Design and synthesis of C10 modified and ring-truncated canthin-6-one analogues as effective membrane-active antibacterial agents

Article abstract of DOI:10.1016/j.bmcl.2018.06.001

A series of canthin-6-one analogues were designed and synthesized in order to study their antibacterial activity and structure–activity relationships. Compound 22 showed a broad spectrum of antibacterial activity and exhibited better bactericidal effect (

Full text of DOI:10.1016/j.bmcl.2018.06.001

Accepted Manuscript
Design and synthesis of C10 modified and ring-truncated canthin-6-one ana-
logues as effective membrane-active antibacterial agents
Jiangkun Dai, Wenjia Dan, Yunyun Zhang, Mengfan He, Junru Wang
PII:
S0960-894X(18)30478-5
https://doi.org/10.1016/j.bmcl.2018.06.001
BMCL 25882
DOI:
Reference:
Bioorganic & Medicinal Chemistry Letters
To appear in:
Received Date:
Revised Date:
Accepted Date:
29 March 2018
25 April 2018
1 June 2018
Please cite this article as: Dai, J., Dan, W., Zhang, Y., He, M., Wang, J., Design and synthesis of C10 modified and
ring-truncated canthin-6-one analogues as effective membrane-active antibacterial agents, Bioorganic & Medicinal
Chemistry Letters (2018), doi: https://doi.org/10.1016/j.bmcl.2018.06.001
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Design and synthesis of C10 modified and ring-truncated
canthin-6-one analogues as effective membrane-active
antibacterial agents
Jiangkun Daia, Wenjia Dana, Yunyun Zhanga, Mengfan Hea, Junru Wanga,b,*
aCollege of Chemistry & Pharmacy, Northwest A&F University, 22 Xinong Road,
Yangling 712100, Shaanxi, China.
bState key Laboratory of Drug Research, Shanghai Institute of Materia Medica,
Chinese Academy of Sciences, 555 Zu Chong Zhi Road, Zhangjiang Hi-Tech Park,
201203, Shanghai, China.
These two affiliation address are equal.
Corresponding Author
*E-mail: wangjunru@nwafu.edu.cn; wangjr07@163.com.
1

ABSTRACT: A series of canthin-6-one analogues were designed and synthesized in
order to study their antibacterial activity and structure–activity relationships.
Compound 22 showed a broad spectrum of antibacterial activity and exhibited better
bactericidal effect (8-fold superiority against Staphylococcus aureus and 2-fold
superiority against Ralstonia solanacearum) than fosfomycin sodium and propineb
with a minimum inhibitory concentration value of 2 μg/mL. Moreover, it showed low
cytotoxicity, stimulation on germination rates and good “drug-like” properties.
Membrane-active mechanism was further studied by fluorescent microscopy, scanning
electron microscopy, cytoplasmic β-galactosidase leakage assay and evaluation of the
molecular docking. The results showed that 22 may exert its bactericidal effect by
damaging bacterial cell membranes and influencing the membrane formation, both of
which could lead to cell death. The in vivo antibacterial assay with a protective
efficacy of 68% demonstrated the potential of C ring-truncated canthin-6-one 22 as a
new bactericide.
Keywords: canthin-6-one; synthesis; antibacterial activity; structure–activity
relationships; mechanism.
2

