Me-Phe-OH (161 mg, 400 μmol),
Fmoc-Phe-OH (155 mg, 400 μmol), HATU (152 mg, 400 μmol), and
DIEA (139 μL, 800 μmol). Reaction time was 22 h. The crude peptide
was purified with RP-HPLC and the pure fractions containing the
dipeptide were collected, pooled, and lyophilized to give 4 as a white
solid in 9.8 mg (30.1% yield). HPLC purity: C18 column, 94.6%;
Fluophase, >92.1%. Subsequent NMR analysis of the product showed
minor impurities. Thus, an analytical sample from the NMR solution
was again analyzed by RP-HPLC and showed the product as a major
peak along with two extra minor peaks. Repurification of the product
was done using RP-HPLC. The major peak (corresponding to the
dipeptide product) and the bigger of the two minor peaks were
isolated and characterized by NMR and MS. The latter compound
turned out to be the corresponding diketopiperazine (DKP).
However, once again a minor portion of the pure isolated dipeptide
converted into the DKP in the NMR solution. Thus, the formation of
Phe-N-Me-Phe-NH2 into some amount of the DKP seems to be
1
unavoidable. H NMR for 4 (CD3OD) δ 2.82 (s, 3H), 2.95−3.15 (m,
4H), 4.51 (dd, J = 6.5, 7.6 Hz, 1H), 5.21 (dd, J = 7.4, 8.6 Hz, 1H),
7.13−7.39 (m, 10H). 13C NMR (CD3OD) δ 32.69, 35.51, 37.98,
53.31, 60.31, 127.89, 129.03, 129.66, 130.09, 130.25, 130.59, 135.18,
138.41, 170.28, 173.66. HRMS (M + H+): 326.1866, C19H23N3O2
requires 326.1869.
1H NMR for the corresponding diketopiperazine of 4 (CD3OD) δ
1.41 (dd, J = 9.1, 13.5 Hz, 1H), 2.65 (dd, J = 3.8, 13.5 Hz, 1H), 2.78
(dd, J = 4.7, 14.2 Hz, 1H), 2.98 (s, 3H), 3.05 (dd, J = 4.7, 14.2 Hz,
1H), 3.98 (dd, J = 3.8, 9.1 Hz, 1H), 4.24 (t, J = 4.7 Hz, 1H), 6.95−7.03
(m, 2H), 7.15−7.41 (m, 8H). LC−MS: (M + H+) 309; (2M + H+)
617.
Peptide 5 (Phe-α-Me-Phe-NH2). The analogue was prepared
according to the general procedure using Rink amide MBHA resin
(157 mg, 100 μmol), Fmoc-Phe-OH (155 mg, 400 μmol), Fmoc-α-
Me-Phe-OH (161 mg, 400 μmol), HATU (152 mg, 400 μmol), and
DIEA (139 μL, 800 μmol). Reaction time was 20 h. The crude peptide
was purified by RP-HPLC to give 5 as a white solid in 17.3 mg (53.2%
yield). HPLC purity: C18 column, >99%; Fluophase, >99%. 1H NMR
(CD3OD) δ 1.46 (s, 3H), 2.97 (dd, J = 9.5, 14.2 Hz, 1H), 3.14 (dd, J =
5.6, 14.2 Hz, 1H), 3.25 (d, J = 13.4 Hz, 1H), 3.37 (d, J = 13.4 Hz, 1H),
4.10 (dd, J = 5.6, 9.5 Hz, 1H), 7.1−7.4 (m, 10H). 13C NMR (CD3OD)
δ 23.9, 38.6, 41.5, 55.9, 61.8, 127.99, 128.9, 129.2, 130.2, 130.4, 131.7,
135.9, 137.5, 168.9, 178.2. HRMS (M + H+): 326.1866, C19H23N3O2
requires 326.1869.
Peptide 1 (H-Phe-Phe-NH2). The analogue was prepared
according to the general procedure using Rink amide MBHA resin
(313 mg, 200 μmol), Fmoc-Phe-OH (310 mg, 800 μmol), HBTU
(303 mg, 800 μmol), and DIEA (278 μL, 1600 μmol). Reaction time
was 18 h. The crude peptide was purified by RP-HPLC to give 1 as a
white solid in 23 mg (37% yield). HPLC purity: C18 column, >99%;
Fluophase, >98%. Spectral data were in agreement with data from
previous reports.23
Peptide 2 (N-Me-Phe-Phe-NH2). The analogue was prepared
according to the general procedure using Rink amide MBHA resin
(157 mg, 100 μmol), Fmoc-Phe-OH (155 mg, 400 μmol), Fmoc-N-
Me-Phe-OH (161 mg, 400 μmol), HATU (152 mg, 400 μmol), and
DIEA (139 μL, 800 μmol). Reaction time was 18 h. The crude peptide
was purified by RP-HPLC to give 2 as a white solid in 14.8 mg (45%
yield). HPLC purity: C18 column, >99%; Fluophase, >98%. 1H NMR
(CD3OD) δ 2.11 (s 3H), 2.79 (dd, J = 10.4, 13.9 Hz, 1H), 3.05−3.15
(m, 2H), 3.21 (dd, J = 5.0, 13.9 Hz, 1H), 3.86−3.93 (m, 1H), 4.77
Peptide 6 (Phe-Phe-NHMe). The analogue was prepared
according to the general procedure using indole AM resin (0.63
mmol/g) (159 mg, 100 μmol), Fmoc-Phe-OH (155 mg, 400 μmol),
HATU (152 mg, 400 μmol), and DIEA (139 μL, 800 μmol). Reaction
time was 22 h. The crude peptide was purified by RP-HPLC to give 6
as a white solid in 25.7 mg (79.1% yield). HPLC purity: C18 column,
>99%; Fluophase, >99%. H NMR (CD3OD) δ 2.65 (bs, 3H), 2.95
(dd, J = 7.6, 13.7 Hz, 1H), 3.03 (dd, J = 8.2, 14.2 Hz, 1H), 3.07 (dd, J
= 7.4, 13.7 Hz, 1H), 3.23 (dd, J = 5.7, 14.2 Hz, 1H), 4.09 (dd, J = 5.7,
1
H
dx.doi.org/10.1021/jm400209h | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
8.2 Hz, 1H), 4.55 (dd, J = 7.4, 7.6 Hz, 1H), 7.16−7.38 (m, 10H). 13
C
separated by RP-HPLC to give the pure diastereomers 9a and 9b as
white solids in 4.7 mg (29% yield) and 4.5 mg (28% yield),
respectively.
