Antibacterial activity of berberine-NorA pump inhibitor hybrids with a methylene ether linking group

Article abstract of DOI:10.1016/j.bmc.2009.04.028

Conjugation of the NorA substrate berberine and the NorA inhibitor 5-nitro-2-phenyl-1H-indole via a methylene ether linking group gave the 13-substituted berberine-NorA inhibitor hybrid, 3. A series of simpler arylmethyl ether hybrid structures were also

Full text of DOI:10.1016/j.bmc.2009.04.028

Bioorganic & Medicinal Chemistry 17 (2009) 3866–3872
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Antibacterial activity of berberine-NorA pump inhibitor hybrids
with a methylene ether linking group
a
a
b
b
Siritron Samosorn a, , Bongkot Tanwirat , Nussara Muhamad , Gabriele Casadei , Danuta Tomkiewicz ,
Kim Lewis b, Apichart Suksamrarn c, Therdsak Prammananan d, Karina C. Gornall e,
Jennifer L. Beck e, John B. Bremner e
*
a Department of Chemistry, Faculty of Science, Srinakharinwirot University, Bangkok 10110, Thailand
b Department of Biology and Antimicrobial Discovery Center, Northeastern University, Boston, MA 02115, USA
c Department of Chemistry, Faculty of Science, Ramkhamhaeng University, Bangkok 10240, Thailand
d BIOTEC Central Research Unit, National Center for Genetic Engineering and Biotechnology, Pathumthani 10120, Thailand
e School of Chemistry, University of Wollongong, NSW 2522, Australia
a r t i c l e i n f o
a b s t r a c t
Article history:
Conjugation of the NorA substrate berberine and the NorA inhibitor 5-nitro-2-phenyl-1H-indole via a
methylene ether linking group gave the 13-substituted berberine-NorA inhibitor hybrid, 3. A series of
simpler arylmethyl ether hybrid structures were also synthesized. The hybrid 3 showed excellent anti-
bacterial activity (MIC Staphylococcus aureus, 1.7 lM), which was over 382-fold more active than the par-
ent antibacterial berberine, against this bacterium. This compound was also shown to block the NorA
Received 24 February 2009
Revised 14 April 2009
Accepted 15 April 2009
Available online 19 April 2009
efflux pump in S. aureus.
Keywords:
Ó 2009 Elsevier Ltd. All rights reserved.
13-Substituted berberine
Antibacterial
NorA efflux pump inhibitors
Dual action
Hybrid
1. Introduction
bacterial hybrid SS14 (2),8 which is a covalently linked combina-
tion of berberine (1) and 2-phenyl-5-nitro-1H-indole, INF55, via
The problem of resistance by human pathogenic bacteria to
antibacterial agents is a serious and growing health issue world-
wide.1 Bacteria employ a number of resistance mechanisms to
counter antibacterial challenge and one of these is the overexpres-
sion of transmembrane protein-based efflux pumps which can
pump out various antibacterials from within the cells, thus lower-
ing the antibacterial concentration to sub-lethal levels.2,3 The ef-
flux pump mechanism is a very significant and widespread one
occurring in a range of human pathogenic bacterial hosts including
Staphylococcus aureus, a common cause of bacteremia, sepsis and
wound infections. In S. aureus, the multidrug resistance (MDR)
pump NorA, for example, can efflux structurally diverse antibacte-
rials like berberine and ciprofloxacin.4
a methylene linking group at the 13 position in 1 and the ortho po-
sition of the indolic 2-phenyl ring (Fig. 1). Such a hybrid has the po-
tential advantage over two separately administered drugs of
synchronous or near-synchronous delivery to the respective bio-
logical target sites. Compound 1 is a natural antibacterial agent
and a preferred substrate of the NorA pump, while INF55 is a syn-
thetic NorA inhibitor which potentiates the antibacterial activity of
1 by reducing the efflux of 1 from bacterial cells.9 Compound 2 is
the first hybrid of an anti-infective/multidrug resistance (MDR)
pump inhibitor. It was shown to be a superior antibacterial with
over 200-fold greater activity than 1 itself against an S. aureus over-
expressing NorA strain and is insensitive to MDR efflux.8
Some studies have been undertaken on structure–activity rela-
tionships for INF55. A series of 2-arylbenzo[b]thiophenes10 related
to INF55 were reported as NorA inhibitors suggesting that the in-
dole-NH is not essential for NorA inhibitory activity. Replacement
of the nitro group in INF55 with other electron-withdrawing
groups such as the cyano group gave a product which retained
potentiation activity similar to INF55.11 However, other electron-
withdrawing groups, like the sulfonamide group, resulted in a
complete loss of activity.
One strategy to attempt to contain the threat of resistant bacte-
ria is to block a major resistance mechanism at the same time as
blocking another bacterial biochemical pathway. One way this ap-
proach could be realized is by combining two drugs with the two
actions proposed in the same molecule,5–7 as is done with the anti-
* Corresponding author. Tel.: +66 2 6495000x8216; fax: +66 2 2592097.
E-mail addresses: siritron@swu.ac.th, siritron@gmail.com (S. Samosorn).
0968-0896/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2009.04.028

S. Samosorn et al. / Bioorg. Med. Chem. 17 (2009) 3866–3872
3867
the appropriate commercially available arylmethyl bromide gave
the salts 4–7 in low to moderate yields; isolated yields were af-
fected by the extensive chromatography required to isolate the
pure salts. The structures of all the final substituted berberine salts
were substantiated by spectroscopic means, with the low field
chemical shift for the H8 berberine proton singlet, the absence of
a signal for H13 in the berberine nucleus, and the downfield singlet
signal for the methylene ether moiety in the linker group being
particularly diagnostic in the 1H NMR spectra.
