Phe-Gly Dipeptidomimetics Designed for the Di-/Tripeptide Transporters PEPT1 and PEPT2: Synthesis and Biological Investigations

Article abstract of DOI:10.1021/jm031022+

A series of five Phe-Gly dipeptidomimetics containing different amide bond replacements have been synthesized in a facile way from the readily available unsaturated ketoester 1, and their affinities for the di-/tripeptide transporters hPEPT1 (Caco-2 cells

Full text of DOI:10.1021/jm031022+

1060
J . Med. Chem. 2004, 47, 1060-1069
P h e-Gly Dip ep tid om im etics Design ed for th e Di-/Tr ip ep tid e Tr a n sp or ter s
P EP T1 a n d P EP T2: Syn th esis a n d Biologica l In vestiga tion s
J on Våbenø,† Tore Lejon,| Carsten Uhd Nielsen,‡ Bente Steffansen,‡ Weiqing Chen, Hui Ouyang,
Ronald T. Borchardt, and Kristina Luthman*,†,§
Department of Medicinal Chemistry, Institute of Pharmacy, University of Tromsø, N-9037 Tromsø, Norway,
Department of Chemistry, University of Tromsø, N-9037 Tromsø, Norway, Department of Pharmaceutics,
The Danish University of Pharmaceutical Sciences, Universitetsparken 2, DK-2100 Copenhagen, Denmark,
Department of Pharmaceutical Chemistry, The University of Kansas, Lawrence, Kansas 66047,
and Department of Chemistry, Medicinal Chemistry, Go¨teborg University, SE-412 96 Go¨teborg, Sweden
Received September 5, 2003
A series of five Phe-Gly dipeptidomimetics containing different amide bond replacements have
been synthesized in a facile way from the readily available unsaturated ketoester 1, and their
affinities for the di-/tripeptide transporters hPEPT1 (Caco-2 cells) and rPEPT2 (SKPT cells)
were tested. The compounds contained the amide bond isosteres ketomethylene (2a ), (R)- and
(S)-hydroxyethylidene (3a and 4a ), and (R)- and (S)-hydroxyethylene (5a and 6a ) to provide
information on the conformational and stereochemical requirements for hPEPT1 and rPEPT2
affinity. The affinity studies showed that for rPEPT2 there is no significant difference in affinity
between the ketomethylene isostere 2a and the natural substrate Phe-Gly (Ki values of 18.8
and 14.6 µM, respectively). Also the affinities for hPEPT1 are in the same range (Ki values of
0.40 and 0.20 mM, respectively). This corroborates earlier findings that the amide bond as
such is not essential for binding to PEPTX, but the results also reveal possible differences in
the binding of ketomethylene isosteres to hPEPT1 and rPEPT2. The trans-hydroxyethylidene
and hydroxyethylene isosteres proved to be poor substrates for PEPTX. These results provide
new information about the importance of flexibility and of the stereochemistry at the C4-position
for this class of compounds. Furthermore, the intracellular uptake of 2a -4a in Caco-2 cells
was investigated, showing a 3-fold reduction of the uptake of 2a in the presence of the
competetive inhibitor Gly-Pro, indicating contribution from an active transport component.
No active uptake of 3a and 4a was observed. Transepithelial transport studies also indicated
active transport of 2a across Caco-2 monolayers.
In tr od u ction
finities in the mM-range, while PEPT2 is referred to as
the high affinity/low capacity transporter, with sub-
strate affinities in the µM-range.13 The transport across
the membrane has been shown to be a tertiary active
process, driven by an H+-gradient. PEPT1 has also been
shown to play a role in the absorption of a wide variety
of peptidomimetic drugs, such as â-lactam antibiotics14
(penicillins and cephalosporins) and angiotensin con-
verting enzyme (ACE) inhibitors.15,16 Another thera-
peutic application of PEPT1 is the transport of amino
acid prodrugs, such as valacyclovir17 and valgancyclo-
vir,18 the L-valyl esters of acyclovir and gancyclovir,
respectively. Other examples include the L-valyl ester
of zidovudine (L-Val-AZT),19 and amino acid prodrugs
of methyldopa.20,21 Likewise, dipeptides have been used
as pro-moieties, and examples of this strategy include
bisphosphonate prodrugs22 and benzyl alcohol model
prodrugs.23
To ensure efficient amino acid absorption and con-
servation in the body, the intestine and kidneys are
equipped with specific transport systems for di- and
tripeptides. The di-/tripeptide transporter PEPT1 was
first cloned from rabbit,1 and both PEPT12 and PEPT23
(PEPTX) were cloned from human in 1995. PEPT1 is
expressed in the intestinal brush border membrane,2
while both PEPT1 and PEPT2 are expressed in the
luminal membrane of the renal proximal tubule.4,5 The
main physiological role of intestinal PEPT1 is absorp-
tion of di- and tripeptides resulting from the enzymatic
breakdown of dietary protein, while renal PEPT1 and
PEPT2 both reabsorb di- and tripeptides from kidney
filtrate.6,7 The existence of peptide transporters, distinct
from PEPT1 and PEPT2, in the basolateral membrane
of both intestinal enterocytes8-11 and renal tubule cells12
have been suggested; however, none of these have yet
been cloned. PEPT1 is often referred to as the low
affinity/high capacity transporter, with substrate af-
The dipeptide prodrug approach has mainly been
focused on the use of metabolically stable dipeptides (i.e.
derivatives containing D-amino acids, imino acids, â-ami-
no acids, etc.) as drug carriers. But as PEPT1 has a
preference for the natural L,L-dipeptide skeleton,24 such
modifications will often make the resulting compound
a poorer substrate. The possibilities for drug/prodrug
delivery by targeting PEPT1 are promising, and bio-
* To whom correspondence should be addressed. Phone: +46-31-
772-2894, fax: +46-31-772-3840, e-mail: luthman@chem.gu.se.
† Department of Medicinal Chemistry, University of Tromsø.
| Department of Chemistry, University of Tromsø.
‡ The Danish University of Pharmaceutical Sciences.
The University of Kansas.
§ Go¨teborg University.
10.1021/jm031022+ CCC: $27.50 © 2004 American Chemical Society
Published on Web 01/06/2004

Phe-Gly Dipeptidomimetics
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 4 1061
Sch em e 1
F igu r e 1. Structures of the synthesized Phe-Gly dipeptido-
mimetics.
pharmaceutical applications of this transport system
have been extensively reviewed.25-29 A peptidomimetic
approach, where the scissile amide bond is replaced with
different metabolically stable isosteric moieties, has the
advantage of keeping the important structural features
of the molecule comparable to that of the natural
peptide (i.e. correct stereochemistry, correct distance
between the N- and C-termini, etc.). However, there is
still relatively limited information available on structure
affinity (and hence, structure transport) relationships
regarding PEPTX, and this approach requires more
knowledge on which structural modifications of a pep-
tide that are allowed without losing affinity for PEPTX.
There are a few examples of compounds lacking an
amide bond and still being substrates, e.g. compounds
such as arphamenine A,30 4-aminophenylacetic acid (4-
APAA),31 δ-amino levulinic acid,32 and ω-amino fatty
acids,33 indicating that the amide bond is not a prereq-
uisite for transport. On the other hand, little is known
about other amide bond isosteres and the importance
of substrate flexibility.
Here we report the design, synthesis and biological
evaluation of five Phe-Gly dipeptidomimetics (2a -6a ,
Figure 1) in which the amide bond has been replaced
by isosteric moieties. To obtain structure-affinity in-
formation these replacements have been designed to
mimic important properties of the amide bond, such as
geometrical and/or electrostatic and hydrogen bond-
ing properties. This series of compounds contains the
ketomethylene isostere 2a , and the trans-hydroxy-
ethylidene (3a and 4a ) and hydroxyethylene (5a and 6a )
isosteres. Thus, the peptidomimetics 2a -6a (Figure 1)
are all of comparable size with the natural Phe-Gly
dipeptide, have the correct stereochemistry, and a
similar distance between the N- and C-termini. The
compounds have been evaluated for their affinities and
transport properties in vitro using the well-established
Caco-2 (human PEPT1 transporter; hPEPT1) and SKPT
(rat PEPT2 transporter; rPEPT2) cell assays.34-37
Catalytic hydrogenation of the corresponding methyl
ester of 3 in EtOH has previously been reported to give
the lactone as the only product.39 This was not observed
using the tert-butyl ester 3 in EtOAc, making this a very
convenient strategy for obtaining both 5 and 6. Depro-
tection of the N- and C-termini of 2-6 with TFA in CH2-
Cl2, followed by purification by reversed phase HPLC,
afforded the trifluoroacetate salts of 2a -6a .