Healthcare systems, farming, and the food production industry are the main
sectors driving antibiotic consumption.1–3 Staphylococcus aureus and
methicillin-resistant S. aureus are the leading causes of bacterial infections in humans
with symptoms ranging from simple skin infections to severe necrotizing fasciitis and
pneumonia.4 Bacillus cereus could cause food poisoning, such as a diarrheal
syndrome and an emetic syndrome, both through the production of distinct toxins.5
Bacillus subtilis and Ralstonia solanacearum are major components of plant
pathogens.6 We could see that all of these diseases caused by bacteria constitute a
major threat to humans’ life and property. The emergence of the antibiotic crisis has
further brought great challenges to normal production and life. “The cost in terms of
lost global production between now and 2050 would be an enormous 100 trillion USD
if we do not take action,” as a UK Government report states.7 Therefore, we need to
develop new drugs to replace the ones that no longer work.
Figure 1. Synthetic tactics of C10 modified and ring-truncated canthin-6-one analogues.
Canthin-6-one alkaloids, a subclass of β-carboline alkaloids, are reported to have
antibacterial activity.8 Recently, our group also reported the antibacterial activity of
canthin-6-one and 10-methoxycanthin-6-one with minimum inhibitory concentration
values (MICs) of 7.81 and 3.91 μg/mL against B. cereus, respectively (Figure 1).9,10
The results indicated that C10 modified canthin-6-one analogues may be a good option
for improving the antibacterial activity. In addition, little attention has been paid to the
C3,10 or C5,10 disubstitution, C10 modified with D ring-rejiggered and ring-truncated
3

(ABC ring and ABD ring) canthin-6-one analogues for seeking highly antibacterial
agents. In the present study, such analogues were synthesized and evaluated for their
in vitro antibacterial activity. In addition, the structure–activity relationships,
preliminary antibacterial mechanism, drug feasibility and the in vivo protective
efficacy were also studied.
The synthesis of the target compounds 3–11, 13–14, 16–17 and 20–26 is
illustrated in Figures S1–S4. Compound 27 has been early prepared and reported by
us.11 All spectral and analytical data were consistent with the assigned structures.
Specifically, the successful synthesis of the canthin-6-one analogues was further
confirmed by the crystal structure of compound 21 (CCDC:1820037; Figure S5).
In the 13C nuclear magnetic resonance (NMR) spectra of compound 21, the
signal of the ketone group was detected around δ = 187.68 ppm and the signal of the
lactam group was detected around δ = 167.23 ppm. The signals of the methylene
1
hydrogen atoms were observed at δ = 3.17 ppm and δ = 3.10 ppm in the H NMR
spectra, respectively. Moreover, the signals of the phenyl hydrogen atoms appeared in
the aromatic region. After oxidation, the signals of the methylene group were
disappeared in the NMR spectra of compound 22, which indicated that the oxidation
reaction was successful. In addition, the signal of [M+H]+ could be found at 198.0556
Da in the high-resolution mass spectroscopy (HRMS) of compound 22 (error = 3.53
ppm), which conformed to the theoretical value within the allowable error range (error
< 5 ppm).
Compounds (3–11, 13–14, 16–17 and 20–27) were evaluated for their in vitro
antibacterial activity against four Gram-positive bacteria and one Gram-negative
bacterium, with fosfomycin sodium, cefotaxime sodium and propineb as the positive
controls (Table 1). The partial antibacterial activity of compounds 4, 5 and 7 has been
reported by us.10 However, the activity against S. aureus and MRSA of these three
compounds has not been reported. Compared with the positive control fosfomycin
4

sodium (MIC = 16 μg/mL) against S. aureus, three analogues (21, 22 and 23)
exhibited superior activity with a peak MIC lower than 2 μg/mL. It was worth
mentioning that compound 22 has showed about 16-fold superiority than
canthin-6-one (MIC = 31.25 μg/mL) against S. aureus.9 Two compounds (21 and 22)
displayed better activity against MRSA than the positive controls. Specifically,
compound 22 was presented as potent antibacterial lead against MRSA (2 μg/mL),
4-fold and 8-fold more promising than the commercial drugs fosfomycin sodium and
cefotaxime sodium, respectively. Six compounds (5, 7, 8, 21, 22 and 23) displayed
equal or superior activity against B. cereus compared with the positive controls
cefotaxime sodium and propineb (MIC = 16 μg/mL). For B. subtilis, compounds 21
and 22 displayed equal or superior activity compared with the agrochemical fungicide
propineb. Compared with the positive control propineb (MIC = 8 μg/mL) against R.
solanacearum, four analogues (8, 10, 21 and 22) exhibited equal or better activity
with a peak MIC lower than 4 μg/mL. Interestingly, 10-methoxycanthine (10), 2-fold
more promising than propineb against R. solanacearum, exhibited some selectivity for
the bacterial. Generally, compound 22 was considered to be the preferred
canthin-6-one analogue which showed a broad spectrum of antibacterial activity and
exhibited better bactericidal effect than fosfomycin sodium and propineb.
5

