Membrane-active antimicrobial poly(amino-modified alkyl) β-cyclodextrins synthesized: Via click reactions

Article abstract of DOI:10.1039/c7md00592j

The emergence of drug-resistant bacteria has led to the high demand for new antibiotics. In this report, we investigated membrane-active antimicrobial β-cyclodextrins. These contain seven amino-modified alkyl groups on a molecule, which act as functional

Full text of DOI:10.1039/c7md00592j

MedChemComm
View Article Online
View Journal
RESEARCH ARTICLE
Membrane-active antimicrobial poly(amino-
modified alkyl) β-cyclodextrins synthesized via
click reactions†
ab
Cite this: DOI: 10.1039/c7md00592j
Miho Nonaka,b Shingo Okuno,b Ryogo Mitsuhashi,b
*
Hatsuo Yamamura,
c
Hisato Kato,
Takashi Katsu,c Kazufumi Masuda,d Koichi Tanimoto,e
Haruyoshi Tomitaf and Atsushi Miyagawaab
The emergence of drug-resistant bacteria has led to the high demand for new antibiotics. In this report,
we investigated membrane-active antimicrobial β-cyclodextrins. These contain seven amino-modified alkyl
groups on a molecule, which act as functional moieties to permeabilize bacterial cell membranes. The
polyfunctionalization of cyclodextrins was achieved through a click reaction assisted by microwave irradia-
tion. A survey using derivatives with systematically varied functionalities clarified the unique correlation of
the antimicrobial activity of these compounds with their molecular structure and hydrophobicity/hydrophi-
licity balances. The optimum hydrophobicity for the compounds being membrane-active was specific to
bacterial strains and animal cells; this led to specific compounds having selective toxicity against bacteria
including multidrug-resistant pathogens. The results demonstrate that cyclodextrin is a versatile molecular
scaffold for rationally designed structures and can be used for the development of new antibiotics.
Received 21st November 2017,
Accepted 17th January 2018
DOI: 10.1039/c7md00592j
rsc.li/medchemcomm
branes, and their multiple hydrophobic groups interact with
the apolar lipid acyl chains in the membranes. This leads to
Introduction
Antimicrobial resistance is a clear and present danger facing
humanity. It is estimated that at least 700 000 people die every
year from antibiotic-resistant infections.1 In Europe, 25 000
deaths each year are caused by drug-resistant bacteria.2 To
combat antimicrobial resistance, the World Health Organiza-
tion and national governments have encouraged the develop-
ment of new antibiotics that utilize alternative mechanisms of
action; many trials have been performed worldwide on these
new antibiotics. It is within this context that antimicrobial
peptides have become attractive as novel agents against drug-
resistant bacteria.3–6 The multiple positive charges of the pep-
tides interact with the negatively charged bacterial cell mem-
membrane disruption and a broad-spectrum antimicrobial ac-
tivity.7,8 Resistance to membrane-active antimicrobial peptides
develops more slowly than that to conventional drugs, as cell
membrane alteration can be metabolically expensive.5,9,10 De-
spite these advantages, applications of the peptides are limited
by their high cytotoxicity and hemolytic activity as well as high
production costs. Furthermore, their synthetic complexity
means that it is difficult to introduce functional moieties into
them that would allow their antimicrobial activity to be maxi-
mized and their toxicity to be minimized.
To solve the problems related to antimicrobial peptides
and mimic their favorable biological activity, we have
designed antimicrobial oligosaccharides.11–14 A molecular
scaffold that we found to be useful was a cyclic oligosaccha-
ride called cyclodextrin (CD). The CD has a cone structure
(about 1 nm in diameter) whose size is comparable to that of
cyclic antimicrobial peptides gramicidin S and polymyxin B.
The molecule is rimmed by hydroxyl groups which can be
chemically modified. In order to mimic the polycationic re-
gions of peptides, we synthesized γ-CD derivatives containing
polyamino groups on the C6 positions (eight groups on the
molecule) which strongly disrupted the bacterial mem-
branes.11 By adding benzyl groups as hydrophobic moieties
to the amino groups, we were able to enhance membrane
permeabilization and inhibit bacterial proliferation in a man-
ner similar to that by natural peptides. These suggest that
a Life and Applied Chemistry, Graduate School of Engineering, Nagoya Institute of
Technology, Gokiso-cho, Showa-ku, Nagoya 466-8555, Japan.
E-mail: cyamamura.hatsuo@nitech.ac.jp
b Materials Science and Engineering, Graduate School of Engineering, Nagoya
Institute of Technology, Gokiso-cho, Showa-ku, Nagoya 466-8555, Japan
c School of Pharmacy, Shujitsu University, 1-6-1 Nishigawara, Naka-ku, Okayama-
shi, Okayama 703-8516, Japan
d Graduate School of Clinical Pharmacy, Shujitsu University, 1-6-1 Nishigawara,
Naka-ku, Okayama-shi, Okayama 703-8516, Japan
e Laboratory of Bacterial Drug Resistance, Graduate School of Medicine, Gunma
University, 3-39-22 Showa-machi, Maebashi, Gunma 371-8511, Japan
f Department of Bacteriology and Laboratory of Bacterial Drug Resistance,
Graduate School of Medicine, Gunma University, 3-39-22 Showa-machi, Maebashi,
Gunma 371-8511, Japan
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c7md00592j
This journal is © The Royal Society of Chemistry 2018
Med. Chem. Commun.