NMR (CD3OD) δ 26.2, 38.5, 39.2, 55.5, 56.6, 127.9, 128.8, 129.5,
130.1, 130.2, 130.5, 135.5, 138.1, 169.4, 173.1. HRMS (M + H+):
326.1870, C19H23N3O2 requires 326.1869.
Peptide 9a. HPLC purity: C18 column, >98.4%; Fluophase,
1
Peptides 7a and 7b (cis-3-Phenylpyrrolidine-Phe-NH2). The
analogues were prepared according to the general procedure using
Rink amide MBHA resin (157 mg, 100 μmol), Fmoc-Phe-OH (155
mg, 400 μmol), racemic Fmoc-cis-3-phenylpyrrolidine-2-carboxylic
acid (62 mg, 150 μmol), HBTU (152 mg, 400 μmol or 57 mg, 150
μmol), and DIEA (278 μL, 1600 μmol or 52 μL, 300 μmol). Reaction
time was 24 h. The crude peptide mixture was purified, and the
diastereomers were separated by RP-HPLC to give the pure
diastereomers 7a and 7b as white solids in 5.4 mg (32% yield) and
9.3 mg (55%), respectively.
Peptide 7a. HPLC purity: C18 column, >99%; Fluophase, >99%.
1H NMR (CD3OD) δ 2.35−2.50 (m, 2H), 2.77 (dd, J = 7.5, 13.7 Hz,
1H), 3.00 (dd, J = 6.5, 13.7 Hz, 1H), 3.38 (m, J = 7.4, 9.7, 11.4 Hz,
1H), 3.73 (m, J = 3.6, 7.8, 11.4 Hz, 1H), 3.88 (ddm, J = 7.4, 9.3 Hz,
1H), 4.33 (dd, J = 6.5, 7.5 Hz, 1H), 4.48 (d, J = 9.3 Hz, 1H), 7.15−
7.40 (m, 10H). 13C NMR (CD3OD) δ 31.63, 39.41, 46.62, 48.23,
55.53, 64.82, 127.82, 129.21, 129.41 (2C), 129.93, 130.34, 137.60,
137.99, 167.12, 173.84. HRMS (M + H+): 338.1875, C20H23N3O2
requires 338.1869.
Peptide 7b. HPLC purity: C18 column, >99%; Fluophase, >99%.
1H NMR (CD3OD) δ 2.20 (dd, J = 5.9, 13.6 Hz, 1H), 2.41−2.49 (m,
2H), 2.50 (dd, J = 8.6 13.6 Hz 1H), 3.42 (ddm, J = 8.7, 11.2 Hz, 1H),
3.75 (dd, J = 5.4, 11.2 Hz 1H), 3.89 (dm, J = 8.8 Hz, 1H), 4.11 (dd, J =
5.9, 8.6 Hz, 1H), 4.55 (d, J = 9.3 Hz, 1H), 6.94−6.99 (m, 2H), 7.16−
7.36 (m, 8H). 13C NMR (CD3OD) δ 31.50, 38.58, 46.76, 48.38, 56.02,
64.65, 127.88, 129.19, 129.51, 129.64, 129.84, 130.18, 137.57, 137.70,
167.44, 178.22. HRMS (M + H+): 338.1870, C20H23N3O2 requires
338.1869.
Peptides 8a and 8b (Phe-cis-3-phenylpyrrolidine-NH2). The
analogues were prepared according to the general procedure using
Rink amide MBHA resin (157 mg, 100 μmol), racemic Fmoc-cis-3-
phenylpyrrolidine-2-carboxylic acid (62 mg, 150 μmol), Fmoc-Phe-
OH (155 mg, 400 μmol), HBTU (152 mg, 400 μmol or 57 mg, 150
μmol), and DIEA (278 μL, 1600 μmol or 52 μL, 300 μmol). Reaction
time was 24 h. The crude peptide mixture was purified, and
diastereomers were separated by RP-HPLC. Because of a difficult
separation, diastereomer “8a” was isolated as white solid in 14.2 mg
(42% yield) in a 2:1 ratio (8a/8b) and the pure diastereomer 8b in
11.2 mg (33% yield). The synthesis and purification have been
reproduced serveral times with the same outcome.
>97.8%. H NMR (CD3OD) δ 1.31 (dd, J = 1.6, 7.1 Hz, 3H), 2.95
(dd, J = 8.4, 14.3 Hz, 1H), 3.13−3.21 (m, 2H), 3.93 (ddd, J = 1.6, 5.0,
8.4 Hz, 1H), 4.65 (dd, J = 1.6, 9.7 Hz, 1H), 7.16−7.40 (m, 10H). 13C
NMR (CD3OD) δ 19.9, 38.4, 42.9, 55.2, 59.6, 128.0, 128.8, 128.9,
129.6, 130.2, 130.5, 135.3, 144.0, 169.1, 175.2. HRMS (M + H+):
326.1863, C19H23N3O2 requires 326.1869.
Peptide 9b. HPLC purity: C18 column, >99%; Fluophase, >99%.
1H NMR (CD3OD) δ 1.29 (dd, J = 0.9, 7.0 Hz, 3H), 2.42 (dd, J = 8.9,
14.5 Hz, 1H), 2.57 (dd, J = 5.0, 14.5 Hz, 1H), 3.13 (ddm, J = 6.7, 9.9
Hz, 1H), 3.95 (dd, J = 5.0, 8.9 Hz, 1H), 4.66 (dd, J = 0.9, 9.9 Hz, 1H),
6.96−7.01 (m, 2H), 7.17−7.33 (m, 8H). 13C NMR (CD3OD) δ 19.6,
38.3, 43.4, 55.4, 59.4, 128.1, 128.8, 128.9, 129.7, 130.1, 130.3, 135.3,
144.1, 169.1, 175.5. HRMS (M + H+): 326.1868, C19H23N3O2 requires
326.1869.