Table 1. Antibacterial activity (MIC in μg/mL) of canthin-6-one analogues.
Gram-positive bacteria
Gram-negative bacteria
No.
3
Structure
S. aureus MRSA B. cereus B. subtilis
R. solanacearum
64
64
>64
>64
>64
>64
>64
>64
>64
>64
16
4
>64
>64
>64
64
4
8
5
6
7
>64
>64
>64
16
>64
>64
>64
>64
64
>64
>64
32
8
>64
>64
>64
>64
>64
>64
8
8
>64
>64
>64
>64
>64
>64
32
>64
4
9
10
11
13
14
>64
>64
>64
>64
>64
32
>64
>64
>64
>64
>64
>64
>64
64
64
>64
>64
16
17
>64
>64
>64
8
>64
4
>64
16
4
>64
4
>64
8
20
21
22
23
2
2
2
4
8
32
16
16
32
6

32
>64
>64
>64
>64
>64
>64
>64
>64
>64
>64
>64
>64
>64
>64
>64
>64
24
25
26
27
>64
>64
>64
F.S.a
C.S.a
P.a
-
-
-
16
1
8
4
2
2
4
8
2
8
16
16
16
16
16
aPositive controls, F.S. = Fosfomycin sodium, C.S. = Cefotaxime sodium, P. = Propineb.
Figure 2. Structure–activity relationships of canthin-6-one analogues against bacteria.
Based on the antibacterial activity data, the structure–activity relationships were
carefully investigated (Figure 2). Hydrogenation is detrimental. For hydrogenated
canthin-6-ones, the activity of N-benzyl compound 3 is better than that of the
debenzyl compound 4. Compared with the MIC of compound 4 against S. aureus and
MRSA (> 64 μg/mL), compound 3 exhibited a better MIC (64 μg/mL). Nitrogen
oxide at N3 and ethyl substituent at C5 are unfavorable. Compounds 6 and 17 are
inactive under the tested range (64–1 μg/mL). The rejiggered D ring could change the
selectivity against bacteria. For example, 10-methoxycanthine (10) exhibited a
significant MIC (4 μg/mL) against R. solanacearum. Compound 16, D ring-rejiggered
analogue with an N-benzyl group at position 5, showed weak antibacterial activity.
For substituents of fully aromatic canthin-6-ones at position 10, the activity against B.
cereus trend is: Methoxy > Propargyl > Hydroxyl > Triazole. For C ring-truncated
canthin-6-one analogues, α,β-unsaturated ketone fragment is favorable for improving
7

the antibacterial activity. For example, compound 22 showed 4-fold superiority than
compound 21 against S. aureus and B. cereus. De-ketone products (25 and 26) were
inactive under the tested range which also indicated that the ketone carbonyl group
was important. The weak or inactive activity of D ring-truncated canthin-6-one
analogues highlighted the importance of D ring. It is obviously to find that the activity
trend of the highly bioactive compounds: ABD ring > ABCD ring > ABC ring.
Many natural and synthetic agents often have relatively low application because
of their toxicity.12 The highly bioactive C ring-truncated canthin-6-one analogue 22
was evaluated for their cytotoxicity against Hela (cervical cancer), SGC-7901 gastric
adenocarcinoma), A549 (lung cancer) and L02 (normal human hepatocyte) cells
(Table S1). The cell viabilities were more than 89% under an effective antibacterial
dose (4 μg/mL) and the selectivity indices were greater than 4.0 for the four cells. The
results suggested that the compound 22 was low cytotoxicity.
Compound 22 was also evaluated for allelopathic activity against turnip
(Raphanus sativus) and wheat (Triticum aestivum L.) seeds by determining the
germination rates with respect to the positive control, glyphosate, a broad-spectrum
systemic herbicide.13 As shown in Table S2, compound 22 showed weak stimulation
effect on germination rates under an effective antibacterial dose (4 μg/mL), which
indicated that compound 22 will not inhibit plant growth when it acts as bactericides.
In 1997, Lipinski et al. published what is widely regarded as the key paper
defining physicochemical and structural properties profiles for the optimal oral
availability of drugs.14 In 2003, Clarke and Delaney reported that most herbicides and
fungicides also adhere to the Lipinski rule.15 As shown in Table S3, the logP value
(2.03) is lower than 5, the molecule weight (197.19) is lower than 500, the number of
hydrogen bond acceptors (3) is lower than 10 and the number of hydrogen bond
donors (0) is lower than 5. The physicochemical properties therefore conform to the
Lipinski rule, which demonstrates that compound 22 has good “drug-like” properties.
8

