Design, synthesis, and antifungal activities in vitro of novel tetrahydroisoquinoline compounds based on the structure of lanosterol 14α-demethylase (CYP51) of fungi

Article abstract of DOI:10.1016/j.bmcl.2006.08.001

Novel tetrahydroisoquinoline compounds were designed by coupling structure-based de novo design based on the structure of lanosterol 14α-demethylase (CYP51). The chemical synthesis and the antifungal activities in vitro of them were reported. The results exhibited that all of the lead compounds showed potent antifungal activities, in which compounds 6 and 7 had equal or stronger antifungal activities against five test fungi than that of fluconazole. The studies presented here provided the antifungal lead compounds. The affinity of the lead molecules for CYP51 was mainly attributed to their non-bonding interaction with the apoprotein, which was different from the azole antifungal agents.

Full text of DOI:10.1016/j.bmcl.2006.08.001

Bioorganic & Medicinal Chemistry Letters 16 (2006) 5285–5289
Design, synthesis, and antifungal activities in vitro
of novel tetrahydroisoquinoline compounds based on the structure
of lanosterol 14a-demethylase (CYP51) of fungi
Ju Zhu, Jiaguo Lu, Youjun Zhou,* Yaowu Li, Jun Cheng and Canhui Zheng
Department of Medicinal Chemistry, School of Pharmacy, Second Military Medical University, Shanghai 200433, China
Received 27 February 2006; revised 16 July 2006; accepted 1 August 2006
Available online 14 August 2006
Abstract—Novel tetrahydroisoquinoline compounds were designed by coupling structure-based de novo design based on the struc-
ture of lanosterol 14a-demethylase (CYP51). The chemical synthesis and the antifungal activities in vitro of them were reported. The
results exhibited that all of the lead compounds showed potent antifungal activities, in which compounds 6 and 7 had equal or stron-
ger antifungal activities against five test fungi than that of fluconazole. The studies presented here provided the antifungal lead com-
pounds. The affinity of the lead molecules for CYP51 was mainly attributed to their non-bonding interaction with the apoprotein,
which was different from the azole antifungal agents.
Ó 2006 Elsevier Ltd. All rights reserved.
Lanosterol 14a-demethylase (CYP51) is one of the key
enzymes of sterol biosynthesis in fungi1 and is a prime
target for development of antifungal drugs. The devel-
opment of inhibitors of lanosterol 14a-demethylase in
fungi has provided azole antifungal drugs used in clini-
cal therapy, such as ketoconazole, fluconazole, vorico-
nazole, and itraconazole (Fig. 1).2 These azole
antifungal drugs are classified as imidazoles and tria-
zoles on the basis of whether they have two or three
nitrogens in the five-membered azole ring. Azole anti-
fungal agents inhibit the CYP51 by a mechanism in
which the heterocyclic nitrogen atom (N-3 of imidazole
and N-4 of triazole) binds to the heme iron atom in the
binding site of the enzyme. The resulting ergosterol
depletion and the accumulation of precursor 14a-meth-
ylated sterols interfere with the function of ergosterol as
a membrane component. They disrupt the structure of
the plasma membrane, making it more vulnerable to fur-
ther damage, and alter the activities of several mem-
brane-bound enzymes.3,4
antifungal agents are generally toxic and are hampered
in the treatment of deep-seated mycoses and life-threat-
ening systemic infections because of their ability to coor-
dinate with the heme of a lot of host cytochrome P450
enzymes, particularly mammalian CYP3A4.5,6 Cases of
fatal hepatotoxicity have been reported.7–12 The azole
ring has been demonstrated to be one of the most impor-
tant pharmacophores for antifungal activity, and both
toxicity and activity of azole antifungal agents are main-
ly attributed to the coordination binding of the nitrogen
atom of the azole ring to the iron atom of heme.13,14 All
of these findings urged us to discover novel non-azole
lead compounds with more structural specificity for
the fungal enzyme to separate their activity from
toxicity.
