Synthesis and chain-dependent antifungal activity of long-chain 2H-azirine-carboxylate esters related to dysidazirine

Article abstract of DOI:10.1016/j.bmcl.2010.01.068

Analogues of the antifungal marine natural product (E)-dysidazirine were prepared and evaluated in broth ro-dilution assays against a panel of fungal pathogens. A simple structure-activity relationship was developed which provides insight into the mechanism of action of long-chain 2H-azirine carboxylates.

Full text of DOI:10.1016/j.bmcl.2010.01.068

Bioorganic & Medicinal Chemistry Letters 20 (2010) 2029–2032
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Synthesis and chain-dependent antifungal activity of long-chain
2H-azirine-carboxylate esters related to dysidazirine
Colin K. Skepper a, Doralyn S. Dalisay a, , Tadeusz F. Molinski a,b,
*
a Department of Chemistry and Biochemistry, San Diego, 9500 Gilman Drive, La Jolla, CA 92093, USA
b Skaggs School of Pharmacy and Pharmaceutical Sciences University of California, San Diego, 9500 Gilman Drive, La Jolla, CA 92093, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Analogues of the antifungal marine natural product (E)-dysidazirine were prepared and evaluated in
broth ro-dilution assays against a panel of fungal pathogens. A simple structure–activity relationship
was developed which provides insight into the mechanism of action of long-chain 2H-azirine
carboxylates.
Received 18 December 2009
Revised 7 January 2010
Accepted 11 January 2010
Available online 1 February 2010
Ó 2010 Elsevier Ltd. All rights reserved.
Keywords:
Antifungal
Oxazole
Azole
Marine natural product
2H-Azirines are highly strained, weakly aromatic heterocycles
rarely found in Nature. In fact, only nine natural products contain-
ing the 2H-azirine ring have been described to date. Azirinomycin1
(1) was isolated from Streptomyces aureus while the other
compounds constitute a small family of long-chain 2H-azirine car-
boxylates from marine sponges. The first of this family, (E)-dysid-
azirine (2), was isolated in 1988 from Dysidea fragilis, collected in
Fiji, and shown to exhibit potent antifungal activity against Can-
including C. albicans (ATCC, UCD-FR1, and 96-489), less susceptible,
non-albicans species such as Candida glabrata and Candida krusei,
and two strains of Cryptococcus neoformans, var. grubii and gatti
(Fig. 1).
In order to elucidate the relationship between structure and
antifungal activity of long-chain 2H-azirine carboxylates we car-
ried out the first synthesis of (Z)-dysidazirine8 and a short series
of structural analogues ((À)-9?13, Fig. 2), along with (E)-dysidaz-
irine. We report here the structure-activity relationships for (À)-2,
(À)-3, ( )-3, and (À)-9?13 that reveal structural determinants for
antifungal activity, including terminal branching [(À)-9], chain
length [(À)-10 and (À)-11], C4–C5 unsaturation [(À)-12, (À)-13]
and C2 configuration [( )-3].
dida albicans and Saccharomyces cerevisiae (4
assay).2 Faulkner and co-workers later characterized (Z)-dysidazi-
rine (3) and two -dibromovinylidene azirine-2-carboxylate
lg/disk, disk diffusion
x
methyl esters, 4 and 5 (E- and Z-antazirines) from D. fragilis and re-
ported them as inactive ‘against a standard panel of microorgan-
isms.3 Motualevic acid F, the parent carboxylic acid of ent-4, was
reported by Bewley and co-workers along with related amides
from Siliquariaspongia sp. with antibacterial activity against Staph-
ylococcus aureus and MRSA.4 In our search5 for antifungal com-
pounds from marine organisms with efficacy against
Fluconazole-resistant strains of Candida and other pathogenic
yeast,6 we found three new antazirines (6–8) along with 4 and 5
from a sample of D. fragilis collected in Pohnpei, Micronesia that
showed cytotoxicity in a crude screen.7 Compounds 4–8 were
moderately cytotoxic (HCT-116 cells), but to our surprise they
were completely inactive against a panel of fungal pathogens
CO2Me
CO2Me
N
N
N
CO2H
1 azirinomycin
2 (−)-(E)-dysidazirine
3 (Z)-dysidazirine
CO2Me
CO2Me
N
N
Y
Y
X
X
4 X = Br, Y = Br (+)-(E)-antazirine
6 X = Br, Y = Cl
7 X = Cl, Y = Cl
5 X = Br, Y = Br (+)-(Z)-antazirine
8 X = Br, Y = Cl
* Corresponding author. Tel.: +1 858 534 7115; fax: +1 858 822 0386.
E-mail address: tmolinski@ucsd.edu (T.F. Molinski).
Present address: Department of Chemistry, University of British Columbia,
Vancouver, BC, Canada V6T 1Z1.
Figure 1. Naturally occurring 2H-azirines.
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.01.068