To investigate the importance of the orientation of the
carboxylate group, we originally intended to synthesize
the fully deprotected, zwitterionic forms of the two
unsaturated ketones trans-1 and cis-1 (trans- and cis-
1a , respectively). To allow a one-step deprotection, the
Boc-group and the tert-butyl ester were used as protect-
ing groups. However, all attempts to deprotect 1 using
TFA in CH2Cl2 failed. 1H NMR spectra of the crude
product indicated degradation, probably due to rear-
rangements taking place in acidic media.
In an attempt to circumvent the problem of degrada-
tion of the unsaturated ketones, we decided to synthe-
size the methyl ester derivatives 7a and 8a . The methyl
ester is more resistant to acidic hydrolysis and would
supposedly make the final products more stable. At the
same time the methyl group is fairly small and would
provide compounds expected to act as substrates when
tested for affinity to PEPTX. Compounds 7 and 8 were
therefore synthesized analogously to the synthesis of 1,
but when an acidic N-deprotection was attempted (using
TFA or MeOH/HCl(g)), we observed the same instability
problems as with the tert-butyl esters.
Resu lts a n d Discu ssion
Ch em istr y. The synthetic strategy is outlined in
Scheme 1. We have earlier described the synthesis of
the key intermediate 1, and the chemo- and diastereo-
selective reduction of 1 to the unsaturated alcohols 3
and 4.38 Catalytic hydrogenation of the double bond in
1, 3, and 4 with H2/Pd in EtOAc gave the saturated
analogues 2, 5, and 6, respectively, in high yields.

1062 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 4
Våbenø et al.
Rearrangements of this kind of compound under basic
conditions are known from the literature,40 and it seems
as if a similar rearrangement can take place also in
acidic media. Due to the failure to deprotect 7 and 8,
and their limited use as such due to solubility problems,
these compounds were not included in the biological
investigations.
Biologica l In vestiga tion s. The synthesized deriva-
tives were tested for their affinity for PEPTX using
Caco-2 and SKPT cell assays, and also for intracellular
uptake and transepithelial transport via hPEPT1 (Ca-
co-2 cells). Initial affinity screening of 2a -6a for
rPEPT2 in SKPT cells was performed in a relevant
concentration range (0-1 mM)13 (results not shown).
The assay was based on inhibition of [14C]-Gly-Sar
uptake, and the results of this screen indicated that
compound 2a was a good substrate for PEPTX, whereas
3a -6a were poor substrates. Approximations of the
affinities of 3a , 5a , and 6a (for compound 4a , see below)
were therefore based on inhibition of [14C]-Gly-Sar
uptake performed at a single high concentration (6 mM)
for each compound, on both hPEPT1 in Caco-2 cells and
rPEPT2 in SKPT cells. From the inhibition values, the
concentrations required to inhibit the maximum uptake
by 50% (IC50 values) can be approximated (eq 2, see
Experimental section). The inhibitory constant (Ki)
values can then be calculated from the IC50 values. A
single interaction for the competitive interaction with
Gly-Sar was obtained for all substrates investigated
(Hill-factor n ≈ 1), and the maximum inhibition of [14C]-
Gly-Sar uptake, Umax, approached 1.0 (100%).
A. Affin ity of 2a -6a for r P EP T2 in SKP T Cells.
The affinities of 2a -6a for rPEPT2 (SKPT-cells) were
determined by their ability to inhibit the transport of
[14C]-Gly-Sar, a standard substrate for PEPTX. Phe-Gly
and Phe-Ala were included in order to investigate the
direct effect of an amide bond replacement.41 For control
purposes also the cephalosporin cephalexin, another
standard substrate for PEPTX, was included. The
Michaelis-Menten constant (Km) of Gly-Sar for rPEPT2
was first determined to enable calculation of Ki values
of the compounds of interest (Figure 2). This was
necessary because the SKPT cell assay used in this
study is newly modified,42 see Experimental section. The
Km value of Gly-Sar was 110.4 ( 29.9 µM, with a
maximum uptake rate (Vmax) of 30.0 ( 3.5 pmol × cm-2
× min-1, which are comparable to the values obtained
by Bravo et al.42
F igu r e 2. Uptake rate of Gly-Sar in SKPT-cells as a function
of concentration.
F igu r e 3. Inhibition of [14C]-Gly-Sar uptake by Phe-Ala (O),
Phe-Gly (b), 2a (]), and cephalexin ([) in SKPT-cells. The
lines describe the best fit to the experimentally obtained points
using eq 2. Error bars represent SE.
Ta ble 1. IC50 and Ki Values for rPEPT2 (SKPT cells)
compound
IC50 (µM) ( SE
Ki (µM)
Phe-Gly
Phe-Ala
2a
3a
4a
5a
6a
17.2 ( 1.0
15.5 ( 0.9
22.2 ( 2.4
4111a
14.6
13.1
18.8
3480a
1349
7451a
6921a
35.2
1594 ( 360b
8802a
8176a
cephalexin
41.6 ( 5.1b
a
b
Estimated values. Significantly different from Phe-Gly (p <
0.05).
substrates Phe-Gly and Phe-Ala, and a higher affinity
than cephalexin (Figure 3). An approximate 80-fold
reduction in affinity is observed for compound 4a (curve
not shown), compared to 2a (Table 1). The results for
the other three compounds (Figure 4) show that 5a and
6a have approximately the same affinity, whereas 3a
has a somewhat higher affinity. The values correspond
to an approximate 200-fold reduction in affinity for 3a
compared to 2a , and a 400-fold reduction for both 5a
and 6a .
As compound 2a , and to some extent compound 4a ,
showed promising results in the initial screening, full
affinity characterizations of these compounds in the
concentration ranges 0-250 and 0-5000 µM, respec-
tively, were undertaken. For the SKPT experiments, the
Hill-factor n was in the range of 1.02-1.42, and Umax
was between 0.97 and 1.02. The results obtained
demonstrate that the ketomethylene isostere 2a has an
affinity for rPEPT2 similar to those of the natural
B. Affin ity of 2a -6a for h P EP T1 in Ca co-2 Cells.
Affinity determinations of 2a -6a for hPEPT1 in con-

Phe-Gly Dipeptidomimetics
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 4 1063
F igu r e 4. Relative reduction of [14C]-Gly-Sar uptake in SKPT-
cells in the presence of 6 mM of 3a , 5a , and 6a , respectively.
Error bars represent SE.
F igu r e 6. Relative reduction of [14C]-Gly-Sar uptake in Caco-2
cells in the presence of 6 mM of 3a , 4a , 5a , and 6a ,
respectively. Error bars represent SE.
F igu r e 7. Intracellular uptake (in percent of total amount
added) of 0.5 mM 2a -4a in Caco-2 cells in absence (gray) and
presence (black) of 5 mM Gly-Pro. Error bars represent SE.
F igu r e 5. Inhibition of [14C]-Gly-Sar uptake by Phe-Ala (b),
Phe-Gly (O), and 2a ([) in Caco-2 cells. The lines describe the
best fit to the experimentally obtained points using eq 2. Error
bars represent SE.
attempts were made to estimate IC50 values for these
two compounds.
C. In tr a cellu la r Up ta k e Stu d ies of 2a -4a in
Ca co-2 Cells. As affinity studies only give information
about binding to the transporter, intracellular uptake
studies in Caco-2 cells were carried out to investigate
the actual translocation by hPEPT1. To demonstrate
active transport, these experiments were carried out in
absence and presence (10-fold excess) of the known
competetive PEPTX inhibitor Gly-Pro. The results
(Figure 7) show a 3-fold reduction (p < 0.01) in the
uptake of 2a in the presence of the inhibitor, indicating
that this compound is transported by hPEPT1. For 3a
and 4a the uptake is low, and no difference in uptake
is observed when Gly-Pro is added, indicating that no
active uptake is taking place.
