Intramolecular azo-bridge as a cystine disulfide bond surrogate: Somatostatin-14 and brain natriuretic peptide (BNP) analogs

Article abstract of DOI:10.1016/j.bmc.2010.12.014

Cystine disulfide bond is a common feature in numerous biologically active peptides and proteins and accordingly its replacement by various surrogates presents a potential route to obtain analogs with improved pharmacokinetic characteristics. The purpose

Full text of DOI:10.1016/j.bmc.2010.12.014

Bioorganic & Medicinal Chemistry 19 (2011) 798–806
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Intramolecular azo-bridge as a cystine disulfide bond surrogate:
Somatostatin-14 and brain natriuretic peptide (BNP) analogs
Gil Fridkin a, , Theodosia Maina b, Berthold A. Nock b, Dan Blat c, Vered Lev-Goldman c, Yosef Scolnik c,
⇑
Aviva Kapitkovski c, Yelena Vachutinsky a, Yoram Shechter a, Yaakov Levy d
a Department of Chemical Biology, The Weizmann Institute of Science, Rehovot 76100, Israel
b Molecular Radiopharmacy, I/R-RP, NCSR ‘Demokritos’ 15310 Ag, Paraskevi Attikis, Athens, Greece
c Department of Organic Chemistry, The Weizmann Institute of Science, Rehovot 76100, Israel
d Department of Structural Biology, The Weizmann Institute of Science, Rehovot 76100, Israel
a r t i c l e i n f o
a b s t r a c t
Article history:
Cystine disulfide bond is a common feature in numerous biologically active peptides and proteins and
accordingly its replacement by various surrogates presents a potential route to obtain analogs with
improved pharmacokinetic characteristics. The purpose of the present study was to assess whether an
azo-bridge can serve as such a surrogate. In view of the marked clinical significance of somatostatin
and the brain natriuretic peptide (BNP) we choose these peptides as a model. Three cyclic-azo somato-
statin analogs and three cyclic-azo BNP analogs were effectively prepared in solution through azo bond
formation between p-amino phenylalanine and His or Tyr residues that were positioned in the peptide
sequences in place of the native Cys residues. The peptides binding affinities to the sst2 and ANP-receptor
(NPR-A) expressed on rat acinar pancreating carcinoma AR4-2J cell membranes and HeLa cells, respec-
tively, were examined. The somatostatin analogs displayed good to moderate affinities to the rat sst2
in the nM range with best results obtained with peptide 2, that is, IC50 = 8.1 nM. Molecular dynamics sim-
ulations on these peptides suggests on a correlation between the observed binding potencies and the
degree of conformational space overlapping with that of somatostatin. The BNP analogs exhibited binding
affinities to the NPR-A in the nM range with best results obtained with BNP-1, that is, IC50 = 60 nM.
Ó 2010 Elsevier Ltd. All rights reserved.
Received 16 September 2010
Revised 26 November 2010
Accepted 4 December 2010
Available online 13 December 2010
Keywords:
Azo-cyclization
Cyclic-azo peptides
Somatostatin
BNP
1. Introduction
non-peptide,8 SST analogs, were synthesized and evaluated for their
therapeutic use. With regard to peptides, in addition to ‘traditional’
Somatostatin (SST), also known as somatotropin release-
inhibiting factor (SRIF) is a natural cyclic tetradecapeptide hormone
originally discovered in and isolated from bovine hypothalamus.1
SST was found to be expressed in brain and in most peripheral
organs and tissues, including the gut, kidneys, submandibular
glands, prostate and placenta. SST exhibits a wide range of diverse
physiological activities, such as inhibition of growth hormone,
glucagon and insulin secretion.2 These effects are elicited after its
binding to a family of G-protein coupled receptors (GPCR) of which
at least five subtypes (sst1–5) have been cloned.3 SST receptors
(sst1–5) are over-expressed in various pathologies, most impor-
tantly, in cancers, for example, neuroendocrine tumors, breast carci-
noma, neuroblastoma, pituitary adenoma, kidney, ovarian and
thyroid cancer.4,5 In view of the above SST holds significant potential
for clinical applications and consequently, numerous peptide,6,7and
structure-function studies involving sequence shortening, amino
acids substitution and incorporation of
D
-amino acids,6 emphasis
was also directed to the replacement of the intramolecular cystine
S–S bridge by other cyclization means. Cyclization was thus induced
for example, by amide bond formation,9,10 by disulfide formation
using penicillamine,11 or b-mercapto-b,b-pentamethylene propi-
onic acid,12 through lanthionine-related mean,13 by RCM,14 via
backbone cyclization15 and by backbone-metal-cyclization (BMC).16
The brain natriuretic peptide (BNP) is a natural occurring disul-
fide cyclic peptide hormone, which was originally isolated from the
brain.17 However, this 32-amino acid peptide, which belongs to the
natriuretic peptides (NP) family, is considered a circulating hor-
mone and is produced mainly in the cardiac ventricles.17 The NP
have a key role in the regulation of sodium and water balance
and accordingly show cardiovascular, renal, hemodynamic, endo-
crine and mitogenic effects.18 These diverse NP actions are medi-
ated through interactions with their receptors, which are located
on numerous tissues such as vasculature, renal artery and the cen-
tral nervous system. At least 3 different subtypes of NP receptors
(NPR) have been identified: ANP-receptor (NPR-A), BNP receptor
⇑
Corresponding author at present address: Department of Organic Chemistry,
The Israel Institute for Biological Research, Ness-Ziona, PO Box 19, 74100, Israel.
