M.
Liu
Y.-C. Tang
K.-Q. Fan
X. Jiang
L.-H. Lai
Y.-H. Ye
Authors' affiliations:
M. Liu, Y.-C. Tang and Y.-H. Ye, Key Laboratory of
Bioorganic Chemistry and Molecular Engineering,
Ministry of Education, College of Chemistry and
Molecular Engineering, Peking University,
Beijing, China
K.-Q. Fan and L.-H. Lai, State Key Laboratory for
Structural Chemistry Studies of Stable and
Unstable Species, Institute of Physical Chemistry,
College of Chemistry and Molecular Engineering,
Peking University, Beijing, China
X. Jiang, Accelrys, San Diego, CA, USA
Correspondence to:
Prof. Yun-hua Ye
College of Chemistry and Molecular Engineering
Peking University, Beijing
100871 China
E-mail: yhye@pku.edu.cn
Dates:
Received 29 September 2004
Accepted 31 October 2004
Present address: Mian Liu, Department of Chemistry,
University of Minnesota, 207 Pleasant Street S.E., MN
55455, USA.
*This paper is dedicated to the memory of Dr. Murray
Goodman, an outstanding and enthusiastic scientist,
professional colleague, and friend.
To cite this article:
Liu, M., Tang, Y.-C., Fan, K.-Q., Jiang, X., Lai, L.-H. &
Ye, Y.-H. Cyclization of several linear penta- and
heptapeptides with different metal ions studied by CD
spectroscopy.
J. Peptide Res., 2005, 65, 55–64.
DOI 10.1111/j.1399-3011.2004.00199.x
Copyright Blackwell Munksgaard, 2004
Cyclization of several linear
penta- and heptapeptides
with different metal ions
studied by CD spectroscopy*
Key words: CD spectra; cyclic peptide; cyclization; DEPBT; linear
peptide; metal ions
Abstract: A cyclic pentapeptide c(Tyr-Leu-Ala-Gly-Pro) (I), which
was isolated and identified from Pseudostellaria heterophylla
medicinal herbs, and two cyclic heptapeptides, c(Gly-Tyr-Gly-Gly-
Pro-Phe-Pro) (II) and c(Gly-Ile-Pro-Tyr-Ile-Ala-Ala) (III), which were
isolated and identified from Stellaria yunnanensis Franch (M),
were synthesized by using 3-(diethoxyphosphoryloxy)-1,2,3-
benzotriazin-4(3 H)-one (DEPBT) as a coupling reagent in solution,
and mediated by different metal ions, from their linear peptide
precursors H-Tyr-Leu-Ala-Gly-Pro-OH (I-1) and H-Ala-Gly-Pro-Tyr-
Leu-OH (I-2), H-Gly-Tyr-Gly-Gly-Pro-Phe-Pro-OH (II-1) and H-Gly-Ile-
Pro-Tyr-Ile-Ala-Ala-OH (III-1), respectively. The results show that
alkali metal ions can improve the cyclization yields and/or the
cyclization rates of linear peptide precursors, such as Na+ ion is
favorable for the cyclization of linear pentapeptides and Cs+ ion is
favorable for the cyclization of linear heptapeptides, while some
bivalent and trivalent metal ions, such as Mg2+, Ca2+, Zn2+, Fe2+,
Ni2+ and Cr3+ reduced/inhibited both the cyclization yields and the
cyclization rates of the linear peptide precursors. The circular
dichroism spectra of I-1, II-1 and III-1 with different metal ions were
studied to elucidate the changes in their secondary structures. It is
shown that Cs+ can induce and stabilize the type I b-turn
conformation in the linear heptapeptide II-1 and the type II b-turn
conformation in the linear heptapeptide III-1.
Abbreviations: CD, circular dichroism; DEPBT,
3-(diethoxyphosphoryloxy)-1,2,3-benzotriazin-4(3 H)-one; DIEA,
diisopropyl ethyl amine; DMF, N, N-dimethylformamide; EDCI,
N-ethyl-N¢-(3-dimethylaminopropyl) carbodiimide; RP-HPLC,
reversed-phase high-performance liquid chromatography; TFA,
trifluoroacetic acid; UV, ultraviolet; vdw, van der Waals.
55
Liu et al. Cyclization of peptides with metal ions
Introduction
Cyclic peptides, being constrained conformationally, have
been applied as synthetic targets for potential drug leads,
building blocks for the synthesis of macromolecules,
conformational analysis and the synthesis of peptide lib-
raries (1, 2). Conventional methods for peptide cyclization
generally involve a partially or fully protected linear pre-
cursor which is then cyclized in organic solvents through
various combinations of orthogonal protecting groups and
on- or off-resin cyclization schemes (3). Recently, the
thioester method has been successfully applied to the
cyclization of cyclic peptides and proteins (4). It is
unnecessary to protect side chains of the linear peptide
precursors and coupling reagent is not required during
