Structure-activity relations of azafluorenone and azaanthraquinone as antimicrobial compounds

Article abstract of DOI:10.1016/j.bmcl.2004.12.059

Antimicrobial activities of two azafluorenones, four 1-azaanthraquinones, five 2-azaanthraquinones, and one 2-azaquinone were tested. Several azaanthraquinones possessed broad, potent activity, while the azafluorenones demonstrated weak activity. The following structure-activity relationship was postulated: (1) activity decreased in the order 2-azaanthraquinones > 1-azaanthraquinones > azafluorenones; and (2) a hydroxyl group at the peri-carbonyl group enhanced activity. In addition, correlations among reduction potential, hydrophobic parameter, and antimicrobial activity were discussed.

Full text of DOI:10.1016/j.bmcl.2004.12.059

Bioorganic & Medicinal Chemistry Letters 15 (2005) 1079–1082
Structure–activity relations of azafluorenone and
azaanthraquinone as antimicrobial compounds
Junko Koyama,a,* Izumi Morita,a Norihiro Kobayashi,a Toshiyuki Osakai,b
Yoshinosuke Usukic and Makoto Taniguchic
aFaculty of Pharmaceutical Sciences, Kobe Pharmaceutical University, Higashinada-ku, Kobe 658-8558, Japan
bFaculty of Science, Kobe University, Nada, Kobe 657-8501, Japan
cGraduate School of Sciences, Osaka City University, Sugimoto, Sumiyoshi-ku, Osaka 558-8585, Japan
Received 27 October 2004; accepted 9 December 2004
Available online 21 January 2005
Abstract—Antimicrobial activities of two azafluorenones, four 1-azaanthraquinones, five 2-azaanthraquinones, and one 2-azaquin-
one were tested. Several azaanthraquinones possessed broad, potent activity, while the azafluorenones demonstrated weak activity.
The following structure–activity relationship was postulated: (1) activity decreased in the order 2-azaanthraquinones > 1-azaanthra-
quinones > azafluorenones; and (2) a hydroxyl group at the peri-carbonyl group enhanced activity. In addition, correlations among
reduction potential, hydrophobic parameter, and antimicrobial activity were discussed.
Ó 2005 Elsevier Ltd. All rights reserved.
Quinones (anthraquinones, naphthoquinones, and het-
eronaphthoquinones) are important, naturally occur-
ring pigments widely distributed in nature, which
demonstrate antimicrobial and anticancer activities.
An azafluorenone alkaloid, onychine (1), occurring in
Cleistopholis patens (Annonaceae), was active against
Candida, with a minimum inhibitory concentration
(MIC) of 3.12 lg/mL against C. albicans B311 in
yeast–nitrogen broth.1 An azaanthraquinone alkaloid,
cleistopholine (3), was isolated from the same plant.2
Onychine and cleistopholine are related to diazafluor-
anthene eupolauridine through a common hypothetical
biosynthetic precursor derived from oxoaporphine
liriodenine.3
iological pH 7.2. A definite correlation between stan-
dard redox potential and inhibition (log IC50) of EBV-
EA activation was reported previously.7,8 In addition,
introduction of the electronic property7,8 as an addi-
tional parameter supports the correlation.
This paper presents preliminary results of the structure–
activity relations of several azafluorenones and
azaanthraquinones.