View Article Online
Research Article
MedChemComm
the amino groups and the hydrophobic benzyl moieties on
the CD cooperatively work to be antimicrobial. Furthermore,
we developed a microwave (MW)-assisted Huisgen 1,3-dipolar
cycloaddition method that could be used for the poly-
functionalization of CDs15 and prepared γ-CD derivatives that
contained alkylamino groups that could act as membrane-
active functionalities.12,13 Additional research conducted
using glucose, maltose, maltooctaose and amylose found that
the antimicrobial activity of these molecules depends on the
number of functional groups on the molecular scaffold.14 We
therefore present in this paper a systematic survey of β-CD
derivatives that possess a series of amino-modified alkyl
groups (seven groups per CD molecule) as antimicrobial
functional groups. This study is expected to provide greater
insight into the correlation of the structure–antimicrobial ac-
tivity of CD derivatives for the sake of developing novel anti-
biotics to combat pathogens.
(Schemes 1 and 2). Compounds 1–14 had R1–NH–CH2 moie-
ties linked to the triazole rings (Scheme 1); in other words,
they were secondary amines possessing alkylamino groups.
The alkyl groups (R1) varied from butyl to heptyl groups and
included isomeric linear, branched, and ring structures.
Compounds 13 and 14 incorporated aromatic benzene moie-
ties. The CD secondary amines 1–14 were prepared by an
MW-assisted Huisgen 1,3-dipolar cycloaddition of per-2,3-
acetylated β-CD heptaazide (39)16 using the corresponding
Boc-protected propargyl-alkylamines 20–33 (Scheme 1). The
alkynes were obtained by alkylation of the Boc-protected
propargylamines with the corresponding haloalkanes. Click
reactions using the alkynes were completed in 10 min by MW
heating (120 °C), which attached the seven amino-modified
alkyl groups onto the β-CD molecules. Deprotection of the
acetyl groups and the Boc groups produced the desired CD
secondary amines 1–14 as trifluoroacetic acid (TFA) salts.
Compounds 15 and 17–19 had 1-aminoalkyl moieties on the
triazoles, which are, more specifically, 1-aminoheptyl (15),
1-amino-2-cyclohexylethyl (17), 1-amino-2-phenylethyl (18),
and 1-amino-3-phenylpropyl (19) moieties, respectively
(Scheme 2). Whereas the amino groups of 15 and 17–19
existed on the α carbon linked to the triazole ring, the
Results and discussion
Chemistry
We prepared β-CD derivatives 1–19, which were hepta-
modified with 19 types of amino-modified alkyl groups
Scheme 1 Preparation of β-CD secondary amines 1–14 via MW-assisted click reactions using alkynes 20–33.
Med. Chem. Commun.
This journal is © The Royal Society of Chemistry 2018

View Article Online
MedChemComm
Research Article
Scheme 2 Preparation of β-CD primary amines 15–19 using alkynes 34–38 to introduce primary amine-modified alkyl groups into them using an
MW-assisted click reaction.
7-aminoheptyl CD (16) had terminal amino groups. 15–19
were primary amines and isomers of the corresponding sec-
ondary amines shown in Scheme 1. The primary aminoalkyl
groups were introduced onto the CDs through the click reac-
tion with the corresponding alkynes 34–38 (Scheme 2). The
alkynes were prepared from the corresponding N-Boc amino
acids according to chemistry with the Ohira–Bestmann modi-
fication of the Gilbert–Seyferth reaction.17–19 Each of the
amino acids reacted with N,O-dimethylhydroxylamine to pro-
duce the corresponding Weinreb amides. The reduction of
the amides with LAH produced aldehydes that were subse-
quently converted into the alkynes 34–38 via reaction with
dimethyl-1-diazo-2-oxopropylphosphonate through Horner–
Wadsworth–Emmons-type reactions, loss of nitrogen, and
rearrangement of the resulting alkenylidenecarbenes. The
MW-assisted Huisgen reaction of the azide 39 with the al-
kynes 34–38, followed by acetyl and Boc group deprotection
promoted the per-functionalization of the β-CD molecules to
efficiently produce 15–19. These demonstrated that the click
reaction is applicable to both the poly-primary and secondary
amino modification of CD molecules. The good results of the
syntheses of the per-functionalized β-CD derivatives as well
as those of γ-CD12,13 and other sugar derivatives14 confirm
that MW-assisted click reactions are advantageous for sugar
molecule polyfunctionalization.
Antimicrobial activity
The inhibition of bacterial proliferation by the CDs was eval-
uated using minimum inhibitory concentration (MIC) values
against Gram-positive Staphylococcus aureus and Bacillus
subtilis strains and Gram-negative Escherichia coli, Salmonella
typhimurium, and Pseudomonas aeruginosa strains (Table 1).
Unmodified native β-CD was not antimicrobial (>113 μM). In
contrast, most of the CD secondary amines 1–14 were found
to effectively inhibit the growth of the two Gram-positive bac-
teria (MIC values = 1–11 μM). The only two exceptions were
cyclopropylmethyl-modified derivative 2 (MIC value >45 μM
against S. aureus and B. subtilis) and cyclopentylmethyl deriv-
ative 9 (MIC value 21 μM against B. subtilis). The benzene-
containing CD derivatives 13 and 14 were also found to be
antimicrobial as well as the aliphatic group-containing CDs.
The CD primary amines (15–19) were also found to inhibit
the growth of Gram-positive bacteria. Their MIC values were
almost comparable to those of their corresponding secondary
amine isomers (more specifically, 15 and 16 corresponded to
6, 17 corresponded to 12, 18 corresponded to 13, and 19
corresponded to 14). The derivative 15, whose n-heptyl moiety
was internally aminated, and the terminally aminated 16
showed similarly good antimicrobial activity against the two
Gram-positive bacteria. Against the Gram-negative bacteria,
This journal is © The Royal Society of Chemistry 2018
Med. Chem. Commun.