cyclization. In addition to these
classic cyclization
methods, cyclic peptides can also be obtained by enzy-
matic cyclization (5).
It is known that several natural cyclic peptides, e.g.
valinomycin, anatamanide, enniatine etc., act as ionph-
ores in vivo. They can form stable ion complexes with
metal ions. It has been reported that LiCl–N, N-dimeth-
ylformamide (DMF) solvent system was used for driving
the cyclization, as lithium salts could mediate the solu-
bilization of peptides in organic solvents (6). Small cyclic
peptides, such as a cyclic hexapeptide (l-prolyl-glycyl)3
could form a 2 : 1 complex with Ca 2+
ion, which is
sandwiched between the two peptide molecules (7). This
cyclic hexapeptide shows selectivity for Li+ and Na+ over
K+ and larger cations; among these bivalent metal ions of
corresponding radii, it shows selectivity in binding Ca
over Ba2+ (8). It was reported that different-membered
cyclic tetrapeptides or macrolactams were synthesized
from non-activated dipeptide ester precursors. The cycli-
zation was carried out in methanol in the presence of
sodium methoxide as base and with Ni2+, Pd2+ or Cu2+ as
template (9).
We focused our efforts on starting with the side-chain of
tyrosine unprotected peptides as linear precursors to syn-
thesize cyclic peptides by using an organophosphorus
coupling reagent 3-(diethoxyphosphoryloxy)-1,2,3-benzo-
triazin-4(3 H)-one (DEPBT) (10, 11). DEPBT can be used
under mild peptide coupling conditions and recently it has
been successfully used for amide-bond formation in the
total synthesis of ())-tamandarin B (12) and Teicoplanin
Aglycon (13). Here, we further studied the effects of these
metal ions on the conformations of these linear penta- and
heptapeptides, and how these metal ions act on the cycli-
zation.
56 J. Peptide Res. 65, 2005 / 55–64
Experimental Procedure
Instruments and materials
Samples were detected by RP-HPLC on HP1100 System
(Agilent Technologies, CA, USA). Semi-preparative reversed-
phase high-performance liquid chromatography (RP-HPLC)
was carried out on a Waters 600E LC System, Water 486
monitor (Waters Corporation, MA, USA). Far-UV CD spectra
were recorded on a Jobin-Yvon (Paris, France) CD 6 spec-
trameter and the data obtained was the average of three scans
corrected by subtracting a spectrum of the appropriate solu-
tion in the absence of peptide recorded under identical con-
ditions. Each scan in the range 190–250 nm or 200–260 nm
was obtained by taking data points every 0.5 nm, with integ-
ration time 0.5 s and a 2-nm band width. A cell with a path
length of 0.1 cm was used. Amino acids used in this paper are
all of l-configuration if not mentioned elsewhere. DMF was
dried over calcium oxide and redistilled in vacuum before use.
Cyclization of linear pentapeptide I-1 and I-2 by DEPBT with
different metal ions
H-Tyr-Leu-Ala-Gly-Pro-OH (I-1) (5.19 mg, 10 lmol) and
DEPBT (3.3 mg, 11 lmol) were dissolved in DMF (10 mL) at
room temperature. Et3N (2.5 lL) was added to the mixture
with stirring (pH 8). Then, the following salts, each of them in
25 lL H2O, were added to each of the above solution con-
secutively. (a) LiClÆH2O (3.0 mg); (b) NaCl (2.9 mg); (c) KCl
2+
(3.7 mg); (d) CsCl (8.4 mg); (e) MgCl2Æ6H2O (10.2 mg); (f)
CaCl2 (5.5 mg); (g) BaCl2Æ2H2O (12.2 mg); (h) ZnCl2 (6.8 mg);
(i) CrCl3Æ6H2O (13.3 mg). At 2, 4 and 24 h of the reaction time,
20 lL of the reaction mixture was taken out and injected into
analytical RP-HPLC. The procedure for the cyclization of
linear pentapeptide H-Ala-Gly-Pro-Tyr-Leu-OH (I-2) is the
same as the cyclization of I-1, while only three alkali metal
ions (Na+, K+ and Cs+) in the form of chloride salts, (a) NaCl
(2.9 mg), (b) KCl (3.7 mg) and (c) CsCl (8.4 mg) were screened.
Compound I was confirmed by Fast Atom Bombardment
Mass Spectrometry (FAB-MS) m/z 502 (M + H)+.
Cyclization of linear heptapeptide II-1 by DEPBT with different
metal ions
H-Gly-Tyr-Gly-Gly-Pro-Phe-Pro-OH (II-1) (2.9 mg, 4.18 lmol)
and DEPBT (1.4 mg, 4.8 lmol) were dissolved in DMF
(2.0 mL) at room temperature. Diisopropyl ethyl amine
 Liu et al. Cyclization of peptides with metal ions