Onychine (1),9 4-methyl-1-azafluorenone (2),9 and cleisto-
pholine (3)10 were synthesized by constructing cyclo-
alkenopyridines using the oxidative thermal rearrange-
ment as previously reported. 9-Hydroxy-4-methyl-1-
azaanthraquinone (4), 6,9-dihydroxy-4-meth- yl-1-aza-
Benzo[g]isoquinoline-5,10-dione (7) showed potent terato-
genic and embryotoxic activity in Acheta domesticus4
and possesses antifungal activity.5,6
anthraquinone
(5),
and
9-hydroxy-3-methyl-
1-azaanthraquinone (6) were prepared by the Diels–
Alder reaction followed by oxidation with 2,3-di-
chloro-5,6-dicyano-1,4-benzoquinone. The Diels–Alder
reaction of isoquinoline-5,8-dione (13) and cyclohexa-
1,3-diene followed by oxidation with Ag2O and thermal
elimination of ethylene gave 2-azaanthraquinone (7) in
30% yield from 13. 6-Methoxy-(8) and 9-methoxy-2-aza-
anthraquinone (9) were synthesized from 13 and 1-meth-
oxy-1,3-cyclohexadiene, respectively, in the same
manner as above. Demethylation of 8 and 9 with BBr3
afforded 6-hydroxy-(10) and 9-hydroxy-2-azaanthra-
quinone (11) in 80% and 78% yields, respectively.
In studies on the structure–activity relations of drugs,
standard redox potential is one of the most important
parameters for estimating physiological activity. Cyclic
voltammetry was employed to determine the standard
redox potentials of 1-and 2-azaanthraquinones at phys-
*
Corresponding author. Tel.: +81 441 7549; fax: +81 441 7550; e-mail:
j-koyama@kobepharma-u.ac.jp
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.12.059

1080
J. Koyama et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1079–1082
Thieno[2,3-g]isoquinoline-4,9-dione (12) was prepared
by tandem-directed metalation reaction11 of N,N-dieth-
ylisonicotinamide and 3-thiophenecarboxaldehyde (Fig.
1).
erence, and counter electrodes, respectively. Before
recording each voltammogram, pretreatment of the
working electrode was conducted as previously de-
scribed.6 The test solution was 2:1 (v/v) phosphate buffer
(pH 7.2)-ethanol containing 0.1 mM test compound,
and was degassed with prepurified N2 gas prior to the
voltammetric measurements. The electrolytic cell was
water-jacketed to maintain the temperature at
25 0.1 °C.
Viabilities of compounds (1, 3–9, 11: 32nmol) for Raji
cells were 70%, 70%, 40%, 60%, 30%, 60%, 60%, 60%,
and 60%, respectively.8,12,13
Compounds 1–12 were tested against the representative
strains of Gram-positive bacteria (Staphylococcus aureus
NCTC 8530, Bacillus subtilis IFO 3007), Gram-negative
bacteria (Escherichia coli IFO 3545, Proteus vulgaris
IFO 3851, Pseudomonas aeruginosa IFO 3080), yeasts
(Saccharomyces cerevisiae IFO 0203, Schizosaccharomyces
pombe IFO 0342, Candida utilis OUT 6020, Rhodotorula
rubra IFO 0001), and filamentous fungi (Aspergillus
niger ATCC 6275, Penicillium chrysogenum IFO 4626,
Rhizopus chinensis IFO 4768, Mucor mucedo IFO
7684). The MIC was determined by the 2-fold broth
dilution method. All compounds were tested at an initial
concentration of 100 lg/mL. Each bacterium was cul-
tured in a test tube with 5 mL of peptone medium con-
taining the test compound at 37 °C for 2days. Yeast and
filamentous fungi were cultured on wort at 25 °C for 2
days. The growth of bacteria and yeast was measured
by optical density at 600 nm, and that of filamentous
fungi was estimated visually.
The antimicrobial activities of compounds 1–12 are
summarized in Table 1. Onychine (1) did not possess a
strong anticandidal activity as reported by Clark and
co-workers.1 All compounds tested were inactive against
E. coli. In general, the azafluorenones (1, 2) possessed
weak activity. The 2-azaanthraquinones (7–11) exhib-
ited stronger activity than did the 1-azaanthraquinones
and were active against all microorganisms except for
E. coli and P. vulgaris. Hydroxyl derivatives of the 1-
azaanthraquinones 4–6 were more active than was 3.
