Endomorphin-1 analogs with enhanced metabolic stability and systemic analgesic activity: Design, synthesis, and pharmacological characterization

Article abstract of DOI:10.1016/j.bmc.2006.12.007

We synthesized four new analogs of endomorphin-1 by systematic chemical modifications. To identify the best possible drug candidates for clinical pain management and to investigate the potential contribution of these alterations to the biological activity, their pharmacological properties were determined. All of the analogs showed significantly enhanced metabolic stability. The fact that centrally mediated analgesia following peripheral administration was observed with one of the analogs suggested the approach design undertaken here had validity in the development of endomorphin-1 as a successful opioid drug for the clinic.

Full text of DOI:10.1016/j.bmc.2006.12.007

Bioorganic & Medicinal Chemistry 15 (2007) 1694–1702
Endomorphin-1 analogs with enhanced metabolic stability and
systemic analgesic activity: Design, synthesis, and
pharmacological characterization
Hongmei Liu,a, Bangzhi Zhang,a, Xuefeng Liu,a Changlin Wang,a
Jingman Nib,*,à and Rui Wanga,c,
*
aState Key Laboratory of Applied Organic Chemistry, and Institute of Biochemistry and Molecular Biology,
School of Life Sciences, Lanzhou University, Lanzhou 730000, People’s Republic of China
bSchool of Pharmacy, Lanzhou University, Lanzhou 730000, People’s Republic of China
cKey Laboratory of Preclinical Study for New Drugs of Gansu Province, Lanzhou 730000, People’s Republic of China
Received 12 October 2006; revised 6 December 2006; accepted 7 December 2006
Available online 12 December 2006
Abstract—We synthesized four new analogs of endomorphin-1 by systematic chemical modifications. To identify the best possible
drug candidates for clinical pain management and to investigate the potential contribution of these alterations to the biological
activity, their pharmacological properties were determined. All of the analogs showed significantly enhanced metabolic stability.
The fact that centrally mediated analgesia following peripheral administration was observed with one of the analogs suggested
the approach design undertaken here had validity in the development of endomorphin-1 as a successful opioid drug for the clinic.
ꢀ 2006 Elsevier Ltd. All rights reserved.
1. Introduction
It is widely accepted that the mediation of opioid anal-
gesic effects occurs exclusively within the central nervous
system (CNS). Unfortunately, exogenous application of
native opioid peptides in general has a limited in vivo
efficacy, owing to their poor metabolic stability and lim-
ited delivery to the CNS. Therefore, in spite of the
strong analgesic activities of endomorphins, some of
their unfavorable pharmacological properties have hin-
dered their development for clinical use. So far,
although a large number of structural analogs of endo-
morphins have been synthesized,5–8 only a very few are
reported to be able to gain access to the CNS and pro-
duce analgesia after peripheral administration.9–11 Very
recently, potent and long-lasting analgesic effects were
observed with five endomorphin-1 analogs developed
by our research group when given subcutaneously.12
Many peptides are potential neuropharmaceuticals, and
the study of naturally occurring peptides provides a ra-
tional and potentially powerful approach in the design
of peptide therapeutics. In 1997, two potent endogenous
opioid peptides, endomorphin-1 (Tyr-Pro-Trp-Phe-
NH2) and endomorphin-2 (Tyr-Pro-Phe-Phe-NH2),
which have high affinity and selectivity for the l-opioid
receptor, were isolated from bovine.1 Since their first
description, these two neuropeptides have been exten-
sively characterized pharmacologically. They are found
to elicit equipotent analgesia to morphine but without
some of its undesirable side effects. Furthermore, there
are evidences that endomorphin-1 has a more favorable
therapeutic profile than endomorphin-2 and other l-opi-
oids,2–4 thus providing a new structural motif for more
effective analgesics with increased therapeutic potential.
Previous studies have shown that it is the presence of an
active saturable brain-to-blood efflux system at the
blood–brain barrier (BBB) that limits the entry of endo-
morphin-1 into the brain.13,14 Although specific trans-
port mechanisms are perhaps the best targets to
enhance their bioavailability to the brain, appropriate
physicochemical modification to endomorphins remains
a simple and viable method that may ultimately provide
a vital increment in their CNS permeability, thus
Keywords: Endomorphin-1; Systematic chemical modification; Meta-
bolic stability; Antinociceptive activity.
*
Corresponding authors. Tel.: +86 931 8912567; fax: +86 931 8911255;
e-mail addresses: lynjm2004@sohu.com; wangrui@lzu.edu.cn
Both authors contributed equally to this work.
à
Co-corresponding author. Tel.: +86 931 8915683.
0968-0896/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2006.12.007

H. Liu et al. / Bioorg. Med. Chem. 15 (2007) 1694–1702
1695
modifying their pharmacological properties following
peripheral administration. Presently, two analogs of
endomorphin-1 with modifications at both the amino-
and the carboxyl-terminus have been synthesized and
extensively investigated in attempts to develop novel
peptide analgesics as well as to determine their pharma-
cological properties. Their corresponding analogs
without C-terminal modification together with endo-
morphin-1 were also synthesized for comparison pur-
pose. The modifications introduced into the molecular
structure included Na-guanidino-addition and O-meth-
ylation on Tyr1, D-Pro-Gly substitution in position 2,
and C-terminal chloro-halogenation. Several of their
pharmacological properties based on the design ratio-
nale were determined and compared.