View Article Online
Research Article
MedChemComm
Table 1 Minimum inhibitory concentration (MIC, μM) of CD derivatives against both Gram-positive S. aureus and B. subtilis strains and Gram-negative
E. coli, S. typhimurium, and P. aeruginosa strainsa
Gram-positive strains
Gram-negative strains
Carbon
Compounds
numberb
log Pc
S. aureus
B. subtilis
E. coli
S. typhimurium
P. aeruginosa
Secondary amines
1
2
3
4
5
6
7
8
5
6
−0.90
−1.40
−0.48
−0.57
−1.00
−0.06
−0.15
−0.08
−0.58
0.36
11
>45
3
11
>45
1
1
3
3
3
3
21
5
>44
>45
43
>43
>43
5
10
10
21
20
>44
>45
>43
>43
>43
10
10
42
>42
20
>44
>45
>43
>43
>43
10
5
42
>42
40
3
5
3
3
3
1
5
7
8
9
10
11
12
13
14
0.27
3
3
5
3
10
3
5
5
5
21
10
10
10
21
10
10
20
21
10
−0.16
−0.26
−0.13
9
7
8
9
3
Primary amines
15
16
17
18
19
−0.18
−0.43
−0.35
−0.67
−0.25
3
3
5
5
3
5
5
6
10
>42
10
21
20
10
>42
10
>41
>40
1
21
>42
20
>41
>40
1
10
21
20
23
7
Polymyxin B
1
28
Gramicidin Sd
56
a
b
c
MICs were determined by the liquid microdilution method. Number of carbons in the substituent on a triazole ring. Values of the
corresponding substituted glucose. d See ref. 22.
we were able to observe unique antimicrobial activities. Among
the secondary amines, 6, 7, 11, 12, and 14 exhibited good activ-
ity. More specifically, the 4-methylpentyl CD 7 had a MIC value
against P. aeruginosa, which is comparable to that of antimicro-
bial peptide polymyxin B. Increasing the size of the cycloalkyl
moiety led to increased antimicrobial activity [order of the MIC
values for such compounds: cyclopropyl (2) ≈ cyclobutyl (5) >
cyclopentyl (9) > cyclohexyl (12)]. Among the linear alkyl groups
(1, 3, 6, and 10), the medium-size hexyl (6) group demonstrated
the best activity. The benzene-containing CDs 13 and 14 showed
good antimicrobial activity, and their MIC values were slightly
larger than that of the cyclohexane-containing derivative (12).
The primary amines 15 and 17 exhibited good antimicrobial ac-
tivity, and their MIC values were almost the same as those of
their secondary amine isomers 6 and 12, respectively. However,
the activities of the phenyl-modified primary amines 18 and 19
were slightly less than those of the secondary amine isomers 13
and 14, respectively. Interestingly, the terminally aminated
n-heptyl compound (16) showed much less activity than its
internally aminated isomer (15) against the three Gram-negative
bacteria. From our results, it appears that the activity of the CDs
against the Gram-positive bacteria was stronger than that
against the Gram-negative bacteria; the results are comparable
to those found for the membrane-active antimicrobial peptide
gramicidin S.20
branes of animal cells (Table 2). The n-heptyl-aminomethyl
CD (10), which was found to have strong antimicrobial activ-
ity (MIC value against S. aureus, 5 μM), exhibited significant
hemolysis (116% at 20 μM), whereas the less antimicrobial
cyclopropylmethyl-aminomethyl CD (2) (MIC value against S.
aureus, >45 μM) was a lot less hemolytically active (≃0% at
20 μM). Because membrane-active peptides such as gramici-
din S are often known to disrupt both bacterial and red blood
cell membranes,21,22 it is reasonable for the antimicrobial
and hemolytic activities of individual CD compounds to cor-
relate. As such, the MIC values should be inversely propor-
tional to the observed hemolysis values, as was found to be
the case for CDs 2, 6, 10–12, 15, and 17–19 (Fig. 1). However,
we did find that some of the CD derivatives deviated from
such correlations, in that they exhibited strong antimicrobial
activity but weak hemolytic activity, which means that they
exhibited selective toxicity against bacteria: 3–5, 8, 9, 13, and
16 were derivatives that exhibited such traits. It is noteworthy
that a ratio of the concentration producing 50% hemolysis di-
vided by MIC as a therapeutic index for the β-CDs 3–5, 8, 9,
13, and 16 may be >10 for Gram-positive bacteria.
Antimicrobial activity against drug-resistant Gram positive
bacteria
The antimicrobial activity of the CDs against clinically iso-
lated multidrug-resistant Gram-positive bacteria was exam-
ined (Table 3). Against methicillin-resistant Staphylococcus
aureus (MRSA), the CD derivatives, n-pentyl- (3),
3-methylbutyl- (4), 2-ethylbutyl- (8), and cyclopentylmethyl- (9)
Hemolytic activity
We examined the hemolysis of red blood cells from rabbits
with the CDs to determine whether the CDs disrupt the mem-
Med. Chem. Commun.