O
O
O2N
Cl
N
N
H
INF55
OCH3
OCH3
5
6
4
Berberine (1)
O
O
Br
N
8
O
14
Br
OCH3
OCH3
OCH3
5"
4"
13
O
N
O
12
O2N
3" HN
NH
NO2
OCH3
6'
7'
2.2. Antibacterial activity and cell uptake studies of 3
3
CH2
SS14 (2)
The antibacterial activity of 3 was evaluated only against Gram-
positive bacteria with three strains of S. aureus and three antibiotic
resistant strains of Enterococcus faecalis, since it was known from
our previous studies8 that the related hybrid 2 had no antibacterial
activity against Gram-negative bacteria (concentration up to
CH2
O
O
Br
N
4; R =
5; R =
6; R =
CN
CH2
N
OCH3
CH2
RO
7; R =
OCH3
150 lM against Escherichia coli) possibly due to the INF55-insens-
tive RND pumps present. The results are given in Table 1. 3 was
more potent as an antibacterial than 2 with MIC values up to
1.8-fold greater, as well as being over 382-fold greater in activity
than 1 itself against NorA overexpressing S. aureus strain K2378.
Figure 1. Structures of antibacterial berberine (1), NorA inhibitor INF55, dual-
action hybrid SS14 (2), and 13-O-arylmethyl-substituted berberine derivatives 3–7.
The MIC of 3 was essentially the same, 1.7 lM, in a mutant strain
In order to gain a deeper understanding of the dispositional and
structural requirements of the efflux pump blocking moiety in ber-
berine-based hybrid compounds, the effect of the methylene ether
linking group on antibacterial activity in the analogue 3, of 2, was
investigated, together with the related derivatives 4–7. These last
four compounds involved simplification of the indolic inhibitor
moiety with aryl- or heteroarylmethyl and naphthylmethyl groups
to assess aromatic ring criteria in this region. The results are now
reported in this paper.
of S. aureus lacking the NorA MDR pump (K1758), in the wild type
(8325-4), and in the strain overexpressing the NorA pump (K2378).
3 thus seems to block not only the NorA pump but also the addi-
tional MDR pumps14,15 in S. aureus. Furthermore 3 had excellent
antibacterial activity (MIC 3.4
is known for its high levels of ‘intrinsic antibiotic resistance’16,17
and was resistant to 1 with an MIC of 650–1300 M.
lM, Table 1) against E. faecalis which
l
It is also pertinent that 3 can be transported rapidly into the bac-
terial cells, which can be seen in the uptake assay results (Fig. 2).
Both hybrids 2 and 3 show a higher level of accumulation than 1
alone or in the presence of the MDR inhibitor INF55 (Fig. 2a–c). No
difference in uptake of both hybrids was observed in norA-deleted
(Fig. 2b) and overexpressed (Fig. 2c) strains of S. aureus, which sug-
gests an inhibition of the NorA pump and also is consistent with the
antibacterial activity data.
The increase in berberine fluorescence on binding to DNA is
used to monitor berberine accumulation in the cell.18 However,
berberine hybrids 2 and 3 contain an additional group which could
affect their fluorescence properties and/or binding to DNA. The
fluorescence spectra of 2 in the presence or absence of DNA (results
not shown) indicated that the fluorescence properties were chan-
ged due to the attachment of INF55. The maximum fluorescence
in the absence of DNA shifted from 560 nm in the case of 1 to
530 nm for 219 and 3. Interestingly the intrinsic fluorescence of 3
was around 36 times higher than that of 1 at the same concentra-
tion. The presence of DNA has a different effect on the hybrids,
with 3 showing no effect on the fluorescence spectra, while the
fluorescence maximum of 2 increased around 3.5 times. Impor-
tantly 2 and 3 showed no significant difference in maximum fluo-
rescence due to different DNA concentrations.
2. Results and discussion
2.1. Synthesis of 3–7
The synthesis of the 13-O-arylmethyl-substituted berberine
derivatives involved three steps (Scheme 1). Treatment of com-
mercially available berberine (1) with acetone and sodium hydrox-
ide gave 8-acetonyldihydroberberine (8) and subsequent oxidation
with potassium permanganate followed by deprotonation with
aqueous sodium hydroxide solution then afforded the key syn-
thetic precursor phenol betaine 9.12 Finally O-alkylation of the
betaine 9 with 2-(2-bromomethyl-phenyl)-5-nitro-1H-indole8,13
provided 3 in modest yield. In the same way O-alkylation of 9 with
O
N
CH2COCH3
OCH3
O
8
a
1
OCH3
8
i)
b
ii) c
iii) d
2.3. Pump inhibitory studies
O
O
Br-
O
O
e
In order to assess the MDR pump inhibitory activity of the stud-
ied compounds, the uptake and efflux of ethidium bromide, EtBr, in
the presence of the berberine hybrids or NorA inhibitor were
examined with the results presented in Figures 3 and 4. EtBr is a
known substrate for the major MDR pump, NorA, in S. aureus.14,20
The increased accumulation of EtBr in the presence of the hybrid
suggests an inhibition of MDR transporter(s) (Fig. 3a–c). Hybrid 3
potentiates EtBr uptake around 1.7 times stronger than 2 in all
N
N
OCH3
OCH3
OCH3
OCH3
-O
RO
3-7
9
Scheme 1. Reagents and conditions: (a) acetone, NaOH (s), 50 °C; (b) KMnO4 (aq),
acetone, –20 °C; (c) HCl (concd), MeOH, reflux; (d) 5% NaOH (aq), DCM, rt; (e) RBr,
CH3CN, 60 °C.