Because the crystal structure of CYP51 of fungi has not
been obtained, in previous studies, we constructed a
three-dimensional model of CYP51 from Candida albi-
cans15 and investigated the active site of CYP51, then,
we designed non-azole antifungal lead compounds based
on the constructed three-dimensional model of CYP51
by coupling structure-based de novo design. The non-
azole lead molecules containing benzopyran ring were
successfully designed and synthesized. The results of
their biological evaluation showed that the benzopyran
compounds are the novel non-azole inhibitors specific
for CYP51 of fungi.16 Although the antifungal activities
of the designed lead compounds are much lower, these
Because CYP51 is a member of the cytochrome P450
superfamily which exists in fungi and mammals, azole
Keywords: CYP51 inhibitor; Antifungal activity; Tetrahydroisoquino-
line; Design; Lanosterol 14a-demethylase (CYP51).
*
Corresponding author. Tel.: +86 021 25070383; e-mail:
zhouyoujun2006@yahoo.com.cn
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.08.001

5286
J. Zhu et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5285–5289
O
CH3
OH
O
N
N
N
CH2
O
N
N
N
CH2
CH2
N
N
CH2
O
N
N
F
Cl
F
Cl
Ketconazole
Fluconazole
F
O
OH
N
N
O
N
N
N
N
N
N
N
N
CH2
O
N
N
CH2
F
O
N
Cl
F
Cl
Itraconzole
Voriconazole
Figure 1. Azole antifungal drugs used in clinical therapy.
findings encouraged us to make further efforts to explore
more potent non-azole antifungal lead compounds spe-
cific for fungi.
The lead molecules were constructed by manually link-
ing some of the MCSS minima.17 The new bond was
constructed in such a way that there was no introduction
of significant internal strain in the candidate ligand.
During the fragment connection step, the synthetic
accessibility of the generated structures was taken into
account. The newly formed ligand molecules were subse-
quently energy-minimized in the rigid protein to regular-
ize the internal coordinates using the CVFF force field
in the Discover 95.0 program within Insight II. The
Affinity module within Insight II was then applied to de-
fine the lowest energy position for the generated mole-
cules by using a Monte Carlo and simulated annealing
combined protocol. All of the atoms within a defined
In this paper, the non-azole lead molecules were further
designed according to the previous method,16 and the
antifungal activity in vitro of the lead compounds was
evaluated.
In previous study, we found that except for the site
coordinating with the heme, the key regions in the
active site of CYP51 for ligand non-covalent binding
can be divided into four subsites by the investigation
of the three-dimensional molecular model of CYP51
from C. albicans15,16: S1–S4. The S1 subsite was the
hydrophilic hydrogen-bonding region; the S2 subsite
was the hydrophobic region; the S3 subsite was the
narrow hydrophobic cleft formed by the residues in
the helix B0-meander1 loop and the N terminus of
helix I; The S4 subsite adjacent to the b6-1/b1-4 sheet
is another important hydrogen-bonding region in the
active site.
˚
radius (5 A) of the lead molecules were allowed to move.
The solvation grid supplied with the Affinity program
was used.18 The resulting structure was accepted if it
passed the Metropolis criterion and then a check of
the rms distance of the new structure versus the struc-
ture found so far. The final conformations were ob-
tained through a simulation annealing procedure from
500 to 300 K, and then 2000 rounds of energy minimiza-
tion were performed to reach a convergence, where the
resulting interaction energy values were used to define
a rank order. Each energy-minimized final docking posi-
tion of the lead molecules was evaluated using the inter-
action score function in the LUDI module.19,20
Because the toxicity of azole antifungal agents is mainly
attributed to the coordination binding of the nitrogen
atom of the azole ring to the iron atom of heme, during
the process of inhibitor design in the present study, only
those functional groups forming non-covalent bonds
with the amino acid residues in the active site of
CYP51 were introduced, avoiding the connection with
the iron atom of heme. By connecting the highest scor-
ing minimum in each subsite, the lead structures were
designed.
Figure 2 shows some of the designed lead molecules con-
taining the tetrahydroisoquinoline ring.
At position 2 of the lead structure, long lipophilic alkyl
side chains were introduced to interact with the hydro-
HO
HO
HO
HO
HO
HO
N
N
N
(CH2)7CH3
(CH2)11CH3
(CH2)9CH3
Compound 4
Compound 6
Compound 7
Figure 2.

