In vitro characterization of human peptide transporter hPEPT1 interactions and passive permeation studies of short cationic antimicrobial peptides

Article abstract of DOI:10.1021/jm1015704

The present study assesses the permeation of cationic antimicrobial di- and tripeptides derived from lactoferricin via interaction with the human intestinal peptide transporter hPEPT1 and via passive routes. While some tested peptides displayed moderate affinity (0.6 and 2.7 mM) for interaction with hPEPT1, none served as substrate for hPEPT1 expressed by Xenopus laevis oocytes. It is shown that structural strategies employed to generate sufficient biological activity and metabolic stability such as introduction of large hydrophobic unnatural amino acids and different C-terminal modifications counteracted hPEPT1 mediated uptake. Most of the included peptides were nevertheless shown to permeate at rates suggesting moderate to excellent human oral absorption in the applied phospholipid vesicle-based passive permeation assay. Although the main factor governing passive permeation appears to be the hydrophobicity, peptide structure was also important and the overall permeation behavior was difficult to predict. Comparisons with a theoretical prediction model were also performed.

Full text of DOI:10.1021/jm1015704

ARTICLE
pubs.acs.org/jmc
In Vitro Characterization of Human Peptide Transporter hPEPT1
Interactions and Passive Permeation Studies of Short Cationic
Antimicrobial Peptides
Gøril Eide Flaten,† Gabor Kottra,‡ Wenche Stensen,§ Geir Isaksen,||,^ Rasmus Karstad,|| John S. Svendsen,||,§
Hannelore Daniel,‡ and Johan Svenson*,||
†Department of Pharmacy, University of Tromsø, N-9037, Tromsø, Norway
‡Molecular Nutrition Unit, Technical University of Munich, D-85350 Freising, Germany
§Lytix Biopharma AS, N-9294 Tromsø, Norway
Department of Chemistry, University of Tromsø, N-9037, Tromsø, Norway
^The Norwegian Structural Biology Centre and The Centre for Theoretical and Computational Chemistry, University of Tromsø,
N-9037, Tromsø, Norway
S Supporting Information
b
ABSTRACT:
The present study assesses the permeation of cationic antimicrobial di- and tripeptides derived from lactoferricin via interaction with
the human intestinal peptide transporter hPEPT1 and via passive routes. While some tested peptides displayed moderate affinity
(0.6 and 2.7 mM) for interaction with hPEPT1, none served as substrate for hPEPT1 expressed by Xenopus laevis oocytes. It is shown
that structural strategies employed to generate sufficient biological activity and metabolic stability such as introduction of large
hydrophobic unnatural amino acids and different C-terminal modifications counteracted hPEPT1 mediated uptake. Most of the
included peptides were nevertheless shown to permeate at rates suggesting moderate to excellent human oral absorption in the
applied phospholipid vesicle-based passive permeation assay. Although the main factor governing passive permeation appears to be
the hydrophobicity, peptide structure was also important and the overall permeation behavior was difficult to predict. Comparisons
with a theoretical prediction model were also performed.
’ INTRODUCTION
natural origin, violate these rules.1,5 Many orally bioavailable
antibiotics belong to this group, and their modes of uptake are
therefore difficult to predict. For the amino cephalosporin
antibiotics it has been established that they manage to utilize
the human intestinal di- and tripeptide transporter (hPEPT1)
to gain entry into the circulatory system.6 Although the physio-
logical role of hPEPT1 is to transport di- and tripeptides formed
during the luminal digestion, it also transports a variety of
peptidomimetic drugs and prodrugs.7-10 hPEPT1 has the
unique feature to accept almost all naturally occurring di- and
tripeptides accounting for more than 8000 substrates, yet each
substrate displays different kinetic properties.11-13
The bioavailability of an orally administered drug depends on
its solubility, the ability to cross the intestinal mucosa, and its
metabolic stability. Most drugs access the circulatory system
either via passive diffusion or through interactions with carrier
proteins in the intestinal epithelium. While diffusion has histori-
cally been considered the main route of absorption for small
drugs, recent studies suggest that selective transporters may play
a prominent role in intestinal absorption.1-3 Regardless of entry
mode, the same biophysical forces, i.e., hydrogen bond formation
and hydrophobic interactions, dictate the interaction with carrier
proteins and/or membrane phospholipids. Lipinski’s empirical
“rule of 5” provides a means for predicting whether a compound
is likely to display poor or good absorption via the passive
diffusion route.4 Several classes of drugs, however, mainly of
Received: December 10, 2010
Published: March 14, 2011
r
2011 American Chemical Society
2422
dx.doi.org/10.1021/jm1015704 J. Med. Chem. 2011, 54, 2422–2432
|

Journal of Medicinal Chemistry
ARTICLE
Truncated antimicrobial peptides based on bovine lactoferricin
are emerging as a class of compounds with potential to treat infec-
tions caused by several types of drug resistant microorganisms.14
Synthetic strategies such as truncation, capping, and introduction
of unnatural amino acid residues have been employed to trans-
form the original 25-mer into rapidly prepared druglike di- and
tripeptides with high activity and selectivity.15,16 From a drug
design perspective they represent a class of antimicrobial pep-
tides clearly within reach of pharmaceutical use with selected
compounds currently in phase II clinical trials (Lytix Biopharma
AS). It has recently been shown that several of these peptides can
be designed with improved stability toward proteolytic degrada-
tion in the gastrointestinal tract.17-19 Such increased half-life
opens up for potential oral bioavailability which would further
add to the attractiveness of these compounds.20,21 Their small
size in combination with the required pharmacophore suggests
that they are potential candidates for either tertiary active transport
via hPEPT1 or the passive permeation route.
To investigate if these compounds are substrates for transe-
pithelial transport, a library of 17 antimicrobial di- and tripeptides
was designed using synthetic strategies previously shown to
generate peptides with increased antimicrobial effect and meta-
bolic stability. hPEPT1 expressed in Xenopus laevis oocytes was
used to investigate whether the peptides serve as substrates of the
route via hPEPT1, while the passive diffusion of the peptides was
studied using the phospholipid vesicle-based permeation assay
(PVPA) as passive permeation model (Figure 1). Only a limited
number of tested peptides were able to interact with hPEPT1 but
in no case was electrogenic transport observed, demonstrating
that the structural alterations used to generate improved anti-
microbial potency prevent the peptides from acting as hPEPT1
substrates. Experiments to investigate passive diffusion, however,
revealed that several synthetic strategies can be employed to
generate potent peptides with excellent permeation properties
via this route. The included peptides are all active membrane
disruptors, and the corresponding challenges associated with the
analysis of their transport properties are further discussed. A
comparison with computationally generated theoretical para-
meters for oral uptake in humans is also included.