D. Tr a n sep ith elia l Tr a n sp or t Stu d ies of 2a -4a
acr oss Caco-2 Cell Mon olayer s. Transepithelial trans-
port studies were subsequently carried out to investigate
the overall transport across the cell monolayer. These
experiments were also carried out in the absence and
presence (20-fold excess) of Gly-Pro. As the transepi-
thelial transport is a result of several processes, it is
difficult to separate the individual transport compo-
nents, which means that these results must be treated
with caution. As shown in Figure 8, the inclusion of Gly-
Pro in the incubation media caused a significant reduc-
tion (p < 0.01) in the apparent permeability coefficients
Ta ble 2. IC50 and Ki values for hPEPT1 (Caco-2 cells)
compound
IC50 (mM) ( SE
Ki (mM)
Phe-Gly
Phe-Ala
2a
3a
4a
0.20 ( 0.02
0.18 ( 0.02
0.40 ( 0.02b
26.8a
0.20
0.18
0.40
26.4a
39.8a
40.5a
a
b
Estimated values. Significantly different from Phe-Gly (p <
0.01).
fluent mature Caco-2 cell monolayers were carried out
in the same manner, also including Phe-Gly and Phe-
Ala. For compound 2a , a full characterization was
undertaken over a concentration range of 0-2 mM. For
the Caco-2 experiments, the Hill-factor n was in the
range of 0.83-1.21, and Umax was between 0.84 and
1.05. The results (Figure 5) demonstrate that there is a
2-fold difference in Ki values between the natural
substrates (Phe-Gly and Phe-Ala) and the ketomethyl-
ene isostere 2a (Table 2). The unsaturated alcohols 3a
and 4a inhibit Gly-Sar uptake to a small extent at a
concentration of 6 mM, whereas the saturated ana-
logues 5a and 6a have a negligible effect at this
concentration (Figure 6). As the values for 5a and 6a
were in the same range as the standard error (SE), no

1064 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 4
Våbenø et al.
could vary depending on the relative importance of
passive diffusion and carrier-mediated transport of the
compound under study, its relative substrate activity
for transporters in the apical and basolateral mem-
branes of Caco-2 cells, and the relative affinity of Gly-
Pro for these transporters. Surprisingly, 5a and 6a
showed high Papp values in these studies (results not
shown). However, considering their polar nature and
their low affinities for hPEPT1, this is most likely due
to paracellular transport.
Str u ctu r e-Affin ity Rela tion sh ip s. 3D structures
of mammalian peptide transporter proteins have not yet
been established, and the only way to extract structure-
affinity information for the transporter is by ligand-
based design. Several models for binding to the trans-
porter have been suggested. Recently, a template model
was proposed in a review by Rubio-Aliaga and Daniel,46
but essentially the same template was earlier proposed
by Bailey et al.47 This latter model also indicated the
relative orientation of the groups of the peptide consid-
ered important for interaction with PEPT1. According
to this model, there are four important structural
features for dipeptides: (1) the positively charged N-
terminus, (2) the negatively charged C-terminus, (3) the
amide carbonyl oxygen, and (4) the interaction of the
R2 side chain with a proposed hydrophobic pocket.48 In
addition, Bailey et al.47 suggest that the bond to the
terminal carboxylate group of a dipeptide should be
oriented 60° relative to the plane of the amide bond.
The amide NH functionality does not seem to be
involved in any direct interaction with the transporter.
Although this model is based on data obtained from
different test systems, it currently remains the most
comprehensive model for describing affinity for PEPT1.
It should be emphasized that none of the four ‘critical’
features highlighted in the model of Bailey et al. are
absolute prerequisites for affinity, as substrates lacking
one of these may still have affinity. The model therefore
in essence describes an ideal substrate.49
F igu r e 8. Transepithelial transport of 0.5 mM carnosine and
2a -4a across Caco-2 monolayers in absence (gray) and pres-
ence (black) of 10 mM Gly-Pro. Error bars represent SE.
(Papp) for all four compounds (carnosine, 2a , 3a , 4a ). For
2a , the transepithelial transport data (Figure 8) are
consistent with the uptake data (Figure 7), which
suggested that transport of this compound across the
apical membrane of Caco-2 cell monolayers was at least
partially transporter-mediated. From the data shown
in Figure 8 it is difficult to draw any conclusions about
the mechanism(s) (passive diffusion and/or active trans-
port) by which 2a permeates the cells’ basolateral
membrane. The transepithelial transport data (Figure
8) for 3a and 4a , particularly the inhibitory effect of Gly-
Pro on their transport, are not consistent with the
cellular uptake data shown in Figure 7, as these uptake
data suggest that 3a and 4a are entering the cells across
the apical membrane by passive diffusion (i.e. their
uptake is not inhibited by Gly-Pro). A possible explana-
tion for this apparent discrepancy is that while the
apical uptake of these two compounds proceeds via
passive diffusion, their efflux out of the cell across the
basolateral membrane is partially mediated by the
basolateral peptide transporter. Even if the basolateral
transporter is suggested to be distinct from the apical
transporter hPEPT1,9 it is likely to be sensitive to
inhibition by intracellular Gly-Pro.43 Considering the
polar nature of all three compounds studied (i.e. 2a , 3a ,
and 4a ), the paracellular pathway of transepithelial
transport may also be contributing significantly to the
Papp values reported in Figure 8, particularly when the
transport experiments were done in the presence of Gly-
Pro.
The replacement of the enzymatically labile amide
bond of Phe-Gly with the stable ketomethylene isostere
of 2a results in a loss of H-bond donor properties as well
as in an increased flexibility, since the partial double
bond character of the amide is lost. From the results it
is seen that this replacement reduces the apparent
affinity for hPEPT1 approximately 2-fold, whereas the
apparent affinity for rPEPT2 is not significantly changed.
When going from the planar carbonyl group of 2a to
the tetrahedral alcohol of 3a -6a , additional molecular
properties are altered, most importantly the geometry,
but also the H-bonding properties and charge distribu-
tion of the molecule. The affinity studies showed a
pronounced reduction in affinity for the alcohols 3a -
6a compared to the ketone 2a , both for hPEPT1 and
rPEPT2. One would expect higher desolvation energies
for the alcohols compared to the ketone, but this is not
sufficient to explain this large difference.
L-Carnosine is a well-known substrate for hPEPT1;
however, there are conflicting results in the literature
regarding the transepithelial transport of this com-
pound. Due to its stability, L-carnosine has been widely
used as a reference compound,24,44 and according to the
results of these studies, the inhibitory effect of Gly-Pro
on the transepithelial transport of L-carnosine shown
in Figure 8 could be expected. In contrast, a recent
kinetic study showed that the transepithelial transport
of L-carnosine across Caco-2 cell monolayers followed
apparent simple nonsaturable kinetics, indicative of
passive transport, but the study confirmed at the same
time that the transport across both the apical and
basolateral membrane was carrier-mediated.45 The
above-mentioned studies illustrate the complexity in
giving a mechanistic interpretation of the data shown
in Figure 8. The results derived from these experiments
Regarding the hydrogen bonding potential, the alco-
hols 3a -6a have both H-bond donor and acceptor
properties. However, the position of the donor (OH-
group) make them less appropriate as replacements for
the NH group of the amide bond if interaction with a
H-bond acceptor in the transporter is of importance. All
the peptidomimetics have an oxygen-containing func-

Phe-Gly Dipeptidomimetics
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 4 1065
tionality at C4, which can act as a H-bond acceptor.
However, the orientation of the oxygen atom in space
will be quite different, and the planar carbonyl group
of 2a obviously gives this compound more favorable
binding properties compared to the alcohol derivatives.
It is difficult to dissect the precise molecular property
that causes this dramatic change, since this difference
possibly originates from a combination of geometrical
and electrostatic properties. In any case the results
demonstrate the importance of a planar, polarized C4-O
bond.
The affinities of the alcohols were lower than what
we consider interesting for pro-moieties useful for active
transport by hPEPT1 (∼15 mM). For hPEPT1 there
seems to be no difference in affinity between the (4S)-
alcohols 4a and 6a and their respective (4R)-epimers
3a and 5a , whereas the unsaturated (4S)-alcohol 4a
shows an approximate 2.5-fold increase in affinity for
rPEPT2 compared to its (4R)-epimer 3a . This indicates
that even though these compounds are not ideal sub-
strates, the orientation of the oxygen at C4 has some
implication for the affinity.
According to the model of Bailey et al., the more
flexible dipeptidomimetics 2a , 5a , and 6a should be
preferred as substrates as they may adopt the required
conformation, positioning the N- and C-termini in the
correct spatial arrangement. However, the results ob-
tained showed that the hydroxyethylene isosteres 5a
and 6a generally had lower affinity than their unsatur-
ated counterparts, both for hPEPT1 and rPEPT2. This
was unexpected, since the trans-double bond in the
allylic alcohols 3a and 4a would prevent these deriva-
tives from adopting the required torsional angle of
approximately 60°. Although these compounds are not
ideal substrates for PEPTX, it can be concluded from
these experiments that the torsional freedom is not a
prerequisite for binding.
There are striking differences between the low affin-
ity/high capacity hPEPT1 and the high affinity/low
capacity rPEPT2. For the natural substrates Phe-Gly
and Phe-Ala, there is an approximate 10-fold reduction
in affinity for hPEPT1 compared to rPEPT2, and this
effect is even more pronounced for 2a , which has a 20-
fold reduction. It should be born in mind that hPEPT1
and rPEPT2 in this study are of human and rat origin,
respectively, and that species differences in part can
contribute to this observation. But this could also
indicate that PEPT1 discriminates between substrates
to a larger extent than PEPT2.
eral transporter. For compound 2a , which shows active
uptake, the results from the transepithelial transport
studies imply an overall active transport component.