(DIEA) (1.2 lL, 6.9 lmol) was added to the mixture with medicinal herbs (15), and two cyclic heptapeptides, c(Gly-
stirring. Then, the following salts, each of them in 20 lL Tyr-Gly-Gly-Pro-Phe-Pro) (II) (16) and c(Gly-Ile-Pro-Tyr-Ile-
H2O, were added to each of the above solution consecu- Ala-Ala) (III) (17), which were isolated and identified from
tively. (a) LiClÆH2O (1.20 mg); (b) NaCl (1.17 mg); (c) KCl Stellaria yunnanensis Franch (M), were synthesized in
(1.49 mg); (d) CsCl (3.37 mg); (e) MgCl2Æ6H2O (4.07 mg); (f) solution with DEPBT as a coupling reagent from their linear
CaCl2 (2.22 mg); (g) BaCl2Æ2H2O (4.88 mg); (h) saturated peptide precursors I-1 and I-2, II-1 and III-1, respectively.
aqueous ZnCl2 (10 lL); (i) FeCl2Æ4H2O (3.98 mg); (j) NiCl2- These linear penta- and heptapeptides were cyclized suc-
Æ6H2O (4.75 mg); (k) CrCl3Æ6H2O (5.33 mg); (l) H2O (20 lL, as cessfully by DEPBT in DMF (1–2 mm) at room temperature.
control). At 1, 3, 5 and 24 h of reaction time, 20 lL of the The cyclization reactions were monitored by analytical
reaction mixture was taken out and injected into analytical RP-HPLC at 2, 4 and 24 h reaction times or 1, 3, 5 and 24 h
RP-HPLC. Compound II was confirmed by FAB-MS m/z 676 reaction times. In order to study the effects of different
(M + H)+. The procedure of cyclization of linear heptapep- metal ions on the yields of cyclization, different metal ions
tide H-Gly-Ile-Pro-Tyr-Ile-Ala-Ala-OH (III-1) (14) is the (several alkali metal ions, bivalent alkaline earths and
same as the cyclization of II-1. Compound III was confirmed transition metal ions) were used as additives in addition to
by High Resolution Mass Spectometry (HRMS) calcd for the coupling reagent DEPBT and the base DIEA or Et3N.
C34H51N7O8 686.3872, found 686.3865. Actually, the effects of nine different metal ions, Li+, Na+,
K+, Cs+, Mg2+, Ca2+, Ba2+, Zn2+ and Cr3+ on the cyclization
of linear pentapeptide I-1, three different metal ions, Na+,
Cyclization of linear heptapeptide II-1 with EDCI as a coupling
K+ and Cs+ on the cyclization of linear pentapeptide I-2, 11
reagent
different metal ions, Li+, Na+, K+, Cs+, Mg2+, Ca2+, Ba2+,
Zn2+, Ni2+, Fe2+ and Cr3+ on the cyclization of linear hep-
II-1 (0.7 mg, 1.0 lmol) was dissolved in DMF (0.45 mL) at
tapeptide II-1, and nine different metal ions, Li+, Na+, K+,
room temperature; N-ethyl-N¢-(3-dimethylaminopropyl)
Cs+, Mg2+, Ca2+, Ba2+, Ni2+, and Cr3+ on the cyclization of
carbodiimide (EDCI) (0.24 mg, 1.25 lmol) and DIEA
(0.35 lL, 2.5 lmol) were added at 0
C with stirring. Then,
the following salts were added to each of the above solution
Table 1. Cyclization yields (%) of I-1 and I-2 induced by different
consecutively. (a) CsCl (1.89 mg in 15 lL H2O); (b) NiCl2- metal ionsa
Æ6H2O (2.38 mg in 10 lL H2O); (c) H2O (15 lL as control). Reaction time (h)
After 30 h, EDCI (0.24 mg, 1.25 lmol) and DIEA (0.35 lL,
Metal ionsb 2 4 24
2.5 lmol) were supplemented to the reaction. At 1, 3, 5, 24
and 48 h of the reaction time, 20 lL of the reaction mixture Li+ 8 17 37
+
was taken out and injected into analytical RP-HPLC. Na 10 20 39 (80)c
+
K 12 19 38 (69)c
Cs+ 12 21 36 (69)c
Analytical RP-HPLC conditions Mg 2+
6 8 22
2
Ca 9 14 27
The column used was a Microsorb column (C18; Ba2+ 10 11 31
150 · 3.9 mm), the buffer A, 0.1% TFA in water, buffer B, Zn 2+
– d
– d
–d
0.1% TFA in acetonitrile. The wavelength was set at Cr 3+
– d
– d
–d
275 nm and the flow rate was 1 mL/min with a linear gra- None 5 9 22 (67)c
dient of 5–50% buffer B over 30 min.
a
Yields were calculated based on
analytical RP-HPLC files. Wave-
length, 275 nm; DEPBT was a coup-
Results and Discussion ling reagent; HPLC analysis
conditions were in experimental
procedure.
b
Concentration of each metal ion
Effects of different metal ions on the cyclization of linear was 5 eq. to the linear precursor
except for Zn2+ ion, which was in
pentapeptides I-1, I-2 and linear heptapeptides II-1 and III-1
saturated aqueous solution.
c
Cyclization yields of I-2 given in
parentheses.
A cyclic pentapeptide c(Tyr-Leu-Ala-Gly-Pro) (I), which was d
Cyclic pentapeptide I was not
found.
isolated and identified from Pseudostellaria heterophylla