3868
S. Samosorn et al. / Bioorg. Med. Chem. 17 (2009) 3866–3872
Table 1
Minimum inhibitory concentrations in
l
M of 1–7 against Gram-positive and Gram-negative human pathogenic bacteria
Strain
1
2a
3
Fold
4
Fold
5
Fold
6
Fold changec
7
Fold changec
changec
changec
changec
S. aureus K1758
D
norA
40a
3.1
3.1
1.7 24
5.5
7
12.0
3
91.4
Twofold less
active
2
4
4
7
191.2
Fivefold less
active
S. aureus 8325-4wild-type
S. aureus K2378 NorA++
325a
1.7 191
1.7 >382
10.9
21.9
10.9
21.9
30
>30
30
24.0
95.8
12.0
24.0
14
7
27
14
182.8
>182.8
91.4
>191.2 <2
>191.2 >3
191.2
191.2
>650a 3.1
325
E. coli BW25113
E. coli ECM1668 marR1642
acrAB
D
TolC
>150 ntb
>150 nt
—
—
2
2
325
15
45.7
D
E. coli K12 wild-type
E. coli ECM1642 marR1642
arcAB+
E. faecalis OG1RF
E. faecalis MMH594
E. faecalis V583
>2600 >150 nt
>2600 >150 nt
—
—
>174.8
>174.8
—
—
>191.6
>191.6
—
—
>182.8
>182.8
—
—
>191.2
>191.2
—
—
650a
6.3
3.4 191
3.4 >191
3.4 >191
nt
nt
nt
—
—
—
nt
nt
nt
—
—
—
nt
nt
nt
—
—
—
nt
nt
nt
—
—
—
>650a 6.3
>650a 6.3
a
Values are retrieved from Ball et al.8
Not tested.
Each compound is compared to 1.
b
c
Figure 2. Accumulation of 1 (ꢀ), 1 in the presence of INF55 (e), 2 (4), and 3 (N) by three strains of S. aureus. (a) wild-type strain, 8325-4; (b) NorA deletion strain, K1758; and
(c) NorA overexpression strain K2378. Accumulation was measured by an increase in fluorescence at 517 nm and expressed as relative fluorescence units (RFUs). Graphs are
representative of at least three independent experiments.
Figure 3. The uptake of ethidium bromide alone (ꢀ) and in the presence of INF55 (s), 2 (
D), and 3 (N): (a) by S. aureus 8325-4, wild-type; (b) by S. aureus K1758, deleting
NorA pump; (c) by S. aureus K2378, overexpressing NorA pump. Graphs are representative of at least three independent experiments.
strains. Similar EtBr uptake in the S. aureus NorA knockout strain
K1758 and wild type 8325-4 suggests that the hybrids effectively
inhibit the NorA pump. However 3 and INF55 show a 2- and 2.6-
fold increase respectively in EtBr uptake in the strain with a NorA
knockout, which implies that these compounds can also inhibit
other MDR transporters in S. aureus in addition to NorA.
In order to elucidate the specificity towards the NorA pump, we
studied the efflux levels in the overexpressing NorA S. aureus strain
K2378 (Fig. 4). The result shows that the efflux of EtBr from this
strain is inhibited by both berberine hybrids, although 3 is a little
more potent than 2.
tive and Gram-negative bacteria, S. aureus and E. coli, respectively
(Table 1). All compounds in this series had stronger activity than 1.
Compounds 4 and 5 showed strong antibacterial activity whereas
6 and 7 had much lower activity in S. aureus. No antimicrobial activ-
ity against the wild type E. coli or strain overexpressing the ArcAB
pump was observed. This, together with the high activity against
RND deletion strains (ECM1668 and BW25113) suggests that those
compounds, as well as berberine are effluxed via the RND pump Ac-
rAB/TolC. The antibacterial activity of all compounds against the
antibiotic resistant strains of E. faecalis was not pursued.
The effect on EtBr efflux in the presence of the NorA inhibitor,
INF55, or 4 and 5 in an overexpressing NorA S. aureus K2378 was
also studied (Fig. 5). Compounds 4 and 5 had an inhibitory effect
on the MDR pumps, whereas 6 and 7 did not show significant ef-
fects through MDR inhibition, a result also commensurate with
the antibacterial trends observed.
2.4. Antibacterial activity of 4–7
The simplified hybrid structural series 4–7 do not contain the
INF55 moiety, therefore they were tested against both Gram-posi-

S. Samosorn et al. / Bioorg. Med. Chem. 17 (2009) 3866–3872
3869
interaction may also have some significance for the antibacterial
action of our 13-O-arylmethyl berberine derivatives. To obtain fur-
ther information on this binding, and taking into account com-
pound availability, the DNA binding affinity, selectivity and
stoichiometry of the most active compound from the simplified
hybrid structure series, 4, was examined. Binding to DNA was as-
sessed using negative ion electrospray ionization mass spectrome-
try (ESI-MS). Two different 16-mer double-stranded (ds) DNA
sequences (D1 and D2) and a DNA-sequence with a 10 base pair
double-stranded region and stretches of contiguous adenines at
the ends of each strand to produce a non-H-bonded forked struc-
ture (F10), were assessed. In addition, binding to an 8-mer tetra-
meric G-quadraplex (Q1) and a quadruplex constructed from two
16-mer strands (Q3)22 was assessed. The quadruplex forms were
included for selectivity purposes and to enable a comparison with
related previous work.23
The results for binding to dsDNA (Fig. 6) showed that 4, which
contains a planar intercalating naphthyl moiety can bind to all
DNA types tested (D1, D2 and F10). While a greater number of ber-
berine molecules bound to the dsDNA (e.g., 6–7 for D2), binding of
4 was still significant with one molecule bound to each of the DNA
sequences. The percentage of DNA bound to 4 as judged from the
ESI mass spectra was approximately 30%, 20% and 15% for D1, D2
and F10, respectively.
Figure 4. The efflux of ethidium bromide by S. aureus K2378, overexpressing NorA
pump, alone (j) and in the presence of INF55 (e), 2 ( ) and 3 (N). Graphs are
representative of at least three independent experiments.
D
Compound 4 had a greater preference for Q1 and Q3 with al-
most 100% of Q1 and 67% of Q3 bound to 4 under the same condi-
tions (not shown). As noted previously,23 the berberine derivative
2 also showed a clear preference for binding to Q1 DNA over D1
or D2 DNA.