J. Zhu et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5285–5289
5287
phobic S3 subsite. The 6,7-phenolic hydroxyl groups
were introduced to form H bonds with the residue in
the S4 subsite. The tetrahydroisoquinoline ring was
selected to act as the skeleton to link the functional
groups in the S2 and S3 subsites, and also because of
the synthetic accessibility. After energy minimization,
the lowest energy position in the active site for the gen-
erated lead molecule was determined by the flexible li-
gand docking procedure in the Affinity module of the
Insight II program and scored by the score-3 function
in LUDI module. The calculated interaction energies
and LUDI scores of each molecule are given in Table 1.
line (III) was refluxed with different alkyl halides to form
the intermediates (IV). Cleavage of the methoxy group of
IV in HBr/CH3COOH provided target compounds (V).24
In vitro antifungal activity was measured by means of
the minimal inhibitory concentrations (MIC) using the
serial dilution method in 96-well microtest plates. Test
fungal strains were obtained from the ATCC or were
clinical isolates. The MIC determination was performed
according to the national committee for clinical labora-
tory standards (NCCLS) recommendations with RPMI
1640 (Sigma) buffered with 0.165 M Mops (Sigma) as
the test medium. The MIC value was defined as the low-
est concentration of test compounds that resulted in a
culture with turbidity less than or equal to 80% inhibi-
tion when compared with the growth of the control. Test
compounds were dissolved in DMSO serially diluted in
growth medium. The yeasts were incubated at 35 °C
and the dermatophytes at 28 °C. Growth MIC was
determined at 24 h for Candida species, at 72 h for Cryp-
tococcus neoformans, and at 7 days for filamentous fun-
gi. In vitro antifungal activity of the lead compounds
was shown in Table 1.
The mode of action of lead molecules 6 with the active site
of CYP51 of C. albicans is shown in Figure 3. After flexi-
ble ligand docking and subsequent molecular dynamics
simulated annealing studies, the 6,7-phenolic hydroxyl
groups of lead molecule 6 form H bonds with the residue
Tyr69, Ser378, Val509 of S4 subsite. At position 2 of the
lead structure, long lipophilic alkyl side chains were inter-
acted with the hydrophobic S3 subsite. The tetrahydroiso-
quinoline ring of lead molecule 6 was interacted with the
hydrophobic S2 subsite. No interactions were found be-
tween the lead molecule and the heme.
The results of in vitro antifungal activities of the novel
compounds showed that all of the lead molecules exhib-
ited potent antifungal activities against six pathogenic
fungi, and compounds 6, 7, exhibited equal or stronger
antifungal activities against five test fungi than that of
control drug fluconazole except C. albicans. The anti-
fungal activities of compounds 1, 2, 3, 4, 5, and 9 were
lower than those of compounds 6, 7, and 8, suggesting
The lead molecules and several of their derivatives were
synthesized (Scheme 1): 2-(3,4-dimethoxyphenyl)ethyla-
mine(I) was submitted to Pictet–Spengler reaction using
HCHO in acidic ethanol to yield 6,7-dimethoxy-1,2,3,4-
tetrahydroisoquinoline hydrochloride (II)21 which was
neutralized with aqueous NaOH to provide free
base(III).23 The 6,7-dimethoxy-1,2,3,4-tetrahydroquino-
Table 1. The calculated interaction energies and LUDI scores of the lead molecules
a
Compound
EVdw
EElec
ETotal
Score
4
6
7
ꢀ44.9629
ꢀ45.9949
ꢀ49.7315
ꢀ11.62
ꢀ56.5829
ꢀ61.5157
ꢀ64.8372
422
460
606
ꢀ15.5208
ꢀ15.1056
Calculated interaction energies (kcal/mol) for the complexes of the lead compounds with the active site of CYP51 of Candida albicans and Ludi
scores.
a ETotal = Evdw + Eelec
.
Figure 3. The mode of action of lead molecules 6 with the active site of CYP51 of Candida albicans.