J. Peptide Res. 65, 2005 / 55–64 57

Liu et al. Cyclization of peptides with metal ions

Table 2. Cyclization yields (%) of linear heptapeptide II-1 induced compared with the 67% cyclization yield of I-2 without
by different metal ionsa
any metal ions (Table 1).
Reaction time (h) In the case of the cyclization of linear heptapeptide II-1
Metal ions b
1 3 5 24 with different metal ions as additives, the results are con-
sistent with the cyclization of linear pentapeptides I-1 and
Li+ 21 39 52 68
I-2, except that, the bigger univalent metal ion Cs+, which
Na+ 36 57 58 71
enhanced the cyclization yields of II-1 from 60 to 78%, in
K+ 34 55 66 71
place of Na+, seems to be a more suitable ion for the cyc-
Cs+ 39 66 73 78
lization of longer linear heptapeptide II-1 (Table 2). The
Mg2+ 2 10 12 29
other alkali metal ions, Li+, Na+, and K+, enhanced the
Ca2+ 6 12 21 41
cyclization yields of II-1 from 60 to 68%, 60 to 71%, and 60
Ba2+ 22 38 51 64
to 71%, respectively, after 24 h cyclization time. Similar to
Zn2+ –c –c 2 4
the cyclization of I-1, bivalent metal ions, Mg2+ and Ca2+,
Ni2+ –c 2 3 12
decreased the cyclization yields of II-1 from 60 to 29% and
Fe2+ –c –c –c –c
60 to 41%, respectively, after 24 h reaction time, while the
Cr3+ –c –c –c –c
cyclization yields of II-1 were decreased drastically when
none 20 38 49 60
Zn2+ or Ni2+ was added. Furthermore, no cyclic heptapep-
a, b
Same as corresponding footnotes tide II was detected by analytical RP-HPLC when Fe2+ or
in Table 1.
c
Cyclic pentapeptide II was not found. Cr3+ was added to the reaction system (Table 2). Influence
of different concentrations of Cs+ was also studied. With 5,
2 and 0.5 eq. of Cs+, the cyclization yields of II-1 were 78, 67
Table 3. Cyclization yields of III-1 induced by different metal ionsa and 62%, respectively, compared with the 60% cyclization
Reaction time(h) yield of II-1 without any metal ions.
In order to investigate if there were some similar effects
Metal ionsb 1 3 5 24
of these metal ions on the cyclization of linear peptides, we
Li+ 24 37 47 60 synthesized another linear heptapeptide III-1 in solution
Na+ 18 35 44 59 and also studied the influence of different metal ions on its
K+ 19 34 41 57 cyclization systematically (14). From Table 3, alkali metal
Cs+ 22 38 51 66 ions, Li+, Na+, K+ and Cs+ also enhanced the cyclization
Mg2+ <5 9 16 22 yields to some extent, from 55 to 60, 59, 57 and 66%,
Ca2+ <5 <5 <5 5 respectively, compared with the cyclization of III-1 without
Ba2+ 19 36 42 59 any metal ions (Table 3), while the cyclization yields of
Ni2+ –c –c –c Trace III-1 were decreased drastically when bivalent metal ions,
Cr3+ –c –c –c –c except for Ba2+, were added to the cyclization reaction. No
None 17 32 39 55 cyclic heptapeptide III was detected by analytical
a, b
Same as corresponding footnotes in
RP-HPLC, when trivalent metal ion Cr3+ was added to the
Table 1. cyclization reaction.
c
Cyclic pentapeptide III was not found.
From the above results (Tables 1–3), it can be concluded
that (a) univalent Na+ ion was favorable for the cyclization
linear heptapeptide III-1 have been studied (14, 18). The of linear pentapeptides, such as I-1 and I-2, while univalent
results are summarized in Tables 1–3. Cs+ ion was favorable for the cyclization of linear hepta-
In the presence of Na+, the cyclization yields of linear peptides, such as II-1 and III-1. (b) Some bivalent metal
pentapeptides I-1 and I-2 were enhanced from 22 to 39% ions, i.e. Mg2+ ion and Ca2+ ion, decreased the cyclization
and 67 to 80% after 24 h cyclization time (Table 1). Inter- yields of linear heptapeptides II-1 and III-1 to some extent
estingly, we could not detect any cyclic pentapeptide by after 24 h reaction time (31 and 19% lower for II; 33 and
analytical RP-HPLC when Zn2+ or Cr3+ was added to each 50% lower for III, respectively, than without any metal
cyclization reaction. The influence of Na+ concentration ions). (c) Among the other bivalent metal ions, Ni2+, Zn2+
was also studied. With 10 , 5 and 2 eq. of Na+, the cycliza- and Fe2+ inhibited the cyclization seriously or completely.
tion yields of I-2 were 81, 80 and 76%, respectively, (d) Bivalent Ba2+ ion could promote cyclization, but not as

58 J. Peptide Res. 65, 2005 / 55–64

 Liu et al. Cyclization of peptides with metal ions

effective as univalent Na+ ion for the linear pentapeptide I-1 2.5×10–4

and I-2 and univalent Cs+ ion for the linear heptapeptide 2.0×10–4

Molar ellipticity (deg cm2 dmol–1)

II-1 and III-1, while it did not reduce/inhibit the cyclization 1.5×10–4

1.0×10–4
as the other bivalent metal ions did. (e) Trivalent Cr 3+
ion
5.0×10–5
inhibited the cyclization totally.
0.0
Preliminary study shows if the ratios of radius to charge –5.0×10–5 YLAGP
(charge density) of the metal ions studied are above 0.071 –1.0×10 YLAGP + 1eq.Na+
(included), the metal ions are in favor of the cyclization –1.5×10–4 YLAGP + 5eq.Na+
YLAGP + 10eq.Na+
involved. These metal ions are Cs+, K+, Na+, Li+ and Ba2+. If –2.0×10–4

–2.5×10–4
these ratios are below 0.071 (not included), the metal ions 210 220 230 240 250 260 nm
are unfavorable for the cyclization involved. These metal
Figure 2. CD spectrum of YLAGP (I-1) at different concentration of Na+.
ions are Ca2+, Mg2+, Ni2+ and Cr3+ (14).

–5
5.0×10
LP none
CD spectra 0.0
LP + 5eq.Cs+

Molar ellipticity (deg cm2 dmol–1)

–5.0×10–5 LP + 5eq.Na+
LP + 5eq.Ca2+
Circular dichroism has been used extensively to give useful –1.0×10
–4

LP + 5eq.Cr3+
information about protein structure, the extent and rate of –1.5×10
–4