3. Conclusions
In summary, an increase of linker chain length in the hybrid 2
by introduction of an oxygen atom produced hybrid 3 which had
stronger antibacterial activity and MDR pump inhibitory potency
than 2. Removal of the indole moiety of 3 produced compounds
with lower dual activity. Structural simplification of 3 gave mole-
cules with a loss of some antibacterial activity. Substitution of
the aryl ring in the 13-substituent with electron-withdrawing
groups (6 and 7) was unfavorable for both antibacterial and NorA
inhibitory activities. Although the antibacterial activity of 4 and 5
was lower than that of 3, they still retained NorA pump inhibitory
activity. This suggested that the indole nucleus is not necessary for
Figure 5. The efflux of ethidium bromide by S. aureus K2378, overexpressing NorA
pump, alone (j) and in the presence of INF55 (e), and 4 (ꢀ), 5 (ꢀ), 6 (
D) and 7 (s).
Graphs are representative of at least three independent experiments.
2.5. DNA binding of 4
The antibacterial activity of 1 is believed to be mediated, at least
in part, through a DNA binding process.21 It is thus possible such an
Figure 6. Binding of berberine (1) (left) and 4 (right) to the dsDNA sequences D1, D2 and F10. Ratio of 4: dsDNA was 9:1. The negative ion ESI mass spectra show the -5 and -6
charge states. (d) dsDNA alone; (}) dsDNA + 1 or 4. Each ion to the right of the ion for dsDNA alone corresponds to another molecule of 1 bound to the DNA. Only one
molecule of 4 was bound to each DNA sequence.

3870
S. Samosorn et al. / Bioorg. Med. Chem. 17 (2009) 3866–3872
NorA pump inhibition, but is required for high antibacterial activ-
ity of the hybrid 13-substituted berberines.
147.4 (C-13), 146.9 (C-3a), 144.5 (C-9), 141.9 (C-8), 141.6 (C-50),
139.7 (C-7a’), 138.4 (2C, C-13a, C-20), 133.6 (C-200)a, 132.7 (C-300),
132.6 (C-100)a, 132.4 (C-13b), 130.1 (C-500), 129.6 (C-600), 128.2 (C-
12a), 127.5 (C-400), 127.0 (C-3a’), 123.9 (C-12)b, 121.6 (C-8a),
119.3 (C-4a), 117.2 (C-11)b, 117.1 (C-60), 116.6 (C-40), 113.1 (C-
70), 111.3 (C-14), 108.1 (C-4), 104.9 (C-30), 102.2 (OCH2O), 74.1
(CH2O), 62.2 (OCH3), 57.0 (C-6), 56.1 (OCH3), 28.0 (C-5). HRMS
(ES); m/z calcd for C35H28N3O7 [M]+: 602.1927; found: 602.1910.
4. Experimental
4.1. Chemistry
Solvents were removed under reduced pressure using a rotary
evaporator. Berberine (chloride salt), benzyl chloride, 4-cyanoben-
zyl bromide, 2-(bromomethyl)pyridine, and 2-(bromomethyl)-
naphthalene were purchased from Sigma-Aldrich Chemical Co.
and were used as supplied. Melting points were obtained using a
Griffin melting point apparatus and are uncorrected. Thin layer
chromatography (TLC) on aluminum backed sheets of Merck Silica
Gel 60 F254 plates were used to follow the progress of chemical reac-
tions. Preparative TLC was performed on 20 ꢀ 20 cm plates. Com-
pounds were detected by examination under UV light. Column
chromatography was performed under medium pressure on silica
gel 60 (230–400 mesh). All solvent proportions were vol/vol. NMR
spectra were obtained on a Varian Unity 300 MHz spectrometer,
where proton (1H) and carbon (13C) spectra were obtained at
300 MHz and 75 MHz, respectively, or on a Varian Inova 500 spec-
trometer, where the 1H and 13C were obtained at 500 MHz and
126 MHz, respectively. Spectra were recorded in CDCl3 (unless
otherwise indicated) and were referenced to the residual non-deu-
terated solvent signal or TMS. Hydrogen and carbon assignments
were also made using gradient correlation spectroscopy (gCOSY),
gradient heteronuclear single quantum correlation (gHSQC) and
gradient heteronuclear multiple bond correlation (gHMBC) spectro-
scopic techniques. Superscript letters refer to interchangeable
chemical shift assignments. Positive ion high resolution electro-
spray mass spectra, HRMS (ES), were obtained with a Micromass
Qtof2 mass spectrometer using a cone voltage of 30 V and polyeth-
ylene glycol (PEG) as an internal reference. Compounds for testing
were >95% pure on the basis of TLC and 1H NMR analysis. Com-
pounds 912 and 2-(2-bromomethyl phenyl)-5-nitro-1H-indole8
were synthesized according to previous methods.