Peptide 10 (Phe-homoPhe-NH2). The analogue was prepared
according to the general procedure using Rink amide MBHA resin
(157 mg, 100 μmol), Fmoc-Phe-OH (155 mg, 400 μmol), Fmoc-
homo-Phe-OH (161 mg, 400 μmol), HBTU (152 mg, 400 μmol), and
DIEA (139 μL, 800 μmol). Reaction time was 18 h. The crude peptide
was purified by RP-HPLC to give 10 as a white solid in 21.8 mg (67%
yield). HPLC purity: C18 column, >99%; Fluophase, >99%. 1H NMR
(CD3OD) δ 1.91−2.15 (m, 2H), 2.60−2.77 (m, 2H), 3.06 (dd, J = 8.5,
14.2 Hz, 1H), 3.31 (dd, J = 5.9, 14.3 Hz, 1H), 3.19 (dd, J = 5.9, 8.5 Hz,
1H), 4.41 (dd, J = 5.4, 8.5 Hz, 1H), 7.14−7.40 (m, 10H). 13C NMR
(CD3OD) 33.0, 35.4, 38.6, 54.4, 55.6, 127.2, 128.9, 129.4, 129.5, 130.2,
130.6, 135.6, 142.4, 169.6, 175.8. HRMS (M + H+): 326.1873,
C19H23N3O2 requires 326.1869.
Peptide 11 (Phe-Phe(3-F)-NH2). The analogue was prepared
according to the general procedure using Rink amide MBHA resin
(157 mg, 100 μmol), Fmoc-Phe-OH (155 mg, 400 μmol), Fmoc-3-
Fluoro-Phe-OH (162 mg, 400 μmol), HBTU (152 mg, 400 μmol),
and DIEA (139 μL, 800 μmol). Reaction time was 18 h. The crude
peptide was purified by RP-HPLC to give 11 as a white solid in 12.9
mg (40% yield). HPLC purity: C18 column, 98.9%; Fluophase, 98.7%.
1H NMR (CD3OD) δ 2.98 (s, 3H), 3.00 (dd, J = 7.6, 14.0 Hz, 1H),
3.02 (dd, J = 8.4, 14.3 Hz, 1H), 3.12 (dd, J = 7.6, 14.0 Hz, 1H), 3.25
(dd, J = 5.7, 14.3 Hz, 1H), 4.09 (dd, J = 5.7, 8.4 Hz, 1H), 4.66 (dd, J =
7.6, 7.6 Hz, 1H), 7.07−7.21 (m, 4H), 7.26−7.41 (m, 5H). 13C NMR
(CD3OD) 19.6, 36.5, 38.6, 54.7, 55.5, 127.0, 128.0, 128.9, 130.2, 130.5,
130.8, 131.4, 135.5, 136.1, 137.9, 169.3, 175.3. HRMS (M + H+):
326.1866, C19H23N3O2 requires 326.1869.
Peptide 8a. HPLC purity (2:1 ratio, 8a/8b): C18 column, >99%;
1
Fluophase, >99%. H NMR for 8a (CD3OD) δ (2:1 ratio of 8a/8b,
Peptide 12 (Phe-Phe(2-Me)-NH2). The analogue was prepared
according to the general procedure using Rink amide MBHA resin
(157 mg, 100 μmol), Fmoc-Phe-OH (155 mg, 400 μmol), Fmoc-2-
methyl-Phe-OH (161 mg, 400 μmol), HBTU (152 mg, 400 μmol),
and DIEA (139 μL, 800 μmol). Reaction time was 18 h. The crude
peptide was purified by RP-HPLC to give 12 as a white solid in 12.4
mg (38% yield). HPLC purity: C18 column, >99%; Fluophase, >99%.
1H NMR (CD3OD) δ 2.98 (dd, J = 8.4, 13.9 Hz, 1H), 3.01 (dd, J =
8.5, 14.3 Hz, 1H), 3.15 (dd, J = 6.3, 13.9 Hz, 1H), 3.26 (dd, J = 5.4,
14.3 Hz, 1H), 4.08 (dd, J = 5.4, 8.5 Hz, 1H), 4.66 (dd, J = 6.3, 8.4 Hz,
1H), 6.92−7.00 (m, 1H), 7.01−7.13 (m, 2H), 7.26−7.40 (m, 6H). 13C
NMR (CD3OD) 38.6, 38.7, 55.5, 55.7, 114.6 (J = 21.4 Hz), 117.0 (J =
21.7 Hz), 126.2 (J = 2.77 Hz), 128.9, 130.2, 130.5, 131.2 (J = 8.29
Hz), 135.4, 141.0 (J = 7.56 Hz), 164.3 (J = 244.4 Hz), 169.5, 174.9.
HRMS (M + H+): 330.1621, C18H20N3O2F requires 330.1618.
[2,4-DehydroPro]SP1−7. The precursor peptide for tritium-
labeling [2,4-dehydroPro]SP1−7 was prepared by standard solid-
phase peptide synthesis techniques using Fmoc/tert-butyl protection
and purified as described above. Tritium labeling of the precursor was
performed by Amersham Biosciences (Cardiff, U.K.) and resulted in
370 MBq (10 mCi) of [3H]SP1−7 with a specific activity of 3.11 TBq/
mmol (84 Ci/mmol).
major diastereomer reported) 2.18 (m, J = 1.03, 6.3, 12,3 Hz, 1H),
2.67 (m, J = 8.6 10.6, 12.3 Hz, 1H), 3.04 (dd, J = 7.7, 14.4 Hz, 1H),
3.35 (dd, J = 6.0, 14.4 Hz, 1H), 3.59−3.74 (m, 3H), 4.38 (dd, J = 6.0,
7.7 Hz, 1H), 4.75 (d, J = 8.7 Hz, 1H), 7.11 (m, 1H), 7.21−7.53 (m,
9H). 13C NMR (CD3OD) δ (2:1 ratio of 8a/8b, major diastereomer
reported) 29.56, 38.30, 47.75, 47.95, 54.06, 65.61, 128.59, 129.17,
129.43, 129.44, 130.13, 130.71, 135.51, 137.49, 168.40, 174.46. HRMS
(M + H+): 338.1865, C20H23N3O2 requires 338.1869.