Tel.: +972 8 9381741; fax: +972 8 9381548.
E-mail addresses: gilf@iibr.gov.il, GilFridkin@hotmail.com (G. Fridkin).
0968-0896/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2010.12.014

G. Fridkin et al. / Bioorg. Med. Chem. 19 (2011) 798–806
799
(NPR-B) and CNP receptor (NPR-C). However, most of the NP bio-
logical effects are mediated through interactions with the NPR-A
and NPR-B. BNP, specifically, acts via the NPR-A and NPR-C, simi-
larly to ANP.16
3 (7/3 ratio, respectively) which could be purified by HPLC. The
possibility that peptides 2 and 3 (Scheme 1 and Fig. 3A), which
showed identical mass, represent cis and trans isomers was ex-
plored by attempting to isomerise them using heat,23 or irradia-
tion.24 As no shift was observed in the UV–vis absorption spectra
of these peptides after these procedures, we concluded that the
peptides are not cis/trans isomers but rather regioisomers presum-
ably obtained through imidazole modification at both positions 2
and 5. Since peptide 2 was obtained in excess we assume it
corresponds to formation of the azo-bridge via position 2 of the
imidazole, which seems more approachable than position 5, which
is relatively sterically hindered by the adjacent b-methylene group.
The fact that the azo-His somatostatin did not exhibit spectral
changes following irradiation may be due to internal hydrogen
bonding involving the imidazolic proton and the vicinal nitrogen
of the azo bond. Such association may contribute to the stability
of a preferred conformer.25 In the case of a phenyl-azo-phenyl
structure this possibility does not exist.
Recently we have introduced a novel approach for the synthesis
of cyclic peptides which is based on the formation of an intramo-
lecular azo bond between diazotized p-amino phenylalanine and
tyrosine or histidine residues located on the peptide backbone.19
Linear peptide sequences were used to develop this concept.19,20
Azo-cyclization, was however, never evaluated as an alternative
mode of ring formation in mimics of naturally-occurring cyclic
peptides. The azo bridge holds several interesting characteristics,
such as hydrophobicity, rigidity and pH and light sensitivity, which
may have marked effect on cyclic peptides activity, selectivity and
metabolic stability. The light sensitivity feature was specifically
utilized for the photomodulation of peptide conformations. For
example, an azobenzene moiety that was introduced into a cyclic
peptide derived from the somatostatin active core provided the
ability to modulate between peptide conformations upon applying
UV irradiation/heat.21 Nevertheless, in these short peptide mimics
which present only part of the native somatostatin sequence, the
azo bond did not directly replace the cystine disulfide bond.
The purpose of the present study was to assess whether the azo
bond can directly and effectively replace the abundant and impor-
tant cystine disulfide bond in naturally-occurring cyclic peptide se-
quences. In view of the marked clinical significance of SST and BNP,
and thus the need to diversify and enlarge the existing SST and
BNP-peptide arsenal, we choose these peptides as a model.
2.2. Binding affnity of the somatostatin analogs
The binding affinity of the three cyclic-azo peptides to the sst2
expressed on rat acinar pancreating carcinoma AR4-2J cell mem-
branes was next determined.26 As shown in the comparative dis-
placement curves (Fig. 4), the peptides displayed good to
moderate affinities for the rat sst2 (Table 1). Peptide
2
(IC50 = 8.1 nM) and 3 (IC50 = 44.0 nM), that is, the two azo-His iso-
mers, inhibited the binding of [125I-Tyr3]octreotide in a monopha-
sic and dose-dependent manner, while peptide 5 (IC50 >600 nM),
that is, azo-Tyr, marginally interacted with the receptor.
2. Results and discussion
2.3. Molecular dynamics of the somatostatin analogs
2.1. Synthesis of somatostatin analogs
In an attempt to understand this difference in binding affinities
we have performed molecular dynamics simulations on these pep-
tides aiming at capturing their available conformational space.
Accordingly, SST-14 and its five synthesized variants (Table 1)
were simulated at high temperatures; for each peptide five hun-
dred conformations were sampled, and were gradually cooled
down to represent conformations at very low temperatures. These
conformations, which may represent the most accessible basins of
the potential energy landscape, were then mapped using the Prin-
cipal Component Analysis (PCA) method27 to allow the projection
of the multidimensional conformational space of the peptides on
the same collective coordinate. The projected conformational
spaces show that the cyclization of SST significantly shrinks the
available conformational space of the peptide (Fig. 5). Peptides 2
and 3, which have the same sequence but differ only in the atoms
of His14 through which the azo-bridge is formed, have similar con-
formational spaces as the overlap between their projected confor-
mational spaces is about 90%. In addition, it was found that cyclic
peptides 2 and 3 hold a smaller conformational space than peptide
5 and present better overlapping, that is, of 73% and 66%, respec-
tively, with the conformational space sampled by SST-14 as com-
pared to peptide 5, which showed 48% overlap (Fig. 5). Peptides
2 and 3 differ from peptide 5 only in the identity of the residue
at position 14, yet show marked difference in binding activity.
