Novel gratisin derivatives with high antimicrobial activity and low hemolytic activity

Article abstract of DOI:10.1016/j.bmcl.2010.10.122

The substitution of each constituent amino acid residue of gratisin (GR) with Ala residue indicated that each side chain structure of the constituent amino acid residues affect largely the antibiotic and hemolytic activities of GR. Among them, the substit

Full text of DOI:10.1016/j.bmcl.2010.10.122

Bioorganic & Medicinal Chemistry Letters 21 (2011) 440–443
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Novel gratisin derivatives with high antimicrobial activity
and low hemolytic activity
Makoto Tamaki a, , Yukie Imazeki a, Aya Shirane a, Kenta Fujinuma a, Mitsuno Shindo b,
⇑
Masahiro Kimura b, Yoshiki Uchida b
a Department of Chemistry, Toho University, Funabashi, Chiba 274-8510, Japan
b Department of Health and Nutrition, Osaka Shoin Women’s University, Higashi-Osaka, Osaka 577-8550, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 9 June 2010
Revised 14 October 2010
Accepted 25 October 2010
Available online 30 October 2010
The substitution of each constituent amino acid residue of gratisin (GR) with Ala residue indicated that
each side chain structure of the constituent amino acid residues affect largely the antibiotic and hemo-
lytic activities of GR. Among them, the substitution of Pro residues at positions 5 and 50 with a cationic
amino acid residues (Lys and Arg) results the high antibiotic activity and the low toxicity against human
0
blood cells. Thus, we have found a novel position on the scaffold of GR at Pro5,5 residues whose modifi-
cation will significantly lower the unwanted hemolytic activity and enhance the desired antibiotic
activity.
Keywords:
Gratisin analogues
Synthesis
Ó 2010 Elsevier Ltd. All rights reserved.
High antimicrobial activity
Low hemolytic activity
Structure–activity relationship
A remarkable increase of multidrug-resistant bacteria has be-
come a serious threat to public health threat.1 Among efforts to
cope with this problem, the amphiphilic antibiotics such as gram-
analogues, which have various amino acid residues in place of
0
7
D
-Tyr6,6
at
.
The results indicated that the structural modifications
0
D
-Tyr6,6 residues of GR are beneficial to identification of novel
0
0
0
0
0
icidin
tyrocidine
S
(GS),2,3 cyclo(-Val1,1 -Orn2,2 -Leu3,3
-
D
D
-Phe4,4 -Pro5,5 -)2,
antibiotic candidates without hemolytic activity.
A
(TA),2,4 cyclo(-Val1-Orn2-Leu3-
-Phe4-Pro5-Phel6-
In the present studies, we performed a systematic approach to
identify the structural determinants of its antibiotic and hemolytic
properties that induced to dissociation of the two closely associ-
ated properties.