Peptide 8b. HPLC purity: C18 column, >99%; Fluophase, >99%.
1H NMR (CD3OD) δ 1.92 (m, J = 1.2, 6.2, 13,2 Hz, 1H), 2.58 (m, J =
8.5, 10.7, 11.8, 13.2 Hz, 1H), 2.73 (m, J = 6.2, 9.6, 10.9 Hz, 1H), 3.17
(d, J = 7.6 Hz, 2H), 3.31 (m, J = 8.6 Hz, 1H), 3.84 (m, J = 1.2, 8.5, 9.7
Hz, 1H), 4.42 (t, J = 7.6 Hz, 1H), 4.53 (d, J = 8.6 Hz, 1H), 7.11 (m,
1H), 7.21−7.45 (m, 10H). 13C NMR (CD3OD) δ 30.07, 37.70, 47.64,
47.72, 54.16, 65.54, 128.94, 129.25, 129.42, 129.49, 130.31, 130.85,
135.36, 137.79, 168.49, 173.78. HRMS (M + H+): 338.1865,
C20H23N3O2 requires 338.1869.
Peptides 9a and 9b (Phe-β-Me-Phe-NH2). The analogues were
prepared according to the general procedure using Rink amide MBHA
resin (157 mg, 100 μmol), Fmoc-Phe-OH (155 mg, 400 μmol),
(2R,3R)/(2S/3S)-racemic-Fmoc-β-methylphenylalanine (60 mg, 150
μmol), HBTU (152 mg, 400 μmol or 57 mg, 150 μmol), and DIEA
(278 μL, 1600 μmol or 52 μL, 300 μmol). Reaction time was 24 h.
The crude peptidemixture was purified and the diastereomers were
Animal Experiment and Membrane Preparation. The
preparations of receptor membranes were conducted using spinal
cords from male Sprague−Dawley rats. The rats (Alab AB, Sollentuna,
I
dx.doi.org/10.1021/jm400209h | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Sweden), weighing 200−250 g, were housed in groups of four in air-
conditioned rooms (22−23 °C and a humidity of 50−60%) under an
artificial light−dark cycle and had free access to food and water. Prior
to tissue sampling the rats were allowed to adapt to the laboratory
environment for 1 week. The animals (n = 15) were killed by
decapitation, and the spinal cords were rapidly removed and quickly
frozen. Tissues were then kept at −80 °C until analyzed. The animal
experiment was approved by the local ethical committee in Uppsala,
Sweden.
The frozen spinal cords weighing approximately 250−300 mg/
animal were thawed and placed on ice before being homogenized for
30 s in 30 volumes of ice-cold 50 mM Tris-HCl buffer (pH 7.4)
containing 5 mM KCl and 120 mM NaCl, using a Polytron
homogenizer. The homogenate was then centrifuged at 40000g for
20 min at 4 °C, and the supernatant was discarded. The resulting pellet
was resuspended in 30 volumes of ice-cold 50 mM Tris-HCl buffer
(pH 7.4) containing 300 mM KCl and 10 mM EDTA. Following
incubation on ice for 30 min the sample was centrifuged at 40000g for
20 min at 4 °C. The pellet obtained was diluted and homogenized in
30 volumes of ice-cold 50 mM Tris-HCl containing 0.02% BSA, 5
mM, EDTA, 3 mM MnCl2, and 40 μg bacitracin and recentrifuged
twice as described above. The final pellet was resuspended and
homogenized in 5 volumes of ice-cold 50 mM Tris-HCl buffer (pH
7.4) and immediately frozen at −80 °C until used in binding studies.
The protein concentration of the membrane suspension was
determined according to the method of Lowry48 using bovine serum
albumin as protein standard.
Radioligand Binding Assay. Assessment of the binding affinity
for the various compounds analyzed in this study was carried out using
the analogue [3H]SP1−7 as tracer. Assays were performed in tubes
containing 50 μL of spinal cord membrane suspension (200 μg of
protein) and 0.9 nM [3H]SP1−7 (specific activity, 3.11 TBq/mmol (84
Ci/mmol)) in a final volume of 0.5 mL of 50 mM Tris binding buffer
(pH 7.4) containing, 3 mM MnCl2, 0.2% BSA, and peptidase
inhibitors (40 μg/mL bacitracin, 4 μg/mL leupeptin, 2 μg/mL
aprotinin, and 4 μg/mL phosphoramidon). The amounts of total and
unspecific binding in percent of the total radioactivity added were
approximately 7% and 1.7%, respectively. The competition experi-
ments were determined at six different concentrations varying between
0.01 nM and 1 μM unlabeled compounds. Nonspecific binding was
determined in the presence of 1 μM SP1−7. Samples were incubated for
60 min at 4 °C before being terminated by rapid filtration under
vacuum with a Brandel 24-sample cell harvester through Whatman
GF/C glass fiber filters treated with a solution containing 50 mM Tris
(pH 7.4), 0.3% polyethylenimine (PEI), and 0.5% Triton X-100 at 4
°C overnight. Filters were washed twice with 3 mL of cold 50 mM
Tris-HCl (pH 7.4) complemented with 0.1 mg/mL BSA and 3 mM
MnCl2. The filters were air-dried for about 60 min before the bound
radioactivity was determined using a liquid scintillation counter
(Beckman LS 6000IC) at 63% efficiency in 5 mL of counting
scintillant. The specific binding was determined as the difference
between total and unspecific binding. All assays were run at six
concentrations in triplicate, and each assay was repeated at least three
times at different days except for compound 8a, which was tested at six
concentrations in triplicate on the same day. The binding curve for 8a
is shown in Figure 5. Data and statistics from the competition
experiments were analyzed in the GraFit program (Erithacus Software,
U.K.).