This journal is © The Royal Society of Chemistry 2018

View Article Online
MedChemComm
Research Article
Table 2 Hemolysis induced by the CD derivativesa
Hemolysis (%)
Concentration of CD (μM)
Carbon
Compounds
numberb
log Pc
50
20
10
≃0
5
2
1
Primary amines
1
2
3
4
5
6
7
8
9
10
11
12
13
14
5
6
−0.90
−1.40
−0.48
−0.57
−1.00
−0.06
−0.15
−0.08
−0.58
0.36
≃0
≃0
≃0
≃0
≃0
≃0
≃0
≃0
4
≃0
≃0
≃0
≃0
≃0
≃0
5
≃0
1
≃0
≃0
≃0
≃0
64
23
8
≃0
≃0
≃0
≃0
25
15
8
≃0
1
≃0
≃0
110
56
12
8
137
119
107
11
≃0
≃0
7
7
8
7
3
1
5
1
1
5
2
9
2
1
116
108
78
3
110
99
38
2
85
73
13
4
39
30
≃0
4
0.27
10
≃0
2
−0.16
−0.26
−0.13
9
7
8
9
40
23
12
7
3
3
Secondary amines
15
16
17
18
19
−0.18
−0.43
−0.35
−0.67
−0.25
118
3
118
4
113
4
117
4
109
3
107
6
101
4
89
3
37
1
31
4
10
3
13
5
113
94
47
17
6
4
a
Rabbit erythrocytes were incubated with a CD derivative at 37 °C for 20 min, and hemolysis was estimated by measuring the absorbance at
b
540 nm. Lysolecithin (50 μM) was used to determine the 100% level of hemolysis. Number of carbon in the substituent on a triazole ring.
c
Values of the corresponding substituted glucose.
aminomethyl CDs which were less hemolytic, exhibited MIC
values of around 20 μM. Furthermore, there were CD deriva-
tives that were good at inhibiting the growth of clindamycin
(ClDM)-resistant Streptococcus agalactiae. While the MIC
value of ClDM against the bacterium was ca. 300 μM and that
of erythromycin (EM) was >174 μM, the MIC values of 3 and
8 were notably in single digits (3 μM) and those of 4 and 9
were around 10 μM. The MIC values for Enterococcus faecium
[MICs of EM, >174 μM; gentamicin (GM), >536 μM; kanamy-
cin (KM), >528 μM; and streptomycin (SM), 440 μM] were al-
most similar to those for MRSA. In contrast, the antimicro-
bial activities of the CDs against vancomycin-resistant
Enterococcus (VRE) [MIC values of the following antibiotics
against this bacterium are vancomycin (VCM), >174 μM; am-
picillin (ABPC), 366 μM; EM, >174 μM; ciprofloxacin (CPFX),
>386 μM; GM, >536 μM; and KM, >528 μM] were a lot less.
Only the 2-ethylbutyl-modified CD (8) exhibited MIC values at
around 20 μM. None of the CDs was found to result in MIC
values of less than 40 μM against E. faecalis [the MIC values
of the antibiotics minocycline (MINO) and SM were 65 and
110 μM, respectively]. The evaluation against the drug-
resistant clinical isolates exhibited that the β-CDs 3, 4, 8, and
9 maintained activity against resistant pathogens. These are
interesting results and may be a foundation for further drug
development.
Bacterial membrane disruption
The above observed antimicrobial activity of β-CDs can be at-
tributed to membrane disruption. Bacteria contain large con-
centrations of K+ (100–500 mM) within their cells, and per-
meabilization of their membranes causes efflux of the
ion.24,25 To examine the cytoplasmic membrane disruption
abilities of β-CDs, we treated Gram-positive S. aureus with the
n-alkyl-aminomethyl β-CD 1, 3, 6 and 10 and monitored the
K+ efflux from the bacterial cells using a K+ electrode, where
Fig. 1 Relationship between the antimicrobial activity of β-CD deriva-
tives against S. aureus and their hemolytic activity. The antimicrobial
activities of the derivatives in the area demarcated by the blue-dashed
line correlated with their hemolytic activities, whereas those contained
within the red-dashed circle had high antimicrobial activities but low
hemolytic activities.
This journal is © The Royal Society of Chemistry 2018
Med. Chem. Commun.

View Article Online
Research Article
MedChemComm
Table 3 Minimum inhibitory concentration (MIC, μM) of CD derivatives against Gram-positive clinically isolated MRSA, E. faecium, E. faecalis, VRE, and
S. agalactiae strainsa
Drug-sensitive strain
Drug-resistant strains
Methicillin-sensitive
Methicillin-resistant
Vancomycin-resistant
Compounds
S. aureus
S. aureus
S. agalactiae
E. faecium
E. faecalis
Enterococcus
Primary amines
1
3
4
5
6
7
8
9
10
11
12
13
22
5
5
11
10
10
5
>44
21
21
>43
21
21
21
21
>40
40
40
>44
3
11
>43
10
5
>44
21
21
>43
10
10
10
21
>40
20
20
>44
>43
>43
>43
>42
>42
>42
>42
>40
>40
>40
>41
40
>44
>43
>43
>43
42
42
21
>42
>40
40
3
5
10
>40
20
40
21
10
40
20
10
5
40
41
40
41
20
41
20
14
5
Secondary amines
15
16
17
18
19
10
21
10
21
20
10
42
20
41
20
10
42
20
21
20
21
21
20
21
20
42
>42
40
>41
40
21
>42
20
>41
40
a
The MICs were determined by the agar dilution method according to CLSI recommendations.23
changes to cell viability were simultaneously measured.11,12
The n-pentyl-modified derivative (3) that exhibited good activ-
ity against resistant pathogens showed significant K+ efflux
(Fig. 2), while the n-butyl CD 1 showed very less efflux. Native
β-CD showed no efflux. Also, it was found that increases in
the K+ efflux were proportional to decreases in the cell viabil-
ity. These results provided direct evidence that 3 disrupted
the cytoplasmic membrane of S. aureus and subsequently
killed the bacteria. The CD concentrations producing 50% K+
efflux (ED50) and those producing a 50% decrease in bacterial
cell survival (LC50) for β-CDs 1, 3, 6, and 10 are summarized
in Table 4. All these derivatives showed good correlation be-
tween K+ efflux and cell viability. We therefore concluded that
the point of action of the CDs was the cytoplasmic mem-
branes of the Gram-positive bacteria, and these CDs were
bactericidal.