The overlap between the conformational space of peptides 3 and
5 is of 72% (Fig. 5C and D). The relative bioactivity27a (calculated
by [Àlog(IC50)]peptide/[Àlog(IC50)]SST-14) of peptides 2, 3, and 5 is
86%, 79%, and 66%, respectively. Namely, the relative bioactivity
correlates well with the degree of overlap between the conforma-
tional space sampled by these peptides and by SST-14. Accordingly,
we propose that the identity of the residues can affect the flexibil-
ity of the peptide, the allowed conformational space and thus the
ability to obtain the bioactive conformation.
To directly evaluate the above question we have incorporated
our cyclization ‘tools’ namely, p-amino phenylalanine (Pap) and
His/Tyr residues in place of the original cysteines at positions 3
and 14, respectively, and maintained the rest of the peptide se-
quences identical to native SST-14 (Table 1). Linear peptides were
assembled on solid-support using a peptide synthesizer, while
azo-cyclization was performed in solution (Scheme 1). Successful
azo-cyclization, with no apparent formation of dimers or oligo-
mers, was confirmed by means of HPLC (Fig. 1), mass spectrometry,
amino acid analyses (Table 2), and by UV–vis absorption studies of
the azo chromophor (Fig. 2). These studies suggested that the
peptides were obtained as a mixture of cis/trans isomers.22 Inter-
estingly, the synthesis of the His-derived azo-bridged SST analog
resulted in the generation of two isomers termed peptides 2 and
Table 1
Cyclic-azo somatostatin peptides and their binding affinity to sst2
b
Peptide Sequencea
IC50
(nM)
13-Cys14
A1-G2-Cys3-K4-N5-F6-F7-W8-K9-T10-F11-T12-S
SST-14
0.4 0.1
S
S
A-G-Pap-K-N-F-F-W-K-T-F-T-S-His
A-G-Pap-K-N-F-F-W-K-T-F-T-S-His
1
2
ND
8.1 1.7
A-G-Pap-K-N-F-F-W-K-T-F-T-S-His
3
4
5
44.0 6.9
ND
A-G-Pap-K-N-F-F-W-K-T-F-T-S-Tyr
A-G-Pap-K-N-F-F-W-K-T-F-T-S-Tyr
>600
ND—not determined.
a
Pap—p-amino phenylalanine; the connected line between Pap and His/Tyr
represents the azo bridge.
b
Peptide concentrations producing 50% inhibition of
[
125I-Tyr3]octreotide
binding.

800
G. Fridkin et al. / Bioorg. Med. Chem. 19 (2011) 798–806
NH2
N
NH
OH
NH
O
3
OH
OH
O
O
O
O
O
O
H
N
H
H
N
H
N
H
H
H2N
N
N
N
N
H
N
H
N
H
N
H
N
H
N
H
N
H
O
O
O
O
O
O
O
OH
3
O
NH2
H2N
NH2
1) NaNO2/HCl
pH ~ 1
2) KHCO3
pH ~ 8
NH2
NH
3
O
O
O
O
H
H
N
H
H
H2N
N
N
O
N
O
O
NH2
N
N
N
N
H
H
H
H
O
O
O
O
O
NH
H2N
N
N
N
OH
HN
HO
OH
O
H
NH
N
N
N
H
H
O
O
OH
Scheme 1. Synthesis of peptide 2, a SST-14 mimic, by azo cyclization through p-amino phenylalanine and an histidine residue; diazotation at C-2 of the imidazole.
2.4. Synthesis of BNP analogs
to the case of GnRH, in the case of BNP the distance between the
Tyr/His and Pap residues was the major parameter affecting the
propensity of the peptide to close by either residue. Evidently, in
the synthesis of both GnRH and BNP the products with the larger
rings were obtained in significant higher yields. In addition, it ap-
pears that like in the case of somatostatin, once a His residue is
present azo-cyclization can occur by either the imidazole 2
(BNP-2) or 5 position (BNP-3).
Following the same rational applied in the somatostatin mimics
design, in the case of BNP the Tyr and Pap residues were incorpo-
rated in place of the native Cys residues at positions 10 and 26,
respectively, while the rest of the peptide sequence was left un-
changed (Table 3). In this manner the effect of azo-cyclization
could be evaluated directly. The 32-amino acid long linear peptide
was assembled, as before, on solid-support using a peptide synthe-
sizer, while azo-cyclization was performed in solution following
similar experimental protocols as mentioned above (Scheme 2).
The crude peptide was then purified by HPLC to furnish three dif-
ferent fractions that hold the desired molecular weight (Table 3).