LP–linear peptide
structural changes and ligand binding (19). In the far UV –2.0×10
–4

region (240–180 nm), which corresponds to peptide bond –2.5×10

–4

absorption, the CD spectrum can be analyzed for studying –3.0×10

–4

the regular secondary structural features, while the con- –3.5×10

–4

formations of most short peptides in aqueous solution are –4.0×10

–4

200 210 220 230 240 250 nm

random coils. As to the linear pentapeptide I-1, it also has
no stable secondary structure in this far UV region (20), but
it has a positive peak around 232 nm and the effects of
different metal ions on its conformations were limited to
the range from 210 to 220 nm (Fig. 1).
Different sodium concentrations also affected the con-
formations of linear pentapeptide I-1 (Fig. 2). It seems the
conformation changed greater with higher sodium concen-
tration, such as 10 eq. sodium ions.
But as to the CD spectra of linear heptapeptide III-1, it
shows a character of type II b-turn, a negative peak at
+ +
225 nm (Fig. 3). When 5 eq. Na ion or 5 eq. Cs ion was
added to the methanol solution, which contained the linear
Figure 3. CD spectrum of GIPYIAA (III-1) with different metal ions.
heptapeptide III-1, the intensity of negative absorbance at
225 nm was enhanced. Especially when Cs+ ion was added,
the intensity of the negative peak was increased by 30%,
compared with 20% increment by adding Na+ ion. As to the
cyclization results (Table 3), the cyclization yields of III-1
were increased up to 11% by adding 5 eq. Cs+ ion and 4% by
adding 5 eq. Na+ ion.
The intensity of negative peak of III-1 at 225 nm was also
enhanced by 22 and 30% with the increase of the Cs+ ion
concentration from 1 to 5 eq. (Fig. 4), which shows that

–5
5.0×10
YLAGP
–4
2.5×10 0.0
YLAGP + 5eq.Na+
Molar ellipticity (deg cm2 dmol–1)

2.0×10–4 YLAGP + 5eq.Cs+

Molar ellipticity (deg cm2 dmol–1)

–5.0×10–5
1.5×10–4 YLAGP + 5eq.Ca2+ L none
–4
–1.0×10 L + 1eq.Cs
1.0×10 –4 YLAGP + 5eq.Cr3+
–1.5×10
–4 L + 5eq.Cs
5.0×10–5
–4
0.0 –2.0×10

–5 –4
–5.0×10 –2.5×10
–4
–1.0×10 –3.0×10
–4

–1.5×10–4 –4
–3.5×10
–2.0×10–4
–4
–4.0×10
–2.5×10–4 200 210 220 230 240 250 nm

210 220 230 240 250 260 nm

Figure 4. CD spectrum of GIPYIAA (III-1) at different concentration
Figure 1. CD spectrum of YLAGP (I-1) with different metal ions. of Cs+.

J. Peptide Res. 65, 2005 / 55–64 59

Liu et al. Cyclization of peptides with metal ions

–4
4
–4.1×10
1.0×10
–4
–4.2×10
Molar ellipticity (deg cm2 dmol–1)

0.0

Molar ellipticity (deg cm2 dmol–1)

–4
–4.3×10
–4
–1.0×10 –4
–4.4×10

–4
–2.0×10
–4 –4.5×10

–4
–4
C none –4.6×10
–3.0×10
C + 1eq.Cs –4
–4.7×10
–4
C + 5eq.Cs
–4.0×10
–4
–4.8×10
–4
–5.0×10 –4
–4.9×10

200 210 220 230 240 250 260 nm –4

–5.0×10

Figure 5. CD spectrum of c(GIPYIAA) (III) at different concentration of 10 20 30 40 50 ºC

Cs+. Figure 7. Temperature-swing curve of GIPYIAA (III-1) at 224 nm.

from 1 to 5 eq., compared with the peak intensity without

0.0
any metal ions. CD spectrum of linear heptapeptide III-1 at
Molar ellipticity (deg cm2 dmol–1)

–4
increasing temperature from 10 to 50
C show the intensity
–1.0×10
at 224 nm decreased (Fig. 6).
10 ºC
–2.0×10
–4
20 ºC The thermodenaturation curve of the linear heptapeptide
30 ºC
40 ºC
III-1 monitored by CD signal at 224 nm from 10 to 50
C
–4
–3.0×10 50 ºC shows a linear decrement in intensity rather than a sigmoid
decrement. This result suggests no significant structural
–4
–4.0×10
change in this temperature range in methanol and indicates
210 220 230 240 250 260 nm no stable conformation is formed by III-1 under this con-
Figure 6. CD spectrum of GIPYIAA (III-1) at different temperature. dition (Fig. 7).

the cesium ion could stabilize the type II b-turn confor-

mation on the one hand and induce the formation of the Proposed inhibition mechanism of some bivalent metal ions and
conformation, which is beneficial to the cyclization of the Cr3+ on the cyclization of linear peptides
linear peptides on the other.
Further CD studies were carried out on the cyclic hep- Some transition metal ions, Ni2+, Zn2+, Cd2+ and Cu2+, were
+
tapeptide III with different concentrations of Cs ion, found to enhance the helicity of small peptides in aqueous
which has little effect on its conformation in the far UV solution, as they could form stable complexes with the side-
region, as cyclic peptide has relatively stable conforma- chains of histidine, cysteine (21) or amino-diacetic acid (22).
tions, compared to its linear precursor, and keeps the sta- In our experiments, the linear peptides have no metal-
bility by intramolecular actions, such as hydrogen bonds complexing side chains, but these metal ions (Fe2+, Ni2+,
between O