Con clu sion s
A series of structurally related Phe-Gly dipeptidomi-
metics were shown to provide important information on
the structure-affinity and structure-transport rela-
tionships for PEPTX transport systems. The affinity
studies show that the metabolically stable ketomethyl-
ene isostere 2a has an affinity comparable to the natural
substrate Phe-Gly, whereas the hydroxyethylidene and
hydroxyethylene isosteres are poor substrates for both
isoforms of the peptide transporter. As compound 2a
also shows active uptake into Caco-2 cells, this implies
that this is a promising pro-moiety for a prodrug
approach targeting hPEPT1, and such studies are
presently ongoing in our laboratory.
Exp er im en ta l Section
Ch em istr y. Gen er a l. Starting materials and reagents were
purchased from Sigma, Fluka, Aldrich, and Merck KGaA, and
were used as received. 1H and 13C NMR spectra were recorded
on a Varian Mercury Plus operating at 400 and 100 MHz,
respectively. Chemical shifts are given relative to the solvent
signal (δ 7.26 and δ 77.0 for CDCl3, and δ 3.49 and δ 49.0 for
CD3OD, respectively). TLC was performed using silica gel
coated aluminum sheets (silica gel 60 F254, Merck), with
visualization by UV detection and/or 5% phosphomolybdic acid
hydrate (PMA) in EtOH, followed by heating. Flash chroma-
tography was performed on silica gel (silica gel 60 (0.040-
0.063 mm), Merck) under nitrogen pressure. THF was distilled
from Na/benzophenone ketyl. Other solvents were of analytical
or synthetic grade and were used without further purification.
Melting points were taken with a Bu¨chi Melting Point B-540
apparatus and are uncorrected. Optical activity was measured
using a Perkin-Elmer 241 Polarimeter. Elemental analyses
were done by Mikro Kemi AB, Uppsala, Sweden. In HPLC a
Gilson 115 UV detector (254 nm) and a Waters 510 pump were
used. Normal phase analytical chromatography for compounds
1-8 was performed on a LiChrosorb Si 60 (5 µm) 4 × 250 mm
analytical column from Merck KgaA (flow rate 1.0 mL/min).
Reversed phase analytical chromatography for compounds 2a -
6a was performed on a Symmetry Shield RP18 (3.5 µm) 2.1 ×
150 mm analytical column from Waters (isocratic: 12% CH3-
CN, 0.1% TFA; flow rate 0.2 mL/min). Normal phase prepara-
tive chromatography of compounds 3 and 4 was performed on
a Prep Nova-Pak HR Silica 6 µm 60 Å 25 × 100 mm column
from Waters (flow rate 6.0 mL/min). Reversed phase prepara-
tive chromatography of compounds 2a -6a was performed on
a Prep Nova-Pak HR C18 6 µm 60 Å 25 × 100 mm column
from Waters (isocratic: 20% CH3CN, 0.1% TFA; flow rate 7.0
mL/min). The syntheses of 1, 3, 4,38 and 739 have been
described earlier.
Str u ctu r e-Tr an spor t Relation sh ips for h P EP T1.
The ketomethylene isostere 2a is the only compound
that shows active intracellular uptake (Figure 7). Since
these experiments were done at a concentration of 0.5
mM, one would not expect the hydroxyethylidene isos-
teres 3a and 4a to show any significant active uptake
because of their weak affinities.
Gen er a l P r oced u r e for Red u ction of th e Dou ble Bon d
(syn th esis of 2, 5, a n d 6). A 130 mg amount of unsaturated
compound was dissolved in EtOAc (15 mL), 10% Pd/C (15 mg)
was added, and the mixture was stirred under H2 atmosphere
overnight. The mixture was filtered through Celite to remove
the catalyst, and the filtrate was concentrated under reduced
pressure. Flash chromatography of the residue gave the pure
product.
In the case of active transepithelial transport the
basolateral transporter also comes into play. From the
transepithelial transport studies (Figure 8) it appears
as compounds 3a and 4a are actively transported, but
as they are not actively taken up into the cell (Figure
7), the overall transport cannot be purely carrier-
mediated. Thus, it is difficult to draw conclusions
regarding the structural requirements for the basolat-
ter t-Bu t yl (5S)-5-[N-(ter t-Bu t oxyca r b on yl)a m in o]-4-
oxo-6-p h en yl-h exa n oa te (2). Flash chromatography [Et2O:
pentane 2:1] gave 2 (83 mg, 64%) as a white solid; mp 78-81
°C; [R]D +25.5° (c 1.94, CHCl3); analytical HPLC: tR 6.61 min
1
(1% EtOH in iso-hexane); H NMR (CDCl3) δ 7.29-7.15 (m,
5H), 5.11 (d, 1H, J ) 7.2 Hz), 4.53 (app q, 1H), 3.12 (dd, 1H,
J ) 14.0, 6.0 Hz,), 2.93 (dd, 1H, J ) 14.2, 7.0 Hz), 2.69 (t, 2H,
J ) 6.8 Hz), 2.47 (t, 2H, J ) 6.6 Hz), 1.42 (s, 9H), 1.38 (s, 9H).
13C NMR (CDCl3) δ 207.8, 172.0, 155.5, 136.5, 129.5 (2 C:s),

1066 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 4
Våbenø et al.
128.8 (2 C:s), 127.2, 80.9, 80.0, 60.4, 37.7, 35.4, 29.2, 28.5 (3
C:s), 28.3 (3 C:s). Anal. (C21H31NO5) C, H, N.
Hz), 4.29-4.27 (m, 1H), 3.54-3.49 (m, 1H), 3.09 (dd, 1H, J )
13.6, 8.0 Hz), 2.95 (dd, 1H, J ) 13.8, 7.0 Hz). 13C NMR (CD3-
OD) δ 168.1, 146.3, 135.7, 129.3 (2 C:s), 128.9 (2 C:s), 127.4,
124.0, 67.9, 56.1, 35.8. Anal. (C12H15NO3‚TFA) H, N; C: calcd,
50.15; found, 49.7.
(4R, 5S)-5-Am in o-4-h yd r oxy-6-p h en yl-h exa n oic Acid
(5a ). Compound 5 (98 mg) gave 5a (76.6 mg, 88%) as a white
solid; mp 70 °C (dec); [R]D -31.1° (c 1.0, MeOH); analytical
HPLC: tR 5.7 min; 1H NMR (CD3OD) δ 7.41-7.28 (m, 5H),
4.76-4.71 (m, 1H), 3.92-3.87 (m, 1H), 3.10 (dd, 1H, J ) 14.4,
6.0 Hz), 2.89 (dd, 1H, J ) 14.4, 8.4 Hz), 2.74-2.53 (m, 2H),
2.45-2.37 (m, 1H), 2.18-2.08 (m, 1H). 13C NMR (CD3OD) δ
176.8, 134.9, 129.2 (2 C:s), 129.1 (2 C:s), 127.6, 78.7, 54.4, 33.4,
27.8, 22.5. Anal. (C12H17NO3‚1.5 TFA): C, H, N.
(4S, 5S)-5-Am in o-4-h yd r oxy-6-p h en yl-h exa n oic Acid
(6a ). Compound 6 (98 mg) gave 6a (70 mg, 81%) as a colorless
semisolid; [R]D +19.5° (c 2.16, MeOH); analytical HPLC: tR
5.6 min; 1H NMR (CD3OD) δ 7.40-7.29 (m, 5H), 4.61-4.55
(m, 1H), 3.71-3.65 (m, 1H), 3.07 (dd, 1H, J ) 14.0, 6.8 Hz),
2.97 (dd, 1H, J ) 14.4, 7.2 Hz), 2.71-2.52 (m, 2H), 2.36-2.28
(m, 1H), 2.06-1.96 (m, 1H).13C NMR (CD3OD) δ 176.6, 134.6,
129.4 (2 C:s), 129.1 (2 C:s), 127.6, 79.1, 55.7, 35.3, 27.9, 24.9.
Anal. (C12H17NO3‚0.75TFA) C, N; H: calcd, 5.80; found, 5.2.
ter t-Bu tyl (4R,5S)-5-[N-(ter t-Bu toxyca r bon yl)a m in o]-
4-h yd r oxy-6-p h en yl-h exa n oa te (5). Flash chromatography
[Et2O:pentane 1:1] gave 5 (106 mg, 81%) as a white solid; mp
140-142.5 °C; [R]D -7.1° (c 1.01, CHCl3); analytical HPLC:
t
R 9.40 min (2% EtOH in iso-hexane); 1H NMR (CDCl3) δ 7.31-
7.19 (m, 5H), 4.66 (br d, 1H), 3.85 (br s, 1H), 3.64 (br d, 1H),
3.57 (br s, 1 H), 2.95-2.74 (m, 2H), 2.51-2.37 (m, 2H), 1.87-
1.68 (m, 2H) 1.45 (s, 9H), 1.35 (s, 9H). 13C NMR (CDCl3) δ
174.1, 156.5, 138.3, 129.5 (2 C:s), 128.7 (2 C:s), 126.6, 80.9,
79.9, 73.8, 56.9, 36.1, 32.6, 28.5, 28.3 (6 C:s). Anal. (C21H33
NO5) C, H, N.