Figure 1. Representations of the biological and artificial systems used in
the present study to investigate active and passive peptide uptake. Top left
microscope image shows oocytes from Xenopus laevis with an average
diameter of 1.1 mm. Top right illustrates the localization of cloned GFP
coupled hPEPT1 in the oocyte membrane. The lower part (adapted from
ref 51) illustrates a schematic bisection of the PVPA barrier showing large
fused multilamellar liposomes deposited on top of a mixed cellulose ester
filter with smaller liposomes filling the pores to generate a tight barrier.51
extremely high values within seconds. However, no visible
damage to the oocytes could be seen. Peptides 3 and 11 were
therefore excluded from further analysis of hPEPT1-mediated
transport. Peptide 5 displayed no affinity for binding to hPEPT1.
The four remaining peptides displayed affinity for hPEPT1 with
peptide 14 being the strongest inhibitor. Despite the obvious
capability of these peptides to interact with the substrate binding
domain of hPEPT1, none induced any significant currents and
therefore they failed to show transport. The three inactive
reference dipeptides were all transported via hPEPT1 as shown
in Figure 3. A detailed analysis of the relations between structure
and hPEPT1-mediated transport rate for these peptides can be
found in the section Discussion.
’ RESULTS
Selected Peptides. A combination of new (4, 7, 9, 13, and the
four reference peptides) and previously synthesized peptides
from our extensive library was carefully chosen based on their
properties to be included in the present study. Their structure
and molecular properties are presented in Figure 2 and Table 1.
hPEPT1 Interactions. To evaluate a possible tertiary active
transport, seven di- and tripeptides (1, 2, 3, 4, 5, 11, and 14)
varying in structure and activity were selected. Inactive reference
peptides WR-OH, WR-NHBn, and BipR-OH were synthesized
and included to investigate the substrate preferences for
hPEPT1. Xenopus laevis oocytes expressing hPEPT1 were used
in a two-electrode voltage clamp (TEVC) experiment to assess
whether the peptides can induce a transport current (concentra-
tion range 0.25-1.0 mM) or whether they can inhibit the transport
current of the hPEPT1 substrate glycylsarcosine (Gly-Sar) (1 mM).
Residual transport currents in the presence of the antimicrobial
peptides are expressed as percent of the Gly-Sar current.
Passive Permeation in the PVPA. The ability of the peptides
to passively permeate across membranes was investigated using
the PVPA method, which is a medium-throughput in vitro
method for screening of passive diffusion of drugs across the
intestinal epithelia. This model has proven to give strong
coherence between its in vitro apparent permeability (Papp
)
and in vivo data on oral absorption of commercial drugs in humans.
On the basis of the common classification system of absorption,
drug compounds are classified as poorly absorbed (<30% absorbed
in vivo) when displaying Papp < 0.1 ꢀ 10-6 cm/s, as moderately
absorbed (30-70% absorbed in vivo) when the Papp values are
between 0.1 and 0.9 ꢀ 10-6 cm/s, and to have excellent oral
absorption (>70% in vivo absorption) with Papp >0.9ꢀ 10-6 cm/s.22
The results of the permeability studies and the classifications are
The results of the hPEPT1 experiments are summarized in
Tables 2 and 3. The very active peptides 3 and 11 increased, even
at very low concentration, the oocyte membrane conductance to
2423
dx.doi.org/10.1021/jm1015704 |J. Med. Chem. 2011, 54, 2422–2432

Journal of Medicinal Chemistry
ARTICLE
Figure 2. Structure of the 17 included antibacterial peptides. Four additional inactive reference peptides (not shown in figure) were also included.
shown in Table 4 and Figure 4. In addition, a correlation between
the permeation data and the antimicrobial activity was clear as
illustrated in Figure 5.
In contrast to the hPEPT1-mediated transport studies, most of
the peptides were able to cross the phospholipid vesicle-based
barrier at rates indicative of a moderate or excellent oral absorption.
2424
dx.doi.org/10.1021/jm1015704 |J. Med. Chem. 2011, 54, 2422–2432

Journal of Medicinal Chemistry
ARTICLE
It was clear that the influence of the properties dictating the
permeation across the membrane varied depending on the subtle
design differences of the peptides. Peptide 8 proved difficult to
work with and was excluded because of membrane disruption at
the concentrations employed. The result from the computer
generated oral uptake model is also presented in Table 4.
rational drug design.26,27 We describe for the first time the active
and passive transport properties of such peptides as a basis for
systemic drug delivery.
The current study employed pharmaceutically relevant pep-
tides fulfilling, or nearly so, the pharmacophore, i.e., at least two
units of hydrophobic bulk in conjunction with two positive charges
(exemplified in Figure 6).28 Both the bulk and the cationic charge
are provided by a range of different natural and unnatural amino
acid subunits. Several of the peptides display low MIC below 10
μg/mL and can be regarded as very active as shown in Table 1.
Although structurally similar, the peptides range from inactive to
highly active by design. This is further reflected in their differ-
ences in hydrophobicity as described by the theoretical ClogP
values calculated using QikProp and their retention times on a
C18 column which serves as an experimental estimate of their
hydrophobicity. It has previously been shown that for some
peptides there is a close correlation between antimicrobial activity
and a certain degree of C18 retention, and the same is seen (see
Supporting Information) for the tested peptides.29
’ DISCUSSION
Selected Peptides. Antimicrobial peptides are produced by
most living organisms as a first line of defense to slow microbial
intruders before the onset of the adaptive immune system.23
They display very favorable destructive properties and a mode of
action that is little likely to cause bacterial resistance.24,25 Never-
theless, their applicability as realistic future pharmaceuticals is
currently somewhat limited by their size (generally between 10
and 50 amino acid residues) and low bioavailability. The heavily
truncated lactoferricin based peptides used here represent an
exception, and their pharmacophore is now actively used in
The included peptides can be divided into three separate
sublibraries: active dipeptides (1-3), amidated tripeptides (4-
7), and highly active benzylamidated tripeptides (8-17). The
last also contains peptides with confused sequences (16 and 17).
Four structurally similar inactive di- and tripeptides were also
included as references to provide positive controls.
hPEPT1 Interactions. As a member of the proton dependent
oligopeptide transporter family the physiological role of hPEPT1
is to enable transepithelial transport of di- and tripeptides. The
708 amino acid residue protein is mainly expressed in the intestine,
and several QSAR models have been developed to explain and
predict the transport of the numerous substrates.9,11,30-32 None
of the compounds included in the present study have previously
been evaluated as substrates for hPEPT1-mediated transport.