Subjecting these samples to amino acid analyses revealed that
the main fraction (>80%) termed BNP-1 is the desired cyclic pep-
tide, which is obtained through azo bond formation between the
Tyr10 and Pap26 residues (Scheme 2). The second fraction however,
obtained in ꢀ15% termed BNP-2 is derived from azo-cyclization
between the Pap26 residue and the native His32 residue located at
the peptide C-terminus (Fig. 6). The feasibility of azo-cyclization
between these rather close residues, which is evident by the ab-
sence of an His and the presence of a Tyr residue in the amino acid
analysis, was further supported by the amino acid analysis of the
third fraction (BNP-3, <5%) which gave the same results. The for-
mation of two cyclic azo peptides from one linear sequence hold-
ing both Tyr and His residues was observed by us previously in
the synthesis of azo cyclic GnRH analogs.20 It appears that similar
2.5. Binding affinity of the BNP analogs
The binding affinity of the two major cyclic-azo BNP peptides
(BNP 1,2) as well as native BNP to the NPR-A expressed on HeLa
cell membranes was next examined using a compettion binding
assay with [125I]ANP as the radioligand. As can be seen in Table 3
BNP-1 was found to be the most active binder with an IC50 value
of 60 nM, which is only 10 times less active than native BNP
(6 nM). BNP-2 was however, a four times less active binder than
BNP-1 exhibiting an IC50 value of 240 nM. As mentioned before,
BNP-1 and 2 differ in the location of the azo bridge. While in
BNP-1 it is located between Tyr10 and Pap26 which is the exact po-
sition of the native disulfide bond in BNP, in BNP-2 the bridge is
placed between the Pap and His residues located in positions 26
and 32, respectively. It is reasonable to hypothesize that due to
BNP-1 similarity to BNP it may obtain conformations that closely
resemble the bioactive conformation exhibited by the latter while
for BNP-2 this process is less favoured.

G. Fridkin et al. / Bioorg. Med. Chem. 19 (2011) 798–806
801
Figure 1. HPLC spectra of somatostatin peptides. (A) crude peptide 1, the linear precursor of cyclic peptides 2 and 3. (B) Crude cyclic peptides 2 and 3. (C) Purified peptide 3.

802
G. Fridkin et al. / Bioorg. Med. Chem. 19 (2011) 798–806
Table 2
Amino acid composition of synthesized somatostatin peptidesa
Peptide No.
Ala
Gly
Lys
Asp
Phe
Thr
Ser
Tyr
His
1
2
3
4
5
(1) 0.99
(1) 0.95
(1) 0.97
(1) 0.99
(1) 1.00
(1) 1.00
(1) 1.00
(1) 1.00
(1) 1.00
(1) 1.00
(2) 2.04
(2) 1.94
(2) 1.85
(2) 2.00
(2) 1.88
(1) 1.04
(1) 0.90
(1) 0.90
(1) 1.06
(1) 1.02
(3) 3.14
(3) 2.93
(3) 2.96
(3) 3.14
(3) 3.08
(2) 1.99
(2) 1.91
(2) 1.90
(2) 1.99
(2) 1.95
(1) 0.93
(1) 0.83
(1) 0.87
(1) 0.90
(1) 0.88
—
—
—
(1) 1.04
(1) 0.07
(1) 0.06
—
(1) 0.99
(1) 0.07
—
a
p-Aminophenylalanine and tryptophan were not determined.
Figure 2. UV absorption spectra of the cyclic azo-bridged somatostatin analogs in different pH conditions.
Spectra-Physics SP-8800 liquid chromatography system equipped
with an Applied Biosystems 757 variable wavelength absorbance
detector. HPLC pre-packed columns used were: Lichrocart, contain-
3. Conclusion
The feasibility of utilizing azo-bridge as a cystine disulfide bond
surrogate was shown through the synthesis and binding studies of
SST-14 and BNP mimetics. As the purpose of the present study was
to conceptually evaluate if these peptides preserve substantial bio-
logical potency, no attempts were made to optimize the formation
of a specific conformer. Thus the affect of parameters like concen-
tration, pH and reaction temperature were not investigated. It is
reasonable to assume that the nature of the azo bond as well as
its location on the imidazole ring may vary from peptide to pep-
tide depending, for example, on intramolecular/intermolecular
interactions such as charge-charge, hydrophobic and hydrogen
bonding. Numerous biologically active peptides of high therapeu-
tic-diagnostic potential possess cystine disulfide bonds, accord-
ingly, we believe that the present findings with somatostatin
and BNP have diverse and wide implications beyond these model
peptides.
ing Lichrospher RP-18 (250 Â 10 mm; 7
l
m) for semi-preparative
purification and Lichrospher 100 RP-18 (250 Â 4 mm; 5
lm) for
analytical purposes (Merck, Darmstadt, Germany). HPLC purifica-
tion and analyses were achieved by using gradients formed from
0.1% TFA in water as solvent A and 0.1% TFA in 75% aqueous MeCN
as solvent B. Eluent composition was 10% B in A for the first 10 min
and increased linearly to 100% B, 50 min after injection time. Flow
rates were 1 ml/min and 10 ml/min for analytical and preparative
purposes, respectively. The column effluents were monitored by
UV absorbance at 214 nm. Mass analyses were performed using
MALDI-TOF and ESI-MS techniques (Bruker-Reflex-Reflection mod-
el, Germany, and VG-platform-II electrospray single quadruple
mass spectrometer, Micro Mass, UK, respectively). For amino acid
composition evaluations peptides were hydrolyzed in 6 N HCl at
100 °C for 24 h under vacuum, and the hydrolyzates were analyzed
with a Dionex Automatic Amino Acid Analyzer. The rat pancreatic
tumor cell line AR4-2J was kindly provided by Dr. R. Kleene (Phi-
lipps University, Marburg, Germany). All culture media were sup-
plied by Gibco BRL, Life Technologies (Grand Island, NY) and all
supplements by Biochrom KG SeromedÒ (Berlin, Germany). For pro-
tein measurements the protein microdetermination kit (Procedure
No. P 5656) by SIGMA Diagnostics (St. Louis, USA) was utilized. Io-
dine-125 was purchased from MDS Nordion, SA (Fleurus, Belgium).