Initially, starting with H-Pro-OH/H-D-Pro-OH or Boc-
Phe-NH2/Boc-p-Cl-Phe-NH2, dipeptide/tripeptide seg-
ments at the N-/C-terminus were prepared by using
N,N0-dicyclohexylcarbodiimide
(DCC)/N-hydroxy-
succinimide (HOSu) coupling via a stepwise elongation
of the peptide chain in the C to N direction. Most
previous guanylation of peptides was carried out on
the entire peptide. Herein, we had improved this proce-
dure by combining the step-by-step elongation of the
peptide chain with segment-coupling strategy for a fast
and straightforward synthesis of four guanylated pep-
tides. This strategy not only resulted in excellent yield
but also reduced the risk of wasting materials. The
guanylation was performed on the N-terminal dipeptide
segment by treatment with a mixture of 1-(bis-benzyl-
oxycarbonylguanyl) pyrazole and diisopropylethylamine
(DIEA) in DMF. Then, the guanylated dipeptides and
the C-terminal dipeptides/tripeptides were assembled in
THF with excellent yield. The protecting groups were
removed by catalytic hydrogenation to give the desired
product. All of the purified peptides were characterized
by electrospray ionization (ESI)-MS, RP-HPLC, and
TLC.
2. Results and discussion
2.1. Peptide synthesis
Endomorphin-1 and its four analogs, listed in Table 1,
were synthesized by solution-phase method (Scheme 1).
Table 1. Opioid receptor binding affinities and functional bioactivities of endomorphin-1 and its analogs
Peptide no.
AA2
AA4
Ki(l) (nM)
Ki(d) (nM)
Ki(d)/Ki(l)
IC50 (nM)
MVD
GPI
1
L-Pro
p-ClPhe
Phe
538 177
792 165
919 108
1062 121
4.55 0.16a
1368 104
4789 640
1478 316
>10,000
2.5
6.0
845 173
988 259
>1000
1288 268
1343 345
>10,000
2
L-Pro
3
D-Pro-Gly
D-Pro-Gly
Tyr-Pro-Trp-Phe-NH2
p-ClPhe
Phe
1.6
>9.4
1121a
4
Endomorphin-1
>1000
11.4 1.13a
>10,000
25.6 3.47a
5093 660a
a Values are cited from our previous report.12
Scheme 1. Synthesis of analog 3. Reagents: (a) DCC-HOSu; (b) 2.5 N HCl in EtOAc; (c) NaH, CH3I in THF; (d) TFA/CH2Cl2 (v/v = 1:1); (e) 1-(Bis-
benzyloxycarbonylguanyl)pyrazole, DIEA in DMF; (f) H2/Pd–C in MeOH.

1696
H. Liu et al. / Bioorg. Med. Chem. 15 (2007) 1694–1702
We have also investigated another approach for the
preparation of guanylated dipeptide, in which the direct
guanylation on the H-Tyr(Me)-OH was studied. How-
ever, the guanylated-Tyr(Me)-OH was proved to be un-
able to couple with H-Pro-OH/H-D-Pro-OH. Therefore,
we did not consider this route for obtaining guanylated
dipeptide.
ence to the corresponding non-halogenated analogs,
which was distinct from the results observed previous-
ly.12 Overall, these data led to the hypothesis that be-
sides the undesired effect of cationization on the
receptor binding ability of peptides, the impact of
O-methylation at the N-terminus on l-affinity was more
significant than that of the C-terminal modification.
2.2. Radioligand binding and in vitro bioactivity assays
2.3. Metabolic stability
The affinity and selectivity of endomorphin-1 and its
four analogs were evaluated by radioligand binding as-
say using rat brain membranes and by bioassays using
guinea pig ileum (GPI) and mouse vas deferens
(MVD) preparations. These results are summarized in
Table 1. In the radioligand binding assay, all of the four
analogs showed low affinities to l-opioid receptor, and
showed no substantial d-opioid receptor binding activi-
ty, with Ki(d) higher than 1000 nM. In agreement with
predictions from the binding data, these analogs showed
low GPI and MVD potencies.
Using in vitro metabolic stability studies, we have as-
sessed the impact of these particular chemical modifi-
cations on the stability of peptides. Table
2
summarizes the half-lives determined for all of the test
peptides in twice washed 15% mouse brain membrane
homogenate and 100% mouse serum. All of the ana-
logs showed a significant increase in brain and serum
stability over the parent, with the half-lives in brain
exceeding 3 h. As expected, the introduction of
D-Pro significantly increased the half-lives of peptides
in both biological media, with effects being more sig-
nificant in the serum. This finding paralleled other re-
ports indicating that the introduction of D-amino acid
in general increased the stability of peptides. An addi-
tional observation of note was that the half-lives of
those halogenated analogs were shorter than those of
the non-halogenated forms in the brain homogenates,
whereas in the serum, the former were found to be
more stable than the latter.
It is believed that the activity of opioid peptides requires
an amino-terminal tyrosine residue that contains a free
hydroxyl group.15 However, recent studies suggest that
the para hydroxyl group on the aromatic nucleus of
Tyr1 is not an absolute requirement for the receptor
interaction.16–20 In our previous report, we described
the synthesis of a variety of analogs of endomorphin-1
and their pharmacological properties.12 In the present
study, the four analogs of endomorphin-1 in which the
alterations in the molecular structure were essentially
similar to four of the compounds disclosed previously
except for the p-methoxyphenylalanine moiety at the
cationized N-terminal. By comparing the four pairs of
compounds, our data indicated that O-methylation de-
creased both the l-opioid receptor affinity and function-
al bioactivity but with variable effects on d-opioid
receptor affinity; namely, for the two pairs of analogs
with halogenation modification at the C-terminus,
O-methylation increased the d-opioid receptor affinity
by approximately 2-fold, whereas for the other two pairs
of corresponding non-halogenated forms, it resulted in a
decrease in d-opioid receptor affinity. These data sug-
gested that this additional group may prevent or hinder
the binding of peptide into its receptor site, thus being
unfavorable for ligand-receptor interaction at the l-opi-
oid receptor. On the other hand, although only four
analogs were analyzed here, it appeared that with the
same modified Tyr in position 1, the C-terminal chlo-
ro-halogenation slightly increased the l-affinity in refer-
In the face of enhancing the metabolic stability of endo-
morphin-1, the unique approach undertaken here relied
upon utilizing N-terminal cationization by guanidino-
addition and D-Pro substitution in position 2. Previous
studies have shown that cationization can improve the
CNS entry of many peptides which is related to the pres-
ence of anionic sites on the endothelial cell membranes
of the BBB and the ability of cationized peptides to
use absorptive-mediated endocytosis (AME) to cross
the BBB.21 The highly basic nature of the guanidino
function may be an important factor that controls the
translocation of peptides. Besides, guanidino-addition
to Tyr1 may result in an enhanced resistance against
enzymatic degradation, as seen previously.9,12 In 2002,
Paterlini et al. showed in a study that drastic loss of bio-
logical activity was observed by replacement of L-Pro
with D-Pro in endomorphin-1 which may be due to a
resulted change in peptide conformation.22 Similarly,
Okada et al. found that the inversion of chirality at
Pro2 in endomorphin-2 also significantly attenuated
the l affinity,23 but this modification resulted in an
Table 2. Stability half-lives in brain and serum, octanol/buffer coefficients and percentage of protein binding in Ringer’s solution for endomorphin-1
and its analogs
Peptide no.