4.1.1.2. 9,10-Dimethoxy-13-(2-naphthylmethyleneoxy)-5,6-dihy-
drobenzo[g]-1,3-benzodioxolo[5,6-a]quinolizinium bromide
(4). Compound 9 was treated with 2-(bromomethyl)naphthalene
according to the general procedure to give the desired product 4 as
a yellow solid, yield 34%; mp 175–177 °C. 1H NMR (500 MHz,
CDCl3/CD3OD) d: 10.11 (s, 1H, H-8), 8.00 (d, J = 9.5 Hz, 1H, H-12),
7.92 (s, 1H, H-14), 7.80–7.88 (m, 3H, H-40, H-60, H-80)a, 7.81 (d,
J = 9.5 Hz, 1H, H-11), 7.72 (s, 1H, H-10), 7.48–7.58 (m, 2H, H-50, H-
70)a, 7.36 (d, J = 8.5 Hz, 1H, H-30), 6.72 (s, 1H, H-4), 6.02 (s, 2H,
OCH2O), 6.00 (s, 2H, CH2O), 5.08 (t, J = 5.5 Hz, 2H, H-6), 4.32 (s, 3H,
OCH3), 4.07 (s, 3H, OCH3), 3.17 (t, J = 5.5 Hz, 2H, H-5). 13C NMR
(126 MHz, CDCl3/CD3OD) d: 151.1 (C-10), 150.0 (C-13), 149.8 (C-
3a)a, 147.2 (C-14a)a, 145.8 (C-9), 143.0 (C-8), 133.3 (C-4a’)b, 132.9
(C-8a’)b, 132.0 (C-4a)c, 131.8 (C-20), 131.6 (C-13a), 129.8 (C-12a),
128.4 (C-10, C-40)d, 128.0 (C-80), 127.6 (C-60)d, 126.7 (C-50)d, 126.5
(C-70)d, 126.1 (C-30), 125.5 (C-11), 122.8 (C-8a), 118.4 (C-13b)c,
118.0 (C-12), 109.0 (C-14), 108.1 (C-4), 101.9 (OCH2O), 77.7
(CH2O), 62.8 (OCH3), 57.1 (C-6), 56.9 (OCH3), 27.8 (C-5). HRMS
(ES); m/z calcd for C31H26NO5 [M]+: 492.1811; found: 492.1825.
4.1.1.3. 13-Benzyloxy-9,10-dimethoxy-5,6-dihydrobenzo[g]-
1,3-benzodioxolo[5,6-a]quinolizinium bromide (5). Compound
9 (1 mmol) was treated with benzyl chloride (10 mmol) and sodium
bromide (10 mmol) initially, for bromide exchange purposes, and
then the general procedure was used to give the desired product 5
as a yellow solid, yield 29%; mp 141–143 °C. 1H NMR (300 MHz,
CDCl3) d: 10.39 (s, 1H, H-8), 7.95 (s, 1H, H-14), 7.92 (d, J = 9.3 Hz,
1H, H-11), 7.79 (d, J = 9.3 Hz, 1H, H-12), 7.34–7.37 (m, 3H, H-20, H-
40), 7.27–7.30 (m, 2H, H-30), 6.83 (s, 1H, H-4), 6.09 (s, 2H, OCH2O),
5.24 (t, J = 6.0 Hz, 2H, H-6), 4.89 (s, 2H, CH2O), 4.35 (s, 3H, OCH3),
4.07 (s, 3H, OCH3), 3.26 (t, J = 6.0 Hz, 2H, H-5). 13C NMR (75 MHz,
CDCl3) d: 151.2 (C-10)a, 149.9 (C-3a)b, 149.8 (C-13), 147.4 (C-14a)b,
146.2 (C-9)a, 143.7 (C-8), 134.4 (C-10), 132.2 (C-4a)c, 131.3 (C-13a),
129.7 (C-12a), 129.1 (C-40), 128.8 (2C, C-20), 128.7 (2C, C-30), 125.4
(C-11), 122.9 (C-8a), 118.5 (C-13b)c, 118.0 (C-12), 108.9 (C-14),
108.3 (C-4), 102.0 (OCH2O), 77.2 (CH2O), 63.1 (OCH3), 57.0 (OCH3),
56.9 (C-6), 28.1 (C-5). HRMS (ES); m/z calcd for C27H24NO [M]+:
442.1654; found; 442.1660.
4.1.1. General procedure for the preparation of 3–7
A solution of the phenolbetaine 9 (1 mmol) and the aryl-
methylbromide (2–10 mmol) in dry acetonitrile (1–2 mL) was
heated at 60 °C for 2–6 h under a nitrogen atmosphere. The reac-
tion mixture was then concentrated by evaporation of the CH3CN.
The residue was chromatographed on silica gel (6–10% MeOH in
DCM), followed by preparative TLC (multiple development, silica
gel, 5% MeOH in DCM) of the main fraction from the column. Sub-
sequently, the polar fraction was precipitated from 2% MeOH in
DCM and then recrystallized from EtOH to give the desired
product.
4.1.1.4. 13-(4-Cyanobenzyloxy)-9,10-dimethoxy-5,6-dihydro-
benzo[g]-1,3-benzodioxolo[5,6-a]quinolizinium bromide
(6). Compound
9 was treated with 4-cyanobenzyl bromide
4.1.1.1. 9,10-Dimethoxy-13-[2-(5-nitro-1H-indol-2-yl)-benzyl-
oxy]-5,6-dihydrobenzo[g]-1,3-benzodioxolo[5,6-a]quinolizini-
according to the general procedure to give the desired product 6
as a yellow solid, yield 40%; mp 132–134 °C. 1H NMR (500 MHz,
CDCl3/CD3OD) d: 10.03 (s, 1H, H-8), 7.88 (d, J = 9.5 Hz, 1H, H-12),
7.83 (s, 1H, H-14), 7.81 (d, J = 9.5 Hz, 1H, H-11), 7.69 (d, J = 8 Hz,
2H, H-30), 7.52 (d, J = 8 Hz, 2H, H-20), 6.87 (s, 1H, H-4), 6.08 (s,
2H, OCH2O), 5.05 (t, J = 6.0 Hz, 2H, H-6), 4.99 (s, 2H, CH2O), 4.32
(s, 3H, OCH3), 4.08 (s, 3H, OCH3), 3.33 (t, J = 6.0 Hz, 2H, H-5). 13C
NMR (126 MHz, CDCl3/CD3OD) d: 151.2 (C-10), 150.1 (C-3a)a,
149.7 (C-13), 147.4 (C-14a)a, 145.9 (C-9), 143.4 (C-8), 140.0 (C-
10), 132.5 (2C, C-30), 132.4 (C-4a), 131.3 (C-13a), 129.5 (C-12a),
128.8 (2C, C-30), 125.5 (C-11), 122.8 (C-8a), 118.3 (CN), 118.1 (C-
13b), 117.6 (C-12), 112.6 (C-40), 108.4 (C-4, C-14), 102.0 (OCH2O),
75.8 (CH2O), 62.7 (OCH3), 57.0 (C-6), 56.9 (OCH3), 27.8 (C-5). HMRS
(ES); m/z calcd for C28H23N2O5 [M]+: 467.1607; found: 467.1618.