0
D
-Phe7-Asn8-Gln9-Tyr10-), and gratisin (GR),2b,5 cyclo(-Val1,1
-
0
0
0
0
0
Orn2,2 -Leu3,3
-
D
-Phe4,4 -Pro5,5
-
D
-Tyr6,6 -)2 are attractive targets for
drug discovery. The antibiotic actions result from an interaction
of these antibiotics with the cell membrane of the target microor-
ganisms. These antibiotics then adopt an antiparallel b-sheet con-
formation with amphiphilicity, which disrupts cell membrane.2–5
In addition, so far, no resistance has been found for the antibiotics,
because it requires significant alteration of the lipid composition of
the cell membrane.6 However, GS and TA not only affect bacterial
membranes but also mammalian cells such as erythrocytes. The
high hemolytic activities of the peptides prevent their direct use
in combating the microbial resistance.2–4 The efforts to increase
its therapeutic index, namely, minimizing its hemolytic activity
while maintaining its high antibiotic activity, have been de-
voted.2–4 For example, single substitution of Gln6 of the natural
TA with a cationic amino acid residue results in significant increase
of a therapeutic index.4c On the other hand, a therapeutic index of
GR and its analogues has not been studied. Recently, we measured
the hemolytic activities against human erythrocytes of GR and GR
The syntheses of GR and its analogues were performed as shown
in Scheme 1. The protected linear precursors were prepared by
using Boc-solid phase peptide synthesis on oxime resin (loading of
oxime group: 0.35 mmol/g resins).8 Leu residue as a C-terminal
amino acid residue was used based on the propensity of the biosyn-
thetic precursor of GS, TA, and GR to form a conformation highly
favorable for head–tail cyclization.1,2,4 The formations of the cyclic
peptides by the cyclization-cleavage of the linear precursors on
oxime resin were performed in 1,4-dioxane with 2 equiv of triethyl-
amine and acetic acid for 1 day at room temperature.7 All the mask-
ing groups were removed by thioanisole (50 equiv) and TFMSA
(5 equiv) in TFA. The cyclic products obtained were purified by
means of sephadex LH-20 column chromatography, followed by
recrystallization. The purity and identity assessment of the products
were confirmed by thin-layer chromatography, high performance
liquid chromatography (HPLC), and fast-atom bombardment (FAB)
mass spectrometry before determination of the biological activities.
First, GR peptides 1–6 having Ala residue in place of each con-
stituent amino acid residue of GR were examined, in order to find
⇑
Corresponding author.
E-mail address: tamaki@chem.sci.toho-u.ac.jp (M. Tamaki).
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.10.122

M. Tamaki et al. / Bioorg. Med. Chem. Lett. 21 (2011) 440–443
441
Figure 1. Dose dependence curves of hemolysis (%) induced by GR and its Ala-
substituted analogues 1–6.
0
0
Scheme 1. Syntheses of GR and its Ala-substituted analogues 1–6. Reagents and
conditions; (a) Boc-amino acid (3 equiv), BOP (3 equiv), HOBt (3 equiv), NEt3
(6.5 equiv) in DMF 90 min, deprotection 25% TFA/DCM 30 min; (b) NEt3 (2 equiv),
AcOH (2 equiv) in 1,4-dioxane, 24 h; (c) thioanisole (50 equiv) and TFMSA (5 equiv)
in TFA, 2 days.
7–10 with various amino acid residues in place of Val1,1 and Pro5,5
residues were prepared (Fig. 2). Syntheses of these GR analogues
were accomplished with the method illustrated in Scheme 1. After
the characterization of the cyclic peptide products, they were sub-
jected to the same biological activity determination.
0
[Lys1,1 ]-GR (7) showed no antibiotic activities against both
novel antibiotic candidates with high antimicrobial and low hemo-
lytic activities. The results are summarized in Table 1 and Fig. 1.
Gram-positive and Gram-negative organisms, and very low hemo-
lytic activity against human blood cells (Table 2 and Fig. 3). The re-
sults indicated that the presence of amino acid residues with
hydrophobic side chains in positions 1 and 10 is important for
exhibiting both antibiotic and hemolytic activities of GR. Next,
0
0
0
[Ala2,2 ]-GR (2), [Ala3,3 ]-GR (3), and [
D
-Ala4,4 ]-GR (4) showed al-
most no antibacterial activities against both Gram-positive and
Gram-negative bacteria. The hemolytic values of 2, 3, and 4
showed considerable decrease in compared with that of GR. The re-
0
0
the biological activities of GR peptides [Ser5,5 ]-GR (8), [Lys5,5 ]-GR
0
0
0
sults indicated that the presences of Orn2,2 , Leu3,3 , and -Phe4,4
D
0
(9), and [Arg5,5 ]-GR (10) having amino acid residues with the neu-
0
residues in GR is necessary to exhibit the antibiotic activity and
hemolytic activity, in other words, its modifications are unlikely
to improve the antibacterial activity of GR. On the other hand,
tral hydroxyl and cationic side chains in place of Pro5,5 residues
0
were compared with that of [Ala5,5 ]-GR (5) and GR (Table 2 and
0
Fig. 3). [Ser5,5 ]-GR (8) with the neutral hydroxyl chains showed
0
0
0
[Ala1,1 ]-GR (1), [Ala5,5 ]-GR (5), and [
D
-Ala6,6 ]-GR (6) showed 1/2
1/2–1/8 activity of GR against Gram-positive bacteria, and no activ-
ity toward Gram-negative bacteria. In addition, 8 showed very low
hemolytic activity against human blood cells. The biological activ-
ities of 8 are similar to that of Ala-substituted analogue (5). The
results suggested that the presence of Pro residues at the 5,50 posi-
tion of GR is not necessary for exhibiting the antibiotic activity
against Gram-positive bacteria, but important for exhibiting the
activities against Gram-negative bacteria and human blood cell.