The antimicrobial mechanisms of canthin-6-one analogues are complex and
elusive. Membrane-active mechanism may be a good breakthrough. The most potent
compound 22, was thus examined for its ability to compromise bacterial membranes
of the representative bacterium S. aureus.
Figure 3. Fluorescence micrographs of S. aureus treated or not treated with 1 μg/mL 22 for 1 h.
(A1) no treatment, Hoechst stained; (A2) no treatment, PI stained; (A3) no treatment, merge graph;
(B1) S. aureus treatment with 22, Hoechst stained; (B2) S. aureus treatment with 22, PI stained;
(B3) S. aureus treatment with 22, merge graph.
Two dyes, Hoechst (2'-(4-ethoxyphenyl)-5-(4-methyl-1-pipe-razinyl)-2,5'-bi-1H-
-benzimidazoletrihydrochl-oride) and PI (propidium iodide), were used to
differentiate between cells with either an intact or a damaged membrane (Figure 3).16
Hoechst can easily permeate the membrane of intact cells and show blue fluorescence
regardless of cell viability. In contrast, PI is a DNA intercalator but lacks cell
permeability which fluoresces in red only when cell membranes are disrupted. As
shown in Figure 3A, S. aureus exhibited blue fluorescence in the absence of
compound 22, whereas no fluorescence was showed in the PI channel, indicating the
membranes of S. aureus were intact. However, after S. aureus was incubated with 22
for 1 h, they were stained by both Hoechst and PI, suggesting that the membranes of S.
aureus were damaged (Figure 3B).
SEM (scanning electron microscopy) of S. aureus revealed morphological
9

changes in the bacterial cell surface (Figure 4). The surfaces of cells in the untreated
group (Figure 4A) was relatively smooth and regular, whereas when treated with
compound 22 (Figure 4B) there was shrinkage. Increased permeabilization of the
membrane may explain the leakage of cytoplasmic material.17
Figure 4. SEM of S. aureus cells: (A) blank group; (B) treated group.
In normal conditions, cytoplasmic β-galactosidase cannot pass through the cell
membrane of bacteria. However, if the cell membrane is damaged, cytoplasmic
β-galactosidase could be detected due to its leakage through the damaged
membrane.18 In the present study, M9 lactose medium was used to induce the
production of cytoplasmic β-galactosidase in S. aureus. Compound 22 was added to
the culture medium and the enzymatic activity of β-galactosidase measured. As shown
in Figure S6, β-galactosidase activity in the culture medium was detected at 10 min
after treatment with 22, and this activity kept increasing during the test period. No
β-galactosidase activity was detected in the culture medium of the control bacterial
cells (not treated with 22). These results indicate that compound 22 increases the cell
membrane permeability of S. aureus, which leads to the leakage of β-galactosidase.
The Surflex-Dock scoring function is a weighted sum of non-linear functions
based on the binding affinities of protein-ligand complexes coupled with their
crystallographically determined structures. The function's primary terms involve
hydrophobic and polar complementarity, with additional terms for entropic and
solvation effects. Surflex-Dock scores are expressed in −log10(Kd) units to represent
binding affinities.19
10