[Tyr3]octreotide was a kind gift from the IAEA whereas SST-14 was
purchased from Sigma–Aldrich (Athens, Greece). Radioiodination of
[Tyr3]octreotide was performed following a modification of a pub-
lished protocol.26,28 For radioactivity measurements an automatic
4. Experimental
4.1. Materials and methods
All chemicals and reagents were of analytical grade. Resins,
Fmoc-protected amino acid derivatives and all reagents for solid-
phase synthesis were obtained from Novabiochem (Laufelingen,
Switzerland). Side-chain protecting groups employed were as fol-
lows: Arg, 2,2,4,6,7-pentamethyl-dihydrobenzofuran-5-sulfonyl
(Pbf); His, Trt; Trp and p-amino phenylalanine, BOC; Ser and Tyr;
tert-butyl (t-Bu). Reverse-phase HPLC was performed on
a

G. Fridkin et al. / Bioorg. Med. Chem. 19 (2011) 798–806
803
A
NH2
NH
3
O
O
O
O
H
N
H
N
H
N
H
N
H2N
NH2
N
H
N
H
N
H
N
H
O
O
N
O
O
O
O
O
O
NH
H2N
N
HN
N
OH
OH
O
H
N
NH
HO
N
H
N
H
O
O
OH
B
NH2
NH
3
O
O
O
O
H
N
H
H
H
N
H2N
N
N
NH2
N
H
N
H
N
N
H
H
O
O
O
O
O
O
O
O
NH
H2N
N
N
HO
OH
OH
O
H
N
NH
HO
N
H
N
H
O
O
OH
Figure 3. Schematic representation of somatostatin cyclic azo mimics. (A) Peptide 3. (B) Peptide 5.
4.2. Somatostatin peptide synthesis
100
75
50
25
0
The pre-cyclic linear peptides were prepared on solid-support
(Wang resin, Novabiochem, Laufelingen, Switzerland) with an
APEX 396 synthesizer (Advanced ChemTech, Louisville, KY, USA)
using Fmoc chemistry, following the company’s protocols. All syn-
thesized peptides were simultaneously deprotected/cleaved from
the resin using a solution of TFA/triethylsilane/water (17:1:1).
After 2 h at room temperature, the cleaved mixtures were filtered
and the peptides were precipitated from the solutions with perox-
ide-free dry methyl t-butyl ether at 0 °C. Precipitated peptides
were washed with cold ether, dissolved in water and lyophilized.
The crude peptides (>85% pure) were directly used, without further
purification for azo-cyclization as described before.19,20 Briefly, the
corresponding lyophilized peptide powders (5–50 mg) were dis-
10-13 10-12 10-11 10-10 10-9 10-8 10-7 10-6 10-5
[SS analog] (M)
Figure 4. Competitive inhibition of [125I-Tyr3]octreotide binding to sst2 expressed
on AR4-2J cell membranes by increasing concentrations of SST-14 (Ã), peptide 2 (h),
peptide 3 (Ç), and peptide 5 (o).
solved in 0.1 M HCl (50 lL per mg peptide) and the solutions were
cooled to 0 °C. A cooled aqueous sodium nitrite solution (0.1 M)
was then added portion-wise to achieve a 1:1 peptide/nitrite molar
ratio and the reaction mixtures were allowed to stand with occa-
sional mixing for 10 min at 0 °C. They were then added, portion-
wise, to an ice-cold 0.1 M KHCO3 solution (2 ml per 1 mg peptide),
and the reaction mixtures, following adjustment of their pH to 8.0
by addition of 1 M K2CO3, were kept for 3–4 h in the cold, followed
by overnight incubation at room temperature. The homogenous
well-type gamma counter was used [NaI(Tl) crystal, Canberra Pack-
ard Auto-Gamma 5000 series instrument]. A Brandel M-48 Cell Har-
vester (Adi Hassel Ingenieur Büro, Munich, Germany) was
employed in the binding experiments.

804
G. Fridkin et al. / Bioorg. Med. Chem. 19 (2011) 798–806
Figure 5. A joint projection of the conformational space of SST-14 and some of its variants on two principal axes. Each point represents a single conformation among the 500
conformations that represent the conformational space of each peptide. (A) Comparison between the conformational space of peptides 1, 3, and SST-14. (B) Comparison
between the conformational space of peptides 1, 5, and SST-14. (C) Comparison of the conformational space of peptides 3 and 5 using the first two principal axes. (D)
Comparison of the conformational space of peptides 3 and 5 using the first and third principal axes.