-
ter t-Bu tyl (4S,5S)-5-[N-(ter t-Bu toxyca r bon yl)a m in o]-
4-h yd r oxy-6-p h en yl-h exa n oa te (6). Flash chromatography
[Et2O:pentane 1:1] gave 6 (110 mg, 84%) as a white solid; mp
54-57 °C; [R]D -7.2° (c 1.49, CHCl3); analytical HPLC: tR 9.40
1
min (2% EtOH in iso-hexane); H NMR (CDCl3) δ 7.30-7.18
(m, 5H), 4.98 (br d, 1H), 3.75-3.69 (m, 1H), 3.56 (br d, 1H),
3.42 (br s, 1 H), 2.94-2.84 (m, 2H), 2.42-2.27 (m, 2H), 1.86-
1.63 (m, 2H) 1.40 (s, 9H), 1.39 (s, 9H). 13C NMR (CDCl3) δ
174.3, 156.4, 138.7, 129.6 (2 C:s), 128.7 (2 C:s), 126.5, 81.1,
79.5, 71.2, 56.4, 38.9, 32.7, 29.9, 28.6 (3 C:s), 28.2 (3 C:s). Anal.
(C21H33NO5) C, H, N.
Cell Assa ys. A. Ma ter ia ls. Gen er a l. Caco-2 cells for the
affinity, uptake, and transport assays were obtained from the
American Type Culture Collection (Manassas, VA). SKPT-0193
cl.2 (SKPT) cells were donated by Dr. Matthias Brandsch
(Biozentrum, Halle, Germany) with kind permission of Dr.
Ulrich Hopfer (Case Western Reserve University, Cleveland,
OH). All chemicals for buffer preparations were of laboratory
grade, obtained from Life Technologies and Sigma. Transwell
plates (12-well and 6-well) were from Corning Costar Corpora-
tion. Affin ity Stu d ies. Phe-Gly and Phe-Ala were from
Bachem, and [14C]-Gly-Sar (specific activity: 49.94 mCi/mmol)
and [3H]-mannitol (specific activity: 51.50 mCi/mmol) were
from New England Nuclear. Transepithelial electrical resis-
tance (TEER) was measured in tissue resistance measurement
chambers (Endohm) with a voltohmmeter (EVOM), both from
World Precision Instruments. The shaking plate used for cell
culture experiments was a KS 10 DIGI shaker from Edmund
Bu¨hler. Radioactivity was determined in a Packard TriCarb
liquid scintillation counter, using Ultima Gold scintillation
fluid from Packcard. Up ta k e a n d Tr a n sp or t Stu d ies. Gly-
Pro and L-carnosine were obtained from Sigma, and [14C]-
mannitol from Amersham. The centrifuge was from IEC and
the sonicator from Cole Parmer Instrument Company. For
HPLC, a Shimadzu SPD-6A UV detector (210 nm), a Shimadzu
LC-6A pump, and a Vydac 218TP54, 5 µM, 250 × 4.6 mm
analytical column were used.
Meth yl (5S)-5-[N-(ter t-Bu toxyca r bon yl)a m in o]-4-oxo-
6-p h en yl-(E)-2-h exen oa te (7) a n d Meth yl (5S)-5-[N-(ter t-
B u t o x y c a r b o n y l)a m i n o ]-4-o x o -6-p h e n y l-(Z )-2-h e x -
en oa te (8). The two isomers were synthesized analogously to
1,38 using methyl glyoxylate instead of tert-butyl glyoxylate.
Starting from 1.09 g (2.94 mmol) of dimethyl [(3S)-3-[N-(tert-
butoxycarbonyl)amino]-2-oxo-4-phenyl-butyl]phosphonate, flash
chromatography of the crude product [Et2O:pentane 2:1] gave
pure 7 (520 mg) and 8 (132 mg) as light yellow solids in a total
yield of 66%. For spectroscopical data on 7, see ref 39. 8: mp
91-93 °C; [R]D -44.8° (c 1.13, CHCl3); analytical HPLC: tR
1
7.70 min (1% EtOH in iso-hexane); H NMR (CDCl3) δ 7.31-
7.18 (m, 5H), 6.44 (d, 1H, J ) 12.0 Hz), 6.06 (d, 1H, J ) 11.6
Hz), 5.13 (br d, 1H, J ) 8.4 Hz), 4.72 (app q, 1H), 3.75 (s, 3H),
3.25 (dd, 1H, J ) 14.2, 5.8 Hz), 3.01 (dd, 1H, J ) 14.2, 7.4
Hz), 1.39 (s, 9H). 13C NMR (CDCl3) δ 201.9, 166.1, 155.5, 139.9,
136.7, 129.6 (2 C:s), 128.8 (2 C:s), 127.1, 126.4, 80.2, 60.8, 52.5,
37.4, 28.5 (3 C:s). Anal. (C18H23NO5) C, H, N.
Dep r otection . Gen er a l P r oced u r e for 2-6. The pro-
tected compound was dissolved in CH2Cl2 (2 mL), and TFA
(0.4 mL) was added. The mixture was stirred at room tem-
perature for 2 h, and Et2O (10 mL) was added and evaporated
three times. The crude mixture was dissolved in 3 mL of mobile
phase and purified with reversed phase preparative HPLC.
The combined fractions were evaporated under reduced pres-
sure. Freeze-drying of the aqueous residue gave the final
products.
(5S)-5-Am in o-4-oxo-6-p h en yl-h exa n oic Acid (2a ). Com-
pound 2 (73 mg) gave 2a (56.2 mg, 87%) as a white solid; mp
115 °C (dec); [R]D +41.7° (c 0.95, MeOH); analytical HPLC: tR
5.9 min; 1H NMR (CD3OD) δ 7.41-7.31 (m, 5H), 4.46 (dd, 1H,
J ) 8.6, 5.8 Hz), 3.40 (dd, 1H, J ) 14.6, 5.4 Hz), 3.00 (dd, 1H,
J ) 14.4, 8.4 Hz,), 2.88-2.72 (m, 2H), 2.62 (app t, 2H). 13C
NMR (CD3OD) δ 204.5, 174.6, 134.5, 129.3 (2 C:s), 129.1 (2
C:s), 127.8, 59.8, 35.8, 34.4, 27.3. Anal. (C12H15NO3‚TFA) C,
H, N.
(4R, 5S)-5-Am in o-4-h yd r oxy-6-p h en yl-(E)-2-h exen oic
Acid (3a ). Compound 3 (62 mg) gave 3a (44 mg, 81%) as a
colorless semisolid; [R]D +22.0° (c 1.09, MeOH); analytical
HPLC: tR 5.5 min; 1H NMR (CD3OD) δ 7.37-7.26 (m, 5H),
6.95 (dd, 1H, J ) 15.6, 4.4 Hz), 6.18 (dd, 1H, J ) 15.6, 2.0
Hz), 4.55-4.53 (m, 1H), 3.71-3.67 (m, 1H), 2.95 (dd, 1H, J )
14.4, 6.0 Hz), 2.86 (dd, 1H, J ) 14.4, 8.8 Hz). 13C NMR (CD3-
OD) δ 168.2, 144.5, 135.8, 129.2 (2 C:s), 128.9 (2 C:s), 127.3,
124.1, 69.1, 56.7, 33.7. Anal. (C12H15NO3‚TFA) C, H, N.
(4S, 5S)-5-Am in o-4-h yd r oxy-6-p h en yl-(E)-2-h exen oic
Acid (4a ). Compound 4 (61 mg) gave 4a (50 mg, 92%) as a
colorless semisolid; [R]D -5.6° (c 0.89, MeOH); analytical
HPLC: tR 5.9 min; 1H NMR (CD3OD) δ 7.38-7.27 (m, 5H),
6.86 (dd, 1H, J ) 15.2, 4.8 Hz), 6.17 (dd, 1H, J ) 15.4, 1.8
B. Cell Cu ltu r e. Caco-2 cells for affinity experiments were
cultured as previously described.50 Briefly, cells were seeded
in culture flasks and passaged in Dulbecco’s Modified Eagle’s
medium (DMEM) supplemented with 10% fetal bovine serum,
penicillin/streptomycin (100 U/mL and 100 µg/mL, respec-
tively), 1% L-glutamine and 1% nonessential amino acids.