2030
C. K. Skepper et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2029–2032
CO2Me
CO2Me
CO2Me
Synthesis of (À)-13, the 4,5-dihydro analogue of 3, began with
generation of the enolate of dioxolane 19 (LDA, HMPA, À40 °C) fol-
lowed by alkylation with 1-bromotetradecane (20) to provide 21 in
low yield (18%) along with an equivalent amount of the product
N
N
N
tBu
(−)-9
(−)-10
(−)-11
from a-alkylation (Scheme 3). Thermolysis of 21 (microwave) gave
CO2Me
CO2Me
the incipient ketene that was captured with methanol to afford O-
methyl b-ketoester 22, and subsequently converted to oxime tosyl-
ate 23 in two steps as before. Treatment of 23 with quinidine led
directly to (À)-13 in 91% yield.
N
N
(−)-13
(−)-12
The antifungal MICs of the synthetic and natural azirines are
compared in Figure 3.10 Each yeast strain was susceptible to (–)-
3 and some synthetic analogues with the notable exception of
Fluconazole-resistant C. krusei. The other Fluconazole-resistant
strains (C. glabrata, C. albicans UCD-FR1 and 96-489) were all sus-
ceptible. Strains of C. neoformans (var. grubii and var. gatti) were
the most susceptible overall.
Figure 2. Targeted synthetic dysidazirine analogues.
The synthesis of (À)-(Z)-dysidazirine and analogues (À)-9 and
(À)-12 has been described previously.8 Photochemical isomeriza-
tion (500 W sunlamp, Pyrex) of synthetic (À)-3 (59% ee) provided
(À)-2 (Scheme 1) in low yield but without epimerization at C2.
Natural dysidazirine and congeneric compounds have all been iso-
lated as non-racemic mixtures of enantiomers. Neat, natural dysid-
azirine spontaneously epimerizes slowly in the dark (t1/2 ꢀ12y,
À20 °C),7 and the lack of racemization of 2 and 3 in the presence
of light excludes a Photochemical mechanism.
The shorter chain analogues (À)-10 and (À)-11 were prepared
in an analogous fashion to (À)-3 (Scheme 2). Addition of the
lithio-acetylide derived from alkyne 14 to methyl malonyl chloride
gave -ketoester 15 in reasonable yield (55%). Treatment of 15 with
NH2OHÁHCl/pyridine led to the corresponding oxime which was
converted to oxime tosylate 16 without purification. Cyclization
in the presence of quinidine under Zwanenburg’s conditions9 pro-
vided 2H-azirine 17 in excellent yield. Partial hydrogenation with
Lindlar’s catalyst provided the truncated dysidazirine analogue
(À)-11. The same sequence applied to alkyne 18 gave analogue
(À)-10.
Both (À)-2 and (À)-3 were consistently among the most active
compounds against all strains, with MIC values in the range of 2–
8 lg/mL. For Candida spp., (–)-3 was notably more active than (–
)-2 (Fig. 3a). Shorter chain analogues (À)-10 and (À)-11 were com-
parable in activity to (–)-3, although Fluconazole-resistant C. albi-
cans UCD-FR1 was susceptible to (À)-10, but not to (À)-11.
Alkynyl analogue (À)-12 retains most of the activity of the nat-
ural products (–)-2 and (–)-3, however the 4,5-dihydro analogue
(À)-13 lost activity across the entire panel of fungal cells. The C2
configuration appears to be less important. Race Z-dysidazirine
[( )-3] shows good activity against all strains, comparable to or
slightly better than the optically enriched homologue (À)-3.
As previously observed,8 the tert-butyl terminus analogue (À)-9
is essentially inactive. Clearly terminal substitution or branching
abrogates antifungal activity, consistent with the lack of activity
for
x
-halo-alkenyl antazirines (4–8).3,7
The foregoing results suggest specific structural attributes
determine antifungal activity rather than non-specific toxicity.
Our original hypothesis regarding a putative sub-cellular target
of dysidazirine, based on data obtained from (À)-3 and (À)-9, is
supported by the wider range of analogues presented here.8 We fa-
vor a model for activity of long-chain azirines involving interaction
with a putative protein target through a 2-point binding motif and
propose the following. We envision the lipid chain occupies a
tightly constrained hydrophobic pocket that is intolerant of bulky
CO2Me
CO2Me
I2, hυ, CH2Cl2
0 °C → rt, 17 h
N
N
25%
(−)-3
(−)-2
ee = 59%
ee = 59%
Scheme 1. Photochemical isomerization of (À)-(Z)-dysidazirine to (À)-(E)-
dysidazirine.
x
-substitution, which helps to explain the loss of activity seen
for 4–8 and (À)-9 despite log Ps and molecular size that are com-
parable to (À)-3. The polar, electrophilic azirine terminus—a potent
Michael acceptor capable of covalent modification of nucleophilic
sites (e.g., cysteine, lysine)—binds at a distal site, possibly interact-
ing with one or more nucleophilic amino acid residues. The
absolute requirement of C4–C5 unsaturation for high antifungal
i. n-BuLi, THF, 0 °C, 1.5 h
O
O
O
O
ii.
OMe
3
14
Cl
OMe
15
3
THF, −78 → 0 °C, 3 h, 55%
i. NH2OH.HCl, pyr,
EtOH, 55 °C, 1 h
TsO
i. quinidine, toluene,
0 °C, 42 h
N
O
i. LDA, HMPA (5 eq)
MeOH, µwave,
O
O
THF, −40 °C 20 min
O
O
120 °C, 20 min
OMe
ii. (Ts)2O, pyr, CH 2Cl2,
DMAP, r.t., 1.5 h
62 %
ii. 5 °C, 5.25 h
iii. r.t., 30 min
91 %
O
16
12
ii. n-C14H29Br (20 )
−40 °C → r.t.
18%
3
O
70 %
21
19
CO2Me
CO2Me
H2
i. NH2OH•HCl, pyr, EtOH
50 °C, 45 min
Lindlar's catalyst
TsO
N
O
O
N
O
N
(−)-11
hexane,
0 °C, 15 min
61 %
3
17
ii. (p-MeC6H4SO2)2O
pyr, DMAP,
CH2Cl2, rt, 1 h
82 %
OMe
OMe
12
12
ee = 61%
22
23
quinidine,
toluene
0 °C, 24 h
CO2Me
CO2Me
N
N
(−)-10
ee = 86%
8
91%
18
(−)-13
ee = 71 %
Scheme 2. Synthesis of truncated analogues (À)-10 and (À)-11.
Scheme 3. Synthesis of 4,5-dihydrodysidazirine [(À)-13].