Table 3
Cyclic-azo BNP peptides and their binding affinity to NPR-A
Sequencea
IC50 (nM)
b
Peptide
BNP
S1-P2-K3-M4-V5-Q6-G7-S8-G9-C10-F11-G12-R13-K14-M15-D16-R17-I18-S19-S20-S21-S22-G23-L24-G25-C26-K27-V28-L29-R30-R31-H32
S1-P2-K3-M4-V5-Q6-G7-S8-G9-Y10-F11-G12-R13-K14-M15-D16-R17-I18-S19-S20-S21-S22-G23-L24-G25-Pap26-K27-V28-L29-R30-R31-H32
S1-P2-K3-M4-V5-Q6-G7-S8-G9-T10-F11-G12-R13-K14-M15-D16-R17-I18-S19-S20-S21-S22-G23-L24-G25-Pap26-K27-V28-L29-R30-R31-H32
S1-P2-K3-M4-V5-Q6-G7-S8-G9-T10-F11-G12-R13-K14-M15-D16-R17-I18-S19-S20-S21-S22-G23-L24-G25-Pap26-K27-V28-L29-R30-R31-H32
6
60
240
ND
BNP-1
BNP-2
BNP-3
ND—not determined.
a
Pap—p-amino phenylalanine; letters in bold represent points of cyclization; in the case of BNP, cyclization is via a disulfide bond while in the case of BNP-1-3 cyclization
is through an azo bridge.
b
Peptide concentrations producing 50% inhibition of [125I]ANP binding.
solutions gradually turned brownish-orange or yellowish upon
standing. Lyophilization resulted in yellowish-brown powdered
crude cyclic peptides, which were further purified by semi-
preparative HPLC. Lyophilized, isolated peptides were evaluated
by analytical HPLC, amino acid analysis and mass spectroscopy.
Analyses revealed high purities (>97%) of the cyclic peptides and
confirmed their identity, respectively. Mass spectrometry (m/z):
Peptide 2; calcd. [M+H]+ = 1744.96; found [M+H]+ = 1745.03. Peptide
3; calcd [M+H]+ = 1744.96; found [M+H]+ = 1744.98. Peptide 5; calcd
[M+H]+ = 1770.99; found [M+H]+ = 1771.42.
scanned for UV–vis absorption over the spectrum of 200–600 nm
(Fig. 2) using a Varian Cary 50 UV–vis spectrophotometer (Palo
Alto, CA, USA).
4.4. Attempted isomerization of the azo bond
The possibility that peptides 2 and 3 represent cis and trans iso-
mers was explored by an attempt to isomerise them using heat
(50 °C for 12 h). Following this treatment the UV–vis absorption
spectra of these peptides were examined, however, no shift was
observed. The two peptides (50 lM) were than dissolved in
4.3. UV–vis absorption studies
100 mM phosphate buffer (pH 7.0) and irradiated for 3 h at 25 °C
using an electronic projector lamp (KL1500, Scott, Germany)
equipped with filters, having a band pass of 320–510 nm with a
kmax of 400 nm. Following that procedure the UV–vis absorption
spectra of the peptides were examined, however, as before, no shift
was observed.
Lyophilized pure peptides were dissolved in double-distilled
water to obtain 50 mM solutions. Different pH values were ob-
tained by titration with minute volumes of concentrated HCl or
NaOH, until the target pH was observed. Each solution was then

G. Fridkin et al. / Bioorg. Med. Chem. 19 (2011) 798–806
805
homogenized using
a
Bioblock Scientific homogenizer (50
S
P
H2N
Tyr
K
M
V
G
R
Q
S
G
F
G
K
M
D
R
I
strokes/5 mL) and the homogenized suspension was centrifuged
at 2600 rpm for 10 min at 4 °C in a J2-MC Beckman centrifuge.
The supernatant was removed and recentrifuged at 26,000 rpm
for 15 min at 4 °C in a CS120GX micro ultracentrifuge (Hitachi).
The pellet was resuspended in homogenization buffer (100
lL/
S
S
S
flask) and stored at À80 °C in aliquots of 100
lL. The protein con-
centration of the samples was determined according to the method
of Lowry31 using bovine serum albumin as the standard.
Pap
HO2C
H
R
R
L
L
S
V
K
G
G
4.5.3. Binding assay
Competition binding experiments were performed according to
a
published protocol26 using 125I-Tyr3]octreotide28 (2200 Ci/
[
mmol) as the radioligand and SST-14 as control peptide. Briefly,
AR4-2J cell membrane homogenate, corresponding to 15 g pro-
tein, was incubated with 30,000 cpm of radioligand in the presence
of increasing concentrations of tested peptide in a total volume of
300 mL of HEPES buffer (50 mM HEPES pH 7.6, 0.3% BSA, 5 mM
1) NaNO2, HCl pH ~1
2) KHCO3 pH ~7.4
l
MgCl2, 10 lM bacitracin) for 40 min at 37 °C. Analysis was per-
formed by non-linear regression according to an one-site model
employing the PRISM™ 2 program (GraphPad Software, San Diego,
CA).
S
P
H2N
Tyr
K
M
V
G
R
Q
S
G
F
G
K
M
D
R
I
4.6. Molecular dynamics
N
N
To capture the conformational space of the peptides and the ef-
fects of the constraints on their flexibility, each system was simu-
lated using molecular dynamics simulations at high temperatures
(1000 K) for 5 ns. Along this high T trajectory, for each peptide
500 conformations were sampled and were gradually cooled down
to represent conformations at very low temperatures. The high T
trajectory allows crossing high barriers in the energy landscapes
of the peptides. The 500 conformations may represent the most
accessible basins of the potential energy landscape. Once 500 con-
formations were selected for each peptide, they were mapped
using the Principal Component Analysis (PCA) method.27
S
S
S
Pap
HO2C
H
R
R
L
L
S
V
K
G
G
Scheme 2. Synthesis of BNP-1, a BNP mimic, by azo cyclization through p-amino
phenylalanine and tyrosine residues (presented in bold).