Caco-2 cells were seeded onto tissue culture treated 12-well
Transwell plates (1.0 cm2, 0.4 µm pore size) at a density of
105 cells × cm-2. Monolayers were grown in an atmosphere of
5% CO2 - 95% O2 at 37 °C. Growth media were replaced every
other day. Dipeptide transport activity reached a steady
maximum level at day 24-30. Affinity experiments were
subsequently performed on day 26-28 after seeding. TEER
at room temperature was measured in each well before the
experiment, and the mean ( standard deviation (SD) was
382.3 ( 41 Ω × cm-2 (n ) 21). The TEER value of a filter
support without cells is typically 27 Ω × cm-2
.
SKPT-cells were cultured as described by Bravo et al.42
Briefly, cells at passage 44 were seeded in culture flasks and
passaged in 1:1 DMEM/Nutrient mixture F-12 (F12). The
culture media were supplemented with 10% fetal bovine
serum, 100 U/mL penicillin, 100 µg/mL streptomycin, 5 mg/
mL insulin, 4 mg/mL dexamethasone, and 5 mg/mL apotrans-
ferrin. When cells reached passages 45-63, they were seeded
onto tissue culture treated 12-well Transwell plates (1 cm2,

Phe-Gly Dipeptidomimetics
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 4 1067
0.4 mm pore size), at a density of 5 × 106 cells/cm2. Monolayers
were grown in an atmosphere of 5% CO2-95% O2 at 37 °C.
Growth media were replaced every other day. Dipeptide
transport activity reached a steady maximum level at day 3-5.
Affinity experiments were subsequently performed on day 3-4
after seeding. TEER at room temperature was measured in
each well before the experiment. The mean ( SD was 5.00 (
0.94 (n ) 15) and 7.56 ( 1.01 (n ) 9) on day 3 and 4,
HBSS. The cells on polyester membrane were harvested into
1.5 mL centrifuge vials, spun down, redissolved in 300 µL of
water, and sonicated for 2 min. The cell preparation was
extracted with 300 µL of CH2Cl2, and a 200 µL sample was
taken from the aqueous phase. After acidification using 100
µL of 0.05 N HCl, the sample was analyzed by HPLC (isocratic;
10% CH3CN; flow rate 1.0 mL/min).
F. Tr a n sep ith elia l Tr a n spor t a cr oss Ca co-2 Cell Mon o-
la yer s. Prior to initiation of the transport experiments, cell
monolayers were washed twice with HBSS (pH 7.35) and
incubated in HBSS at 37 °C for 20 min. Then, 2.5 mL of fresh
HBSS (pH 7.35) was applied to the basolateral side of each
well, and 1.5 mL of peptide analogue solution (500 µM in
EBSS, pH 6.0), with or without 10 mM of Gly-Pro, was applied
to the apical side of each well. Samples were taken from the
basolateral side at intervals up to 90 min and analyzed by
HPLC (isocratic; 10% CH3CN; flow rate 1.0 mL/min).
G. Da ta An a lysis. 1. Tr a n sp or t Kin etics. A. Up ta k e of
Gly-Sa r . Uptake of Gly-Sar in SKPT-cells was corrected for
noncellular uptake, and the cellular uptake as a function of
apical Gly-Sar concentration was fitted to a Michaelis-Menten
type equation:
respectively, giving an overall mean of 5.96 ( 1.57 kΩ × cm-2
.
Caco-2 cells for the uptake and transport experiments were
maintained for 21 days in 6-well Transwell plates as described
earlier.51 For the transport experiments the integrity of the
monolayers was checked using [14C]-mannitol. The Papp values
of [14C]-mannitol across Caco-2 cell monolayers were typically
in the range of 0.1-0.6 × 10-6 cm/s.
C. Glycylsa r cosin e Up ta k e Exp er im en ts. For SKPT-
cells, uptake of [14C]-Gly-Sar was measured in Hanks Balanced
Salt Solution (HBSS) supplemented with 0.05% bovine serum
albumin (BSA). Apical media were buffered with 10 mM 2-[N-
morpholino]ethanesulfonic acid (MES) buffer, and pH was
adjusted to 6.0. Basolateral media were buffered with 10 mM
N-[2-hydroxyethyl]piperazine-N′-[2-ethanesulfonate] (HEPES)
buffer and adjusted to pH 7.4. Cells were placed on a shaking
plate, preheated to 37 °C, and allowed to equilibrate for 15
min in apical and basolateral buffer solutions. The experiment
was started by adding fresh apical buffer containing the
relevant Gly-Sar concentration (0-500 µM) and 0.5 µCi [14C]-
Gly-Sar per well. 0.5 µCi [3H]-mannitol per well was added to
the apical solution, as a marker of extracellular space. Uptake
experiments were terminated after 30 min by gentle suction
of the uptake medium, followed by three washes of the
monolayers with ice-cold HBSS. Following the washing step,
the cells were cut from the Transwell support and placed into
scintillation vials, and 2 mL of scintillation fluid was added.
The cell-associated radioactivity was counted via liquid scintil-
lation spectrometry.
Vmax × [S]n
V )
(1)
n
Km + [S]n
where V ) uptake rate (pmol × cm-2 × min-1), Vmax
)
maximum uptake rate (pmol × cm-2 × min-1), Km ) the
Michaelis-Menten constant (µM), [S] ) Gly-Sar concentration
(µM), and n ) Hill-factor.
B. Affin ity. Affinity for hPEPT1 and rPEPT2 in Caco-2 and
SKPT-cells, respectively, was determined as inhibition of [14C]-
Gly-Sar uptake in the presence of varying concentrations of
the respective compound. The degree of inhibition was fitted
to a Michaelis-Menten type equation:
For Caco-2 cells, a Km value of 1.1 mM obtained in our
laboratory was used for Gly-Sar, and concentration dependent
Gly-Sar uptake was not investigated further.
(1 - (U/U0))max × [I]n
1 - (U/U0) )
(2)
IC50 + [I]n
n
D. Affin ity Stu d ies. The buffers used for affinity studies
were the same as for Gly-Sar uptake, as was the equilibration
time and conditions. The affinity experiment was initiated by
adding 0.5 mL of MES-buffer containing 0.5 µCi of [14C]-Gly-
Sar and various amounts of the respective compounds to the
apical side. For SKPT-cells, 0.5 µCi of [3H]-mannitol per well
was also added to the apical solution, as a marker of extra-
cellular space. A 1.0 mL amount of HEPES buffer was added
to the basolateral side of each cell monolayer. Previous results
in our laboratory showed that uptake in Caco-2 cells was linear
up to 6-7 min of incubation, and thus the incubation time
was 5 min. For SKPT-cells, the uptake has been shown to be
linear for >60 min, and the incubation time was 30 min.
During incubation the plates were circularly and continuously
shaken. In all experiments the investigated compounds were
only applied on the apical side. The temperature was main-
tained at 37 °C. The uptake of 0.5 µCi [14C]-Gly-Sar was
terminated by gentle suction of the uptake medium followed
by three washes of the monolayer with ice-cold HBSS. The
filter supports were cut out, and the cell-associated radioactiv-
ity was determined.
where U ) uptake of [14C]-Gly-Sar, U0 ) uptake of [14C]-Gly-
Sar at zero inhibitor concentration, IC50 ) the concentration
required to inhibit the maximum uptake by 50% (µM or mM),
[I] ) compound concentration (µM or mM), and n ) Hill-factor.
When estimating IC50 values for compounds which were tested
at only one concentration, the Hill-factor, n, was assumed to
equal 1 and the expression (1 - (U/U0))max was assumed to be
1 based on the results obtained for the other compounds. Ki
values were calculated as described by Cheng and Prusoff.52
C. In tr a cellu la r Up ta k e. The percent uptake of peptide
analogue was calculated using:
Qretained
uptake (% ) )
× 100
(3)
Qtotal
where Qretained is the amount retained in the cell monolayer
and Qtotal is the total amount added to each well.
D. Tr a n sep ith elia l Tr a n sp or t. The Papp values of the
compounds were calculated by:
For affinity studies of 3a -6a (Caco-2 cells) and 3a , 5a , and
6a (SKPT-cells), the same procedure was used, however only
at one concentration (6 mM) of the compound.
∆Q/∆t
A × C0
Papp
)
(4)
E. In tr a cellu la r Up ta k e in Ca co-2 Cells. Caco-2 cells
were maintained for 21 days in 6-well Transwell plates. Prior
to the initiation of the uptake study, cell monolayers were
washed twice with HBSS (pH 7.35) and incubated in HBSS
for 20 min at 37 °C. After removing HBSS on both apical and
basolateral sides, 0.5 mM of the peptide analogue in 1500 µL
of Earle’s balanced salt solution (EBSS) (pH 6.0), with or
without the presence of 5 mM of Gly-Pro, was applied to the
apical side. Fresh HBSS was then applied to the basolateral
side of each well. Cell monolayers were incubated for another
60 min at 37 °C and then washed three times with ice-cold
where ∆Q/∆t is the linear appearance rate of mass in the
receiver solution, A is the cross-section area (4.71 cm2), and
C0 is the initial concentration of the donor side at t ) 0.