C. K. Skepper et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2029–2032
2031
NH2 OH
16
14
12
10
8
a
ATCC
UCD-FR1
96-489
OH OH
NH2
24 D-phytosphingosine
OH OH
25 D-sphingosine
NH2
26 Crucigasterin 277
27 Crucigasterin 275 (ΔC17-C18)
OH
OH
H2N
6
28 ΔC3-C4
29 ΔC4-C5
30 ΔC5-C6
NH2
O
OH
HO
4
11
OH
O
NH2
HO
HO
O
2
31 Oceanapiside
OH
0
Figure 4. Structures of sphingoid bases and antifungal long-chain aminoalcohol
marine natural products.
(–)-3 ( )-3 (–)-2 (–)-9 (–)-10 (–)-11 (–)-12 (–)-13
Compound
20
15
10
5
The structure of dysidazirine bears a resemblance to phytosphin-
gosine (24) and sphingosine (25) (Fig. 4)—both 24 and 25 share the
same carbon chain length (C18) as dysidazirine. The long-chain base
24 occurs in plant and fungal cells and acts in multiple capacities,
both as a cellular structural component and a signaling molecule.11
Sphingosine (25) is the primary long-chain base of mammalian
cells; it also plays roles in cell structure and signaling, and exhibits
b
C. glabrata
C. krusei
modest antifungal activity (30 l
g/mL, C. glabrata).12
Several long-chain vicinal aminoalcohol marine natural prod-
ucts (Fig. 4) also exhibit antifungal activity,5 including crucigaste-
rins 275 (26) and 277 (27),13 (R)-1-aminotridecen-2-ol14 (28–30),
and oceanapiside (31),15 an unusual ‘two-headed’ sphingoid base.16
Natural products dysidazirine (À)-3, other long-chain azirines
and 31 resemble sphingosine or its immediate precursor sphinga-
nine (4,5-dihydrosphingosine) and may act as antimetabolites that
competitively inhibit sphingolipid metabolism.
The biosynthesis of 24 and 25 in yeast follows the canonical
sphingolipid pathway that begins with pyridoxal phosphate-
dependent condensation of palmitoyl CoA thioester with serine.
Long-chain azirines may compete for natural substrate in the con-
densation reaction or downstream reactions that tailor 24 and 25.
It is notable that the antifungal properties of 3 and other sphingoid
base analogues are not solely determined by the presence of a long
hydrophobic tail since simple lipids are inactive and other anti-
fungal long-chain fatty acid conjugates appear to act by distinctly
different mechanisms.17
In summary, we have completed the synthesis of a series of
dysidazirine analogues and evaluated each for antifungal activity
against a panel of clinically-relevant yeasts. Several generalizations
can be made: (1) C4–C5 unsaturation is required for high activity.
(2) Activity is not strongly dependent on minor changes in chain
length or the configuration at C2. (3) Terminal branching of the li-
pid chain abolishes activity. Further investigations are underway in
our laboratories to identify high-affinity binding targets of (À)-3.
0
(–)-3 ( )-3 (–)-2 (–)-9 (–)-10 (–)-11 (–)-12 (–)-13
Compound
16
14
12
10
8
c
Crypto. var. grubii
Crypto. var. gatti
6
4
2
0
Acknowledgments
(–)-3 ( )-3 (–)-2 (–)-9 (–)-10 (–)-11 (–)-12 (–)-13
Compound
The NSF CRIF program (CHE0741968) is acknowledged for
funding the purchase of the 500 MHz NMR spectrometers. We
are grateful to the NIH for generous funding of this research (AI
039987).
Figure 3. Antifungal activity of (À)-2, (À)-3, ( )-3, and (À)-9?13 against: (a)
Candida albicans, three strains: ATCC 14503, UCD-FR1, and 96-489 (see Ref. 18); (b)
Candida glabrata and Candida krusei; (c) Cryptococcus neoformans var. grubii and
gatti. Compounds with off-scale MICs were inactive up to 64 lg/mL.
References and notes
activity supports the latter hypothesis. (Z)-Dysidazirine [(À)-3)],
which is expected to be a superior Michael acceptor to (À)-2,
shows enhanced antifungal activity against C. albicans strains.
1. Miller, T. W.; Tristram, E. W.; Wolf, F. J. J. Antibiot. 1971, 24, 48.
2. Molinski, T. F.; Ireland, C. M. J. Org. Chem. 1988, 53, 2103.
3. Salomon, C. E.; Williams, D. H.; Faulkner, D. J. J. Nat. Prod. 1995, 58, 1463.