Table 1. Molecular Properties and Antibacterial Activities of
the Included Peptides
MIC (μg/mL)c
peptide
sequence
TbtR-NH2
mass bulka chargeb S. aureus MRSA E. coli
1
2
3
4
5
6
7
8
9
527.4
411.2
631.5
515.3
552.3
496.3
552.3
773.5
586.4
566.3
642.4
642.4
1
1
2
1
1
1
1
2
2
2
2
2
4
2
3
3
2
þ2
þ2
þ2
þ3
þ3
þ3
þ3
þ3
þ3
þ3
þ3
þ3
þ3
þ3
þ3
þ3
þ3
15
15
ndd >200
2.5 10
>150
BipR-OCH3
TbtR-CH2Bn
RWR-NH2
150
2.5
>200 >200 >200
200 50
ndd
>200 >200 ndd
RBipR-NH2
KBipK-NH2
RDipR-NH2
RTbtR-NHBn
KBipK-NHBn
100
2.5
15
>150 >150
2.5
25
75
5
7.5
Table 3. Transport Generated by the Control Peptides (All
Concentrations 1 mM)
>150
>150
100
10 RFR-NHBn
11 RBipR-NHBn
12 RDipR-NHBn
150
5
flux rate,a
flux rate,a
sequence
Gly-Sar
test substance
flux rateb (%) (n)
25
10
5
>150
75
13 GppBipGpp-NHBn 736.4
<2.5
75
WR-OH
14.5
0.3
43.5 ( 2.2 (3)
1.2 ( 1.0 (3)
10.7 ( 1.5 (3)
14 RWR-NHBn
15 RWGpp-NHBn
16 RGppW-NHBn
17 RRW-NHBn
605.4
653.4
653.4
605.4
25
10
15
25
>150
>150
>150
>150
WR-NHBn
33.4
50
BipR-OH
3.6
50
a Flux rates expressed in pmol/(s cm2) were calculated from the
3
75
transport-associated currents measured at -60 mV membrane potential.
For the calculation, the surface area of an oocyte was estimated at 3.80
mm2 (corresponding to a diameter of 1.1 mm). b Compared to the
transport of Gly-Sar.
a Number of isolated hydrophobic units. b At physiological pH. c Mini-
mal inhibitory concentration determined by mico dilution experiments.
d Not determined.
Table 2. Inhibition of Dipeptide-Induced Current
flux rate,a
flux rate,a
peptide
sequence
Gly-Sar 1 mM
Gly-Sar 1 mM þ test substance
inhibition (%) (n)
IC50 b (mM)
1
TbtR-NH2
BipR-OMe
RWR-NH2
RBipR-NH2
RWR-NHBn
43.8
64.0
72.2
56.4
27.4
34.8
57.2
61.8
56.7
3.8
20.3 ( 6.3 (4)
10.4 ( 2.3 (5)
15.1 ( 6.1 (4)
-3.3 ( 8.3 (4)
87.6 ( 0.5 (4)
0.6
ncc
2.7
ncc
ncc
2
4
5
14
a Flux rates expressed in pmol/(s cm2) were calculated from the transport-associated currents measured at -60 mV membrane potential. For the
3
calculation, the surface area of an oocyte was estimated at 3.80 mm2 (corresponding to a diameter of 1.1 mm). Differences were due to various surface
densities of hPEPT1. b The IC50 values were calculated from the mean inhibitions for those substrates where data could reasonably be fitted to the
Michaelis-Menten kinetics. c Not calculated because of poor fit.
2425
dx.doi.org/10.1021/jm1015704 |J. Med. Chem. 2011, 54, 2422–2432

Journal of Medicinal Chemistry
ARTICLE
Figure 3. (A-C) Current-voltage relations of Gly-Sar (1 mM, pH 6.5) in the absence and presence of increasing concentrations of RWR-NH2 (4),
TbtR-NH2 (1), and RWR-NHBn (14), respectively, and of the inhibitors without added Gly-Sar (in A and B). The values presented are the mean of four
independent measurements. The data shown in Table 2 were calculated from the current values recorded at -60 mV membrane potential. (D)
Current-voltage relations of the substrates WR-OH, WR-NHBn, and BipR-OH in comparison with the transport of Gly-Sar, a known substrate of
hPEPT1. All compounds were analyzed at 1 mM at pH 6.5. The values presented are the mean of three independent measurements. The data shown in
Table 3 were calculated from the current values recorded at -60 mV membrane potential.
Such a study does not therefore only answer important questions
about the likely absorption via hPEPT1 but also provides insight
into the mechanism of transport and substrate specificity of the
peptide transporter, as the peptides contain numerous unnatural
elements. The TEVC technique provides a direct, fast, and
sensitive measurement of the functional properties of this class
of transporters, allowing measurements to be carried out under
well-defined voltage conditions.33
The remaining three peptides (1, 2, and 4) all inhibited the
dipeptide transport by 10-20% when employed at 1 mM. Only
peptides 1 and 4 allowed for proper IC50 calculations and
revealed values of 0.6 and 2.7 mM, respectively. They may
therefore be classified as inhibitors with medium affinity.35 None
of the peptides could, however, induce a transport current,
indicating that they do not serve as substrates and subsequently
cannot be taken up into the enterocytes via hPEPT1.
It has earlier been shown that affinity to hPEPT1 does not
necessarily translate into transport behavior but can also reflect
inhibitory activity.34 Such behavior is seen in the present study
where four of the tested peptides display binding to hPEPT1 with
modest affinities but are not transported as shown in Table 2.
The promising dipeptide 3 and tripeptide 11 could not be
analyzed because of their membrane disruptive activity at all
concentrations employed. Both peptides 3 and 11 are highly
bioactive, and using them at concentrationsof >200 ꢀ their MIC
is impossible with this technique. Of the five remaining peptides
fulfilling the pharmacophore, only peptide 5 did not display any
binding to hPEPT1. Out of the remaining four compounds,
peptide 14 was the only one that produced a considerable
inhibition of the dipeptide-induced current. However, this sub-
stance inhibited to a somewhat lower degree also the current
induced by the glucose analogue R-methyl D-glucopyranoside
(RMDG) via the sodium-coupled glucose transporter SGLT1
used as selectivity control. This suggests that peptide 14 can also
impair membrane protein functions in an unspecific manner.
Moreover, the degree of inhibition as a function of concentration
of peptide 14 was nearly linear; i.e., it did not seem to comply
with the expected protein-mediated kinetics (data not shown).
The three reference peptides all displayed electrogenic trans-
port but at different rates as presented in Table 3 and Figure 3.