S
P
H2N
K
M
V
G
R
G
Y
Q
S
G
F
G
K
M
D
R
I
4.7. BNP peptides synthesis
The linear peptide was assembled and cleaved from the resin as
described above for the somatostatin analogs. For azo cyclization,
S
S
S
the lyophilized peptide (3.6 mg, 1
HCl (0.1 M) and the solution was cooled to 0 °C using an ice bath.
A cooled aqueous sodium nitrite solution (10 L, 0.1 M) was then
lmol) was dissolved in 100 lL
N N
L
His
Pap
HO2C
R
l
R
L
S
V
K
G
added portion-wise to the peptide solution and the reaction mix-
ture was vortexed, left in the cold for 5 min and then added, por-
tion-wise, to an ice-cold KHCO3 solution (30 ml, 0.03 M). The
resulting reaction mixture, following adjustment of its pH to 7.4
by addition of 1 M K2CO3, was then kept for 1 h in the cold and
5 h at room temperature. Lyophilization of the solution resulted
in yellowish powdered crude cyclic peptide, which was purified
by semi-preparative HPLC (a gradient of 0–70% B over 50 min). Iso-
lated peptides were evaluated by analytical HPLC, amino acid anal-
ysis and mass spectroscopy. Analyses revealed high purities (>97%)
of the cyclic peptides and confirmed their identity, respectively.
Figure 6. Schematic representation of BNP-2 and 3. BNP-2, diazotation at C-2 of the
His imidazole; BNP-3, diazotation at C-5 of the His imidazole.
4.5. Receptor binding
4.5.1. Cell culture
Rat acinar pancreatic carcinoma AR4-2J cells predominantly
29
expressing the sst2 were maintained in Ham’s F-12K nutrient
mixture supplemented by 20% (v/v) fetal bovine serum, 2 mM
glutamine, 100 U/mL penicillin and 100 lg/mL streptomycin, and
grown to confluence in humidified air containing 5% CO2 at
37 °C. Sub-culturing was performed once a week employing a
trypsin/EDTA (0.05%/0.02% w/v) solution.
4.8. Receptor binding
4.8.1. Cell culture
NPR-A expressing HeLa cells were grown at 35 °C in a 5% CO2
humidified incubator in Dulbecco’s modified Eagle’s medium
(DMEM) containing 10% fetal calf serum, penicillin and streptomy-
cin (100 U/mL of each). The binding assay was performed on sub-
confluent cell monolayers in 24-well dishes at a density of
3 Â 105 cells/well preincubated for 24 h.
4.5.2. Preparation of AR4-2J cell membranes
AR4-2J cell membranes were collected according to a reported
protocol.30 In brief, AR4-2J cells were grown to confluence,
mechanically disaggregated, washed twice with cold PBS buffer
pH 7.0 and resuspended in homogenization buffer (1 mL/flask)
containing 10 mM Tris pH 7.4 and 0.1 mM EDTA. Cells were

806
G. Fridkin et al. / Bioorg. Med. Chem. 19 (2011) 798–806
10. Veber, D. F.; Freidinger, R. M.; Perlow, D. S.; Paleveda, W. J., Jr.; Holly, F. W.;
4.8.2. Competition binding assay
Strachan, R. G.; Nutt, R. F.; Arison, B. H.; Homnick, W. C.; Randall, M. S.; Glitzer,
M. S.; Saperstein, R.; Hirschmann, R. Nature 1981, 292, 55.
11. Pelton, J. T.; Gulya, K.; Hruby, V. J.; Duckles, S. P.; Yamamura, H. I. Proc. Natl.
Acad. Sci. U.S.A. 1985, 82, 236.
HeLa cell monolayers were washed twice with PBS buffer (pH
7.4) containing 1% Bovine Serum Albumin and then incubated in
the same buffer containing [125I]ANP (100,000 cpm/well) and
increasing concentrations (1–2000 ng/mL) of either native BNP or
azo cyclic BNP derivatives for 2 h at 4 °C. Two wells did not contain
BNP to allow the determination of total binding of [125I]ANP alone.
Non-specific binding was measured in the presence of an excess na-
12. Baranowska, B.; Radzikwska, M.; Wasilewska-Dziubinska, E.; Plonowski, A.;
Roguski, K.; Legowska, A.; Przybylski, J. Neuro. Endocrinol. Lett. 1999, 20, 237.
13. Melacini, Q.; Zhu, G.; Osapay; Goodman, M. J. Med. Chem. 1997, 40, 2252.
14. D’ Addona, D.; Carotenuto, A.; Novellino, E.; Piccand, V.; Reubi, J. C.; Di Cianni,
A.; Gori, F.; Papini, A. M.; Ginaaeschi, M. J. Med. Chem. 2008, 51, 512.