Compound 5, with two hydroxyl groups at the peri-car-
bonyl position, was the most active compound among
the 1-azaanthraquinones. The 2-azaanthraquinones
showed the same activity trend as did the 1-azaanthra-
quinones. 6-Hydroxy (10) and 9-hydroxyquinone (11)
were more active than were 7–9. Among the compounds
prepared, 12 was the most potent (0.39–6.25 lg/mL)
against the microorganisms tested, except for E. coli
and P. vulgaris.
Cyclic voltammetry was performed by a conventional
three-electrode system using a laboratory-constructed
microcomputer-controlled system, in which a potentio-
stat (Hokuto Denko, HA-301) was used to control the
working electrode potential. A plastic-formed-carbon
(PFC) electrode with surface area = 0.071 cm2 (BAS,
PFCE-3), an Ag/AgCl (saturated NaCl) electrode, and
a platinum coil electrode functioned as the working, ref-
Cytotoxicity against human A549 cells was also tested
for two compounds 7 and 12.14 LC50 values of 7 and
12 were 1.3 and 0.16 lg/mL, respectively. Compound 7
had less toxicity to A549 cells and antimicrobial activity
than 12. It was found that there was no relationship
between viability for Raji cells and antimicrobial
activity.
O
O
O
CH3
N
N
N
S
O
(2)
CH3
(1)
(12)
R4
O
(3) R1 = R3 = R4 = H, R2 = CH3
(4) R1 = R3 = H, R2 = CH3 R4 = OH
(5) R1 = H, R2 = CH3 R3 = R4 = OH
N
,
R1
,
R2
O
O
R3
R1
(6) R1 = CH3
,
R2 = R3 = H, R4 = OH
(7) R1 =R2 = H
O
N
(8) R1 = H, R2 = OCH3
(9) R1 = OCH3 , R2 = H
(10) R1 = H, R2 = OH
(11) R1 = OH, R2= H
N
O
R2
O
(13)
Figure 1. Structures of compounds.

J. Koyama et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1079–1082
1081
12
Table 1. Minimum inhibitory concentrationsa of azafluorenones, 1-azaanthraquinones, and 2-azaanthraquinones
1
2
3
4
5
6
7
8
9
10
11
S. aureus
>100
100
>100
100
100
100
50
100
50
25
25
12.5
12.5
12.5
6.25
12.5
6.25
>100
6.25
3.13
100
50
6.25
12.5
3.13
12.5
6.25
12.5
3.13
12.5
6.25
12.5
12.5
1.56
3.13
1.56
3.13
1.56
6.25
100
50
1.56
B. subtilis
E. coli
6.25
>100
50
6.25
25
1.56
3.13
>100
50
6.25
25
3.13
25
6.25
>100
100
50
>100
50
50
>100
>100
100
12.5
12.5
6.25
>100
25
3.13
>100
25
3.13
P. vulgaris
P. aeruginosa
S. cerevisiae
S. pombe
50
25
100
25
>100
100
100
100
100
50
100
25
100
100
3.13
0.20
6.25
3.13
3.13
1.56
6.25
6.25
3.13
0.78
1.56
0.39
3.13
1.56
0.78
0.78
6.25
3.13
12.5
12.5
C. utilis
100
100
100
50
50
50
50
50
12.5
12.5
50
25
50
25
12.5
R. rubra
12.5
12.5
12.5
25
3.13
3.13
1.56
A. niger
100
50
>100
25
>100
25
12.5
12.5
25
P. chrysogenum
R. chinensis
M. mucedo
6.25
>100
>100
>100
100
>100
50
>100
50
12.5
12.5
>100
25
12.5
3.13
12.5
25
a MICs are expressed as lg/mL.