around 20 μM or less. In contrast, against VRE, the CDs in-
cluding 8, whose log P values were from −0.35 to −0.08,
exhibited MIC values at around 20 μM. The hydrophobicity of
CD molecules was also found to affect their abilities to dis-
rupt animal cell membranes. The correlation between red
blood cell hemolysis of the CDs and their log P values is
shown in Fig. 3. The membrane-permeabilizing activity
against red blood cells drastically changed at log P values of
around −0.2, and the more hydrophilic CDs showed almost
no hemolytic activity. This correlation is different from that
observed above for the antimicrobial activity of the CDs.
Therefore, the difference enables CDs with selective toxicity
against Gram-positive bacteria to be identified in such as 3,
4, 8, and 9.
The observed correlation between the membrane activity
and antimicrobial activity with the hydrophobicity of the
β-CDs suggests that β-CD derivatives aminated with suitable
hydrophobic moieties can be membrane active and therefore
antimicrobial. A CD that is too hydrophilic and has small
negative values for log P can interact with a bacterial cell
membrane's anionic surface, but it will not be able to pene-
trate into the apolar hydrocarbon core of the lipid bilayer
membrane. If a CD is too hydrophobic and has a large posi-
tive log P value, then it may disturb the polar surface interac-
tion. The unique activities of the CDs against the bacteria
may be due to the differences between the bacterial cell
membranes and the lipid molecules. The cytoplasmic mem-
brane of Gram-positive bacteria such as S. aureus is rich in
anionic phosphatidylglycerol and cardiolipin, respectively,
whereas that of Gram-negative E. coli is rich in neutrally
charged phosphatidylethanolamine.28 Therefore, the CDs
Relationship between hydrophobicity and antimicrobial and
hemolytic activity
In this study, we observed a unique correlation between the
hydrophobicity of the CD molecules and their antimicrobial
activity and hemolytic activities. The partition coefficient log-
arithm (log P) is a measure of the hydrophobicity of a com-
pound; log P is proportionally related to a compound's hydro-
phobicity. We determined the presence of an optimum
region of log P values of the corresponding substituted glu-
cose of β-CDs that were antimicrobial against resistant patho-
gens (Fig. 3). Against MRSA, ClDM-resistant S. agalactiae and
E. faecium, the CD derivatives (including 3, 4, 8, and 9) with
log P values between −0.6 and −0.06, exhibited MIC values of
Med. Chem. Commun.
This journal is © The Royal Society of Chemistry 2018

View Article Online
MedChemComm
Research Article
Moreover, the unique bacterial selectivity was observed in
Gram-positive resistant pathogens (S. agalactiae > E. faecium
≥ MRSA > VRE > E. faecalis). The CDs may discriminate be-
tween the different bacterial cell membranes. The red blood
cell membrane permeation of the CDs may also be related to
the lipids that the cells contain, such as phosphatidylcholine
and sphingomyelin.28 The neutral lipids seem to prefer the
relatively hydrophobic CDs. Thus, the lowest log P value
(≈−0.2) being hemolytically active may lead to the selective
toxicity of the CDs 3, 4, 8 and 9 against the Gram-positive
pathogens as described above. The antimicrobial and hemo-
lytic activities observed for the CDs in this study showed that
the membrane activity of the CDs depended on their hydro-
phobicity (or a balance of their hydrophobicity and hydrophi-
licity). Aminoglycoside-derived amphiphiles have shown char-
acteristic relationships between hydrophobicity and
antimicrobial activity, and this has led to them having excel-
lent activity against Gram-negative bacteria such as P.
aeruginosa.29,30 The relationship for the CDs in this study is
different from those found for the antibiotics and is unique.
Conclusion
In this study, we developed membrane-active and antimicro-
bial β-CD derivatives containing a variety of amino-modified
alkyl substituents that included primary and secondary amino
groups. The β-CDs exhibited good antimicrobial activities simi-
lar to those found for natural peptides. The activity of the CD
derivatives observed in this study may reflect the characteristic
cell membranes of bacteria, which could lead to selective activ-
ity against bacteria. CD derivatives may therefore be promising
agents against Gram-positive bacteria. In particular, the
n-pentyl- (3), 3-methylbutyl- (4), 2-ethylbutyl- (8), and
cyclopentyl- (9) aminomethyl CDs were found to have strong
antimicrobial activity against drug-resistant Gram-positive bac-
teria such as clindamycin-resistant group B Streptococcus and
far less toxicity against red blood cells. Considering the results
of the γ-CD analogs,12,13 glucose, maltose, maltooctaose and
amylose derivatives,14 the seven and eight alkylamino substitu-
ents on the molecule are qualified for expressing antimicrobial
activity, which suggests that β-CD molecules could be good
molecular scaffolds for antibacterial derivatives. Moreover, the
molecular hydrophobicity may be related to the membrane ac-
tivity and therefore the antimicrobial activity of the CD deriva-
tives. The antimicrobial activity of the CDs described in this
paper is comparable to that found for antimicrobial peptides
such as gramicidin S and polymyxin B; at present it is not supe-
rior to daptomycin. However, the versatile derivatization of
CDs may allow rationally designed structures with desired ac-
tivities to be developed for new antibiotic compounds. The
MW-assisted poly click reaction is a powerful tool for synthe-
sizing these derivatives. The total yield of the antimicrobial de-
rivatives from native CDs here can reach 65% (with five steps),
which suggests that CDs are practical compounds for the de-
velopment of novel antibiotics. CD derivatives such as
1-aminoheptyl- (15) and 1-amino-2-cyclohexylethyl CDs (17)
Fig. 2 Dose-dependence curves for the K+ efflux and the cell viability
of S. aureus on the addition of CDs ((a) 3 and (b) 1). The purple lines
indicate K+ efflux, and the orange lines indicate cell viability. The cells
were incubated with the CD derivatives for 30 min at 37 °C. To
determine the level of 100% K+ efflux, melittin (10 μM for S. aureus)
was used.