15. Gilon, C.; Huenges, M.; Matha, B.; Gellerman, G.; Hornik, V.; Afargan, M.;
Amitlay, O.; Ziv, O.; Feller, E.; Gamliel, A.; Shohat, D.; Wanger, M.; Arad, O.;
Kessler, H. J. Med. Chem. 1998, 41, 919.
16. Fridkin, G.; Bonasera, T. A.; Litman, P.; Gilon, C. Nucl. Med. Biol. 2005, 32, 39.
17. Ogawa, Y.; Nakao, K.; Mukoyama, M.; Hosoda, K.; Shirakami, G.; Arai, H.; Saito,
Y.; Suga, S.; Jougasaki, M.; Imura, H. Circ. Res. 1991, 69, 491–500.
18. Abassi, Z.; Karram, T.; Ellaham, S.; Winaver, J.; Hoffman, A. Pharmacol. Ther.
2004, 102, 223–241.
tive BNP (20 lg/mL) The cells were then washed twice with ice-cold
buffer, and solubilized with 0.2% w/v sodium dodecylsulphate
(SDS). Following incubation (37 °C for 30 min), the solubilized cells
were transferred to counting vials and counted for their radioactive
content (c-counter, Packard Cobra, GMI, USA). All experiments were
performed in duplicates. Specific binding is defined as total radioac-
tive content (in the absence of native BNP) minus that obtained at
high (20 lg/mL) concentration of native BNP. Receptor binding
capacity is determined by calculating the concentration of native
BNP or its derivative that is required to displace 50% of specifically
bound [125I]ANP.
19. Fridkin, G.; Gilon, C. J. Pept. Res. 2002, 60, 104.
20. Fridkin, G.; Rahimipour, S.; Ben-Aroya, N.; Kapitkovsky, A.; Di-Segni, S.;
Rosenberg, M.; Kustanovich, I.; Koch, Y.; Gilon, C.; Fridkin, M. J. Pept. Sci. 2006,
12, 106.
21. Ulysse, L. G., Jr.; Chmielewski, J. Chem. Biol. Drug Des. 2006, 67, 127.
22. Behrendt, R.; Renner, C.; Schenk, M.; Wang, F.; Wachtveitl, J.; Oesterhelt, D.;
Moroder, L. Angew. Chem., Int. Ed. 1999, 38, 2771.
23. Renner, C.; Behrendt, R.; Heim, N.; Moroder, L. Biopolymers 2002, 63, 382.
24. Wilner, I.; Rubin, S. Angew. Chem. 1996, 108, 419. Angew. Chem., Int. Ed. 1996,
35, 367.
25. Bandara, H. M. D.; Friss, T. R.; Enriquez, M. M.; Isley, W.; Incarvito, C.; Frank, H.
A.; Gascon, J.; Burdette, S. C. J. Org. Chem. 2010, 75, 4817.
26. Maina, T.; Nock, B.; Nikolopoulou, A.; Sotiriou, P.; Maintas, G. D.; Cordopatis, P.;
Chiotellis, E. Eur. J. Nucl. Med. Mol. Imaging 2002, 29, 742.
27. (a) Becker, O. M.; Levy, Y.; Ravitz, O. J. Phys. Chem. B 2000, 104, 2123; (b) Levy,
Y.; Jortner, J.; Becker, O. M. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 2188–2193; (c)
Levy, Y.; Becker, O. M. J. Chem. Phys. 2001, 114, 993–1009.
28. Bakker, W. H.; Krenning, E. P.; Breeman, W. A. P.; Koper, J. W.; Kooij, P. P.; Reubi, J.
C.; Klijn, J. G.; Visser, T. J.; Docter, R.; Lamberts, S. J. Nucl. Med. 1990, 31, 1501.
29. Froidevaux, S.; Hintermann, E.; Török, M.; Mäcke, H. R.; Beglinger, C.; Eberle, A.
N. Cancer Res. 1999, 59, 3652.
30. Raynor, K.; Reisine, T. J. Pharmacol. Exp. Ther. 1989, 251, 510.
31. Lowry, O. H.; Rosebrough, N. J.; Farr, A. L.; Randall, J. J. Biol. Chem. 1951, 193, 265.
References and notes
1. Brazeau, P.; Vale, W.; Burgus, R.; Ling, N.; Butcher, M.; Rivier, J.; Guillemin, R.
Science 1973, 179, 77–79.
2. Guillemin, R.; Gerich, J. E. Ann. Rev. Med. 1976, 27, 379.
3. Reisine, T.; Bell, G. I. Endocr. Rev. 1995, 16, 427–442.
4. Reubi, J. C. Endocr. Rev. 2003, 24, 389.
5. Froidevaux, S.; Eberle, A. N. Biopolymers 2002, 66, 161.
6. Grace, C. R. R.; Erchegyi, J.; Koerber, S. C.; Reubi, J. C.; Rivier, J.; Riek, R. J. Med.
Chem. 2006, 49, 4487.
7. Janecka, A.; Zubrzyca, M.; Janecki, T. J. Pept. Res. 2001, 58, 91.
8. Rohrer, S. P. et al Science 1998, 282, 737.
9. Veber, D. F.; Holly, F. W.; Nutt, R. F.; Bergstrand, S. J.; Brady, S. F.; Hirschmann,
R.; Glitzer, M. S.; Saperstein, R. Nature 1979, 280, 512.