Figure 5. Inhibition of [3H]SP1−7 binding by compound 8a.
pair distance of 0.5 Å after superposition. The pharmacophore features
found and assigned in the active compounds were hydrogen bond
acceptors and donors, aromatic rings, and positively charged groups.
The score formula used for pharmacophore hypothesis scoring was the
sum of the vector, site, and volume scores plus a score for rewarding a
high number of matches. Because the number of actives matching the
pharmacophore hypotheses was 4 in all cases, the last term was set to a
constant 1. The other terms in the formula have the range 0−1, which
gives a matching score range of 1−4.
Metabolic Stability. Compounds (1 μM) were preincubated for 5
min at 37 °C with pooled human liver microsomes (0.5 mg/mL;
Xenotech, KS) in 0.1 M potassium phosphate buffer, pH 7.4, prior to
the addition of NADPH (1 mM) to initiate the reaction. Mixtures
were then incubated for 0, 5, 15, and 40 min, and at each time point
the reaction was stopped by the addition of acetonitrile (50% v/v).
Plates were centrifuged at 3500 rpm for 20 min at 4 °C, and the
supernatants were subjected to liquid chromatography/mass spec-
trometry analysis. The natural logarithm of the analytical peak area
ratio (relative to 0 min sample which was considered as 100%) was
plotted against time and analyzed by linear regression. In vitro half-life
(t1/2) and in vitro intrinsic clearance (Clint) were calculated on the
basis of first-order reaction kinetics of the percentage of remaining
compound. Dextromethorphan (3 μM) and midazolam (5 μM) were
used as positive controls for cytochrome P450 enzymes (CYP)
isoforms CYP2D6 and CYP3A4, respectively.
Cell Culture. Caco-2 cells (obtained from American Tissue
Collection, Rockville, MD) were maintained in an atmosphere of
90% air and 10% CO2 as described previously.28 For transport
experiments, 3.0 × 105 cells (passages 98−102) were seeded on
polycarbonate filter inserts (12 mm diameter, pore size 0.4 μm; Costar,
Cambridge, MA) and allowed to grow and differentiate for 21−24
days. The integrity of the monolayers was assessed by measuring the
paracellular marker [14C]mannitol (1.0 μCi/mL, 57.3 mCi/mmol;
Perkin-Elmer Life Sciences, Boston, MA) transport and the trans-
epithelial electrical resistance (TEER) before and after the experi-
ments.
CHO-K1 and CHO-PEPT1 cells (control and stably transfected
cells, respectively, were kind gifts from Dr. Anna-Lena Ungell,
AstraZeneca Molndal, Sweden) were maintained at 37 °C in an
̈
atmosphere of 90% air and 10% CO2, with DMEM containing 4.5 g/L
glucose, 10% fetal bovine serum, 1% nonessential amino acids, and 50
μg/mL gentamicin sulfate (Invitrogen, Carlsbad, CA). For the uptake
experiments, 1.0 × 105 cells/well were seeded in 24-well plates and
allowed to grow in antibiotic-free medium for 2 days.
Computational Analysis. Compounds 1−12 were built in
Maestro49 and imported to Phase.30 The structures were ionized at
pH 7.0 (i.e., protonation of the N-terminus). Conformations were
generated in Phase using MacroModel with 200 search steps per
rotatable bond of mixed MCMM/LMOD with sampling set to
Thorough. Energy minimization was performed using maximum 100
steps of truncated Newton conjugate gradient minimization with
convergence criterion set to 0.05 kJ mol−1 Å−1. The force field OPLS
200550 and the generalized-Born/surface area (GB/SA) model for
water51 as solvation model were used. Conformations within 5 kcal
mol−1 of the lowest found energy minimum were saved, and
redundant conformations were removed based on maximum atom
Transcellular Transport and Uptake Experiments. Stock
solutions (10 mM) of the peptides were prepared in dimethylsulfoxide
(DMSO) and diluted to 100 μM (final DMSO concentration of 1%)
in Hank’s balanced salt solution (HBSS) containing 10 mM MES at
pH 6.0 (HBSS pH 6.0) or containing 10 mM HEPES at pH 7.4
(HBSS pH 7.4). In all experiments, [14C]glycylsarcosine ([14C]GlySar,
1.82 μM, 0.1 μCi/mL, 55 mCi/mmol; ARC, St. Louis, MO) was used
as a PEPT1 substrate control. For inhibition controls, an excess of
unlabeled competitor (10 mM; Sigma-Aldrich, St. Louis, MO) was
used.
J
dx.doi.org/10.1021/jm400209h | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
The intestinal epithelial permeability was determined from transport
rates across Caco-2 cell monolayers. The cell monolayers were gently
rinsed with HBSS, pH 6.0, and left to equilibrate in the same solution
for 30 min at 37 °C. The transport experiments were run for 2 h at 37
°C and were started by the application of the compound solution to
the donor side, which was the apical chamber in the apical to
basolateral (a−b) experiments and the basolateral chamber in the
basolateral to apical (b−a) experiments. Filter inserts were
AUTHOR INFORMATION
Corresponding Author
*Phone: +46-18-4714285. Fax: +46-18-4714474. E-mail: Anja.
Sandstrom@orgfarm.uu.se.
■
Author Contributions
The manuscript was written through contributions of all
authors. All authors have given approval to the final version of
the manuscript.
continuously stirred at 500 rpm on IKA-Schuttler MTS4 to obtain
̈
data that were unbiased by the aqueous boundary layer. The receiver
chambers were sampled in suitable time points, and the samples were
replaced with equal volumes of preheated receiver solution.
For transport studies performed under sink conditions, where less
than 10% of the compound was transported across the Caco-2 cell
monolayers, the apparent permeability coefficients (Papp) were
calculated from the equation
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Dr. Milad Botros for technical assistance regarding
the radioligand binding assay and Rebecka Strand for synthetic
assistance. This work was supported by grants from the
Swedish Research Council (Grants 9478 and 21386 and a grant
to the Chemical Biology Consortium Sweden), Uppsala Berzelii
Technology Centre for Neurodiagnotics, and the Knut and
Alice Wallenberg Foundation. J.M.K. thanks CAPES/MEC/
Brazil for his research fellowship.