The natural dipeptide WR-OH yielded transport currents but at
lower rates than Gly-Sar, possibly because of a lower affinity
(inhibition studies were not performed). Addition of a C-term-
inal benzylamide (WR-NHBn), which is a common strategy for
increasing the antibacterial activity, nearly abolished transport. In
contrast, the replacement of the tryptophan of WR-OH with the
unnatural amino acid biphenylalanine (BipR-OH) retained a low
residual transport activity. The biphenyl side chain of Bip is
similar in accessible areas (625 Å2 compared to 578 Å2) to
tryptophan but is significantly longer and more rigid. It is also
more hydrophobic, and those two factors may explain the
difference in residual transport. Biphenylalanine has previously
been included as N-terminal amino acid in BipL-OH, a peptide
that was shown to be transported at a low rate but displayed a
high affinity, illustrating that the Bip placement within the
sequence could also be a contributing factor.36 To our knowledge
the most similar peptide previously investigated is KWK-OH by
Andersen et al. That peptide displayed an affinity at 3.7 (
0.9 mM for hPEPT1.9 Only very few studies on peptides with
C-terminal arginines have been performed, and DIR-OH and
2426
dx.doi.org/10.1021/jm1015704 |J. Med. Chem. 2011, 54, 2422–2432

Journal of Medicinal Chemistry
ARTICLE
Table 4. Theoretical and Experimental Peptide Hydrophobicity Together with Experimental Apparent Passive Permeability and a
Prediction of Oral Absorption in Humans Based on the Experimental Data and Theoretical Calculations
predicted oral absorption
peptide
sequence
ClogP a
tR b (min)
Papp (10-6 cm/s)
experimentald
theoreticala
1
TbtR-NH2
1.597
1.304
16.3
13.7
18.1
3.15
15.6
13.8
7.6
0.43 ( 0.05
1.06 ( 0.17
5.27 ( 1.83
2.37 ( 0.55
0.01 ( 0.01
0.06 ( 0.03
0.83 ( 0.13
ndc
moderate
excellent
excellent
excellent
poor
7.2
24.2
53.0
0.0
0.0
0.0
0.0
5.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
2
BipR-OCH3
TbtR-CH2Bn
RWR-NH2
3
4.188
4
-2.995
-1.548
-0.852
-2.035
2.590
5
RBipR-NH2
6
KBipK-NH2
RDipR-NH2
RTbtR-NHBn
KBipK-NHBn
RFR-NHBn
poor
7
moderate
ndc
8
19.0
11.0
8.9
9
1.600
0.26 ( 0.07
2.31 ( 0.71
0.01 ( 0.01
0.42 ( 0.10
0.90 ( 0.27
0.77 ( 0.28
0.47 ( 0.11
0.38 ( 0.17
0.96 ( 0.25
moderate
excellent
poor
10
11
12
13
14
15
16
17
-0.816
0.944
RBipR-NHBn
RDipR-NHBn
GppBipGpp-NHBn
RWR-NHBn
RWGpp-NHBn
RGppW-NHBn
RRW-NHBn
reference drugs
metoprolol
11.3
10.6
11.7
9.2
0.389
moderate
excellent
moderate
moderate
moderate
excellent
1.647
-0.904
1.525
9.5
-0.012
-0.941
9.3
9.3
1.884
0.094
3.23 ( 0.78e
0.40 ( 0.05e
0.03 ( 0.01e
excellent
moderate
poor
89.5
63.0
51.5
terbutaline
chlorothiazide
-0.421
a Calculated using QikProp. b Retention time on C18 column. The compound differs too much in retention to allow isocratic elution, and different
gradients have been applied to the separate libraries. The retention times should thus only be compared within the libraries c Not determined. Peptide
disrupted membrane barrier d According to definition in ref 22. e From ref 22.
DRR-OH recently investigated by Omkvist et al. displayed no or
very low affinities (21 mM).37 Previous work (unpublished)
from the Freising laboratory also demonstrated a low affinity or
even lack of binding for peptides carrying a cationic C-terminus.
Furthermore, the data of Vig and co-workers show a nearly
complete shutdown of transport activity upon C-terminal amida-
tion both for the investigated tripeptide GGG-NH2 and for
several dipeptides such as GF-NH2, YG-NH2, and YY-NH2.34
FA-NH2 was the only amidated peptide out of eight in that study
to display a weak affinity for hPEPT1. Thus, both the C-terminal
arginine and its amidation seem to counteract transport via
hPEPT1. The fact that the incorporation of an unnatural hydro-
phobic alternative to W also had a negative effect on transport
strongly indicates that active peptides fulfilling the pharmacophore
for antibacterial activity, schematically illustrated in Figure 6, do
not represent viable substrates for active transport by hPEPT1.
Passive Permeability. No or low active uptake does not
prevent a drug candidate from being orally active. Since the short
peptides are all adhering reasonably to Lipinki's “rule of 5” with
regard to size and polarity, efficient permeation appears feasible.4
To assess this, a passive diffusion study of the entire peptide
library was subsequently initiated using the PVPA method. PVPA
is a recently developed robust model that has been shown to
provide permeability data that correlate well with in vivo data on
fraction absorbed in humans, to the same extent as the Caco-2
cell model and even better than PAMPA models.22 In addition
to the analysis of drugs, PVPA has also be been successfully used
to investigate formulations of poorly soluble drugs containing
surfactants.38 The peptides employed here are all relatively
hydrophobic, and a transcellular uptake is the anticipated passive
route. The passive uptake behavior of larger water-soluble
pharmaceutical peptides through the intestinal epithelium via
the paracellular route has recently been reported, and an alter-
native paracellular transport via the anionic tight junctions
cannot be ruled out for the most polar of the included peptides
in an in vivo situation.39,40
The permeation data were also compared with theoretical
values for uptake generated by QikProp.41,42 Both Qikprop and
the assay are able to predict classification of small drugs according
to in vivo oral absorption. The potential secondary structures of
peptides can, however, have an unexpected impact on the
permeability that is difficult to anticipate as has been seen in
several transport studies.43,44 To verify good compliance with
in vivo data, the passive transport of three commercial drugs
metoprolol (excellent absorption), terbutaline (moderate absorp-
tion), and chlorothiazide (poor absorption) was investigated
using the two methods (Table 4). Both methods showed good
correlation with a slightly too high theoretical uptake for chlorothia-
zide compared to 13-25% in vivo absorption earlier reported.22
The three active dipeptides (1-3) all crossed the phospho-
lipid vesicle-based barrier at substantial rates. Peptide 3 was
efficiently transported across the barrier at 5.27 ꢀ 10-6 cm/s,
predicting a ranking as excellently absorbed in humans. The
predicted absorption of peptide 2 was also classified as “excel-
lent” according to the earlier definition, while peptide 1 only
displayed moderate transport. This pattern is also suggested by
QikProp albeit with slightly lower absorption values. These three
peptides are all rather hydrophobic, as evident by both their C18
2427
dx.doi.org/10.1021/jm1015704 |J. Med. Chem. 2011, 54, 2422–2432

Journal of Medicinal Chemistry
ARTICLE
Figure 4. Graphical representation of the apparent passive permeation rates of the peptides using the PVPA.
control peptide was permeating significantly faster than any other
tested compound across the barrier. Peptides high in arginine are
known to transverse membranes, and several studies have been
devoted to develop polyarginine transmembrane carriers for
challenging drug molecules.45,46 The mechanism behind this
effect is not clear, and suggestions ranging from active transport,
through facilitated carrier mediated transport, to passive trans-
cellular transport have been proposed.45-47 Although the opti-
mal polyarginine chains are four to eight units,48 it seems that a
similar effect is seen for our short peptides containing two
arginines together with a free N-terminus and little hydrophobic
bulk. Clearly two or more mechanisms are simultaneously acting
to decide whether permeation across a membrane of these
peptides is favored or not. The theoretical model did not
anticipate a transport for any of these tripeptides.