5288
J. Zhu et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5285–5289
H3CO
H3CO
H3CO
a
b
NH2
NH HCl
H3CO
HO
HO
H3CO
H3CO
H3CO
H3CO
d
c
HBr
N
NH
N
a
R
Scheme 1. Reagents and conditions: (a) HCHO/HCl/CH3CH2OH, yield 68.7%; (b) NaOH, yield 96.8%; (c) alkyl halides/NaCO3/CH3CH2OH, reflux,
yield 71.3–90.5%; (d) HBr–CH3COOH, reflux, yield 85.3–93.2%.
that the appropriate substitution at position 2 of the
compound was important for antifungal activities. Com-
pound 9 was less potent because the alkyl side chain at
its position 2 was too long to be accommodated by the
hydrophobic S3 subsite. Compounds 1, 2, 3, 4, and 5
were less potent because their alkyl side chain was too
short to interact firmly with the hydrophobic region,
although it was able to enter the active site. The overall
trend for the interaction energy and LUDI scores was in
good qualitative agreement with the antifungal activities
in vitro. All of the results gave us the suggestion to fur-
ther design and synthesize their derivatives (Table 2).
Table 2. In vitro antifungal activity of the lead compounds
HO
N
HO
R
MIC80 (ug/ml)
Compound
R
C. alb.
C. par.
C. neo.
A. fum.
M. can.
T. rub.
1
2
–(CH2)4CH3
–(CH2)5CH3
–(CH2)6CH3
–(CH2)7CH3
–(CH2)8CH3
–(CH2)9CH3
–(CH2)11CH3
–(CH2)13CH3
–(CH2)15CH3
Fluconazole
>200
>200
>200
>200
200
100
100
100
100
25
100
100
100
100
25
>200
>200
>200
>200
200
>200
>200
>200
50
>200
>200
>200
50
3
4
5
50
12.5
6
100
50
12.5
50
3.125
50
50
12.5
12.5
12.5
>200
100
3.125
7
6.25
6.25
12.5
25
8
50
>200
1.562
50
>200
200
>200
9
10
100
50
100
6.25
200
25
Abbreviations: C. alb., Candida albicans; C. par., Candida parapsilosis; C. neo., Cryptococcus neoformans; A. fum., Aspergillus fumigatus; T. rub.,
Trichophyton rubrum; M. can., Microsporum canis.
Figure 4. The mode of action of fluconazole with the active site of CYP51 of Candida albicans.

J. Zhu et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5285–5289
5289
9. Su, F. W.; Perumalswami, P.; Grammer, L. C. Allergy
2003, 58, 1215.
10. (a) Wolf, R.; Wolf, D.; Kuperman, S. J. Clin. Gastroen-
terol. 2001, 33, 418; (b) Jeng, M. R.; Feusner, J. Pediatr.
Hematol. Oncol. 2001, 18, 137.
11. Kauffman, C. A.; Hedderwick, S. A. Drugs Aging 2001,
18, 313.
12. Linnebur, S. A.; Parnes, B. L. Ann. Pharmacother. 2004,
38, 612.
13. Plempel, M. Shinkin to shinkinsho 1982, 23, 17.
14. Fromtling, R. A. Clin. Microbiol. Rev. 1988, 1, 187.
15. Ji, H.; Zhang, W.; Zhou, Y.; Zhang, M.; Zhu, J.; Song, Y.;
Lu, J.; Zhu, J. J. Med. Chem. 2000, 43, 2493.
16. Ji, H.; Zhang, W.; Zhang, M.; Kudo, M.; Aoyama, Y.;
Yoshida, Y.; Sheng, C.; Song, Y.; Yang, S.; Zhou, Y.; Lu,
J.; Zhu, J. J. Med. Chem. 2003, 46, 474.
17. Joseph-McCarthy, D.; Tsang, S. K.; Filman, D. J.; Hogle,
J. M.; Karplus, M. J. Am. Chem. Soc. 2001, 123, 12758.
18. Luty, B. A.; Wasserman, Z. R.; Stouten, P. F. W.; Hodge,
C. N.; Zacharias, M.; McCammon, J. A. J. Comput.
Chem. 1995, 16, 454.