Table 4 ED50 (CD concentration giving 50% K+ efflux, μM) and LC50 (CD
concentration giving 50% bacterial cell viability, μM) for CDs 1, 3, 6, and
10a
S. aureus
K+ efflux
Cell viability
Compounds
ED50/μM
LC50/μM
1
3
>50
20
>50
20
6
15
15
10
30
15
β-CD
>50
2–5
0.2–0.5
>50
2–5
0.2–0.5
Gramicidin Sb
Melittinc
a
Cells were incubated with a CD or a peptide for 30 min at 37 °C.
b
See ref. 26. c See ref. 27.
possessing cationic amino moieties can be more active
against Gram positive bacteria than Gram negative patho-
gens. The outer membrane of Gram-negative bacteria may be
associated with the less antimicrobial activity of the CDs.
This journal is © The Royal Society of Chemistry 2018
Med. Chem. Commun.

View Article Online
Research Article
MedChemComm
Fig. 3 Relationship between the antimicrobial and hemolytic activity of β-CD derivatives 1–19 and their hydrophobicities (the log P values of the
corresponding substituted glucoses). Plots are shown for (a) MRSA, (b) S. agalactiae, (c) E. faecium, (d) VRE and (e) hemolytic activity (CD concen-
tration, 50 μM).
were found to have strong activity against bacteria including
drug-resistant pathogens but significant hemolytic activity;
they could therefore be well-suited for topical use in treating
infectious skin diseases as the case of the peptide gramicidin
S. In addition, the preliminary experiment of resistance devel-
opment showed that the resistance of S. aureus to the n-hexyl-
aminomethyl CD (6) did not develop by multipassage treat-
ment (see the ESI†). These suggest that the antimicrobial CDs
are promising and further studies including resistance devel-
opment experiment of 3, 4, 8, and 9 found here are underway.
30 min under argon. After addition of 1-bromo-2-ethylbutane
(2.14 g, 1.94 × 10−2 mol) the reaction mixture was stirred for
12.5 h. After addition of a small amount of methanol and
then EtOAc the reaction mixture was washed with water and
saturated aq. NaCl. It was dried with anhydrous Na2SO4 and
concentrated in vacuo. Silica gel chromatography (hexane/
AcOEt) yielded the corresponding alkyne 27 as oil (1.22 g,
79.2%). The other derivatives were prepared similarly. Spec-
tral details of 20–33 are given in the ESI.†
General procedure for synthesis of Boc-protected amino-
alkyne 34–38 for preparation of CD primary amines
Experimental section
General chemistry
1-Cyclohexylmethyl-Boc-propargylamine 36 was prepared as
follows. Boc-L-cyclohexylalanine (1.05 g, 3.88 × 10−3 mol),
EDC·HCl (876 mg, 4.56 × 10−3 mol), and HOBt (562 mg, 4.16
× 10−3 mol) were dissolved in CH2Cl2 (15.7 cm3). The solution
was stirred at rt for 30 min under argon. N,O-
Dimethylhydroxylamine·HCl (423 mg, 4.34 × 10−3 mol) and
N-methylmorpholine (460 mg, 4.50 × 10−3 mol) were added
and the solution was stirred for 16 h. The solution was evapo-
rated in vacuo, diluted with EtOAc, and then washed with 1
M dHCl, saturated aq. NaHCO3, and saturated aq. NaCl. It
was dried with anhydrous Na2SO4 and concentrated in vacuo.
Silica gel chromatography (hexane/AcOEt) yielded the corre-
1H spectra were recorded at 30 °C on
a
Bruker
AVANCE400Plus Nanobay and AVANCE500US CryoProbe.
MALDI-TOF mass measurements were carried out with a
JEOL JMS-S3000 spectrometer. The microwave reactor used
was a Biotage Initiator EXP used for the Huisgen reaction.
The log P values of the corresponding substituted glucose of
β-CDs were obtained using CambridgeSoft ChemDraw Profes-
sional 15.1.
General procedure for synthesis of N-alkyl-Boc-
propargylamines 20–33 for preparation of CD secondary
amines
1
sponding Weinreb amide as oil (976 mg); H NMR (400 MHz,
N-2-Ethylbutyl-Boc-propargylamine 27 was prepared as fol-
lows. Boc-propargylamine (1.0 g, 6.45 × 10−3 mol) was stirred
with 55% NaH (570 mg, 1.30 × 10−2 mol) in DMF (30 cm3) for
CDCl3) δ 0.87–2.05 (22H, cyclohexylmethyl, Boc), 3.20 (s, 3H,
CH –N–), 3.78 (s, 3H, CH –O–), 4.70–4.80 (1H, Boc-NH–CH_–),
_
3
3
5.02 (d, J = 9.6, 1H, Boc-N_H_–). The amide (717 mg) was
Med. Chem. Commun.