The third library is composed of the type of modified tripep-
tides known for yielding the most efficient antibacterials with
compounds very similar to those in phase II clinical trials.14 With
the exception of peptide 11, remaining peptides within library 3
were either moderately or excellently transported. Peptide 8
disrupted the integrity of the barriers and could not be analyzed
in high enough concentrations (200 ꢀ MIC). This effect has
previously been seen for an active peptide in our preliminary
study.16 When plotting the retention time against permeation for
the nine peptides (peptide 8 not included) investigated in library
3, a trend of increasing permeability with decreasing retention
time could be seen (Figure 7). The only peptide not adhering to
this trend is 13 with its two unnatural 4-guanidinophenyl side
chains as arginine mimics. The confused sequences 16 and 17
displayed near to identical permeation values compared with
their “native” counterparts 14 and 15, suggesting that the overall
hydrophobicity and not sequence is the main factor for these
peptides. Again, the theoretical model was not able to predict the
transport seen experimentally for this library. The dependence
on hydrophobicity for activity also results in a close inverse
correlation between the passive permeability and the
Figure 5. Passive transport as a function of the antimicrobial effect of
the peptides from library 3 against Staphylococcus aureus ATCC 25923.
Peptide 13 is omitted from the presented regression analysis. R2 with
peptide 13 included is 0.73.
retention times and their ClogP values, and it seems that the very
bulky Tbt unit of peptide 3 is providing the extra hydrophobic
push, setting that peptide apart from the other ones.
Tripeptides 4-7 lack a bulky C-terminal modification and are
thus less hydrophobic than the other tripeptides. This is clearly
reflected in their lower antimicrobial activity as shown in Table 1.
The transport properties of these four peptides are distinctly
different from the dipeptides with the most polar peptide 4
displaying the highest diffusion rate. In fact the most hydro-
phobic peptides 5 and 6 are hardly diffusing through the
liposome barrier at all. To further investigate this unforeseen
trend, an inactive highly polar tripeptide (RAR-NH2, ClogP = -
4.705) was prepared (not included in Table 4) and investigated.
With an apparent permeability value of 12.0 ꢀ 10-6 cm/s this
2428
dx.doi.org/10.1021/jm1015704 |J. Med. Chem. 2011, 54, 2422–2432

Journal of Medicinal Chemistry
ARTICLE
Figure 6. Schematic representation of the structural features employed to influence the properties of the included peptides, here exemplified with peptide 11.
are exerting their effect to different degrees depending on the
peptide type, and the large differences between the different
groups make interlibrary comparisons problematic. Hydropho-
bicity was shown to be the main determinant, while no correla-
tion between molecular mass and transport was evident (see
Supporting Information). It cannot be ruled out that the ability of
the peptides to adopt different configurations in solution and in a
membrane may have a large impact on the passive transport and
also the bioactivity. It has previously been shown that secondary
structure exerts a considerable influence on the paracellular
transport of more polar peptides.39 Such an effect would increase
with peptide size and could explain the unpredicted results and
the fact that QikProp did not anticipate any uptake of the
tripeptides while the smaller reference drugs and dipeptides
fitted well with the predictions. What is nevertheless clear is that
most of the analyzed short cationic antimicrobial peptides in fact
display a significant passive transport correlating to a moderate or
excellent oral absorption and, based on those observations, may
be suited for oral administration.
Figure 7. Passive transport (Papp) correlated with experimental C18
column retention. The transport of the more polar peptides is favored
with the notable exception of peptide 13 incorporating Gpp instead of
Arg as cationic amino acids.
This is the first reported comprehensive study of the uptake of
short membrane active cationic peptides, and while it has
generated detailed insights into their transport modes it also
highlights challenges associated with the methods employed.
The first one is peptide solubility, and the second one is lysis at
the concentrations employed. The peptides are designed to be
selective against anionic membranes but also display lytic proper-
ties against eukaryotic and neutral membranes albeit at signifi-
cantly higher concentrations. Such a general cytotoxic effect is
important to consider even though these and similar peptides have
been shown to display large selectivity factors.14,19 This effect is
apparent in the hPEPT1 study where both peptides 3 and 11 had
to be excluded because of oocyte membrane damage at even the
lowest employed concentration. The PVPA was more tolerant
and allowed for analysis of peptides 3 and 11, while peptide 8
influenced the barrier integrity. An uptake of peptide 8 cannot be
ruled out at lower concentrations, although it is unlikely if it
adheres to the trend of the other peptides in library 3. Even
though the peptides included represent particularly challenging
molecules to evaluate because of their membrane activity, only
one had to be excluded, which indicates that the PVPA is a robust
antimicrobial activity of the active benzylamidated tripeptides
illustrated in Figure 5.
The fact that the polar peptides are more rapidly crossing the
barrier may indicate that they are less likely to dwell inside the
barrier matrix compared to the more hydrophobic active ones.
Such behavior may allow an insight into the mode of action of
this class of peptides which remain elusive and still under debate.
A low Papp, seen for the more bioactive peptides, would, if they
first enter the membrane, permit them to reside longer inside a
biological membrane and could be a contributing factor enabling
or causing the membrane destabilization believed to be the main
route to cell lysis. It is also important to realize that such a trend
may prevent an efficient uptake of the most antibacterial drug
candidates, although several of the active peptides displayed a
significant uptake. The fact that peptides 3 and 13 display both
a high uptake and a high antibacterial effect shows that such a
combination is possible.
Collectively it is clear that the generality of the observations is
low even for such a homogeneous group of peptides. The
physicochemical properties responsible for barrier permeation
2429
dx.doi.org/10.1021/jm1015704 |J. Med. Chem. 2011, 54, 2422–2432

Journal of Medicinal Chemistry
ARTICLE
screening assay. QikProp was used to assess if the theoretical
properties predicted could be generally applicable to these
unusual druglike compounds. Even though the theoretical ClogP
values are not absolute descriptors of hydrophobicity for ioniz-
able compounds, they still correlated reasonably well with the
experimental hydrophobicity determined chromatographically
(see Supporting Information). The peptides ranged significantly
in retention, and each library had to be investigated using a
separate linear gradient. QikProp predicted the oral absorption of
the three reference drugs with some accuracy and was also able to
reasonably predict well the absorption, according to the in vitro
passive permeability, of the active dipeptides 1-3. For libraries 2
and 3, however, no oral absorption was anticipated. This is in
clear contrast to the experimental data and implies that the
general applicability of QikProp is limited, at least with regard to
this compound class where secondary structure may also play a
decisive role.
the final oocyte collection. Oocytes were treated with 2.5 mg/mL
collagenase for 70 min and were separated manually thereafter. They
were incubated in Barth’s solution containing (in mM) NaCl 88, KCl 1,
MgSO4 0.8, CaCl2 0.4, Ca(NO3)2 0.3, NaHCO3 2.4, HEPES 10 (pH
7.5) at 17 ꢀC overnight. Next day stage V/VI oocytes were injected with
about 25 ng of hPEPT1 cRNA in 18-27 nL and incubated for another
3-5 days at 17 ꢀC.49
Electrophysiology. TEVC experiments were performed as de-
scribed previously.33 Briefly, the oocyte was placed in an open chamber
(∼0.5 mL total volume) and continuously superfused (∼3 mL/min) at
room temperature with Barth’s solution in which HEPES was replaced
by MES (pH 6.5). Oocytes were voltage-clamped at -60 mV using a
TEC-05 amplifier (NPI Electronic, Tamm, Germany) and current-
voltage (I/V) relations were measured using short (100 ms) pulses
separated by 200 ms pauses in the potential range of -100 to -20 mV.