um bromide (3). Compound
9
was treated with 2-(2-
bromomethyl phenyl)-5-nitro-1H-indole according to the general
procedure to give the desired bromide salt (3) as a yellow solid,
yield 35%; mp 201 °C (dec.). 1H NMR (500 MHz, CDCl3/CD3OD) d:
9.56 (s, 1H, H-8), 8.31 (s, 1H, H-40), 7.97 (d, J = 8.5 Hz, 1H, H-60),
7.68 (d, J = 8.5 Hz, 1H, H-70), 7.64 (s, 1H, H-14), 7.53 (d, J = 9.5 Hz,
1H, H-12)a, 7.37 (d, J = 7.0 Hz, 1H, H-600), 7.25 (d, J = 7.0 Hz, 1H,
H-500), 7.14 (d, J = 7.6 Hz, 1H, H-300), 7.08 (d, J = 9.5 Hz, 1H, H-11)a,
6.90 (t, J = 7.6 Hz, 1H, H-400), 6.63 (s, 1H, H-4), 6.21 (s, 1H, H-30),
6.12 (s, 2H, OCH2O), 5.18 (s, 2H, CH2O), 4.75 (br s, 2H, H-6), 4.19
(s, 3H, OCH3), 3.64 (s, 3H, OCH3), 2.88 (t, J = 5.8 Hz, 2H, H-5). 13C
NMR (126 MHz, CDCl3/CD3OD) d: 149.7 (C-14a), 149.6 (C-10),

S. Samosorn et al. / Bioorg. Med. Chem. 17 (2009) 3866–3872
3871
4.1.1.5. 9,10-Dimethoxy-13-(2-pyridylmethyleneoxy)-5,6-dihy-
drobenzo[g]-1,3-benzodioxolo[5,6-a]quinolizinium bromide
(7). Compound 9 was treated with 2-(bromomethyl)pyridine
according to the general procedure to give the desired product 7
as a yellow solid, yield 8%; mp 134–136 °C. 1H NMR (300 MHz,
CDCl3) d: 10.34 (s, 1H, H-8), 8.59 (dd, J = 4.8, 0.9 Hz, 1H, H-30),
8.07 (d, J = 9.0 Hz, 1H, H-12), 7.94 (s, 1H, H-14), 7.81 (d,
J = 9.3 Hz, 1H, H-11), 7.76 (td, J = 7.5, 1.8 Hz, 1H, H-50), 7.42 (d,
J = 8.1 Hz, 1H, H-60), 7.29 (td, J = 4.8, 0.9 Hz, 1H, H-40), 6.82 (s, 1H,
H-4), 6.06 (s, 2H, OCH2O), 5.24 (t, J = 5.9 Hz, 2H, H-6), 5.01 (s, 2H,
CH2O), 4.34 (s, 3H, OCH3), 4.07 (s, 3H, OCH3), 3.30 (t, J = 5.9 Hz,
2H, H-5). 13C NMR (75 MHz, CDCl3) d: 154.2 (C-10), 151.2 (C-10),
149.9 (C-3a)a, 149.8 (C-13), 149.6 (C-30), 147.5 (C-14a)a, 146.0 (C-
9), 143.7 (C-8), 137.1 (C-50), 132.1 (C-4a)b, 131.2 (C-13a), 129.5
(C-12a), 125.5 (C-11), 123.7 (C-40), 123.1 (C-60), 122.8 (C-8a),
118.3 (C-13b)b, 118.2 (C-12), 108.8 (C-14), 108.3 (C-4), 102.0
(OCH2O), 76.6 (CH2O), 63.1 (OCH3), 57.1 (C-6), 57.0 (OCH3), 28.1
(C-5). HRMS (ES); m/z calcd for C26H23N2O5 [M]+: 443.1607; found:
443.1626.
at 37 °C for 20 min. After washing twice with ice-cold PBS, cells
were added to a chilled 96-well flat-bottom black microtiter plate
containing ice-cold PBS + 10 mM dextrose at an OD of 0.3 in a total
volume of 200
cells to give final concentrations of 3
(i.e., no ethidium efflux), PBS (without dextrose) containing
30 g/mL of reserpine was added instead of 3. Fluorescence was
l
L. Compounds 2–7 or INF55 were added before
l
M. As a negative control
l
measured with a SpectraMax Gemini XS at excitation/emission
wavelengths of 530/600 nm, respectively.
4.3. DNA binding
4.3.1. Materials
TM
MilliQ water (Millipore, Bedford, USA) was used in all experi-
ments. Ammonium acetate, methanol and acetonitrile were pur-
chased from Ajax Finechem (Seven Hills, Australia).