ΔQ
1
P
=
app
Δt AC0
where ΔQ/Δt is the steady-state flux (mol/s), C0 is the initial
concentration in the donor chamber at each time interval (mol/mL),
and A is the surface area of the filter (cm2). If the fraction transported
exceeded 10%, the Papp coefficients were calculated applying nonsink
conditions from the equation
ABBREVIATIONS USED
■
SP, substance P; NK-1, neurokinin-1; SAR, structure−activity
relationship; EM-2, endomorphin-2; E1, endomorphin-1;
PEPT1, dipeptide/tripeptide transporter; HLM, human liver
microsome; Clint, intrinsic clearance; Papp, apparent perme-
ability coefficient; t1/2, in vitro half-life; DMSO, dimethylsulf-
oxide; HBSS, Hank’s balanced salt solution
⎛
⎞
⎟
⎠
M
VD + VR
M
VD + VR
appA(1/VD+VR)t
e−P
CR(t) =
+ C
−
⎜
R,0
⎝
where CR(t) is the time-dependent drug concentration in the receiver
compartment (μM), M is the amount of drug in the system (nmol),
VD and VR are the volumes of the donor and receiver compartment
(mL), respectively, and t is the time that has elapsed from the start of
the interval (s).52 Papp was obtained from nonlinear regression of the
accumulated dose in the receiver compartment over time, minimizing
the sum of squared residuals in the equation.
Uptake studies with CHO cells were performed at 37 °C in HBSS,
pH 6.0. Initially, cells were gently rinsed and left to equilibrate for 30
min at 37 °C. After the equilibration period, buffer was removed and
replaced by compound diluted in HBSS, pH 6.0, and cells were
incubated for 15 min at 37 °C. Uptake was terminated by buffer
removal followed by two washes with ice-cold phosphate buffer saline
(PBS). Cells were then lysed with acetonitrile/water (60:40, v/v) and
centrifuged at 3500 rpm and 4 °C for 20 min. In the case of
experiments with [14C]GlySar, 1 M sodium hydroxide solution was
used for cell lysis. The results were expressed as (pmol/mg of
protein)/min.
REFERENCES
■
(1) Von Euler, U. S.; Gaddum, J. H. An unidentified depressor
substance in certain tissue extracts. J. Physiol. 1931, 72, 74−87.
(2) Nakanishi, S. Mammalian tachykinin receptors. Annu. Rev.
Neurosci. 1991, 14, 123−136.
(3) Hokfelt, T.; Pernow, B.; Wahren, J. Substance P: a pioneer
̈
amongst neuropeptides. J. Intern. Med. 2001, 249, 27−40.
(4) Lee, C.-M.; Sandberg, B. E. B.; Hanley, M. R.; Iversen, L. L.
Purification and characterization of a membrane-bound substance P-
degrading enzyme from human brain. Eur. J. Biochem. 1981, 114, 315−
327.
(5) Sakurada, T.; Le Greves, P.; Stewart, J.; Terenius, L.
Measurement of substance P metabolites in rat CNS. J. Neurochem.
1985, 44, 718−722.
(6) Hallberg, M.; Nyberg, F. Neuropeptide conversion to bioactive
fragmentsan important pathway in neuromodulation. Curr. Protein
Pept. Sci. 2003, 4, 31−44.
Data Analysis. Caco-2 and CHO cell permeability experiments
were performed in, at least, triplicate and metabolic stability
experiments in duplicate, and all samples were analyzed with liquid
chromatography/mass spectrometry analysis on a ThermoFinnigan
TSQ Quantum Discovery triple-quadrupole instrument (electrospray
ionization, ESI; Thermo Electron Corp. Waltham, MA, U.S.) coupled
to an Acquity ultrahigh performance LC system (Waters, Milford,
MA). For chromatographic separation, an Acquity UPLC C18 column
(1.7 μm; Waters, Milford, MA) and a flow rate of 600 μL/min were
used. During the analysis an amount of 10 μL of the samples was
injected with a gradient with 0.1% formic acid in acetonitrile and 0.1%
formic acid in water. Electrospray ionization was used in positive
mode, and the daughter ions of m/z 120.1 (1, 6, and 12), 134.1 (2, 3,
4, 5, 9, and 11), 146.1 (7 and 8), and 281.2 (10) were used for
quantification of the respective compounds. The internal standard
warfarin (50 nM) was used throughout the analysis.
(7) Kreeger, J. S.; Larson, A. A. Substance P-(1−7), a substance P
metabolite, inhibits withdrawal jumping in morphine-dependent mice.
Eur. J. Pharmacol. 1993, 238, 111−115.
(8) Zhou, Q.; Carlsson, A.; Botros, M.; Fransson, R.; Sandstrom, A.;
̈
Gordh, T.; Hallberg, M.; Nyberg, F. The C-terminal amidated
analogue of the substance P (SP) fragment SP1−7 attenuates the
expression of naloxone-precipitated withdrawal in morphine depend-
ent rats. Peptides 2009, 30, 2418−2422.
(9) Zhou, Q.; Frandberg, P.-A.; Kindlundh, A. M. S.; Le Greves, P.;
̈
Nyberg, F. Substance P(1-7) affects the expression of dopamine D2
receptor mRNA in male rat brain during morphine withdrawal.
Peptides 2003, 24, 147−153.
(10) Wiktelius, D.; Khalil, Z.; Nyberg, F. Modulation of peripheral
inflammation by the substance P N-terminal metabolite substance P1-
7. Peptides 2006, 27, 1490−1497.
Radioactive samples ([14C]GlySar and [14C]mannitol) were
analyzed with a liquid scintillation counter (TopCount NXT,
Perkin-Elmer Life Sciences, Boston, MA). Cellular protein content
was determined using a protein assay kit with bovine serum albumin
(BSA) as standard (Thermo Scientific, Rockford, IL).
(11) Sakurada, C.; Watanabe, C.; Sakurada, T. Occurrence of
substance P(1-7) in the metabolism of substance P and its
antinociceptive activity at the mouse spinal cord level. Methods Find.
Exp. Clin. Pharmacol. 2004, 26, 171−176.