I/V measurements were made immediately before and 20-30 s after
substrate application when current flow reached steady state. The current
evoked at a given membrane potential was calculated as the difference
between the currents measured in the presence and the absence of
substrate. Positive currents denote positive charges flowing out of the
oocyte. The inhibitory effect of the selected substances was tested by
comparison of currents generated by the peptide transporter substrate
Gly-Sar at 1 mM in the absence and presence of the putative inhibitor.
Preparation of the PVPA Barriers. The phospholipid vesicle-
based barriers were prepared and stored according to earlier reported
procedures.22,50 In brief, liposome dispersions extruded through filters
with pore sizes of 800 nm and 800 þ 400 nm, respectively, were
deposited on a filter support by use of centrifugation. The liposomes
were added in consecutive steps, first the smaller liposomes allowing
them to deposit inside the pores of the filter and then the larger ones to
layer on top of the filter. Freeze-thaw cycling was then performed to
promote fusion of the liposomes and hence produce a tight permeation
barrier.
Permeability Experiments. Permeability experiments were per-
formed at room temperature without agitation. Inserts containing the
different peptides dissolved in phosphate buffer, pH 7.4, were moved at
certain time intervals to wells containing equal quantity of fresh
phosphate buffer, pH 7.4.50 UV absorbance (Spectramax 190; Molecular
devices, Molecular Device Corporation, Sunnyvale, CA, U.S.) at wave-
lengths suitable for the different peptides was used to quantify the
amount of peptides in the acceptor compartments. The electrical
resistance across the lipid barriers was measured (Millicell-ERS, Milli-
pore, Billerica, MA, U.S.) immediately after completion of the permea-
tion studies to control the integrity of the barriers. Barriers displaying a
resistance below 1000 Ω were excluded from the study. Each peptide
was tested in at least eight parallels to ensure reliable data.
’ CONCLUSION
The goal of the present study was to assess whether short
cationic antimicrobial peptides are able to be transported by
hPEPT1 which mediates efficient uptake of di- and tripetides in
the mammalian intestine and whether they permeate passively
across cell membranes. It is shown that the strategies employed
to generate peptides with high biological activity, such as the
introduction of large hydrophobic unnatural amino acids and
C-terminal amidation, counteract the transport capability for
hPEPT1-mediated uptake. While some of the included peptides
displayed moderate affinity for the transporter, none of them
revealed significant transport by it. On the other hand, the PVPA
study aimed at investigating the passive permeation properties
revealed high permeation rates, suggesting a moderate to ex-
cellent oral absorption for most peptides. The factor mainly
governing the permeation rate was the hydrophobicity but the
structural features of the compounds are also important, making
predictions difficult. It is thus clear that while the promising
compounds are no proper substrates for hPEPT1-mediated
uptake, they do offer, next to a high bioactivity, the potential
for oral absorption via passive transcellular mechanisms. Those
findings, in combination with their significant metabolic stabi-
lities, make them attractive targets to pursue. A correlation
between membrane permeation rates and biological activity
may also provide insight into the mechanisms by which this class
of compounds acts as antimicrobials that are efficient against
bacteria resistant to conventional antibiotics.
Theoretical Predictions. QikProp (Schr€odinger, New York, NY)
was used to calculate the ClogP values for the peptides and their
theoretical oral absorption.41 All the peptides were within the structural
limits of the program.
’ MATERIALS AND METHODS
Materials. Egg phosphatidylcholine (lipoid E-80) was obtained
from Lipoid, Ludwigshafen, Germany. Filter inserts (Transwell, d =
6.5 mm) and plates were purchased from Corning Inc., Lowell, MA, U.
S., and the mixed cellulose ester filters (0.65 μm pore size) from
Millipore, Billerica, MA, U.S.. Xenopus laevis frogs were supplied by
the African Xenopus Facility, Knysna, R.S.A. Standard chemicals and
solvents were supplied by Sigma-Aldrich.
Peptide Preparation. Peptides (4, 7, 9, 13, and the four reference
peptides) were prepared as described earlier.16,19 See Supporting Informa-
tion for detailed description on chemicals used, preparation, purification,
characterization, and antimicrobial testing of all included peptides.
Oocyte Preparation. Xenopus laevis oocytes were collected under
anesthesia (immersion in a solution of 0.7 g/L 3-aminobenzoic acid
ethyl ester) from frogs that were killed with an anesthetic overdose after
’ ASSOCIATED CONTENT
S
Supporting Information. Synthetic procedures, purity
b
data, HPLC traces, 1H NMR spectra for all the tested peptides,
plots of ClogP vs C18 retention, antimicrobial activity vs C18
retention, and molecular weight vs Papp. This material is available
free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
*Phone: þ 47 776 45505. Fax: þ 47 776 44765. E-mail: johan.
svenson@uit.no.
2430
dx.doi.org/10.1021/jm1015704 |J. Med. Chem. 2011, 54, 2422–2432

Journal of Medicinal Chemistry
ARTICLE
’ ACKNOWLEDGMENT
Expression cloning of a mammalian proton-coupled oligopeptide trans-
porter. Nature 1994, 368, 563–566.
(13) Daniel, H.; Morse, E. L.; Adibi, S. A. Determinants of substrate
affinity for the oligopeptide/Hþ symporter in the renal brush-border
membrane. J. Biol. Chem. 1992, 267, 9565–9573.
(14) Haug, B. E.; Stensen, W.; Kalaaji, M.; Rekdal, O.; Svendsen, J. S.
Synthetic antimicrobial peptidomimetics with therapeutic potential.
J. Med. Chem. 2008, 51, 4306–4314.
Professor Georg Nagel at the University of W€urzburg is
gratefully acknowledged for providing Xenopus oocyte images.
The Norwegian Structural Biology Centre (NorStruct) is sup-
ported by the Functional Genomics Program (FUGE) of the
Research Council of Norway.
(15) Haug, B. E.; Stensen, W.; Stiberg, T.; Svendsen, J. S. Bulky
nonproteinogenic amino acids permit the design of very small and
effective cationic antibacterial peptides. J. Med. Chem. 2004, 47,
4159–4162.