2. Sta tistics An a lysis. Unless stated otherwise, values are
given as means ( SE, and the statistical significance of the
results was determined using two-tailed Students t-test (or
Welch’s corrected t-test when means with different variances
were compared) using GraphPad InStat v3.05. p < 0.05 was
considered significant. Affinity experiments on both Caco-2 and
SKPT-cell monolayers were done in duplicates in each passage

1068 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 4
Våbenø et al.
(15) Moore, V. A.; Irwin, W. J .; Timmins, P.; Lambert, P. A.; Chong,
S.; Dando, S. A.; Morrison, R. A. A rapid screening system to
determine drug affinities for the intestinal dipeptide transporter
2: affinities of ACE inhibitors. Int. J . Pharm. 2000, 210, 29-
44.
(16) For a review with other examples of pharmacologically active
compounds with affinity for/transport by PepT1, see ref 25 and
references therein.
(17) de Vrueh, R. L. A.; Smith, P. L.; Lee, C. P. Transport of L-valine-
acyclovir via the oligopeptide transporter in the human intes-
tinal cell line, Caco-2. J . Pharmacol. Exp. Ther. 1998, 286, 1166-
1170.
(18) Sugawara, M.; Huang, W.; Fei, Y. L.; Leibach, F. H.; Ganapathy,
V.; Ganapathy, M. E. Transport of valganciclovir, a ganciclovir
prodrug, via peptide transporters PEPT1 and PEPT2. J . Pharm.
Sci. 2000, 89, 781-789.
(19) Han, H. K.; de Vrueh, R. L. A.; Rhie, J . K.; Covitz, K. M. Y.;
Smith, P. L.; Lee, C. P.; Oh, D. M.; Sadee, W.; Amidon, G. L.
5′-Amino acid esters of antiviral nucleosides, acyclovir, and AZT
are absorbed by the intestinal PEPT1 peptide transporter.
Pharm. Res. 1998, 15, 1154-1159.
(20) Hu, M.; Subramanian, P.; Mosberg, H. I.; Amidon, G. L. Use of
the peptide carrier system to improve the intestinal absorption
of L-R-methyldopa: Carrier kinetics, intestinal permeabilities,
and in vitro hydrolysis of dipeptidyl derivatives of L-R-methyl-
dopa. Pharm. Res. 1989, 6, 66-70.
(N ) 2), and n was the number of cell passages used (n ) 3).
Uptake and transport experiments were done in triplicates (n
)3).
Ackn owledgm en t. Financial support for this project
was obtained from the Norwegian Research Council
(128256/420), the Swedish Science Research Council
(621-2001-1431), the Knut and Alice Wallenberg Foun-
dation (98.176), and the Danish Medicinal Research
Council via the Center for Drug Design and Transport
and via project grant #22-01-0310. This research has
also been supported by a Marie Curie Fellowship of the
European Community Program “Improving the Human
Research Potential and the Socio-Economic Knowledge
Base” under contract number HPMT-CT-2001-00403.
We thank Kjell H. Halvorsen for excellent assistance
in the synthetic work, Morten Moe for assistance with
preparative and analytical HPLC, Dr. Bengt Erik Haug
for assistance with the freeze-drying, and Susanne N.
Sørensen for doing the cell culturing.
(21) Bai, P. F.; Subramanian, P.; Mosberg, H. I.; Amidon, G. L.
Structural requirements for the intestinal mucosal-cell peptide
transporter: The need for N-terminal R-amino group. Pharm.
Res. 1991, 8, 593-599.
(22) Ezra, A.; Hoffman, A.; Breuer, E.; Alferiev, I. S.; Monkkonen,
J .; El Hanany-Rozen, N.; Weiss, G.; Stepensky, D.; Gati, I.;
Cohen, H.; Tormalehto, S.; Amidon, G. L.; Golomb, G. A peptide
prodrug approach for improving bisphosphonate oral absorption.
J . Med. Chem. 2000, 43, 3641-3652.
(23) Taub, M. E.; Moss, B. A.; Steffansen, B.; Frokjaer, S. Oligopep-
tide transporter mediated uptake and transport of D-Asp(OBzl)-
Ala, D-Glu(OBzl)-Ala, and D-Ser(Bzl)-Ala in filter-grown Caco-2
monolayers. Int. J . Pharm. 1998, 174, 223-232.
(24) Tamura, K.; Bhatnagar, P. K.; Takata, J . S.; Lee, C. P.; Smith,
P. L.; Borchardt, R. T. Metabolism, uptake, and transepithelial
transport of the diastereomers of Val-Val in the human intestinal
cell line, Caco-2. Pharm. Res. 1996, 13, 1213-1218.
(25) Brodin, B.; Nielsen, C. U.; Steffansen, B.; Frøkjær, S. Transport
of peptidomimetic drugs by the intestinal di/tri-peptide trans-
porter, PepT1. Pharmacol. Toxicol. 2002, 90, 285-296.
(26) Bai, J . P. F.; Amidon, G. L. Structural specificity of mucosal-
cell transport and metabolism of peptide drugs: Implication for
oral peptide drug delivery. Pharm. Res. 1992, 9, 969-978.
(27) Yang, C. Y.; Dantzig, A. H.; Pidgeon, C. Intestinal peptide
transport systems and oral drug availability. Pharm. Res. 1999,
16, 1331-1343.
(28) Walter, E.; Kissel, T.; Amidon, G. L. The intestinal peptide
carrier: A potential transport system for small peptide derived
drugs. Adv. Drug Delivery Rev. 1996, 20, 33-58.
(29) Nielsen, C. U.; Brodin, B.; J orgensen, F. S.; Frokjaer, S.;
Steffansen, B. Human peptide transporters: therapeutic ap-
plications. Expert Opin. Ther. Pat. 2002, 12, 1329-1350.
(30) Enjoh, M.; Hashimoto, K.; Arai, S.; Shimizu, M. Inhibitory effect
of arphamenine A on intestinal dipeptide transport. Biosci.
Biotechnol. Biochem. 1996, 60, 1893-1895.
Refer en ces
(1) Fei, Y. J .; Kanai, Y.; Nussberger, S.; Ganapathy, V.; Leibach, F.
H.; Romero, M. F.; Singh, S. K.; Boron, W. F.; Hediger, M. A.
Expression cloning of a mammalian proton-coupled oligopeptide
transporter. Nature 1994, 368, 563-566.
(2) Liang, R.; Fei, Y. J .; Prasad, P. D.; Ramamoorthy, S.; Han, H.;
Yang-Feng, T. L.; Hediger, M. A.; Ganapathy, V.; Leibach, F.
H. Human intestinal H+/peptide cotransporter. Cloning, func-
tional expression, and chromosomal localization. J . Biol. Chem.
1995, 270, 6456-6463.
(3) Liu, W.; Liang, R.; Ramamoorthy, S.; Fei, Y. J .; Ganapathy, M.
E.; Hediger, M. A.; Ganapathy, V.; Leibach, F. H. Molecular
cloning of PEPT 2, a new member of the H+/peptide cotrans-
porter family, from human kidney. Biochim. Biophys. Acta 1995,
1235, 461-466.
(4) Smith, D. E.; Pavlova, A.; Berger, U. V.; Hediger, M. A.; Yang,
T.; Huang, Y. G.; Schnermann, J . B. Tubular localization and
tissue distribution of peptide transporters in rat kidney. Pharm.
Res. 1998, 15, 1244-1249.
(5) Boll, M.; Herget, M.; Wagener, M.; Weber, W. M.; Markovich,
D.; Biber, J .; Clauss, W.; Murer, H.; Daniel, H. Expression
cloning and functional characterization of the kidney cortex high-
affinity proton-coupled peptide transporter. Proc. Natl. Acad. Sci.
U.S.A. 1996, 93, 284-289.
(6) Ganapathy, V.; Leibach, F. H. Carrier-mediated reabsorption of
small peptides in renal proximal tubule. Am. J . Physiol. Renal
Physiol. 1986, 251, F945-953.
(7) Daniel, H.; Morse, E.; Adibi, S. The high and low affinity
transport systems for dipeptides in kidney brush border mem-
brane respond differently to alterations in pH gradient and
membrane potential. J . Biol. Chem. 1991, 266, 19917-19924.
(8) Dyer, J .; Beechey, R. B.; Gorvel, J . P.; Smith, R. T.; Wootton,
R.; Shirazi-Beechey, S. P. Glycyl-L-proline transport in rabbit
enterocyte basolateral-membrane vesicles. Biochem. J . 1990,
269, 565-571.
(9) Saito, H.; Inui, K. Dipeptide transporters in apical and basolat-
eral membranes of the human intestinal cell line Caco-2. Am.