H


N. While the interactions between Cs+ Zn2+ and Cr3+) were strongly coordinated to the carboxyl
ion and amide bonds or other functional groups in the cyclic group (Fig. 8A) of the linear peptides at the C-terminus and
peptide are quite weak, the conformation change of cyclic formed transition metal-ion mediated complexes, which
heptapeptide III caused by Cs+ is very limited. But inter- capped the C-terminus and blocked the attack route of
estingly, with the increased concentration (from 1 to 5 eq.) amino group from the N- to C-terminus of the linear pep-
of the Cs+ ion, the intensity of the negative peak around tides, so the cyclization reaction was terminated at stage 1
227 nm was a little decreased (Fig. 5). (Fig. 8).
Another CD study of linear heptapeptide II-1 shows a Ni2+ (used as an example in Fig. 8) and other transition
negative peak at about 208 nm, which is a character of type metal ions could also form complexes with coupling rea-
I b-turn (spectrum not shown). Alkali metal ions, such as gent DEPBT at site B, which may affect the UV absorbance
+ +
Na , and Cs could also enhance the intensity of the neg- intensity of DEPBT, even its retention time reflected by RP-
ative peak of II-1 from 205 to 215 nm in methanol (spec- HPLC, but these metal ions did not catalytically breakdown
trum not shown). The peak intensity in this range was or coordinate to the coupling reagent. The intensity of UV
enhanced with the increment of the concentration of Cs+ absorbance of II-1 was enhanced at 275 nm with Ni2+ and

60 J. Peptide Res. 65, 2005 / 55–64

 Liu et al. Cyclization of peptides with metal ions

Figure 8. Proposed inhibition mechanism of O O

Ni2+ B
Ni2+ on the cyclization of linear peptide with EtO P O
O N O
DEPBT as a coupling reagent. EtO –PO2(OEt)2
N EtO N N
N P O N

O A EtO O
NH2
H2N Ni2+ et. al. O
O

1 2

O O
Cyclic peptide –
NH2 + OOBt
N O
N
N

kept almost unchanged, although at the presence of the

120 A
coupling reagent DEPBT, as Ni2+ formed stable complex
Comparative UV absorbance

100 Ni+ with II-1 (Fig. 9A). However, II-1 was consumed quickly by
DEPBT at the presence of Cs+ (Fig. 9B).
at 275nm (%)

80 B A-GYGGPFP+5eq.Ni ;
2+

+
In another study, EDCI replaced DEPBT as a coupling
60 B-GYGGPFP+5eq.Cs ;
C-Without any metal ions. reagent for the cyclization of II-1 (Table 4). II-1 almost
40 C remained unchanged at the presence of Ni2+ for 48 h, while
20 the cyclization yields were 19.4% with Cs+ ion and 10.5%
+
Cs as a control after 2 days cyclization.
0
When EDCI was used as a coupling reagent, site A at the
0 10 20 30 40 50
Reaction time(h) linear peptide was the only metal-ligating carboxyl group

Figure 9. Comparative UV absorbance of GYGGPFP (II-1) at 275 nm

and formed complexes with Ni2+ ion, which inhibited
with DEPBT as a coupling agent. the formation of O-acyl urea intermediate (Fig. 10). So, the
amino group at N-terminus could not attack C-terminus
a
Table 4. Cyclization yields (%) of II-1 using EDCI as a coupling of the linear peptide in the reaction system. The experi-
reagent
mental results also indicated that no cyclic peptides were
Reaction time (h) formed.

Metal ionsb 1 3 5 24 48c

Without metal ions 0 1.4 4.3 10.1 10.5

Proposed promotion mechanism of alkali metal ions on the
Cs+ 0 4.4 7.8 15.3 19.4
cyclization of linear penta- and heptapeptides
Ni2+ 0 0 0 0 0

a
Concentration of linear peptide precursor: Alkali metal ions (Li+, Na+, K+, Rb+, or Cs+) were found to
2 · 10)3 mol/L. Calculated yields from HPLC profiles. enhance the helicity of small peptides as well (23), but
b
Concentration of metal ions: 20 · 10)3 mol/L.
c
After reacting 30 h, 0.5 eq. EDCI and 0.5 eq. Et3N unlike the transition metals, they could not form cross-
were added again.
linked structure with peptides that have high affinities to

Figure 10. Proposed inhibition mechanism of

Ni2+ on the cyclization of linear peptides
Ni2+
with EDCI as a coupling reagent.
O CH3
pH=7–8
H2N C + C2H5-N=C=N-CH2CH2CH2-N
A O CH3

H2N COO CH3

C2H5-N=C-NH-CH2CH2CH2-N Cyclic peptide
CH3
O-acyl urea

J. Peptide Res. 65, 2005 / 55–64 61

Liu et al. Cyclization of peptides with metal ions

NH2 Figure 11. Proposed cyclization mechanism

+ by DEPBT as a coupling reagent promoted by
R5 CH Na
Na+ ion on the cyclization of linear penta-
O
C peptides.
NH BH
O
CH O O O B: DIEA or Et3N
C NH
R4 CH C NH CH C NH CH C O
A R3 R2 R1