Cyclic voltammograms of 2, 7, and 12, which were ob-
tained at 20 mVsÀ1, are shown in Figure 2. As repre-
sented by the voltammogram for compound 7, many
compounds, except for 2, 11, and 12, had two reduction
peaks (the cathodic and anodic peak potentials are sum-
marized in Table 2). However, when voltage scan rate
was increased (e.g., up to 200 mVsÀ1), the second reduc-
tion peak became larger than the first reduction peak,
and therefore, should be considered an adsorption wave
due to the reduction of a pyridine derivative adsorbed at
the electrode surface. In contrast, the first reduction
peak usually is proportional to the square root of the
scan rate, indicating that the peak is limited by diffusion
of the pyridine derivative. Unfortunately, the corre-
sponding anodic peak was not detected for some com-
pounds (1, 2, 11), probably because of the instability
of the reduction products. Accordingly, we chose the
first reduction peaks at 20 mVsÀ1, which reflect the
‘‘thermodynamic’’ redox potentials, to examine their
connections with microorganism MICs.
Furthermore, we calculated some electronic properties
of these compounds by the PM3 method using the
CAChe MOPAC program15 and the logarithm of the
octanol–water partition coefficient of the molecule
(log P).16 Table 3 contains the electronic properties
(LUMO energy, HOMO energy, steric energy, and total
energy) and the hydrophobic parameter (log P). Table 4
shows the results of regression analysis between MIC
and first reduction potential (Epc-1). Good to moderate
correlations between MIC and Epc-1 were obtained with
the exception of activity against E. coli, P. vulgaris,
Figure 2. Cyclic voltammograms of azafluorenone and 2-azaanthra-
quinones.
Table 2. The first and second cathodic peak potentials (Epc-1 and Epc-2
)
and the anodic peak potentials (Epa) versus Ag/AgCl (saturated NaCl)
obtained at 20 mVsÀ1 for azafluorenones, 1-azaanthraquinones, and 2-
azaanthraquinones
Table 3. Electronic properties of azafluorenones, 1-azaanthraquinones,
and 2-azaanthraquinones
Compound
EpcÀ1 (V)
EpcÀ2 (V)
Epa (V)
Compound LUMO HOMO Steric
(eV)
Total
energy
log P
(eV)
energy
(kcal/mole) (hartree)
1
À0.388
À0.788
ND
NDa
ND
2
À0.356
3
À0.339
À0.460
À0.488
À0.536
À0.495
À0.432
À0.373
À0.363
À0.456
ND
À0.447
À0.479
À0.524
À0.485
À0.422
À0.355
À0.353
À0.444
ND
1
À1.199
À1.206
À9.434
À9.366
À1.006
À99.191 2.460
À99.195 2.460
À116.660 2.014
À121.723 1.262
À141.146 1.445
À128.913 1.729
À109.484 1.147
À128.872 0.895
À128.859 0.895
À121.732 0.863
À121.729 0.863
À106.178 À0.026
4
À0.338
2
3.830
5
À0.330
3
À1.477 À10.209 À13.581
6
À0.328
4
À1.534
À1.771
À1.659
À9.522 À16.951
À9.009 À12.087
À9.498 À18.283
7
À0.345
5
8
À0.320 (shb)
À0.324 (sh)
À0.320
6
9
7
À1.634 À10.339 À19.074
10
11
12
8
À1.528
À1.533
À1.795
À1.790
À9.641 À15.259
À9.643 À15.293
À9.622 À18.591
À9.609 À18.595
À0.350
9
À0.319
ND
À0.303
10
11
12
a ND: not detected.
b The first reduction peak was obtained as a shoulder at 5 mVsÀ1
À1.914 À10.157 À10.419
.