AA2
AA4
Stability (half-lives, min)
Octanol/buffer coefficient (D)
Protein binding (%)
Brain
Serum
1
L-Pro
p-ClPhe
Phe
190 3.5
258 27
409 34
447 27
21.2 1.8a
48.2 4.3
31.3 1.3
216 26.2
165 21.3
9.4 1.3a
21.2 4.8
65.57 9.59
40.58 3.49
62.52 3.47
45.15 2.94
32.9 4.46a
2
L-Pro
1.63 0.21
43.3 3.56
6.04 0.25
12.5 0.32a
3
D-Pro-Gly
D-Pro-Gly
Tyr-Pro-Trp-Phe-NH2
p-ClPhe
Phe
4
Endomorphin-1
a Values are cited from our previous report.12

H. Liu et al. / Bioorg. Med. Chem. 15 (2007) 1694–1702
1697
analog that produced more potent and longer-lasting
antinociception after ventricular administration.24 Pres-
ently, the introduction of D-Pro in position 2 was mainly
for stability enhancement purpose. However, it must be
considered that the potential biologically active struc-
ture of endomorphins was suggested to be a b-turn con-
formation.25 Since Gly essentially lacks a side chain and
is believed to have a strong turn-forming propensity due
to its flexibility and its small size, herein, we introduced
Gly as a spacer residue in position 3 with the hope of
ensuring a b-turn conformation of the peptide.
than endomorphin-1. It is likely that there will be less
free fraction of analogs available in the circulation to
gain access to the CNS and produce analgesia. Howev-
er, although it is often assumed that plasma protein-
bound drug is not available for transport through the
BBB, studies in the past have shown that many plasma
protein-bound drugs are available for transport through
the brain capillary endothelial wall.31 Therefore, exten-
sive binding of peptides to serum protein might actually
be protecting peptides from enzymatic degradation by
systemic peptidases.
As for the influence of O-methylation on the stability, by
comparison with our previous results, it was found that
this structural modification resulted in a slight increase
of the serum stability but had little effect on the brain
stability.
2.6. Assessment of antinociception
Antinociceptive activities of endomorphin-1 and its
analogs were studied in the tail-flick test in mice after
intracerebroventricular (icv) and subcutaneous (sc)
administration. Figure 1 shows the time course of
the antinociceptive activity of peptides. With icv
administration, although the four analogs produced
a less potent antinociceptive activity compared with
the parent, with an analgesic potency being approxi-
mately 50% of that of endomorphin-1 at the same
2.4. Octanol/buffer distribution
The relative lipophilicity of peptides as determined by
octanol/buffer distribution is shown in Table 2. Analog
3 exhibited the highest lipophilicity. Analogs 1 and 3
were found to be more lipophilic than their correspond-
ing non-halogenated forms (analogs 2 and 4), with D
values being increased by 13- and 7.2-fold, respectively,
indicating that halogenation of a peptide indeed increase
its overall lipophilicity.26–29 Additionally, comparisons
of D values between analogs 1 and 3 as well as between
analogs 2 and 4 revealed that the substitution of L-Pro
in position 2 by D-Pro-Gly also resulted in a slight in-
crease of lipophilicity, which was in agreement with
our previous observations.12
The introduction of hydroxymethyl on Tyr was mainly
for hydrogen bonding reduction and lipophilicity
improvement purpose, which was modeled after the effect
methylation has on morphine (i.e., codeine).30 The effect
of O-methylation on the lipophilicity can be assessed by
comparing the four pairs of compounds in each of which
one member differs from the other only in the N-terminal
Tyr1 of the phenolic hydroxyl by methylation. Our results
indicated that this single modification increased the over-
all lipophilicity by approximately 2- to 5-fold. Herein,
both analog 2 and analog 4 showed significantly de-
creased lipophilicity compared with endomorphin-1,
indicating that the N-terminal cationization drastically
decreased the lipophilicity despite the expected positive
effect of O-methylation, an observation previously report-
ed for similar analogs.9,12 Taken together, we concluded
that C-terminal chloro-halogenation contributed a great-
er part than O-methylation to the overall lipophilicity
enhancement, and these two modifications in combina-
tion effectively overcame the undesired impact of N-ter-
minal cationization on lipophilicity.
Figure 1. Time course of the antinociceptive effect of icv endomorphin-
1, analogs 1 and 3 (A), and analogs 2 and 4 (B) in the mouse tail-flick
test. Groups of mice were administered an injection of 20 nmol/kg of
endomorphin-1 or its analogs. The tail-flick responses were measured
at 5, 10, 15, 20, 25, 30, and 45 min after the injection. Each value
represents the mean with SEM for 10 mice. Control mice treated with
saline only did not show any significant change of nociceptive
threshold, and these data are not shown.