Fatty acids are an important component of the bacterial cell membrane. The
bacterial fatty acid biosynthesis pathway has recently generated much interest for the
development of novel classes of antibacterial agents.20 Biotin carboxylase (BC), one
important portion of the acetyl-CoA carboxylase which catalyzes the first enzymatic
step of fatty acid biosynthesis, is an effective target for antibacterial activity.21,22 In
addition, the compound 22 is similar to the skeleton of the BC inhibitor
imidazo[4,5-b]pyridine to some extent.23 So we further explored the inhibition activity
of compound 22 against BC through molecular docking evaluation, which might
influence the formation of bacterial membranes.
Figure 5. Three-dimensional conformations of compound 22 docked in biotin carboxylase.
A crystal of bacterial biotin carboxylase (PDB ID: 2V58) was selected as a
possible target. The docking evaluation gave a good total score (5.5146) for
compound 22. And the calculated dissociation constant (Kd) value was 3.05×10-6.
These results might rationalize the antibacterial mechanism that the intercalation of 22
disrupts the function of biotin carboxylase, inhibiting the biosynthesis of fatty acids,
further influencing the formation of membrane and ultimately leading to cell death.
As shown in Figure 5, there are two π–alkyl interactions between the planar ABD ring
and the MET302 residue of the target (4.50 and 5.42 Å). Simultaneously, two
hydrogen bonds (1.92 and 2.02 Å) are formed between the carbonyl of D ring and the
GLY364 residue. The ketone carbonyl is adjacent to the ARG366 and PHE363
residues, forming two carbon–hydrogen bonds (2.55 and 2.90 Å). These interactions
also explained the importance of ketone carbonyl group which is beneficial for
11

improving antibacterial activity, conforming to the proposed structure–activity
relationships.
Fluorescent microscopy, scanning electron microscopy and cytoplasmic
β-galactosidase leakage assay indicated that 22 may exert its bactericidal effect by
damaging bacterial cell membranes. The docking evaluation illustrated that 22 might
disrupt the function of biotin carboxylase which could influence the formation of
membrane. These findings suggested dual antibacterial mechanism of
membrane-active compound 22, namely membrane damage and membrane formation.
The in vivo antibacterial activity (protective effect) of the highly bioactive
compound 22 was tested against R. solanacearum on eggplant leaf (Figure S7).24,25
Compound 22 was found to have a good preventative effect against R. solanacearum,
with a protective efficacy of 68%. This demonstrated the potential of the C
ring-truncated canthin-6-one analogue 22 as a new candidate for crop protection.
In summary, we have reported a class of canthin-6-one derivatives as potential
antibiotic agents, and compound 22 was considered to be the highly bioactive
canthin-6-one analogue which exhibited better bactericidal effect than fosfomycin
sodium and propineb. Simultaneously, this study provided a feasible approach,
ring-truncated structural modification, for developing new bactericides based on
polycyclic functional structure. Moreover, the dual membrane-active antibacterial
mechanism, membrane damage and membrane formation, was proposed as one of the
possible targets. To our delight, the in vivo antibacterial assay demonstrated the
potential of the C ring-truncated canthin-6-one 22 as a new candidate for crop
protection.
Acknowledgements
This work was funded by the National Natural Science Foundation of China
(Grant No. 81773603), the Open Foundation of Key Laboratory of Synthetic and
Natural Functional Molecule Chemistry of Ministry of Education (Northwest
12