K
dx.doi.org/10.1021/jm400209h | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(12) Carlsson, A.; Ohsawa, M.; Hallberg, M.; Nyberg, F.; Kamei, J.
Substance P1-7 induces antihyperalgesia in diabetic mice through a
mechanism involving the naloxone-sensitive sigma receptors. Eur. J.
Pharmacol. 2010, 626, 250−255.
(13) Dyck, P. J.; Dyck, P. J. B.; Velosa, J. A.; Larson, T. S.; O’Brien, P.
C.; Grp, N. G. F. S. Patterns of quantitative sensation testing of
hypoesthesia and hyperalgesia are predictive of diabetic polyneurop-
athy. A study of three cohorts. Diabetes Care 2000, 23, 510−517.
(14) Skilling, S. R.; Smullin, D. H.; Larson, A. A. Differential-effects
of C-terminal and N-terminal substance-P metabolites on the release
of amino-acid neurotransmitters from the spinal-cord. Potential role in
nociception. J. Neurosci. 1990, 10, 1309−1318.
(15) Mousseau, D. D.; Sun, X. F.; Larson, A. A. Identification of a
novel receptor mediating substance-P-induced behavior in the mouse.
Eur. J. Pharmacol. 1992, 217, 197−201.
(16) Igwe, O. J.; Kim, D. C.; Seybold, V. S.; Larson, A. A. Specific
binding of substance P aminoterminal heptapeptide [SP(1-7)] to
mouse brain and spinal cord membranes. J. Neurosci. 1990, 10, 3653−
3663.
based on molecular surface properties. J. Med. Chem. 2003, 46, 558−
570.
(30) Phase, version 3.1; Schrodinger, LLC: New York, NY, 2009.
̈
(31) Piercey, M. F.; Dobry, P. J. K.; Einspahr, F. J.; Schroeder, L. A.;
Masiques, N. Use of substance-P fragments to differentiate substance-
P receptors of different tissues. Regul. Pept. 1982, 3, 337−349.
(32) Huston, J. P.; Hasenoehrl, R. U.; Boix, F.; Gerhardt, P.;
Schwarting, R. K. W. Sequence-specific effects of neurokinin substance
P on memory, reinforcement, and brain dopamine activity.
Psychopharmacology 1993, 112, 147−162.
(33) Tomaz, C.; Silva, A. C.; Nogueira, P. J. Long-lasting
mnemotropic effect of substance P and its N-terminal fragment
(SP1-7) on avoidance learning. Braz. J. Med. Biol. Res. 1997, 30, 231−
233.
(34) Rizo, J.; Gierasch, L. M. Constrained peptides: models of
bioactive peptides and protein substructures. Annu. Rev. Biochem.
1992, 61, 387−418.
(35) Turk, J.; Panse, G. T.; Marshall, G. R. alpha-Methyl amino acids.
Resolution and amino protection. J. Org. Chem. 1975, 40, 953−955.
(36) Boyle, S.; Guard, S.; Higginbottom, M.; Horwell, D. C.;
Howson, W.; McKnight, A.; Martin, K.; Pritchard, M. C.; O’Toole, J.;
et al. Rational design of high affinity tachykinin NK1 receptor
antagonists. Bioorg. Med. Chem. 1994, 2, 357−370.
(37) Haskell-Luevano, C.; Toth, K.; Boteju, L.; Job, C.; Castrucci, A.
M. D.; Hadley, M. E.; Hruby, V. J. beta-Methylation of the Phe(7) and
Trp(9) melanotropin side chain pharmacophores affects ligand−
receptor interactions and prolonged biological activity. J. Med. Chem.
1997, 40, 2740−2749.
(17) Botros, M.; Hallberg, M.; Johansson, T.; Zhou, Q.; Lindeberg,
G.; Frandberg, P.-A.; Toemboely, C.; Toth, G.; Le Greves, P.; Nyberg,
̈
F. Endomorphin-1 and endomorphin-2 differentially interact with
specific binding sites for substance P (SP) aminoterminal SP1-7 in the
rat spinal cord. Peptides 2006, 27, 753−759.
(18) Botros, M.; Johansson, T.; Zhou, Q.; Lindeberg, G.; Tomboly,
C.; Toth, G.; Le Greves, P.; Nyberg, F.; Hallberg, M. Endomorphins
interact with the substance P (SP) aminoterminal SP1-7 binding in the
ventral tegmental area of the rat brain. Peptides 2008, 29, 1820−1824.
(19) Rimon, R.; Legreves, P.; Nyberg, F.; Heikkila, L.; Salmela, L.;
Terenius, L. Elevation of substance P-like peptides in the CSF of
psychiatric-patients. Biol. Psychiatry 1984, 19, 509−516.
(38) Brandsch, M. Transport of drugs by proton-coupled peptide
transporters: pearls and pitfalls. Expert Opin. Drug Metab. 2009, 5,
887−905.
(39) Brodin, B.; Nielsen, C. U.; Steffansen, B.; Frokjaer, S. Transport
of peptidomimetic drugs by the intestinal di/tri-peptide transporter,
PepT1. Pharmacol. Toxicol. 2002, 90, 285−296.
(20) Nyberg, F.; Le Greves, P.; Sundqvist, C.; Terenius, L.
Characterization of substance P(1-7) and (1-8) generating enzyme
in human cerebrospinal fluid. Biochem. Biophys. Res. Commun. 1984,
125, 244−250.
(40) Bailey, P. D.; Boyd, C. A. R.; Bronk, J. R.; Collier, I. D.;
Meredith, D.; Morgan, K. M.; Temple, C. S. How to make drugs orally
active: a substrate template for peptide transporter PepT1. Angew.
Chem., Int. Ed. 2000, 39, 505−508.
(21) Fransson, R.; Botros, M.; Nyberg, F.; Lindeberg, G.; Sandstrom,
̈
A.; Hallberg, M. Small peptides mimicking substance P (1-7) and
encompassing a C-terminal amide functionality. Neuropeptides 2008,
42, 31−37.