4.3.1.1. Oligonucleotides. Complementary strands of two duplex
DNA sequences (d(CCTCTCTGGACCTTCC), D1A, Mr 4744.1 Da; d(GG
AAGGTCCAGAGAGG), D1B, Mr 5020.3 Da; d(GCTGCCAAATACCTCC),
D2A, Mr 4786.2 Da; d(GGAGGTATTTGGCAGC), D2B, Mr 4977.3 Da)
and a forked duplex sequence (d(AAAAAAAAAACGTCCGAGCA),
F10A1, Mr 6145.1; d(TGCTCGGACGAAAAAAAAAA), F10A2, Mr
6176.1 Da), and two quadruplex DNA strands (d(TTGGGGGT), Q1,
Mr 2496.7 Da and d(GGGGTTTTGGGG), Q3, Mr 3788.5 Da were pur-
chased from Geneworks (South Australia) as ‘trityl-off’ and purified
by chromatography on a Waters C18 Delta Pak radial cartridge high-
performance liquid chromatography (HPLC) column. The column
was equilibrated with 10 mM ammonium acetate, and the DNA
was eluted from the column using a gradient of 0–60% acetonitrile
in 10 mM ammonium acetate over 30 min (1 mL minꢁ1). The puri-
fied DNA strands were freeze-dried using a Savant Speed-Vac and
4.2. Antibacterial testing
4.2.1. Bacterial strains
The following bacterial strains were used in this study: S. aureus
8325-4 (wild-type), K1758
NorA++ (K1758/pK374:norA, with norA from S. aureus SA1199),24
E. coli K12 (wild-type), BW25113 TolC, EMC1668 marR1642 a-
DnorA (8325-4 D
norA),15 K2378
D
D
crAB, EMC1642 marR1642 acrAB+ and E. faecalis MMH594, OG1RF
and V583.25
4.2.2. Antimicrobial susceptibility
Cells (105 mLꢁ1) were inoculated into broth and dispensed at
L wellꢁ1 in 384 well microtitre plates. MICs were determined
redissolved in MilliQ water giving a concentration of 1–2 mM prior
TM
50
l
in triplicate by serial 2-fold dilution of the test compound. The
MIC was defined as the concentration of the agent that completely
inhibited cell growth during an 18 h incubation at 37 °C. Growth
was assayed with a microtitre plate reader (Spectramax PLUS384;
Molecular Devices) by monitoring absorption at 600 nM.
to storage at –20 °C. The concentrations of the oligonucleotides
were determined by measuring the UV absorbance at 260 nm using
molar absorption coefficients for D1A, D1B, D2A, D2B, F10A1, F10A2,
Q1 and Q3 of 137,620, 194,580, 159,363, 177,368 135,030, 143,190,
85,250, and 129,680 Mꢁ1 cmꢁ1, respectively, obtained from the
website ‘Oligonucleotide Properties Calculator’.
4.2.3. Uptake experiments
Experiments were performed essentially as described previ-
ously,8 in which S. aureus cultures were grown at 37 °C until the
optical density (OD) reached 1.5 (at 600 nm). Cells were pelleted
by centrifugation, washed twice with PBS and resuspended in
PBS containing 10 mM dextrose to obtain an ODꢂ0.8. Cells were
then incubated for 1 h at 37 °C (with aeration) before being washed
twice with PBS (containing 10 mM dextrose) and further diluted to
ODꢂ0.3 in PBS. The assay was performed in 96-well flat-bottom
4.3.1.2. Preparation of duplex and quadruplex DNA. The du-
plexes D1 (CTCTCTGGACCTTCC, GGAAGGTCCAGAGAGG), D2 (GCTG
CCAAATACCTCC, GGAGGTATTTGGCAGC) and the forked sequence
F10 (AAAAAAAAAACGTCCGAGCA, TGCTCGGACGAAAAAAAAAA)
were prepared by annealing the complementary single strands in
0.1 M ammonium acetate as previously described,26 to give a stock
concentration of 1 mM. The quadruplexes Q1, d(TTGGGGGT)4, and
Q3,22 d(GGGGTTTTGGGG)2 were prepared by dissolving an appro-
priate quantity of freeze-dried Q1 or Q3 in 0.15 M ammonium ace-
tate (NH4OAc), pH 7. The solutions were heated to 56 °C for 15 min
and allowed to cool to room temperature. The resulting tetrameric
Q1 (1 mM) and dimeric Q3 (1 mM) were stored at –20 °C.
white microtiter plates in a final volume of 200
lL. Compounds
were added at a concentration of 3 M each. For ethidium bromide
l
experiments, 2 or 3 or INF55 were added first. Fluorescence was
measured using a SpectraMax Gemini XS spectrofluorometer
(Molecular Devices). Uptake experiments with 1, INF55 and hy-
brids 2 and 3 were performed at excitation/emission wavelengths
of 355/517 nm. Experiments with ethidium bromide were per-
formed at excitation/emission wavelengths of 530/600 nm. Back-
ground fluorescence for all compounds in the absence of cells
was subtracted from the raw data.
4.3.1.3. Preparation of drug/DNA complexes. Stock 4 and 1 solu-
tions (1 mM) were prepared in 0.1 M NH4OAc, pH 8.5. Appropriate
volumes of 0.1 M NH4OAc, D1, D2, or F10 and each of the com-
pounds 1 and 4, were mixed to give reaction mixtures containing
10
l
M dsDNA and drug in the dsDNA/drug ratios of 1:1, 1:3, 1:6,
L. The same reaction mix-
1:9 and 1:12. The final volume was 100
l
4.2.4. Efflux experiments
tures were set up for the quadruplex Q1. For these solutions, an
appropriate volume of stock drug solution was freeze-dried using
a Savant Speed-Vac followed by addition of tetrameric Q1 in
0.15 M NH4OAc, pH 7, to give Q1/drug ratios of 1:1, 1:3, 1:6, 1:9
and 1:12 in a final volume of 100 lL. The D1, D2, F10 and Q1
and Q3 reaction mixtures were diluted with an equal volume of
The efflux assay was performed essentially as described previ-
ously8 with minor modifications for the use of ethidium bromide.
S. aureus NorA overexpressing cells (K2378 NorA++) were grown
at 37 °C to an ODꢂ0.9, pelleted, washed twice with PBS and then
resuspended in PBS to an ODꢂ0.8. Cells were then loaded with
3 lM ethidium bromide and 30 lg/mL of reserpine and incubated
0.1 M NH4OAc (or 0.15 M for Q1) prior to ESI-MS analysis.

3872
S. Samosorn et al. / Bioorg. Med. Chem. 17 (2009) 3866–3872
4. Ines Borges-Walmsley, M.; Walmsley, A. R. Trends Microbiol. 2001, 9, 71.
5. Bremner, J. B.; Ambrus, J. I.; Samosorn, S. Curr. Med. Chem. 2007, 14, 103.
6. Barbachyn, M. R., In Ann. Rep. Med. Chem. Macor, J. E., Eds.; Elsevier Academic:
San Diego, 2008; Vol. 43, pp 281–290.