University, China), and the State Key Laboratory of Drug Research (Grant No.
SIMM1705KF-09).
Supplementary data
Supplementary data associated with this article can be found.
References
[1] Laxminarayan R, Duse A, Wattal C, et al. Lancet. Infect. Dis. 2013, 13, 1057–1098.
[2] Li DD, Zhang SS, Song ZH, et al. Eur. J. Med. Chem. 2018, 143, 558–567.
[3] Zhou Y, Zhang SS, He HW, et al. Bioorg. Med. Chem. 2018, 26, 84–95.
[4] Benoit AR, Schiaffo C, Salomon CE, et al. Bioorg. Med. Chem. Lett. 2014, 24, 3014–3017.
[5] Griffiths MW, Schraft H. Foodborne Diseases (Third edition) 2017, 395–405.
[6] Salanoubat M, Genin S, Artiguenave F, et al. Nature 2002, 415, 497–502.
[7] O’Neill J. Government of the UK. http://apo.org.au/node/63983.
[8] Dai JK, Li N, Wang JR, et al. Molecules 2016, 21, 493.
[9] Dai JK, Dan WJ, Li N, et al. Bioorg. Med. Chem. Lett. 2016, 26, 580–583.
[10] Zhao F, Dai JK, Liu D, et al. Molecules 2016, 21, 390.
[11] Li N, Dai JK, Du HT, et al. Spectrochim. Acta. A Mol. Biomol. Spectrosc. 2015, 136,
900–905.
[12] Xu QB, Chen XF, Feng J, et al. Sci. Rep. 2016, 6, 36238.
[13] Zhang Q, Wang SQ, Tang HY, et al. J. Agric. Food. Chem. 2013, 61, 11447–11452.
[14] Lipinski CA, Lombardo F, Dominy BW, et al. Adv. Drug. Deliver. Rev. 1997, 23, 3–25.
[15] Clarke ED, Delaney JS. Chimia 2003, 57, 731–734.
[16] Su M, Xia DL, Teng P, et al. J. Med. Chem. 2017, 60, 8456–8465.
[17] Sharma R, Patel S, Abboud C, et al. Int. J. Antimicrob. Ag. 2017, 49, 224–232.
[18] Miao J, Zhou J, Liu G, et al. Food Control 2016, 59, 609–613.
[19] Pham TA, Jain AN. J. Med. Chem. 2006, 49, 5856–5868.
[20] Wright HT, Reynolds KA. Curr. Opin. Microbiol. 2007, 10, 447–453.
[21] Cronan Jr. JE, Waldrop GL. Prog. Lipid. Res. 2002, 41, 407–435.
13

[22] Miller Richard J, Dunham S, Mochalkin L, et al. P. Natl. Acad. Sci. USA 2009, 106,
1737–1742.
[23] Cheng CC, ShippsJr. GW, Yang ZW, et al. Bioorg. Med. Chem. Lett. 2009, 19, 6507-6514.
[24] Li P, Shi L, Yang X, et al. Bioorg. Med. Chem. Lett. 2014, 24, 1677–1680.
[25] Li DD, Zhang SS, Zong ZH, et al. Eur. J. Med. Chem. 2017, 136, 114–121.
14

Table Caption
Table 1. Antibacterial activity (MIC in μg/mL) of canthin-6-one analogues.
Figure Captions
Figure 1. Synthetic tactics of C10 modified and ring-truncated canthin-6-one analogues.
Figure 2. Structure–activity relationships of canthin-6-one analogues against bacteria.
Figure 3. Fluorescence micrographs of S. aureus treated or not treated with 1 μg/mL 22 for 1 h.
(A1) no treatment, Hoechst stained; (A2) no treatment, PI stained; (A3) no treatment, merge graph;
(B1) S. aureus treatment with 22, Hoechst stained; (B2) S. aureus treatment with 22, PI stained;
(B3) S. aureus treatment with 22, merge graph.
Figure 4. SEM of S. aureus cells: (A) blank group; (B) treated group.
Figure 5. Three-dimensional conformations of compound 22 docked in biotin carboxylase.
15

Highlights
 The peak MIC of the synthesized canthin-6-one analogues is 2 μg/mL.
 The structure–activity relationships are summarized.
 The highly bioactive compound 22 has good drug feasibility.
 The compound 22 might damage bacterial cell membranes and influence the
membrane formation.
16

17