Significant further increase in the activity is achieved when the
cationic amino acid residues are introduced into 5,50 positions of
GR, in comparison to the parent GR and Ala-substituted analogue
(5) at same positions.
activity of GR against Gram-positive bacteria, but showed no anti-
biotic activity against Gram-negative bacteria. In addition, the
hemolytic values of 1, 5, and 6 showed considerable decrease in
compared with that of GR. In the previous studies, we have re-
ported that the antibiotic potency against both Gram-positive
and Gram-negative bacteria can increase and the hemolytic po-
tency for mammalian cell membrane can decrease when cationic
0
D
-amino acid residues are incorporated into positions of
D
-Tyr6,6
0
of GR.7 The results suggested that the modifications of Val1,1 and
0
Pro5,5 residues are likely to improve the antibacterial and hemo-
lytic activity of GR (Table 1 and Fig. 1).
To further exploit the potential of the sensitivity of the biolog-
0
0
ical activities to the Val1,1 and Pro5,5 residues, GR analogues
Table 1
Antibiotic activities of GR and its Ala-substituted analogues 1–6a
A
B
C
D
E
F
GR
1
2
3
4
6.25
12.5
>50
6.25
12.5
>50
100
>100
25
12.5
25
>50
>100
>100
50
6.25
6.25
>50
100
>100
>50
>100
>100
>100
>100
100
>100
>50
>100
>100
>100
>100
100
100
>100
12.5
12.5
>100
12.5
12.5
5
6
12.5
25
a
MIC (minimum inhibitory concentration) was determined by a medium dilu-
tion method with 106 organisms per ml. The microorganisms employed in the
assays were Bacillus subtilis NBRC 3513 (A), Bacillus megaterium ATCC 19213 (B),
Staphylococcus epidemidis NBRC 12933 (C), Staphylococcus aureus NBRC 12732 (D),
Pseudomonas aeruginosa NBRC 3080 (E), and Escherichia coli NBRC 12734 (F).
Figure 2. Primary structures of GR and its analogues (1, 5, and 7–10) with various
0
0
amino acid residues at positions of Val1,1 and Pro5,5
.

442
M. Tamaki et al. / Bioorg. Med. Chem. Lett. 21 (2011) 440–443
Table 2
Antibiotic activities of GR and its analogues (1, 5, and 7–10) with various amino acid
0
0
residues at positions of Val1,1 and Pro5,5 a
A
B
C
D
E
F
GR
1
7
5
8
6.25
12.5
>100
12.5
25
6.25
12.5
>100
25
50
6.25
6.25
12.5
25
>100
50
50
6.25
6.25
6.25
6.25
>100
12.5
12.5
6.25
6.25
100
100
>100
>100
>100
>100
100
>100
>100
>100
>100
25
9
10
6.25
6.25
50
12.5
a
MIC (minimum inhibitory concentration) was determined by a medium dilu-
tion method with 106 organisms per ml. The microorganisms employed in the
assays were Bacillus subtilis NBRC 3513 (A), Bacillus megaterium ATCC 19213 (B),
Staphylococcus epidemidis NBRC 12933 (C), Staphylococcus aureus NBRC 12732 (D),
Pseudomonas aeruginosa NBRC 3080 (E), and Escherichia coli NBRC 12734 (F).