(22) Ohsawa, M.; Carlsson, A.; Asato, M.; Koizumi, T.; Nakanishi, Y.;
(41) Rubio-Aliaga, I.; Daniel, H. Mammalian peptide transporters as
targets for drug delivery. Trends Pharmacol. Sci. 2002, 23, 434−440.
Fransson, R.; Sandstrom, A.; Hallberg, M.; Nyberg, F.; Kamei, J. The
̈
effect of substance P1‑7 amide on nociceptive threshold in diabetic
(42) Vabeno, J.; Lejon, T.; Nielsen, C. U.; Steffansen, B.; Chen, W.
̊
̈
mice. Peptides 2011, 32, 93−98.
Q.; Hui, O. Y.; Borchardt, R. T.; Luthman, K. Phe-Gly dipeptidomi-
metics designed for the di-/tripeptide transporters PEPT1 and PEPT2:
synthesis and biological investigations. J. Med. Chem. 2004, 47, 1060−
1069.
(23) Fransson, R.; Botros, M.; Skold, C.; Nyberg, F.; Lindeberg, G.;
̈
Hallberg, M.; Sandstrom, A. Discovery of dipeptides with high affinity
̈
to the specific binding site for substance P1-7. J. Med. Chem. 2010, 53,
2383−2389.
(43) Vig, B. S.; Stouch, T. R.; Timoszyk, J. K.; Quan, Y.; Wall, D. A.;
Smith, R. L.; Faria, T. N. Human PEPT1 pharmacophore distinguishes
between dipeptide transport and binding. J. Med. Chem. 2006, 49,
3636−3644.
(44) Gupta, D.; Gupta, S. V.; Lee, K. D.; Amidon, G. L. Chemical and
enzymatic stability of amino acid prodrugs containing methoxy, ethoxy
and propylene glycol linkers. Mol. Pharmaceutics 2009, 6, 1604−1611.
(45) Tehler, U.; Nelson, C. H.; Peterson, L. W.; Provoda, C. J.;
Hilfinger, J. M.; Lee, K. D.; McKenna, C. E.; Amidon, G. L.
Puromycin-sensitive aminopeptidase: an antiviral prodrug activating
enzyme. Antiviral Res. 2010, 85, 482−489.
(46) Giacomini, K. M.; Huang, S. M.; Tweedie, D. J.; Benet, L. Z.;
Brouwer, K. L. R.; Chu, X. Y.; Dahlin, A.; Evers, R.; Fischer, V.;
Hillgren, K. M.; Hoffmaster, K. A.; Ishikawa, T.; Keppler, D.; Kim, R.
B.; Lee, C. A.; Niemi, M.; Polli, J. W.; Sugiyama, Y.; Swaan, P. W.;
Ware, J. A.; Wright, S. H.; Yee, S. W.; Zamek-Gliszczynski, M. J.;
Zhang, L. Membrane transporters in drug development. Nat. Rev. Drug
Discovery 2010, 9, 215−236.
(47) Polli, J. W.; Wring, S. A.; Humphreys, J. E.; Huang, L. Y.;
Morgan, J. B.; Webster, L. O.; Serabjit-Singh, C. S. Rational use of in
vitro P-glycoprotein assays in drug discovery. J. Pharmacol. Exp. Ther.
2001, 299, 620−628.
(24) Ohsawa, M.; Carlsson, A.; Asato, M.; Koizumi, T.; Nakanishi, Y.;
Fransson, R.; Sandstrom, A.; Hallberg, M.; Nyberg, F.; Kamei, J. The
̈
dipeptide Phe-Phe amide attenuates signs of hyperalgesia, allodynia
and nociception in diabetic mice using a mechanism involving the
sigma receptor system. Mol. Pain 2011, 7 (85), 1−12.
(25) Houston, J. B. Utility of in-vitro drug-metabolism data in
predicting in-vivo metabolic-clearance. Biochem. Pharmacol. 1994, 47,
1469−1479.
(26) Obach, R. S. Prediction of human clearance of twenty-nine
drugs from hepatic microsomal intrinsic clearance data: an
examination of in vitro half-life approach and nonspecific binding to
microsomes. Drug Metab. Dispos. 1999, 27, 1350−1359.
(27) Baranczewski, P.; Stanczak, A.; Sundberg, K.; Svensson, R.;
Wallin, A.; Jansson, J.; Garberg, P.; Postlind, H. Introduction to in
vitro estimation of metabolic stability and drug interactions of new
chemical entities in drug discovery and development. Pharmacol. Rep.
2006, 58, 453−472.
(28) Hubatsch, I.; Ragnarsson, E. G. E.; Artursson, P. Determination
of drug permeability and prediction of drug absorption in Caco-2
monolayers. Nat. Protoc. 2007, 2, 2111−2119.
(29) Bergstrom, C. A. S.; Strafford, M.; Lazorova, L.; Avdeef, A.;
̈
Luthman, K.; Artursson, P. Absorption classification of oral drugs
L
dx.doi.org/10.1021/jm400209h | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(48) Lowry, O. H.; Rosebrough, N. J.; Farr, A. L.; Randall, R. J.
Protein measurement with the Folin phenol reagent. J. Biol. Chem.
1951, 193, 265−275.
(49) Maestro, version 9.0; Schrodinger, LLC: New York, NY, 2009.
̈
(50) Kaminski, G. A.; Friesner, R. A.; Tirado-Rives, J.; Jorgensen, W.
L. Evaluation and reparametrization of the OPLS-AA force field for
proteins via comparison with accurate quantum chemical calculations
on peptides. J. Phys. Chem. B 2001, 105, 6474−6487.
(51) Still, W. C.; Tempczyk, A.; Hawley, R. C.; Hendrickson, T.
Semianalytical treatment of solvation for molecular mechanics and
dynamics. J. Am. Chem. Soc. 1990, 112, 6127−6129.
(52) Tavelin, S.; Gråsjo, J.; Taipalensuu, J.; Ocklind, G.; Artursson, P.
̈
Applications of epithelial cell culture in studies of drug transport.
Methods Mol. Biol. 2002, 188, 233−272.
M
dx.doi.org/10.1021/jm400209h | J. Med. Chem. XXXX, XXX, XXX−XXX