In summary, the novel tetrahydroisoquinolines based on
lanosterol 14a-demethylase of fungi were discovered by
coupling structure-based de novo design with chemical
synthesis, and their antifungal activities in vitro were
evaluated. The mode of action of the lead molecules
was represented, which was different from that of azoles
(Fig. 4). The lead compounds exhibited potent anti-
fungal activities in vitro. Because the affinity of the lead
molecules for CYP51 was mainly attributed to their
non-bonding interaction with the apoprotein, the studies
presented here afford the opportunity to develop novel
antifungal agents that specifically interact with the resi-
dues in the active site and avoid the serious toxicity aris-
ing from coordination binding with the heme of
mammalian P450s.
Acknowledgment
19. Bohm, H.-J. J. Comput.-Aided Mol. Des. 1994, 8, 243.
20. Bohm, H.-J. J. Comput.-Aided Mol. Des. 1998, 12, 309.
21. Mp 252–253 °C (lit.22 251–252 °C); 1HNMR (DMSO,
300 MHz): 3.04 (t, 2H, J = 6.2 Hz), 3.49 (t, 2H,
J = 6.3 Hz), 3.81 (s, 3H), 3.82 (s, 3H), 4.28 (s, 2H), 6.81
(s, 1H), 6.87 (s, 1H).
The work was supported by the National Natural
Science Foundation of China (Grant No. 30572257).
References and notes
22. Bobbitt, J. M.; Kiely, J. M.; Khanna, K. L.; Ebermann, R.
J. Org. Chem. 1965, 30, 2247.
1. Vanden Bossche, H.; Koymans, L. Mycoses 1998, 41, 32.
2. (a) Kale, P.; Johnson, L. B. Drug Today (Barc) 2005, 41,
91; (b) Sheehan, D. J.; Hitchcock, C. A.; Sibley, C. M.
Clin. Microbiol. Rev. 1999, 12, 40.
3. Georgopapadakou, N. H.; Walsh, T. J. Antimicrob.
Agents Chemother. 1996, 40, 279.
23. Mp 82–84 °C (lit.22 84–85 °C).
1
24. Compound 4ÆHBr, HNMR (DMSO, 500 MHz): 0.85 (t,
J = 6.7 Hz, 3H), 1.27–1.79 (m, 12H), 2.95–2.98 (m, 2H),
3.23 (m, 2H), 3.29–3.35 (t, 2H), 4.11–4.4 (dd,
JA = 133 Hz, JB = 15.1 Hz, 2H), 6.68 (s, 1H), 6.75 (s,
1H), 9.61 (s, 1H); compound 6ÆHBr (DMSO, 300 MHz)
0.86 (t, J = 6.7 Hz, 3H), 1.26 (m, 14H), 1.73 (s, 2H),
2.73–2.83 (m, 2H), 2.96–3.58 (m, 4H), 4.02–4.33 (dd,
JA = 75.2 Hz, JB = 13.6 Hz, 2H), 6.55 (d, J = 9.4 Hz,
2H), 9.05 (d, J = 13.2 Hz, 2H), 10.05 (s, 1H); compound
7ÆHBr (DMSO, 500 MHz) 0.86 (t, J = 6.8 Hz, 3H), 1.25–
1.73 (m, 20H), 2.80 (m, 2H), 2.96 (m, 2H), 3.15–3.65 (m,
2H), 4.09–4.36 (dd, JA = 117.8 Hz, JB = 14.6 Hz, 2H),
6.57 (d, J = 13.6 Hz, 2H), 9.04 (d, J = 16.5 Hz, 2H), 9.78
(s, 1H).
4. Lupetti, A.; Danesi, R.; Campa, M.; Tacca, M. D.; Kelly,
S. Trends Mol. Med. 2002, 8, 76.
5. Slama, J. T.; Hancock, J. L.; Rho, T.; Sambucetti, L.;
Bachmann, K. A. Biochem. Pharmacol. 1998, 55, 1881.
6. Wulff, H.; Miller, M. J.; Hansel, W.; Grissmer, S.;
Cahalan, M. D.; Chandy, K. G. Proc. Natl. Acad. Sci.
U.S.A. 2000, 97, 8151.
7. Legra, S. A.; Bergemer-Fouguet, A. M.; Jonville-Bera, A.
P. Am. J. Med. 2002, 113, 352.
8. Bok, R. A.; Small, E. J. Drug Safety 1999, 20, 451.