Figure 4. CD spectra of GR and its analogues (5 and 8–10) with various amino acid
0
residues at positions of Pro5,5 in methanol.
conformation of these analogues in methanol are similar to each
0
other, but different from that of GR. Thus, the presence of Pro5,5
residues with pyrrolidine ring is important for maintaining the
GR-like conformation. These results suggested that the replace-
0
ments of Pro 5,5 residues may be partially effective through a struc-
tural change in its biological activity of 5 and 8–10.
In the present studies, we synthesized novel GR peptides 9
and 10 with the strong activity against both Gram-positive
and Gram-negative bacteria, which have amino acid residues
0
with cationic side chains in place of Pro5,5 residues. In addition,
we found that the GR peptides 9 and 10 have differential ionic
interaction against the prokaryotic membrane and eukaryotic
membrane. In other words, the dissociations of high antimicro-
bial and low hemolytic activities are caused by the additional
positive charges of 9 and 10. Our findings should be helpful
in finding drug candidates with high antimicrobial and low
hemolytic activities that are capable of combating microbial
resistance. Currently, further synthetic studies of GR peptides
with both strong antibiotic and low hemolytic activities are car-
rying on.
Figure 3. Dose dependence curves of hemolysis (%) induced by GR and its
0
analogues (1, 5, and 7–10) with various amino acid residues at positions of Val1,1
5,50
and Pro
.
0
0
[Lys5,5 ]-GR (9) and [Arg5,5 ]-GR (10) with basic side chains
showed similar activities against Gram-positive microorganisms
and higher activities than that of parent GR against Gram-negative
microorganisms. Among them, the antibiotic activities against
Pseudomonas aeruginosa NBRC 3080 and Escherichia coli NBRC
12734 of 10 are two and eight times higher than that of parent
GR, respectively. In addition, 9 and 10 displayed stronger inhibi-
tory activity against E. coli NBRC 12734 compared to P. aeruginosa
NBRC 3080. Further, the GR analogues 9 and 10 showed very low
toxicity against human erythrocytes compared with that of GR.
Acknowledgments
This work was supported, in part, by Grant-in-Aid for Scientific
Research (No. 22550156) from the Ministry of Education, Culture,
Sports, Science and Technology of Japan. We are grateful to
Professor Ikuo Kato, Hokuriku University, for his helpful discussion.
Thus, it is interesting to note that significant further increase in
0
therapeutic index is achieved when Pro5,5 residues of GR were re-
placed by cationic amino acid residual Lys or Arg, in comparison to
the Ala-substituted analogue at the same position and parent GR.
Supplementary data
Similar results were reported when cationic side chain was intro-
0
-Tyr6,6 residues.7 Recently, Qin et al.
Supplementary data (experimental procedures, TLC, HPLC and
MS data for 1–10) associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2010.10.122.
duced into the positions of
D
reported in studies of TA that the additional positive charge will
greatly strengthen binding affinity of the resulting TA analogues
for the prokaryotic membrane that contains predominantly nega-
tively charged phospholipids, whereas its effect on the analogue’s
affinity for eukaryotic membrane will be much less significant be-
cause of the zwitterionic nature of the phospholipids in mamma-
lian cells.4c Similar differential ionic interaction might also be
References and notes
1. Travis, J. Science 1994, 264, 360.
2. (a) Izumiya, N.; Kato, T.; Aoyagi, H.; Waki, M.; Kondo, M. Synthetic Aspects of
Biologically Active Cyclic Peptides: Gramicidin S and Tyrocidines; Wiley: New York,
1979; (b) Waki, M.; Izumiya, N. In Biochemistry of Peptide Antibiotics; Kleinkauf,
H., Dhren, H., de Gruyter, Von, Eds.; Fed Rep. Ger.: Berlin, 1990; pp 205–244.