2.5. Protein binding
Protein binding also plays a major role in determining
CNS uptake. We assessed the ability of peptides to bind
to BSA in the mammalian Ringer’s solution. As shown
in Table 2, all of the analogs, with the exception of ana-
log 2, were found to bind to the BSA to a greater degree

1698
H. Liu et al. / Bioorg. Med. Chem. 15 (2007) 1694–1702
dose, they showed longer durations of action. The
peak analgesic effect of endomorphin-1 occurred at
5 min after injection, rapidly declined and returned
to the pre-injection level 20 min after injection,
whereas the antinociceptive effects of the analogs
reached their peaks 5–10 min after injection, slowly
declined, and returned to the pre-injection level
30 min after injection of analog 1 and approximately
45 min after injection of analog 2, 3, and 4. We in-
ferred that this may be related to decreased suscepti-
bility to enzymatic degradation.
Although analog 3 had low binding to the l- or d-opioid
receptor in rat brain membrane and was not active in the
GPI and MVD assays, based on its highest lipophilicity
and high stability in brain and serum among all of the
peptides tested, analog 3 was selected for determination
of antinociception after sc injection. Endomorphin-1
was injected as a positive control. Our results revealed
that endomorphin-1 failed to induce any significant
antinociception at a dose of 30 mg/kg, whereas, when
equimolar dose of analog 3 was injected, low but signif-
icant antinociception was observed (Fig. 2). Significant
antinociception could be detected for approximately
1 h. It is not known at present what mechanism is
responsible for the improved analgesia, but we reasoned
that possible explanations for this phenomenon were
that the significantly enhanced serum and brain stability
together with the increased lipophilicity might increase
passive diffusion of peptide into the CNS, thus increas-
ing the bioavailability of the peptide to the brain. Fur-
thermore, it was surprising to find that with similar
in vitro stability and an approximately 4-fold higher
lipophilicity, analog 3 exhibited a less potent and shorter
duration of action in the antinociception test after sc
administration when compared to its non-methylated
correlate.12 We assumed that the lower potency was
partly due to its decreased l-opioid receptor affinity
and functional bioactivity based on the current biologi-
Figure 3. Antagonism of antinociception elicited by sc 30 mg/kg of
analog 3 in the tail-flick assay by naloxone (10 nmol/mouse icv,
administered 5 min before drug) and naloxone methiodide (10 mg/kg
sc, administered 10 min before drug). Each value represents the mean
with SEM for five to seven mice. The vertical line indicates the SEM of
the mean. ***p < 0.001, compared with the drug-injected control.
cal data tested. Besides, we found that despite its higher
lipophilicity, analog 3 had a lower protein binding abil-
ity to BSA in the mammalian Ringer’s solution than its
non-methylated correlate (unpublished data). As illus-
trated in Figure 3, we have also shown that the antino-
ciceptive effect of analog 3 after sc injection can
significantly be antagonized by naloxone, but not by
naloxone methiodide, an opioid antagonist that does
not readily cross the BBB, indicating that the analgesia
is mediated by l-opioid receptor largely through a cen-
tral mechanism.
3. Conclusion
The scope of the present study was to investigate the
pharmacological characteristics of endomorphin-1 and
a series of its analogs designed by systematic structural
modification, and to identify the best possible drug can-
didates for clinical nociceptive pain management as well
as to determine the potential contribution of these alter-
ations to the biological activity, thus furthering our
knowledge of the structural requirements necessary for
pharmacological effects. All of the analogs had low
binding to the l- or d-opioid receptor in rat brain mem-
brane and showed low GPI and MVD potencies. Cati-
onization of endomorphin-1 by guanidino-addition on
Tyr1 led to a significant increase in metabolic stability
in both brain and serum. The introduction of D-Pro-
Gly in place of L-Pro not only drastically enhanced their
stability but also resulted in a slight increase in lipophil-
icity. C-terminal chloro-halogenation produced analogs
that all showed an increase in lipophilicity and in protein
binding affinity over their corresponding non-halogenat-
ed forms. By comparing the four pairs of analogs of
cationized endomorphin-1 in each of which one member
differed from the other only in the N-terminal Tyr1 of
the phenolic hydroxyl by methylation that were previ-
ously reported, we concluded that O-methylation in-
creased the overall lipophilicity and serum stability
with no significant influence on brain stability, while at
Figure 2. Time course of the antinociceptive effect of sc endomorphin-
1 and analog 3 at a dose of 30 mg/kg in the mouse tail-flick test. The
tail-flick responses were measured at 5, 10, 15, 20, 25, 30, 45, and
60 min after the injection. Each value represents the mean with SEM
for six to eight mice. Control mice treated with saline only did not
show any significant change of nociceptive threshold, and these data
are not shown.

H. Liu et al. / Bioorg. Med. Chem. 15 (2007) 1694–1702
1699
the same time decreasing l-opioid receptor affinity and
functional bioactivity but with variable effects on d-re-
ceptor affinity. Antinociception studies showed that icv
administration of the four analogs elicited less potent
but longer-lasting analgesia. Inconsistent with its bind-
ing property, the selected analog 3, Na-amidino-
Tyr(Me)-D-Pro-Gly-Trp-p-Cl-Phe-NH2, was found to
display potent and prolonged antinociceptive activity
upon sc administration through a central mechanism.
Overall, the results indicated that these combined chem-
ical modifications in the molecular structure of endo-
morphin-1 resulted in alterations in receptor binding,
lipophilicity, stability, protein binding affinity, and pos-
sibly some other physicochemical properties, which in
combination contributed to the altered analgesic effects.
Finally, due to its enzyme resistance and antinociceptive
activity after peripheral administration, analog 3 can be
regarded as a promising modified endomorphin model
and may validate further study of developing a peptide
analgesic based on the endomorphin sequence.
ous trifluoroacetic acid (TFA) over 15 min was used.
The flow rate was 0.6 mL/min.
4.3. Synthesis of Boc-p-Cl-Phe-NH2 (3a)
To a stirred suspension of Boc-p-Cl-Phe-OH (900 mg,
3 mmol) and HOSu (380 mg, 3.3 mmol) in anhydrous
THF (15 mL), a solution of DCC (742 mg, 3.6 mmol) in
anhydrous THF (5 mL) was added at 0 ꢁC. The reaction
mixture was stirred for 30 min at the same temperature
and 6 h at room temperature. After N,N0-dic-
yclohexylthiourea (DCU) was discarded by filtration, to
the remained solution was added ammonia (2 mL) at
0 ꢁC. The reaction mixture was stirred for 30 min at the
same temperature and overnight at room temperature.