(16) Svenson, J.; Karstad, R.; Flaten, G. E.; Brandsdal, B. O.; Brandl,
M.; Svendsen, J. S. Altered activity and physicochemical properties of
short cationic antimicrobial peptides by incorporation of arginine
analogues. Mol. Pharmaceutics 2009, 6, 996–1005.
(17) Svenson, J.; Stensen, W.; Brandsdal, B. O.; Haug, B. E.; Monrad,
J.; Svendsen, J. S. Antimicrobial peptides with stability toward tryptic
degradation. Biochemistry 2008, 47, 3777–3788.
(18) Svenson, J.; Vergote, V.; Karstad, R.; Burvenich, C.; Svendsen,
J. S.; De Spiegeleer, B. Metabolic fate of lactoferricin-based antimicrobial
peptides: effect of truncation and incorporation of amino acid analogs on
the in vitro metabolic stability. J. Pharm. Exp. Ther. 2010,
332, 1032–1039.
’ ABBREVIATIONS USED
hPEPT1, human intestinal peptide transporter 1; PVPA, phos-
pholipid vesicle-based permeation assay; TEVC, two-electrode
voltage clamp; RMDG, R-methyl D-glucopyranoside; Gly-Sar,
glycylsarcosine; PAMPA, parallel artificial membrane permeabi-
lity assay; Bn, benzyl; GFP, green fluorescent protein; MIC,
minimal inhibitory concentration; RP-HPLC, reversed phase high
performance liquid chromatography; Bip, biphenylalanine; Dip, di-
phenylalanine; Gpp, L-2-amino-3-(4-guanidinophenyl)propanoic
acid; ATCC, American Type Culture Collection;Papp, apparent
permeability
’ REFERENCES
(19) Karstad, R.; Isaksen, G.; Brandsdal, B. O.; Svendsen, J. S.;
Svenson, J. Unnatural amino acid side chains as S1, S10, and S20 probes
yield cationic antimicrobial peptides with stability toward chymotryptic
degradation. J. Med. Chem. 2010, 53, 5558–5566.
(1) Dobson, P. D.; Kell, D. B. Carrier-mediated cellular uptake of
pharmaceutical drugs: an exception or the rule?. Nat. Rev. Drug Discovery
2008, 7, 205–220.
(2) Dobson, P. D.; Lanthaler, K.; Oliver, S. G.; Kell, D. B. Implica-
tions of the dominant role of transporters in drug uptake by cells. Curr.
Top. Med. Chem. 2009, 9, 163–181.
(3) Sugano, K.; Kansy, M.; Artursson, P.; Avdeef, A.; Bendels, S.; Di,
L.; Ecker, G. F.; Faller, B.; Fischer, H.; Gerebtzoff, G.; Lennernaes, H.;
Senner, F. Coexistence of passive and carrier-mediated processes in drug
transport. Nat. Rev. Drug Discovery 2010, 9, 597–614.
(4) Lipinski, C. A.; Lombardo, F.; Dominy, B. W.; Feeney, P. J.
Experimental and computational approaches to estimate solubility and
permeability in drug discovery and development settings. Adv. Drug
Delivery Rev. 1997, 23, 3–25.
(20) Hamman, J. H.; Enslin, G. M.; Kotze, A. F. Oral delivery of
peptide drugs: barriers and developments. BioDrugs 2005, 19, 165–177.
(21) Giannis, A. Peptidomimetics for receptor ligands discovery,
development, and medical perspectives. Angew. Chem. 1993,
32, 1244–1267.
(22) Flaten, G. E.; Dhanikula, A. B.; Luthman, K.; Brandl, M. Drug
permeability across a phospholipid vesicle based barrier: a novel
approach for studying passive diffusion. Eur. J. Pharm. Sci. 2006,
27, 80–90.
(23) Zasloff, M. Antimicrobial peptides of multicellular organisms.
Nature 2002, 415, 389–395.
(5) Payne, D. J.; Gwynn, M. N.; Holmes, D. J.; Pompliano, D. L.
Drugs for bad bugs: confronting the challenges of antibacterial discov-
ery. Nat. Rev. Drug Discovery 2007, 6, 29–40.
(24) Melo, M. N.; Ferre, R.; Castanho, M. OPINION antimicrobial
peptides: linking partition, activity and high membrane-bound concen-
trations. Nat. Rev. Microbiol. 2009, 7, 245–250.
(6) Bretschneider, B.; Brandsch, M.; Neubert, R. Intestinal transport
of beta-lactam antibiotics: analysis of the affinity at the Hþ/peptide
symporter (PEPT1), the uptake into Caco-2 cell monolayers and the
transepithelial flux. Pharm. Res. 1999, 16, 55–61.
(7) Temple, C. S.; Stewart, A. K.; Meredith, D.; Lister, N. A.;
Morgan, K. M.; Collier, I. D.; Vaughan-Jones, R. D.; Boyd, C. A. R.;
Bailey, P. D.; Bronk, J. R. Peptide mimics as substrates for the intestinal
peptide transporter. J. Biol. Chem. 1998, 273, 20–22.
(25) Yeaman, M. R.; Yount, N. Y. Mechanisms of antimicrobial
peptide action and resistance. Pharmacol. Rev. 2003, 55, 27–55.
(26) Bremner, J. B.; Keller, P. A.; Pyne, S. G.; Boyle, T. P.; Brkic, Z.;
David, D. M.; Garas, A.; Morgan, J.; Robertson, M.; Somphol, K.; Miller,
M. H.; Howe, A. S.; Ambrose, P.; Bhavnani, S.; Fritsche, T. R.;
Biedenbach, D. J.; Jones, R. N.; Buckheit, R. W.; Watson, K. M.; Baylis,
D.; Coates, J. A.; Deadman, J.; Jeevarajah, D.; McCracken, A.; Rhodes,
D. I. Binaphthyl-based dicationic peptoids with therapeutic potential.
Angew. Chem. 2010, 49, 537–540.
(27) Hansen, T.; Alst, T.; Havelkova, M.; Strom, M. B. Antimicrobial
activity of small beta-peptidomimeties based on the pharmacophore
model of short cationic antimicrobial peptides. J. Med. Chem. 2010,
53, 595–606.
(28) Strom, M. B.; Haug, B. E.; Skar, M. L.; Stensen, W.; Stiberg, T.;
Svendsen, J. S. The pharmacophore of short cationic antibacterial
peptides. J. Med. Chem. 2003, 46, 1567–1570.
(29) Haug, B. E.; Stensen, W.; Svendsen, J. S. Application of the
Suzuki-Miyaura cross-coupling to increase antimicrobial potency gen-
erates promising novel antibacterials. Bioorg. Med. Chem. Lett. 2007,
17, 2361–2364.
(8) 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.
(9) Andersen, R.; Jorgensen, F. S.; Olsen, L.; Vabeno, J.; Thorn, K.;
Nielsen, C. U.; Steffansen, B. Development of a QSAR model for binding
of tripeptides and tripeptidomimetics to the human intestinal di-/
tripeptide transporter hPEPT1. Pharm. Res. 2006, 23, 483–492.