DEPBT

R5 R4
NH2 CH NH
C CH C C R5
O

NH
NH O O CH

R4 CH NH2
R3 CH
C O Na+
C O Na+
NH

NH
CH CH O O
OBt
R3 C O R2 C C
NH O O NH CH
CH R1
B C NH CH C OBt C
R2
R1

R4 R4
O NH
CH NH CH C R5
C C
O
NH

NH

O CH
C
O NH2
R3 CH R3 CH
CH R5 +
C O Na
C O
NH
NH O NH
C CH O OBt
CH O
R2 C NH CH R2 C C
R1 NH CH
E O
D
R1

metal ions (21, 22). Alkali metal ions are oxyphilic, so they
prefer carbonyl oxygens for binding (24). It was shown that
Na+ is always coordinated to the carbonyl groups near the
+
C-terminus (22). We propose that Na (used as an example
to represent other alkali ions) was first coordinated to the
oxygen atoms of carbonyl and amide groups near the
C-terminus (Fig. 11A) (23). With the coordination of addi-
tional amide oxygen atoms along the peptide chain
(Fig. 11B), the linear peptide forms stronger turn structure
as revealed by CD spectra (Figs 1–4), ultimately resulted in
a cyclic-like conformation (Fig. 11C,D). We believe that the
ability of alkali metal ions to coordinate to peptides with
moderate affinity and to form a conformation to bring the
linear peptide N- and C-terminus together (25), is the key
reason for the enhanced reaction rate and yield of the linear
peptide cyclization. As to whether the interaction between
the alkali metal ions and the peptide is like the interaction
between ions and crown ether, or the alkali metal ions are
sandwiched between two cyclic peptides (8) remains to be
elucidated.
Molecular modeling study
Based on the proposed promotion mechanism discussed
above, the conformations of the linear peptides just before
cyclization should be able to accommodate the coordi-
nation of metal ion. In another word, if we assume that
the conformation of the linear peptide just before cycli-
zation should be similar to the conformation of the
corresponding cyclic peptide, and the metal ion is
coordinated with carbonyl oxygens, then the diameter of
the cyclic peptide (if its conformation is like a cycle)
should be close to the diameter of the metal ion plus two
van der Waals (vdw) radii of oxygen atom. In the current
study, we modeled the conformations of the cyclic pep-
tides using molecular dynamics (MD) simulations to
measure the time averaged geometric properties, and
compared them to metal ion diameters and vdw radii of
oxygen/nitrogen atoms.
The conformation of the cyclic pentapeptide I was pre-
viously determined by NMR in DMSO solution (15). In this

62 J. Peptide Res. 65, 2005 / 55–64

 Liu et al. Cyclization of peptides with metal ions

study, we used the NMR conformation as the starting
conformation for MD simulations, because we feel that the
time averaged measurement is better than a measurement
on a single conformation, and the solvents used in experi-
ments are different (DMF vs DMSO). For the cyclic hepta-
peptides II and III, as no X-ray or NMR conformations were
available, energy minimized random conformations were
used as the starting conformation. Insight II (Accelrys,
San Diego, CA, USA) was used in the study with the CFF91
forcefield. Distance-dependent dielectric constant was
applied (4r) and MD simulations were performed at 298 K
for 100 ns (1 fs per timestep). Various distances between the
backbone atoms cross the cyclic structures were measured
and averaged by using the trajectories saved every 1 ns from
the MD simulations. The backbone of the cyclic penta-
peptide is a circle with a diameter of 6.0 ± 0.5 Å (standard
deviation). The two cyclic heptapeptides were almost
identical in shape, which is a rectangle or ellipse with the
dimensions of 6.3 ± 0.5 and 9.0 ± 0.3 Å. The diameter of
Na+ is 1.96 Å, and the diameter of Cs+ is 3.30 Å. The vdw
radius of oxygen or nitrogen atom is about 1.5 Å. Therefore,
based on the molecular modeling study, the conformations
of the linear penta- and heptapeptides in the current study
should be able to form a conformation to effectively
coordinate to the metal ion (e.g. Na+ or Cs+).
Conclusion
In conclusion, some alkali metal ions, such as Na+ and Cs+
can enhance the cyclization yields of some linear peptides
and/or the cyclization rates. The smaller Na+ ion is suitable
for the cyclization of linear pentapeptides I-1 and I-2, while
the larger Cs+ ion is better for the cyclization of linear
heptapeptide II-1 and III-1. Metal ion selectivity is an
expected consequence of different sizes of the linear pep-
tides. We suppose that Na+ or Cs+ ion is firstly coordinated
to the carbonyl groups at C-terminus and its proximity,
then, under the interactions of the other carbonyl groups far
away from the C-terminus of the linear peptide precursors
together, draws the N- and C-terminus close to each other.
Therefore, it is favorable for the ring closure. On the con-
trary, bivalent or trivalent metal ions, such as Ni2+ or Cr3+,
which can coordinate strongly to the carboxyl groups in the
linear peptide, prevent the amino and carboxyl groups from
forming a cyclic structure. Therefore, the yields of cyclic
pentapeptide I, cyclic heptapeptide II and III decreased or
no cyclization occurred, when these metal ions were
introduced. Primary CD spectra study shows that different
metal ions and different metal ion concentrations could
affect the conformations of the linear peptides precursors or
cyclic peptide moiety. Na+ or Cs+ may stabilize or induce
the conformations of its correspondent linear peptide pre-
cursors, which are beneficial to the cyclization.
We are currently investigating the extension of this
method to other linear peptide precursors, for example,
linear hexapeptides. Furthermore, longer peptide chains
will be studied to determine whether cyclization can also
be promoted by these alkali metal ions. This method of
adding alkali metal ions, especially, Na+ and Cs+ to enhance
the cyclization of different size of linear peptides, will
probably be useful in many other situations involving
synthesis of cyclic peptides or cyclic depsipeptides beyond
those described here.