Solvents were then evaporated in vacuo and the residue
was solidified from ethanol and H2O to give white powder
(882 mg, 98%). ESI-MS: calculated for C14H19ClN2O3
[M+H]+, 300.11; found, 300.29; TLC Rf (I) 0.35.
4.4. Synthesis of Boc-Trp-p-Cl-Phe-NH2 (3b)
4. Experimental section
4.1. Chemicals
Boc-p-Cl-Phe-NH2 (1.05 g, 3.5 mmol) was dissolved into
2.5 N HCl/EtOAc (20 mL) and stirred for 2 h at room
temperature. The pH of the solution was adjusted to 9
with 2 N NaOH. To a stirred suspension of Boc-Trp-
OH (1.12 g, 3.67 mmol) and HOSu (464 mg, 4.0 mmol)
in anhydrous THF (20 mL), DCC (907 mg, 4.4 mmol)
in THF (5 mL) was added to the solution at 0 ꢁC. The
reaction mixture was stirred for 30 min at the same tem-
perature and 6 h at room temperature. DCU was dis-
carded by filtration. The remained solution was added
to the amine component at 0 ꢁC. This mixture was stir-
red for 30 min at the same temperature and overnight at
room temperature. Solvents were then evaporated in
vacuo and the residue was diluted with EtOAc
(200 mL) and washed with 5% citric acid, saturated
NaHCO3, and brine. The organic phase was dried over
Na2SO4, filtered, and evaporated in vacuo. The residue
was crystallized from EtOAc to give 3b as white powder
(1.58 g, 93%). ESI-MS: calculated for C25H29ClN4O4
[M+H]+, 486.19; found, 486.18; TLC Rf (III) 0.32.
Radioligands [3H]H-Tyr-D-Ala-Gly-MePhe-Gly-ol and
[3H]H-Tyr-c[D-Pen-Gly-Phe-D-Pen]-OH were purchased
from GE Healthcare (Little Chalfont, Buckinghamshire,
UK). Commercial N-Boc/Cbz-protected amino acids
were obtained from GL Biochem (Shanghai) Ltd, Chi-
na. 1H-Pyrazole-1-carboxamidine hydrochloride was
purchased from Aldrich Chemical Co. (Milwaukee,
WI). Naloxone hydrochloride and naloxone methiodide
were purchased from Sigma–Aldrich (St. Louis, MO).
4.2. General methods
Synthesis of all of the peptides was conducted by
conventional solution methods with N-Boc and N-Cbz
protection and DCC/HOSu coupling. N-a-t-Boc-O-
methyl-tyrosine was prepared by simple methylation of
Boc-Tyr-OH with sodium hydride (NaH) and iodome-
thane (CH3I) in THF and used without further purification.
1-(bis-benzyloxycarbonylguanyl) pyrazole was prepared
according to the reported method with Cbz-Cl, Cbz-
OSu, and 1H-pyrazole-1-carboxamidine.32 The first step
in the synthesis of the two of chloro-halogenated ana-
logs involved the preparation of the amide of Boc-p-
Cl-Phe-OH that was easily achieved by DCC/HOSu
coupling with ammonia. TLC was performed on silica
gel 60 F254-precoated glass plates with the following sol-
vent systems: (I) ethyl acetate/petroleum ether (2:2), (II)
ethyl acetate/petroleum ether (2:1), (III) ethyl acetate/pe-
troleum ether (3:1), (IV) ethyl acetate/methanol (5:0.3),
and (V) ethyl acetate/methanol/ammonia (6:2:1). UV
light, I2 vapor, and water were applied to visualize the
TLC spots. Flush column chromatography for the inter-
mediates and final products was performed using silica
gel 60. Analytical HPLC was performed on a reverse
phase Water Delta Pak C18 column (3.9 · 150 mm) with
the absorbance monitored at 280 nm. An eluting system
of a linear gradient of 20–80% acetonitrile in 0.1% aque-
4.5. Synthesis of Boc-Gly-Trp-p-Cl-Phe-NH2 (3c)
The same peptide coupling procedure as for compound
3b for the synthesis of Boc-Gly-Trp-p-Cl-Phe-NH2 was
employed. The Boc-protected tripeptide was solidified
from EtOAc to give 3c as white powder with a yield of
95%. ESI-MS: calculated for C27H32ClN5O5 [M+H]+,
543.21; found, 543.57; TLC Rf (IV) 0.57.
4.6. Synthesis of Boc-Tyr(Me)-D-Pro-OH (3d)
The same peptide coupling procedure as for com-
pound 3b for the synthesis of Boc-Tyr(Me)-D-
Pro-OH was employed, but the saturated NaHCO3
washing in the workup process was skipped. The
residue was purified by flush chromatography eluted
with petroleum ether/EtOAc/acetic acid (50:80:1) to
give 3d with a yield of 78%. ESI-MS: calculated for
C20H28N2O6 [M+H]+, 393.20; found, 393.27; TLC Rf
(III) 0.42.

1700
H. Liu et al. / Bioorg. Med. Chem. 15 (2007) 1694–1702
4.7. Synthesis of Na-(N, N0-Bis-benzyloxycarbonyl) ami-
dino-Tyr(Me)-D-Pro-OH (3e)
4.12. Na-amidino-Tyr(Me)-D-Pro-Gly-Trp-Phe-NH2 (4)
ESI-MS: calculated for C38H45N9O6 [M+H]+, 724.35;
found, 724.28; TLC Rf (V) 0.32; RP-HPLC tR =
9.86 min.