(10) Vabeno, J.; Nielsen, C. U.; Ingebrigtsen, T.; Lejon, T.; Steffansen,
B.; Luthman, K. Dipeptidomimetic ketomethylene isosteres as pro-
moieties for drug transport via the human intestinal di-/tripeptide
transporter hPEPT1: design, synthesis, stability, and biological investiga-
tions. J. Med. Chem. 2004, 47, 4755–4765.
(11) 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.
2000, 39, 506–508.
(12) Fei, Y. J.; Kanai, Y.; Nussberger, S.; Ganapathy, V.; Leibach,
F. H.; Romero, M. F.; Singh, S. K.; Boron, W. F.; Hediger, M. A.
(30) Brandsch, M. Transport of drugs by proton-coupled peptide
transporters: pearls and pitfalls. Expert Opin. Drug Metab. Toxicol. 2009,
5, 887–905.
(31) Biegel, A.; Gebauer, S.; Hartrodt, B.; Brandsch, M.; Neubert, K.;
Thondorf, I. Three-dimensional quantitative structure-activity rela-
tionship analyses of beta-lactam antibiotics and tripeptides as substrates
2431
dx.doi.org/10.1021/jm1015704 |J. Med. Chem. 2011, 54, 2422–2432

Journal of Medicinal Chemistry
ARTICLE
of the mammalian Hþ/peptide cotransporter PEPT1. J. Med. Chem.
2005, 48, 4410–4419.
barrier structure, storage stability and stability towards pH changes. Eur.
J. Pharm. Sci. 2006, 28, 336–343.
(51) Flaten, G. E. The Phospholipid Vesicle-Based Barrier: A Novel
Method for Passive Drug Permeability Screening. Ph.D. Thesis, Uni-
versity of Tromsø, Tromsø, Norway, 2007.
(32) Larsen, S. B.; Omkvist, D. H.; Brodin, B.; Nielsen, C. U.;
Steffansen, B.; Olsen, L.; Jorgensen, F. S. Discovery of ligands for the
human intestinal di-/tripeptide transporter (hPEPT1) using a QSAR-
assisted virtual screening strategy. ChemMedChem 2009, 4, 1439–1445.
(33) Wagner, C. A.; Friedrich, B.; Setiawan, I.; Lang, F.; Broer, S.
The use of Xenopus laevis oocytes for the functional characterization of
heterologously expressed membrane proteins. Cell. Physiol. Biochem.
2000, 10, 1–12.
(34) 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.
(35) Brandsch, M.; Knutter, I.; Leibach, F. H. The intestinal Hþ/
peptide symporter PEPT1: structure-affinity relationships. Eur.
J. Pharm. Sci. 2004, 21, 53–60.
(36) Knutter, I.; Hartrodt, B.; Toth, G.; Keresztes, A.; Kottra, G.;
Mrestani-Klaus, C.; Born, I.; Daniel, H.; Neubert, K.; Brandsch, M.
Synthesis and characterization of a new and radiolabeled high-affinity
substrate for Hþ/peptide cotransporters. FEBS J. 2007, 274, 5905–5914.
(37) Omkvist, D. H.; Larsen, S. B.; Nielsen, C. U.; Steffansen, B.;
Olsen, L.; Jørgensen, F. S.; Brodin, B. A quantitative structure-activity
relationship for translocation of tripeptides via the human proton-coupled
peptide transporter, hPEPT1 (SLC15A1). AAPS J. 2010, 385–396.
(38) Kanzer, J.; Tho, I.; Flaten, G. E.; Magerlein, M.; Holig, P.;
Fricker, G.; Brandl, M. In-vitro permeability screening of melt extrudate
formulations containing poorly water-soluble drug compounds using
the phospholipid vesicle-based barrier. J. Pharm. Pharmacol. 2010,
62, 1591–1598.
(39) Salamat-Miller, N.; Johnston, T. P. Current strategies used to
enhance the paracellular transport of therapeutic polypeptides across the
intestinal epithelium. Int. J. Pharm. 2005, 294, 201–216.
(40) Foger, F.; Kopf, A.; Loretz, B.; Albrecht, K.; Bernkop-Schnurch,
A. Correlation of in vitro and in vivo models for the oral absorption of
peptide drugs. Amino Acids 2008, 35, 233–241.
(41) QikProp, version 3.3; Schr€odinger, LLC: New York, NY, 2010.
(42) Duffy, E. M.; Jorgensen, W. L. Prediction of properties from
simulations: free energies of solvation in hexadecane, octanol, and water.
J. Am. Chem. Soc. 2000, 122, 2878–2888.
(43) Knipp, G. T.; Ho, N. F. H.; Barsuhn, C. L.; Borchardt, R. T.
Paracellular diffusion in Caco-2 cell monolayers: effect of perturbation
on the transport of hydrophilic compounds that vary in charge and size.
J. Pharm. Sci. 1997, 86, 1105–1110.
(44) Knipp, G. T.; Velde, D. G. V.; Siahaan, T. J.; Borchardt, R. T.
The effect of beta-turn structure on the passive diffusion of peptides
across Caco-2 cell monolayers. Pharm. Res. 1997, 14, 1332–1340.
(45) Wright, L. R.; Rothbard, J. B.; Wender, P. A. Guanidinium rich
peptide transporters and drug delivery. Curr. Protein Pept. Sci. 2003,
4, 105–124.
(46) Kirschberg, T. A.; VanDeusen, C. L.; Rothbard, J. B.; Yang, M.;
Wender, P. A. Arginine-based molecular transporters: the synthesis and
chemical evaluation of releasable Taxol-transporter conjugates. Org. Lett.
2003, 5, 3459–3462.
(47) Rothbard, J. B.; Garlington, S.; Lin, Q.; Kirschberg, T.; Kreider,
E.; McGrane, P. L.; Wender, P. A.; Khavari, P. A. Conjugation of arginine
oligomers to cyclosporin A facilitates topical delivery and inhibition of
inflammation. Nat. Med. 2000, 6, 1253–1257.
(48) Futaki, S.; Suzuki, T.; Ohashi, W.; Yagami, T.; Tanaka, S.; Ueda,
K.; Sugiura, Y. Arginine-rich peptides: an abundant source of membrane-
permeable peptides having potential as carriers for intracellular protein
delivery. J. Biol. Chem. 2001, 276, 5836–5840.
(49) Boll, M.; Markovich, D.; Weber, W. M.; Korte, H.; Daniel, H.;
Murer, H. Expression cloning of a cDNA from rabbit small-intestine
related to proton-coupled transport of peptides, beta-lactam antibiotics
and ACE-inhibitors. Pfluegers Arch. 1994, 429, 146–149.
(50) Flaten, G. E.; Bunjes, H.; Luthman, K.; Brandl, M. Drug perme-
ability across a phospholipid vesicle-based barrier-2. Characterization of
2432
dx.doi.org/10.1021/jm1015704 |J. Med. Chem. 2011, 54, 2422–2432