J . Physiol. 1993, 265, G289-G294.
(10) Thwaites, D. T.; Brown, C. D. A.; Hirst, B. H.; Simmons, N. L.
Transepithelial glycylsarcosine transport in intestinal Caco-2
cells mediated by expression of hydrogen ion-coupled carriers
at both apical and basal membranes. J . Biol. Chem. 1993, 268,
7640-7642.
(11) Terada, T.; Sawada, K.; Saito, H.; Hashimoto, Y.; Inui, K.
Functional characteristics of basolateral peptide transporter in
the human intestinal cell line Caco-2. Am. J . Physiol. 1999, 276,
G1435-1441.
(12) Terada, T.; Sawada, K.; Ito, T.; Saito, H.; Hashimoto, Y.; Inui,
K.-I. Functional expression of novel peptide transporter in renal
basolateral membranes. Am. J . Physiol. Renal Physiol. 2000,
279, F851-857.
(13) For Ki values of typical substrates for PEPTX, including peptides;
see ref 29 and references therein.
(31) Temple, C. S.; Stewart, A. K.; Meredith, D.; Lister, N. A.;
Morgan, K. M.; Collier, I. D.; Vaughan-J ones, 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.
(32) Do¨ring, F.; Walter, J .; Will, J .; Focking, M.; Boll, M.; Amasheh,
S.; Clauss, W.; Daniel, H. Delta-aminolevulinic acid transport
by intestinal and renal peptide transporters and its physiological
and clinical implications. J . Clin. Invest. 1998, 101, 2761-2767.
(33) Do¨ring, F.; Will, J .; Amasheh, S.; Clauss, W.; Ahlbrecht, H.;
Daniel, H. Minimal molecular determinants of substrates for
recognition by the intestinal peptide transporter. J . Biol. Chem.
1998, 273, 23211-23218.
(34) Ganapathy, M. E.; Brandsch, M.; Prasad, P. D.; Ganapathy, V.;
Leibach, F. H. Differential recognition of â-lactam antibiotics
by intestinal and renal peptide transporters, PEPT 1 and PEPT
2. J . Biol. Chem. 1995, 270, 25672-25677.
(35) Ganapathy, M. E.; Prasad, P. D.; Mackenzie, B.; Ganapathy, V.;
Leibach, F. H. Interaction of anionic cephalosporins with the
intestinal and renal peptide transporters PEPT 1 and PEPT 2.
Biochim. Biophys. Acta 1997, 1324, 296-308.
(14) Bretschneider, B.; Brandsch, M.; Neubert, R. Intestinal transport
of â-lactam antibiotics: Analysis of the affinity at the H+/peptide
symporter (PEPT1), the uptake into Caco-2 cell monolayers and
the transepithelial flux. Pharm. Res. 1999, 16, 55-61.

Phe-Gly Dipeptidomimetics
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 4 1069
(36) Brandsch, M.; Brandsch, C.; Prasad, P. D.; Ganapathy, V.;
Hopfer, U.; Leibach, F. H. Identification of a renal cell line that
constitutively expresses the kidney-specific high-affinity H+/
peptide cotransporter. FASEB J . 1995, 9, 1489-1496.
(44) Tamura, K.; Lee, C. P.; Smith, P. L.; Borchardt, R. T. Effect of
charge on oligopeptide transporter-mediated permeation of cyclic
dipeptides across Caco-2 cell monolayers. Pharm. Res. 1996, 13,
1752-1754.
(45) Nielsen, C. U.; Supuran, C. T.; Scozzafava, A.; Frokjaer, S.;
Steffansen, B.; Brodin, B. Transport characteristics of L-carnos-
ine and the anticancer derivative 4-toluenesulfonylureido-car-
nosine in a human epithelial cell line. Pharm. Res. 2002, 19,
1337-1344.
(46) Rubio-Aliaga, I.; Daniel, H. Mammalian peptide transporters as
targets for drug delivery. Trends Pharmacol. Sci. 2002, 23, 434-
440.
(47) 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. Engl. 2000, 39, 506-508.
(37) Ganapathy, M. E.; Huang, W.; Wang, H.; Ganapathy, V.;
Leibach, F. H. Valacyclovir: A substrate for the intestinal and
renal peptide transporters PEPT1 and PEPT2. Biochem. Bio-
phys. Res. Commun. 1998, 246, 470-475.
(38) Våbenø, J .; Brisander, M.; Lejon, T.; Luthman, K. Diastereose-
lective reduction of
a chiral N-Boc protected δ-amino-R,â-
unsaturated γ-keto ester Phe-Gly dipeptidomimetic. J . Org.
Chem. 2002, 67, 9186-9191.
(39) Berts, W.; Luthman, K. Synthesis of a complete series of C-4
fluorinated Phe-Gly mimetics. Tetrahedron 1999, 55, 13819-
13830.
(40) Deziel, R.; Plante, R.; Caron, V.; Grenier, L.; LlinasBrunet, M.;
Duceppe, J . S.; Malenfant, E.; Moss, N. Practical and diastereo-
selective synthesis of ketomethylene dipeptide isosteres of the
type AAΨ[COCH2]Asp. J . Org. Chem. 1996, 61, 2901-2903.
(41) The natural dipeptides used in this study (Phe-Gly and Phe-
Ala) are labile to enzymatic hydrolysis in cell culture. As there
is no way to circumvent this problem, it should be kept in mind
that there are potential uncertainties associated with the
calculations. The possible effects on the Ki values of Phe-Gly and
Phe-Ala will be less pronounced in the Caco-2 assay (incubation
time 5 min) than in the SKPT assay (incubation time 30 min).
Regarding Gly-Pro, it has been used in several studies of PEPTX-
mediated transport due to its relatively high stability (Ganapa-
thy, V.; Mendicino, J .; Pashley, D. H.; Leibach, F. H. Carrier-
mediated transport of glycyl-L-proline in renal brush border
vesicles. Biochem. Biophys. Res. Commun. 1980, 97, 1133-1139).
(42) Bravo, S. A.; Nielsen, C. U.; Amstrup, J .; Frokjaer, S.; Brodin,
B. Epidermal growth factor (EGF) decreases PepT2 transport
capacity and expression in the rat kidney proximal tubule cell
line SKPT0193 cl.2. Am. J . Physiol. Renal Physiol. 2003, in
press.
(43) Another explanation is that the different calculation methods
used can influence the results shown in Figures 7 and 8 and
hence the conclusions drawn from these results. Papp is calculated
from the linear part of the curve describing the amount of
substrate transported as a function of time (eq 4) and is a steady-
state value. In contrast, the apical uptake is not a steady-state
value as such, since the uptake is measured under both pre-
steady- and steady-state conditions (eq 3). However, the uptake
experiments will still reflect the contribution from hPEPT1, and
since this seems to be absent, the transepithelial transport
(which suggests the contribution of hPEPT1) must be described
by another mechanism.
(48) Studies have also indicated that substrates may interact with a
hydrophobic pocket in the R1-position (Nielsen, C. U.; Andersen,
R.; Brodin, B.; Frokjaer, S.; Taub, M. E.; Steffansen, B. Dipeptide
model prodrugs for the intestinal oligopeptide transporter.
Affinity for and transport via hPepT1 in the human intestinal
Caco-2 cell line. J . Controlled Release 2001, 76, 129-138).
(49) It should be mentioned that there is
a possibility for the
involvement of the peptide transporter 1 (hPT1) in the transport
of dipeptidomimetics as suggested for valaciclovir (Landowski,
C. P.; Sun, D.; Foster, D. R.; Menon, S. S.; Barnett, J . L.; Welage,
L. S.; Ramachandran, C.; Amidon, G. L. Gene expression in the
human intestine and correlation with oral valacyclovir pharma-
cokinetic parameters. J . Pharmacol. Exp. Ther. 2003, 306, 778-
786); however, the role of hPT1 in transport of peptides and
peptidomimetics is still not well understood.
(50) Nielsen, C. U.; Amstrup, J .; Steffansen, B.; Frokjaer, S.; Brodin,
B. Epidermal growth factor inhibits glycylsarcosine transport
and hPepT1 expression in a human intestinal cell line. Am. J .
Physiol. Gastroint. Liver Physiol. 2001, 281, G191-G199.
(51) Ouyang, H.; Tang, F.; Siahaan, T. J .; Borchardt, R. T. A modified
coumarinic acid-based cyclic prodrug of an opioid peptide: Its
enzymatic and chemical stability and cell permeation charac-
teristics. Pharm. Res. 2002, 19, 794-801.
(52) Cheng, Y. C.; Prusoff, W. H. Relationship between the inhibition
constant (KI) and the concentration of inhibitor which causes
50% inhibition (I50) of an enzymatic reaction. Biochem. Phar-
macol. 1973, 22, 3099-3108.
J M031022+