Dipeptide Boc-Tyr(Me)-D-Pro-OH (250 mg, 0.64 mmol)
was dissolved in TFA/CH2Cl2 (v/v = 1:1, 2 mL) at 0 ꢁC
and stirred for 1 h. Solvents were then evaporated by
EtOAc in vacuo to give Boc-removed dipeptide as color-
less oil (178 mg, 0.61 mmol). Under argon, this amine
component and (benzyloxycarbonylimino-pyrazol-1-yl-
methyl) carbamic acid benzyl ester (193 mg, 0.51 mmol)
were dissolved in DMF (1.5 mL). The pH of the solution
was adjusted to 8 with DIEA at 0 ꢁC. The mixture was
stirred for 72 h at room temperature, and DIEA was
added in the middle of the reaction to keep the pH 8.
The reaction mixture was poured into 5% citric acid,
and the precipitated oil was extracted with EtOAc
(150 mL). The EtOAc solution was washed with brine,
dried over Na2SO4, filtered, and evaporated in vacuo.
The residual oil was purified by flush chromatography
eluted with petroleum ether/ethyl acetate/acetic acid
(10:17:1) to give colorless oil (235 mg, 77%). ESI-MS:
calculated for C32H34N4O8 [M+H]+, 603.24; found,
603.11; TLC Rf (II) 0.43.
4.13. Tyr-Pro-Trp-Phe-NH2 (endomorphin-1)
ESI-MS: calculated for C34H38N6O5 [M+H]+, 611.29;
found, 611.30; TLC Rf (V) 0.63; RP-HPLC tR =
9.19 min.
4.14. Radioligand binding and in vitro bioactivity assays
The affinity and selectivity of peptides had been investi-
gated by radioligand binding and functional in vitro bio-
assay. Competition binding studies of endomorphin-1
and its analogs for the l-opioid receptor and d-opioid
receptor were performed on synaptosomal brain mem-
branes according to the literature method.33 All binding
experiments were performed in 50 mM Tris–HCl buffer,
pH 7.4, at a final volume of 0.5 mL containing
300–500 lg/mL protein. The following conditions were
used for incubations: [3H]Tyr-D-Ala-Gly-MePhe-
Gly-ol, 0.5 nM, 1 h; [3H]Tyr-c[D-Pen-Gly-Phe-D-Pen]-
OH, 1 nM, 3 h. The extent of non-specific binding was
determined in the presence of 10 lM naloxone. All
experiments were carried out in duplicate assays and
repeated at least three times. Ki values were calculated
according to the equation of Cheng and Prusoff.34
4.8. Synthesis of Na-(N, N0-Bis-benzyloxycarbonyl) ami-
dino-Tyr(Me)-D-Pro-Gly-Trp-p-Cl-Phe-NH2 (3f)
The same peptide coupling procedure as for compound
3b for the synthesis of Na-(N, N0-bis-benzyloxycarbonyl)
amidino-Tyr(Me)-D-Pro-Gly-Trp-p-Cl-Phe-NH2 was
employed. The residual oil was purified by flush chroma-
tography eluted with petroleum ether/ethyl acetate/
methanol/acetic acid (20:80:4:1) to give colorless oil with
a yield of 68%. ESI-MS: calculated for C54H56ClN9O10
[M+Na]+, 1049.53; found, 1049.70; TLC Rf (IV) 0.32.
In vitro opioid activities of peptides were tested in the
guinea pig ileum (GPI) and mouse vas deferens
(MVD) bioassays as reported previously.12 Briefly, the
GPI tissue and MVD tissue were mounted in a 10-mL
bath that contained aerated (95% O2, 5% CO2) Krebs-
Henseleit solution at 37 and 36 ꢁC, respectively. Twitch
contractions were evoked by rectangular pulses with the
following parameters: 0.1-Hz, 50-V, 0.5-ms pulse width
for GPI assay; pairs (100-ms pulse distance) of rectangu-
lar impulses (1-ms pulse width, 9 V/cm, i.e., supramaxi-
mal intensity) were repeated by 10 s for MVD assays.
Isometric responses were recorded using a train gauge
transducer linked to a recorder system (model BL-
420E+, Taimeng technology Corporation of Chengdu,
China). Dose–response curves were constructed, and
IC50 values were calculated graphically. The values were
arithmetic means of five to eight measurements.
4.9. Synthesis of Na-amidino-Tyr(Me)-D-Pro-Gly-Trp-p-
Cl-Phe-NH2 (3)
To the solution of protected pentapeptide in methanol
[6:1, w/v (mg/mL)] was added 20% Pd–C. This solu-
tion was vigorously stirred under H2 atmosphere for
2 h at room temperature. The catalyst was removed
by filtration, and the filtrate was concentrated in vac-
uo. The residue was purified by flush chromatography
eluted with ethyl acetate/methanol/ammonia (45:20:8)
to give a white powder with a yield of 51%. ESI-
MS: calculated for C38H44ClN9O6 [M+H]+, 758.31;
found, 758.21; TLC Rf (V) 0.35. RP-HPLC
tR = 10.37 min.
4.15. Metabolic stability
Enzymatic degradation studies of endomorphin-1 and
its analogs were carried out using 15% mouse brain
homogenate and 100% mouse serum that were prepared
as described previously.35 A final protein concentration
of 2.3 mg/mL in 50 mM Tris buffer, pH 7.4, was used
for all incubations. About 10 lL of a 10 mM peptide
stock solution was added to 190 lL of the matrix. To
resuspended 15% brain homogenate or serum were
added peptides to a final concentration of 500 lM.
Incubations were carried out at 37 ꢁC for 0, 60, 120,
180, 240, and 300 min in triplicate. Aliquots of 20 lL
were withdrawn from the incubation mixtures and
4.10. Na-amidino-Tyr(Me)-Pro-Trp-p-Cl-Phe-NH2 (1)
ESI-MS: calculated for C36H41ClN8O5 [M+H]+, 701.29;
found, 701.25; TLC Rf (V) 0.34; RP-HPLC tR =
10.25 min.
4.11. Na-amidino-Tyr(Me)-Pro-Trp-Phe-NH2 (2)
ESI-MS: calculated for C36H42N8O5 [M+H]+, 667.33;
found, 667.24; TLC Rf (V) 0.31; RP-HPLC tR =
9.74 min.