References
1. Al-Obeidi, F., Castrucci, A.M.L., Hadley,
M.E. & Hruby, V.J. (1989) Potent and
prolonged acting cyclic lactam analogues of
a-melanotropin: design based on molecular
dynamics. J. Med. Chem. 32, 2555–2561.
2. Spatola, A.F. & Crozet, Y. (1996)
Rediscovering an endothelin antagonist
(BQ-123): a self-deconvoluting cyclic
pentapeptide library. J. Med. Chem. 39,
3842–3846.
3. Davies, J.S. (2003) The cyclization of
peptides and depsipeptides. J. Peptide Sci. 9,
471–501.
4. Zhang, L. & Tam, J.P. (1999) Lactone and
lactam library synthesis by silver ion-
assisted orthogonal cyclization of
unprotected peptides. J. Am. Chem. Soc. 121,
3311–3320.
5. Jackson, D.Y., Bumier, J.P. & Wells, J.A.
(1995) Enzymatic cyclization of linear
peptide esters using subtiligase. J. Am.
Chem. Soc. 117, 819–820.
6. Robey, F.A. (2000) Selective and facile
cyclization of N-chloroacetylated peptides
from the C4 domain of HIV Gp120 in LiCl/
DMF solvent systems. J. Peptide Res. 56,
115–120.
7. Kartha, G., Aimoto, S. & Varughese, K.I.
(1986) Crystal and molecular structure of
cyclo(L-prolyl-glycyl)3. Int. J. Peptide Res.
27, 112–117.
8. Varughese, K.I., Aimoto, S. & Kartha, G.
(1986) Calcium binding cyclic hexapeptide.
Int. J. Peptide Res. 27, 118–122.
9. Haas, K., Ponikwar, W., Nöth, H. & Beck,
W. (1998) Facile synthesis of cyclic
tetrapeptides from nonactivated peptide
esters on metal centers. Angew. Chem. Int.
Ed. 37, 1086–1089.

J. Peptide Res. 65, 2005 / 55–64 63

Liu et al. Cyclization of peptides with metal ions
10. Fan, C.X., Hao, X.L. & Ye, Y.H. (1996) A
novel organophosphorus compound as a
coupling reagent for peptide Synthesis.
Synth. Commun. 26, 1455–1460.
11. Li, H.T., Jiang, X.H., Ye, Y.H., Fan, C.X.,
Romoff, T. & Goodman, M. (1999)
3-(diethoxyphosphoryloxy)-1,2,3-
benzotriazin-4(3H)-one (DEPBT): a new
coupling reagent with remarkable resistance
to racemization. Org. Lett. 1, 91–93.
12. Joullié, M.M., Portonovo, P., Liang, B. &
Richard, D.J. (2000) Total synthesis of ())
tamandarin B. Tetrahedron Lett. 41,
9373–9376.
13. Boger, D.L., Kim, S.H., Mori, Y., Weng, J.H.,
Rogel, O., Castle, S.L. & McAtee, J.J. (2001)
First and second generation total synthesis
of the teicoplanin aglycon. J. Am. Chem.
Soc. 123, 1862–1871.
14. Liu, M. & Ye, Y.H. (2002) Studies on
synthesis of a cycloheptapeptide and effects
of different metal ions on the cyclization.
Chin. J. Chem. 20, 1347–1353.
15. Morita, H., Kayashita, T., Kobata,
H., Gonda, A., Takeya, K. & Itokawa, H.
(1994) Pseudostellarins A–C, new tyrosinase
inhibitory cyclic peptides from
Pseudostellaria heterophylla. Tetrahedron
50, 6797–6804.
64 J. Peptide Res. 65, 2005 / 55–64
16. Zhao, Y.R., Wang, X.K., Zhou, J., Cheng,
C.X., Huang, X.L. & Wu, H.M. (1995)
Stellarins F and G, two new cyclopeptides
from Stellaria yunnanensis. Chin. J. Chem.
13, 552–557.
17. Zhao, Y.R., Wang, X.K., Zhao, T.F., Zhou, J.,
Huang, X.L. & Wu, H.M. (1995) Two new
cyclopeptides from Stellaria yunnanensis.
Chin. J. Chem. 13, 267–272.
18. Ye, Y.H., Liu, M., Tang, Y.C. & Jiang, X.H.
(2002) Promotion of cyclization of linear
pentapeptides and heptapeptide by different
univalent metal ions. J. Chem. Soc., Chem
Commun. 5, 532–533.
19. Kelly, S.M. & Price, N.C. (2000) The use of
circular dichroism in the investigation of
protein structure and fuction. Curr. Prot.
Peptide Sci. 1, 349–384.
20. Tang, Y.C., Xie, H.B., Tian, G.L. & Ye, Y.H.
(2002) Synthesis of cyclopentapeptides and
cycloheptapeptides by DEPBT and the
influence of some factors on cyclization.
J. Peptide Res. 60, 95–103.
21. Ghadiril, M.R. & Choi, C. (1990) Secondary
structure nucleation in peptides. Transition
metal ion stabilized a-helices. J. Am. Chem.
Soc. 112, 1630–1632.
22. Ruan, F., Chen, Y. & Hopkins, P.B. (1990)
Metal ion enhanced helicity in synthetic
peptides containing unnatural, metal-
ligating residues. J. Am. Chem. Soc. 112,
9403–9404.
23. Kohtani, M., Kinnear, B.S. & Jarrold, M.F.
(2000) Metal-ion enhanced helicity in the
gas phase. J. Am. Chem. Soc. 122,
12377–12378.
24. Cerda, B.A., Hoyau, S., Ohanessian, G. &
Wesdemiotis, C. (1998) Na+ binding to cyclic
and linear dipeptides. Bond Energies,
entropies of Na+ complexation, and
attachment sites from the dissociation of
Na+-bound heterodimers and ab initio
calculations. J. Am. Chem. Soc. 120,
2437–2448.
25. Wyttenbach, T., Witt, M. & Bowers, M.T.
(2000) On the stability of amino acid
Zwitterions in the gas phase: the influence
of derivatization, proton affinity, and alkali
ion addition. J. Am. Chem. Soc. 122,
3458–3464.