7. German, N.; Wei, P.; Kaatz, G. W.; Kerns, R. J. Eur. J. Med. Chem. 2008, 43, 2453.
8. Ball, A. R.; Casadei, G.; Samosorn, S.; Bremner, J. B.; Ausubel, F. M.; Moy, T. I.;
Lewis, K. ACS Chem. Biol. 2006, 1, 594.
9. Markham, P. N.; Westhaus, E.; Klyachko, K.; Johnson, M. E.; Neyfakh, A. A.
Antimicrob. Agents Chemother. 1999, 43, 2404.
10. Fournier dit Chabert, J.; Marquez, B.; Neville, L.; Joucla, L.; Broussous, S.;
Bouhours, P.; David, E.; Pellet-Rostaing, S.; Marquet, B.; Moreau, N.; Lemaire,
M. Bioorg. Med. Chem. 2007, 15, 4482.
11. Ambrus, J. I.; Kelso, M. J.; Bremner, J. B.; Ball, A. R.; Casadei, G.; Lewis, K. Bioorg.
Med. Chem. Lett. 2008, 18, 4294.
12. Iwasa, K.; Nanba, H.; Lee, D.-U.; Kang, S.-I. Planta Med. 1998, 64, 748.
13. Samosorn, S.; Bremner, J. B.; Ball, A.; Lewis, K. Bioorg. Med. Chem. 2006, 14, 857.
14. Hsieh, P.-C.; Siegel, S. A.; Rogers, B.; Davis, D.; Lewis, K. Proc. Natl. Acad. Sci.
U.S.A. 1998, 95, 6602.
4.3.2. Electrospray ionization mass spectrometry (ESI-MS)
Negative ion ESI mass spectra of DNA and the drug/DNA com-
plexes were acquired using a Waters extended mass range Q-ToF
TM
Ultima (Wyntheshawe, UK) mass spectrometer, fitted with a Z-
spray ESI source.27 The capillary, RF lens 1 and collision cell were
at 2.5 kV, 70 V and 4 V, respectively. The cone voltage was 100 V
for experiments involving dsDNA and 150 V for experiments
involving Q1 and Q3, and the transport and aperture in all experi-
ments were each at 5 V. The pressure in the ion optics region was
3 ꢀ 10–6 mbar. Thirty acquisitions were combined and the result-
ing spectrum was baseline subtracted and smoothed using a Sav-
itzky Golay algorithm. The instrument was calibrated using
1 mg/mL cesium iodide.
15. Price, C. T. D.; Kaatz, G. W.; Gustafson, J. E. Int. J. Antimicrob. Agents 2002, 20,
206.
Acknowledgments
16. Lynch, C.; Courvalin, P.; Nikaido, H. Antimicrob. Agents Chemother. 1997, 41,
869.
17. Gilmore, M. S.; Sahm, D. F.; Huycke, M. M. Emerging Infect. Dis. 1998, 4, 239.
18. Stermitz, F. R.; Lorenz, P.; Tawara, J. N.; Zenewicz, L. A.; Lewis, K. Proc. Natl.
Acad. Sci. U.S.A. 2000, 97, 1433.
19. Data obtained together with that for isomers of the berberine hybrid 2 with the
methylene linker group attached to the meta or para position of the 2-phenyl
ring in the INF55 moiety: Kelso, M. J.; Tomkiewicz, D.; Casadei, G.; Bremner, J.
B.; Lewis, K., unpublished results.
20. Nelson, M. L. Curr. Med. Chem.: Anti-Infect. Agent 2002, 1, 35.
21. Amin, A. H.; Subbaiah, T. V.; Abbasi, K. M. Can. J. Microbiol. 1969, 15, 1067.
22. Rosu, F.; Gabelica, V.; Houssier, C.; Colson, P.; De Pauw, E. Rapid Commun. Mass
Spectrom. 2002, 16, 1729.
We wish to thank The Thailand Research Fund (MRG 4980158),
the Commission on Higher Education, Srinakharinwirot University,
Ramkhamhaeng University, Thailand, and the University of
Wollongong, Australia, and Northeastern University, USA for sup-
porting this work. The work was also supported by Grant
1R01AI076372-01A2 from the NIH awarded to K. Lewis. The award
of an APA scholarship to K. Gornall and an Endeavour Award to S.
Samosorn are gratefully acknowledged. We thank Glenn Kaatz for
kindly providing S. aureus strains.
23. Gornall, K. C.; Samosorn, S.; Talib, J.; Bremner, J. B.; Beck, J. L. Rapid Commun.
Mass Spectrom. 2007, 21, 1759.
24. Kaatz, G. W.; Seo, S. M. Antimicrob. Agents Chemother. 1997, 41, 2733.
25. Huycke, M. M.; Spiegel, C. A.; Gilmore, M. S. Antimicrob. Agents Chemother 1991,
35, 1626.
26. Kapur, A.; Beck, J. L.; Sheil, M. M. Rapid Commun. Mass Spectrom. 1999, 13, 2489.
27. Sobott, F.; Hernandez, H.; McCammon, M. G.; Tito, M. A.; Robinson, C. V. Anal.
Chem. 2002, 74, 1402.
References and notes
1. Craft, A. C.; D’Antoni, A. V.; Terzulli, S. L. Med. Sci. Mon. 2007, 13, RA103.
2. Poole, K. Ann. Med. 2007, 39, 162.
3. Lomovskaya, O.; Zgurskaya, H. I.; Bostian, K. A.; Lewis, K. In Bacterial Resistance
to Antimicrobials; Wax, R. G., Lewis, K., Salyers, A. A., Taber, H., Eds., 2nd ed.;
CRC Press-Taylor and Francis Group: Florida, 2008; pp 45–69.