1082
J. Koyama et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1079–1082
Table 4. Correlation coefficients for MIC of azafluorenones, 1-azaanthraquinones, and 2-azaanthraquinones
Epc-1
LUMO
HOMO
Steric energy
Total energy
log P
Multiple regression with Epc-1 and log P
S. aureus
0.863 0.799
0.866 0.794
0.1420.391
0.292
0.229
0.808
0.805
0.624
0.656
0.726
0.737
0.878
0.884
B. subtilis
E. coli
0.703
0.115
0.389
0.494
0.575
P. vulgaris
P. aeruginosa
S. cerevisiae
S. pombe
0.411
0.828
0.495
0.628
0.500
0.536
0.193
0.243
0.280
0.636
0.707
0.665
0.887
0.111
0.554
0.452
0.626
0.687
0.845
0.818
0.785
0.879
0.696
0.906
0.827
0.792
0.814
0.796
0.805
C. utilis
0.7720.914
800.2
0.8620.306
0.406
0.818
0.767
0.598
0.554
0.844
0.886
R. rubra
0.730
0.251
0.898
A. niger
0.288
0.128
0.159
0.275
0.136
0.674
0.450
0.765
0.013
0.471
0.355
0.601
0.581
0.880
0.623
0.880
P. chrysogenum
R. chinensis
M. mucedo
0.622
0.514
0.855
0.868
0.703
0.838
0.7920.795
0.764 0.884
3. Tadic, D.; Cassels, B. K.; Leboeuf, M.; Cave, A. Phyto-
chemistry 1987, 26(2), 537.
4. Walton, B. T.; Ho, C.-H.; Ma, C. Y.; OÕNeill, E. G.; Kao,
G. L. Science 1983, 222, 422.
5. Clark, A. M.; Huddleston, D. L.; Ma, C. Y.; Ho, C.-H.
Pharm. Res. 1984, 1, 269.
6. Okunade, A. L.; Clark, A. M.; Hufford, C. D.; Oguntim-
ein, B. O. Planta Med. 1999, 65, 447.
7. Koyama, J.; Tagahara, K.; Osakai, T.; Tsujino, Y.;
Tsurumi, S.; Nishino, H.; Tokuda, H. Cancer Lett. 1997,
115, 179.
8. Koyama, J.; Morita, I.; Tagahara, K.; Osakai, T.; Hotta,
H.; Yang, M. X.; Mukainaka, T.; Nishino, H.; Tokuda,
H. Chem. Pharm. Bull. 2001, 49, 1214.
S. cerevisiae, A. niger, and R. chinensis. The first reduction
potential at pH 7.2may be set a parameter determining
the antimicrobial activity of the pyridine derivatives.
The more positive Epc-1, the stronger the antimicrobial
activity. Among these electronic properties, LUMO en-
ergy and steric energy demonstrated definite correla-
tions. Good correlation of log P and MIC was also
obtained. The low log P value of the compounds indi-
cates potent antimicrobial activity. Thus, we introduced
log p as an additional parameter, and performed multi-
ple regression analyses, which resulted in slight improve-
ment in correlation, as shown in Table 4.
9. Koyama, J.; Sugita, T.; Suzuta, Y.; Irie, H. Heterocycles
1879, 12, 1017.
10. Koyama, J.; Okatani, T.; Tagahara, K.; Irie, H. Hetero-
cycles 1989, 29, 1649.
11. Watanabe, M.; Snieckus, V. J. Am. Chem. Soc. 1980,
102(4), 1457.
12. Konoshima, T.; Kozuka, M.; Koyama, J.; Okatani, T.;
Tagahara, K.; Tokuda, H. J. Nat. Prod. 1989, 52(5), 987.
13. Koyama, J.; Morita, I.; Kobayashi, N.; Osakai, T.; Hotta,
H.; Takayasu, J.; Nishino, H.; Tokuda, H. Cancer Lett.
2004, 212, 1.
In conclusion, some azaanthraquinones possess potent
antimicrobial activity. Reduction potentials measured
at physiological pH 7.2, log P, and LUMO energy were
useful parameters for estimating the antimicrobial activ-
ity of azafluorenones, 1-azaanthraquinones, and 2-
azaanthraquinones.
References and notes
1. Hufford, C. D.; Liu, S.; Clark, A. M. J. Nat. Prod. 1987,
50, 961.
2. Waterman, P. G.; Muhammad, I. Phytochemistry 1985,
24, 523.
14. Koyama, J.; Morita, I.; Kobayashi, N. Unpublished work.
15. Stewart, J. J. P. J. Comput. Chem. 1989, 10, 209.
16. Ghose, A. K.; Pritchett, A.; Crippen, G. M. J. Comput.
Chem. 1988, 9, 80.