H. Liu et al. / Bioorg. Med. Chem. 15 (2007) 1694–1702
1701
enzyme activity was terminated by precipitating proteins
with 90 lL of glacial acetonitrile, vortexing, and
maintaining the sample on ice for a couple of minutes.
Samples were then diluted with 90 lL of 0.5% acetic acid
to prevent further enzymatic breakdown and centrifuged
at 13,000g for 15 min. The supernatants were collected
for analysis by RP-HPLC as described below. The rate
constants of degradation (k) were obtained by least
square linear regression analysis of logarithmic peptides,
peak areas [ln (At/A0)] versus time courses, using a min-
imum of five points. Degradation half-lives (t1/2) were
calculated from the rate constants as ln2/k.
0.1% trifluoroacetic acid in water; and B, 0.1% trifluoro-
acetic acid in acetonitrile. The column was eluted at a
flow rate of 0.6 mL/min with a linear gradient of
A:B = 80:20 to A:B = 20:80 for 12 min and
A:B = 20:80 to A:B = 80:20 for 3 min.
4.19. Assessment of antinociception
Antinociceptive activities of endomorphin-1 and its ana-
logs were assessed using the 50 ꢁC warm water tail-flick
test in mice after icv (4 lL) and sc (100 lL) administra-
tion. Nociception was evoked by immersing the mouse’s
tail in hot water (50 0.2 ꢁC) and measuring the latency
to withdrawal. Before treatment, each mouse was tested,
and the latency to tail-flick was recorded [control
latency (CL)]. Mice not responding within 5 s were
excluded from further testing; the tail-flick responses
were measured at different times after icv or sc injection
of drugs. The latency to tail-flick was defined as the test
latency (TL); a cut-off of 10 s was adopted; 0.9% saline
was used as control. The antinociceptive response was
expressed as percentage of maximal possible effect
(%MPE), calculated by the following equation:
%MPE = 100 · (TL ꢀ CL)/(10 ꢀ CL). For the study
involving the opioid antagonists, animals were pre-
treated with naloxone (10 nmol/mouse icv) or naloxone
methiodide (10 mg/kg sc) before sc challenge with
peptide (30 mg/kg).
4.16. Octanol/buffer distribution
Partition coefficients for peptides were expressed as the
ratio of peptide found in the octanol phase to that found
in the aqueous phase. Equal volumes of octanol and
0.05 M Hepes buffer in 0.1 M NaCl, pH 7.4, were mixed
and allowed to equilibrate for 12 h. The layers were then
separated and stored at 4 ꢁC. At testing, 50 lg of peptide
was added to 500 lL of the Hepes buffer and mixed with
500 lL of octanol by vortexing for 2 min. The octanol/
buffer solution was centrifuged in a Beckman microfuge
(Beckman Coulter, Fullerton, CA) for 1 min at
4000 rpm. After separation into aqueous and octanol
phases, peptide content of the aqueous phase was quan-
tified by RP-HPLC. A portion of the octanol phase was
lyophilized and reconstituted in methanol before RP-
HPLC analysis. All octanol/buffer distribution studies
were performed in triplicate. The octanol/buffer distri-
bution coefficient (D) was calculated as the ratio of oct-
anol layer to the buffer layer.
Acknowledgments
This work was supported by the National Natural Sci-
ence Foundation of China (Grants 20525206 and
20472026), by the Specialized Research Fund for the
Doctoral Program in Higher Education Institutions,
and by the Chang Jiang Scholar Program of the Minis-
try of Education of China.
4.17. Protein binding
The binding affinity of the peptides to BSA in the mam-
malian Ringer’s solution [117.0 mM NaCl, 4.7 mM
KCl, 0.8 mM MgSO4, 24.8 mM NaHCO3, 1.2 mM
KH2PO4, 2.5 mM CaCl2, 10 mM D-glucose, 3.9% dex-
tran (mol. wt. 70,000), and BSA, 10 gLꢀ1, pH 7.4] was
determined by ultrafiltration centrifugal dialysis. Pep-
tides were added to 0.5 mL of Ringer’s solution pre-
warmed to 37 ꢁC and then ultrafiltered using a
CentrifreeTM micropartition device (cut-off mol.
wt. = 30,000; Millipore, Americon) by centrifugation at
2000g for 10 min to separate the protein-bound drugs
from free drugs. For control samples, drugs were pipet-
ted into filtered blank Ringer’s solution as 100% recov-
ery standards. The total concentration (T) of peptides
introduced into the system and the amount found in
the ultrafiltrate (F) were determined by RP-HPLC anal-
ysis as described below. All protein binding studies were
performed in triplicate. Percentage of peptide bound to
BSA was calculated as: % bound = (T ꢀ F)/T · 100.
References and notes
1. Zadina, J. E.; Hackler, L.; Ge, L.-J.; Kastin, A. J. Nature
(Lond) 1997, 386, 499.
2. Wilson, A. M.; Soignier, R. D.; Zadina, J. E.; Kastin, A.
J.; Nores, W. L.; Olson, R. D.; Olson, G. A. Peptides 2000,
20, 1871.
3. Hung, K.-C.; Wu, H.-E.; Mizoguchi, H.; Tseng, L. F. Eur.
J. Pharmacol. 2002, 448, 169.
4. Czapla, M. A.; Gozal, D.; Alea, O. A.; Beckerman, R. C.;
Zadina, J. E. Am. J. Respir. Crit. Care Med. 2000, 162,
994.
5. Lengyel, I.; Orosz, G.; Biyashev, D.; Kocsis, L.; Al-
´
´
Khrasani, M.; Ronai, A.; To¨mbo¨ly, C.; Furst, Z.; Toth,
¨
G.; Borsodi, A. Biochem. Biophys. Res. Commun. 2002,
290, 153.