This journal is © The Royal Society of Chemistry 2018

View Article Online
MedChemComm
Research Article
dissolved in Et2O (12.2 cm3) and cooled to 0 °C under argon.
LiAlH4 (95 mg, 1.85 × 10−3 mol) in Et2O (3.5 dm3) was added
and the mixture was stirred for 2 h at 0 °C. 5% aq. NaHSO4
was added and diluted with Et2O. The organic layer was
washed with water and saturated aq. NaCl. It was dried with
anhydrous Na2SO4, concentrated in vacuo, and dissolved in
EtOH. DimethylIJ1-diazo-2-oxopropyl)phosphonate (495 mg,
2.58 × 10−3 mol) and K2CO3 (888 mg, 4.69 × 10−3 mol) were
added and the reaction mixture was stirred at rt under argon
for 20 h. The solution was concentrated in vacuo, diluted with
Et2O, and washed with water and saturated aq. NaCl. The or-
ganic layer was dried over Na2SO4 and concentrated in vacuo.
Silica gel chromatography (hexane/AcOEt) yielded the corre-
sponding alkyne 36 (361 mg, 53.8%, 2 steps). The other deriv-
atives were prepared similarly. Spectral details of 34–38 are
given in the ESI.†
broth were diluted 100 times with fresh broth. One loopful (5
μl, about 105 to 104 CFUs (Colony Forming Units) of bacterial
cells) of each dilution was plated using a Microplanter
(Sakuma: Tokyo, Japan) on agar plates containing the drug.
The growth of the bacteria was determined after incubation
of the plates at 37 °C for 18–20 hours.
Hemolytic activity
Rabbit erythrocytes obtained from rabbit blood (NIPPON
BIO-TEST LABORATORIES INC.) were suspended in buffer
(150 mM NaCl/10 mM Hepes-NaOH, pH 7.4) at a final con-
centration of 0.5% hematocrit. After incubation with CDs at
37 °C for 30 min, hemolysis was estimated by measuring the
absorbance at 540 nm. Lysolethicin (50 μM) was used to de-
termine the 100% level of hemolysis.
K+ efflux and cell viability
General procedure for synthesis of CD amines 1–19
Bacterial cells were washed twice with a buffer (100 mM cho-
line chloride/50 mM Mops-Tris, pH 7.2) and suspended in
this buffer at 2 × 109 cells per cm3. The final volume of the
cell suspension was 1 cm3. The cells were incubated with the
CD at 37 °C for 30 min. After incubation, 0.1 cm3 of the cell
suspension was taken, diluted with physiological saline, and
dispersed on an agar plate prepared with 1% polypeptone,
0.5% yeast extract, 0.5% NaCl, and 1.5% agar (pH was ad-
justed by adding 1 M NaOH). The colonies were counted after
being left to stand overnight at 37 °C and the viability of the
cells was determined. The remaining cell suspension was
centrifuged, and the amount of K+ in the supernatant was
measured using a K+-selective electrode. Melittin (10 μM) was
used to determine the 100% level of K+ efflux from S. aureus.
Preparation of 1-amino-2-phenylethyl-modified 18 was
performed as follows. A reaction solution was prepared by
dissolving per-2,3-acetylated β-CD octaazide 39 (19.0 mg, 1.00
× 10−5 mol) in DMSO–H2O (10 : 1) (1.1 cm3) containing the al-
kyne 37 (21.4 mg, 1.25 mol eq. of an azide group), CuSO4
·5H2O (1.75 mg, 0.1 mol eq.), and sodium ascorbate (17.3
mg, 1.25 mol eq.). After MW heating (120 °C, 10 min), ethyl
acetate was added followed by washing with 5% aq. EDTA.
Silica gel column chromatography (CH2Cl2/methanol) gave
the click reaction product (33.0 mg). Deprotection of the ace-
tyl groups with NaOMe–MeOH followed by that of the Boc
group with TFA gave the desired product 18 (76.3%, overall
yield). Other derivatives were prepared similarly. Spectral de-
tails of 1–19 are given in the ESI.†
Conflicts of interest
Bacteria
There are no conflicts of interest to declare.
The microorganisms Bacillus subtilis 168, Staphylococcus au-
reus FDA 209P, Escherichia coli K12 W3110, S. typhimurium
LT2, and Pseudomonas aeruginosa PAO1 were used as drug-
sensitive bacteria. Drug-resistant bacteria MRSA (methicillin-
resistant Staphylococcus aureus) MS29202, Enterococcus
faecium MS29120, Enterococcus faecalis MS29030, VRE (vanco-
mycin resistant Enterococcus) MS29016, and Streptococcus
agalactiae Str.B-1 were clinical isolates. Their MIC values
against conventional antibiotics are shown in the ESI.†
Acknowledgements
This work was supported by JSPS Grants-in-Aid for Scientific
Research grant number 15K07857 and JST A-STEP (Adaptable
and Seamless Technology Transfer Program through Target-
driven R&D) Program grant number AS2711906T.
References
Minimum inhibitory concentration (MIC)
1 J. O'Neill, Review on Antimicrobial Resistance, Antimicrobial
Resistance: Tackling a Crisis for the Health and Wealth of
Nations, supported by the Wellcome Trust and HM
Government, 2014.
2 European Centre for Disease Prevention and Control,
Antimicrobial resistance surveillance in Europe 2015, Annual
report of the European Antimicrobial Resistance Surveillance
Network (EARS-Net), 2017.