3. (a) Kondejewski, L.; Farmer, S. W.; Wiskart, D. S.; Hancock, R. E. W.; Hodges, R. S.
Int. J. Pept. Protein Res. 1996, 47, 460; (b) Stren, A.; Gibbons, W. A.; Craig, L. C.
Proc. Natl. Acad. Sci. U.S.A. 1968, 61, 734; (c) Hull, S. E.; Karlson, R.; Woolfson, M.
M.; Dodson, E. J. Nature (London) 1978, 275, 206; (d) Katus, T.; Kobayashi, H.;
Kirots, H.; Fujita, Y.; Sato, K.; Nagai, U. Biochim. Biophys. Acta 1987, 899, 159; (e)
Kawai, M.; Yamamura, H.; Tanaka, R.; Umemoto, H.; Ohmizo, C.; Higuchi, S.;
Katitsu, T. J. Peptide Res. 2005, 65, 98.
responsible for the increase in therapeutic index when positively
5,50
charged side chains are introduced into the position of Pro
D
and
0
-Tyr6,6 of GR.
Next, CD spectra of 5, 8–10 and GR were measured in methanol,
in order to investigate the structure–activity relationship of 5 and
8–10 (Fig. 4). CD spectra of 5 and 8–10 were observed a curve sim-
ilar to each other, but different from that of GR, indicating that the

M. Tamaki et al. / Bioorg. Med. Chem. Lett. 21 (2011) 440–443
443
4. (a) Zhou, N.; Mascagni, P.; Gibbons, W. A.; Niccolai, N.; Rossi, C.; Wyssbrod, H. J.
Chem. Soc., Perkin Trans. 2 1985, 581; (b) Kohli, R. M.; Walsh, C. T.; Burkar, M. D.
Nature 2002, 418, 658; (c) Qin, C.; Zhong, X.; Bu, X.; Ng, N. L. J.; Guo, Z. J. Med.
Chem. 2003, 46, 4830.
5. (a) Zharikova, G. G.; Myaskovskaya, S. P.; Silaev, A. B. Vestn. Mosk. Univ. Biol.
Pochvoved. 1972, 27, 110; (b) Myaskovskaya, S. P.; Zharikova, G. G.; Silaev, A. B.
Vestn. Mosk. Univ. Biol. Pochvoved. 1973, 28, 123; (c) Tamaki, M. Bull. Chem. Soc.
Jpn. 1982, 57, 3210; (d) Tamaki, M.; Takimoto, M.; Muramatsu, I. Bull. Chem. Soc.
Jpn. 1987, 60, 2101; (e) Watanabe, E.; Sakamoto, Y.; Kikuchi, S.; Tamaki, M.
Peptide Science 2005; Wakamiya, 2006.
6. Hancock, R. The Lancet 1997, 349, 418.
7. Tamaki, M.; Kokuno, M.; Sasaki, I.; Suzuki, Y.; Iwama, M.; Saegusa, K.; Kikuchi,
Y.; Shindo, M.; Kimura, M.; Uchida, Y. Bioorg. Med. Chem. Lett. 2009, 19, 2856.
8. (a) DeGrado, W. F.; Kaiser, E. T. J. Org. Chem. 1980, 45, 1295; (b) DeGrado, W. F.;
Kaiser, E. T. J. Org. Chem. 1982, 47, 3258; (c) Nakagawa, S. H.; Kaiser, E. T. J. Org.
Chem. 1983, 48, 678; (d) Nakagawa, S. H.; Kaiser, E. T. J. Am. Chem. Soc. 1985, 107,
7087; (e) Mihara, H.; Kaiser, E. T. Science 1988, 242, 925; (f) Sasaki, T.; Findeis, M.
A.; Kaiser, E. T. J. Org. Chem. 1991, 56, 3159; (g) Tamaki, M.; Honda, K.; Kikuchi,
S.; Ishii, R. Tetrahedron Lett. 2006, 47, 8475.