´
6. To¨mbo¨ly, C.; Ko¨ver, K. E.; Peter, A.; Tourwe, D.;
Biyashev, D.; Benyhe, S.; Borsodi, A.; Al-Khrasani, M.;
´
´
4.18. RP-HPLC analysis
´
´
Ronai, A. Z.; Toth, G. J. Med. Chem. 2004, 47, 735.
7. Zhao, Q. Y.; Chen, Q.; Yang, D. J.; Feng, Y.; Long, Y.;
Wang, P.; Wang, R. Life Sci. 2005, 77, 1155.
8. Wei, J.; Shao, X.; Gong, M. Z.; Zhu, B. B.; Cui, Y. X.;
Gao, Y. F.; Wang, R. Bioorg. Med. Chem. Lett. 2005, 15,
2986.
Samples from stability, octanol/buffer distribution, and
protein binding studies were analyzed by RP-HPLC
on a Water Delta Pak C18 column (3.9 · 150 mm; Mil-
ford, MA) with the absorbance monitored at 280 nm.
The solvents for analytical HPLC were as follows: A,

1702
H. Liu et al. / Bioorg. Med. Chem. 15 (2007) 1694–1702
9. Hau, V. S.; Huber, J. D.; Campos, C. R.; Lipkowski, A.
W.; Misicka, A.; Davis, T. P. J. Pharm. Sci. 2002, 91,
2140.
10. Spampinato, S.; Qasema, A. R.; Caliennia, M.; Muraria,
G.; Gentiluccib, L.; Tolomellib, A.; Cardillo, G. Eur.
J. Pharmacol. 2003, 469, 89.
23. Okada, Y.; Fukumizu, A.; Takahashi, M.; Shimizu, Y.;
Tsuda, Y.; Yokoi, T.; Bryant, S. D.; Lazarus, L. H.
Biochem. Biophys. Res. Commun. 2000, 276, 7.
24. Shane, R.; Wilk, S.; Bodnar, R. J. Brain Res. 1999, 815,
278.
25. Eguchi, M.; Shen, R. Y. M.; Shea, J. P.; Lee, M. S.; Kahn,
M. J. Med. Chem. 2002, 45, 1395.
26. Abbruscato, T. J.; Williams, S. A.; Misicka, A.; Lipkow-
ski, A. S.; Hruby, V. J.; Davis, T. P. J. Pharmacol. Exp.
Ther. 1996, 276, 1049.
27. Gentry, C. L.; Egleton, R. D.; Gillespie, T.; Abbruscato,
T. J.; Bechowski, H. B.; Hruby, V. J.; Davis, T. P. Peptides
1999, 20, 1229.
28. Weber, S. J.; Greene, D. L.; Sharma, S. D.; Yamamura, H.
I.; Kramer, T. H.; Burks, T. F.; Hruby, V. J.; Hersh, L. B.;
Davis, T. P. J. Pharmacol. Exp. Ther. 1991, 259, 1109.
29. Weber, S. J.; Abbruscato, T. J.; Brownson, E. A.;
Lipkowski, A. W.; Polt, P.; Misicka, A.; Haaseth, R. C.;
Bartosz, H.; Hruby, V. J.; Davis, T. P. J. Pharmacol. Exp.
Ther. 1993, 266, 1649.
11. Krzanowska, E.; Znamensky, V.; Wilk, S.; Bodnar, R. J.
Peptides 2000, 21, 705.
12. Liu, H. M.; Liu, X. F.; Yao, J. L.; Wang, C. L.; Yu, Y.;
Wang, R. J. Pharmacol. Exp. Ther. 2006, 319, 308.
13. Kastin, A. J.; Fasold, M. B.; Smith, R. R.; Horner, K. A.;
Zadina, J. E. Exp. Brain Res. 2001, 139, 70.
14. Somogyvari-Vigh, A.; Kastin, A. J.; Liao, J.; Zadina, J. E.;
Pan, W. H. Exp. Brain Res. 2004, 156, 224.
15. Chang, J.-K.; Fong, B. T. W.; Pert, A.; Pert, C. B. Life Sci.
1976, 18, 1473.
16. Sasaki, Y.; Sasaki, A.; Niizuma, H.; Goto, H.; Ambo, A.
Bioorg. Med. Chem. 2003, 11, 675.
17. Gulya, K.; Pelton, J. T.; Hruby, V. J.; Yamamura, H. I.
Life Sci. 1986, 38, 2221.
18. Mosberg, H. I.; Ho, J. C.; Sobczyk-Kojiro, K. Bioorg.
Med. Chem. Lett. 1998, 8, 2681.
30. Oldendorf, W. H.; Hyman, S.; Braun, L.; Oldendorf, S. Z.
Science 1972, 178, 984.
19. McFadyen, I. J.; Sobczyk-Kojiro, K.; Schaefer, M. J.; Ho,
J. C.; Omnaas, J. R.; Mosberg, H. I.; Traynor, J. R.
J. Pharmacol. Exp. Ther. 2000, 295, 960.
31. Pardridge, W. M. Am. J. Physiol. 1987, 252, E157.
32. Bernatowicz, M. S.; Wu, Y.; Matsueda, G. R. Tetrahedron
Lett. 1993, 34, 3389.
20. Burden, J. E.; Davis, P.; Porreca, F.; Spatola, A. F.
Bioorg. Med. Chem. Lett. 1999, 9, 3441.
21. Kumagai, A. K.; Eisenberg, J. B.; Pardridge, W. M.
J. Biol. Chem. 1987, 262, 15214.
22. Paterlini, M. G.; Avitabile, F.; Ostrowski, B. G.;
Ferguson, D. M.; Portoghese, P. S. Biophys. J. 2000,
78, 590.
33. Simon, J.; Benyhe, S.; Abutidze, K.; Borsodi, A.; Szucs,
¨
´
M.; Toth, G.; Wollemann, M. J. Neurochem. 1986, 46,
695.
34. Cheng, Y. C.; Prusoff, W. H. Biochem. Pharmacol. 1973,
22, 3099.
35. Gillespie, T. J.; Konings, P. N.; Merrill, B. J.; Davis, T. P.
Life Sci. 1992, 51, 2097.