MICs against B. subtilis, S. aureus, E. coli, S. typhimurium,
and P. aeruginosa were determined by the liquid micro-
dilution method, using serially diluted (two-fold) CDs. Cells
(1 × 104) were cultured at 37 °C for 20 h in Mueller–Hinton
broth (0.1 cm3) containing CDs in a 96-well microtiter plate.
The MIC was determined as the lowest concentration of CD
at which cells were unable to grow. The MICs against MRSA,
S. agalactiae, E. faecium, E. faecalis, and VRE were deter-
mined by the agar dilution method according to CLSI recom-
mendations.23 Overnight cultures of the strains grown in
3 A. P. Zavascki, L. Z. Goldani, J. Li and R. L. Nation,
J. Antimicrob. Chemother., 2007, 60, 1206–1215.
4 M. Zasloff, Nature, 2002, 415, 389–395.
This journal is © The Royal Society of Chemistry 2018
Med. Chem. Commun.

View Article Online
Research Article
MedChemComm
5 R. E. W. Hancock and H.-G. Sahl, Nat. Biotechnol., 2006, 24,
1551–1557.
18 E. R. Kiemele, M. Wathier, P. Bichler and J. A. Love, Org.
Lett., 2016, 18, 492–495.
6 F. Van Bambeke, M.-P. Mingeot-Leclercq, M. J. Struelens
and P. M. Tulkens, Trends Pharmacol. Sci., 2008, 29,
124–134.
7 W. C. Wimley, ACS Chem. Biol., 2010, 5, 905–917.
8 I. M. Herzog and M. Fridman, Med. Chem. Commun.,
2014, 5, 1014–1026.
9 M. R. Yeaman and N. Y. Yount, Pharmacol. Rev., 2003, 55,
27–55.
10 K. M. G. O'Connell, J. T. Hodgkinson, H. F. Sore, M. Welch,
G. P. C. Salmond and D. R. Spring, Angew. Chem., Int. Ed.,
2013, 52, 10706–10733.
19 G. J. Roth, B. Liepold, S. Müller and H. J. Bestmann,
Synthesis, 2004, 59–62.
20 V. Kapoerchan, V. E. Spalburg, A. J. de Neeling, R. H. Mars-
Groenendijk, D. Noort, J. M. Otero, P. Ferraces-Casais, A. L.
Llamas-Saiz, M. J. van Raaij, J. van Doorn, G. A. van der
Marel, H. S. Overkleeft and M. Overhand, Chem. – Eur. J.,
2010, 16, 4259–4265.
21 S. Ando, H. Nishikawa, H. Takiguchi, S. Lee and G. Sugihara,
Biochim. Biophys. Acta, 1993, 1147, 42–49.
22 D. L. Lee and R. S. Hodges, Biopolymers, 2003, 71, 28–48.
23 Clinical and Laboratory Standards Institute. Performance
Standards for Antimicrobial Susceptibility Testing; Twenty-
Second Informational Supplement. M07-A9, Wayne, PA 2012.
24 T. Katsu, H. Kobayashi and Y. Fujita, Biochim. Biophys. Acta,
1986, 860, 608–619.
25 T. Katsu, H. Nakagawa and K. Yasuda, Antimicrob. Agents
Chemother., 2002, 46, 1073–1079.
26 K. Yamada, M. Kodaira, S. Shinoda, K. Komagoe, H. Oku, R.
Katakai, T. Katsu and I. Matsuo, Med. Chem. Commun.,
2011, 2, 644–649.
11 H. Yamamura, K. Suzuki, K. Uchibori, A. Miyagawa, M.
Kawai, C. Ohmizo and T. Katsu, Chem. Commun., 2012, 48,
892–894.
12 H. Yamamura, Y. Sugiyama, K. Murata, T. Yokoi, R. Kurata,
A. Miyagawa, K. Sakamoto, K. Komagome, T. Inoue and T.
Katsu, Chem. Commun., 2014, 50, 5444–5446.
13 H. Yamamura, A. Miyagawa, H. Sugiyama, K. Murata, T.
Mabuti, R. Mitsuhashi, T. Hagiwara, M. Nonaka, K.
Tanimoto and H. Tomita, ChemistrySelect, 2016, 3, 469–772.
14 H. Yamamura, T. Mabuchi and A. Miyagawa, Chem. Biol.
Drug Des., 2017, 90, 1012–1018.
27 S. Nakao, K. Komagoe, T. Inoue and T. Katsu, Biochim.
Biophys. Acta, 2011, 1808, 490–497.
15 H. Yamamura, K. Shimohara, R. Kurata, Y. Fujita, K.
Murata, T. Hayashi and A. Miyagawa, Chem. Lett., 2013, 6,
643–645.
16 J. Boger, R. J. Corcoran and J.-M. Lehn, Helv. Chim. Acta,
1978, 61, 2190–2218.
28 K. Lohner and E. J. Prenner, Biochim. Biophys. Acta,
1999, 1462, 141–156.
29 R. I. Benhamou, P. Shaul, I. M. Herzog and M. Fridman,
Angew. Chem., Int. Ed., 2015, 54, 13617–13621.
30 L. Zimmermann, A. Bussiere, M. Ouberai, I. Baussanne, C.
Jolivalt, M.-P. Mingeot-Leclercq and J.-L. Decout, J. Med.
Chem., 2013, 56, 7691–7705.
17 E. Ko, J. Liu, L. M. Perez, G. Lu, A. Schefer and K. Burgess,
J. Am. Chem. Soc., 2011, 133, 462–477.
Med. Chem. Commun.
This journal is © The Royal Society of Chemistry 2018