2032
C. K. Skepper et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2029–2032
4. Keffer, J. L.; Plaza, A.; Bewley, C. A. Org. Lett. 2009, 11, 1087.
5. Molinski, T. F. Curr. Med. Chem.: Anti-Infect. Agents 2004, 3, 197.
6. (a) White, T. C.; Holleman, S.; Dy, F.; Mirels, L. F.; Stevens, D. A. Antimicrob.
Agents Chemother. 2002, 46, 1704; (b) Lyons, C. N.; White, T. C. Antimicrob.
Agents Chemother. 2000, 44, 2296; (c) White, T. C.; Marr, K. A.; Bowden, R. A.
Clin. Microbiol. Rev. 1998, 11, 382; (d) Collin, B.; Clancy, C. J.; Nguyen, M. H. Drug
Resist. Updates 1999, 2, 9; (e) Charlier, C.; Hart, E.; Lefort, A.; Ribaud, P.; Dromer,
F.; Denning, D. W.; Lortholary, O. J. Antimicrob. Chemother. 2006, 57, 384.
7. Skepper, C. K.; Molinski, T. F. J. Org. Chem. 2008, 73, 2592.
8. (a) Skepper, C. K.; Dalisay, D. S.; Molinski, T. F. Org. Lett. 2008, 10, 5269; (b)
Skepper, C. K., Ph.D. Thesis, University of California at San Diego, December
2009.
9. Verstappen, M. M. H.; Ariaans, G. J. A.; Zwanenburg, B. J. Am. Chem. Soc. 1996,
118, 8491.
11. (a) Sims, K. J.; Spassieva, S. D.; Voit, E. O.; Obeid, L. M. Biochem. Cell Biol. 2004,
82, 45; (b) Dickson, R. C.; Lester, R. L. In Sphingolipid Biology; Hirabayashi, Y.,
Igarashi, Y., Merrill, A. H., Jr., Eds.; Springer: New York, 2006; p 141. Part 2,
Chapter 9; (c) Funato, K.; Vallée, B.; Riezman, H. Biochemistry 2002, 41, 15105.
12. Nicholas, G. N.; Li, R.; MacMillan, J. B.; Molinski, T. F. Bioorg. Med. Chem. Lett.
2002, 12, 2159.
13. Jares-Erijman, E. A.; Bapat, C. P.; Lithgow-Bertelloni, A.; Rinehart, K. L.; Sakai, R.
J. Org. Chem. 1993, 58, 5732.
14. Searle, P. A.; Molinski, T. F. J. Org. Chem. 1993, 58, 7578.
15. Nicholas, G. M.; Hong, T. W.; Molinski, T. F.; Lerch, M. L.; Cancilla, M. T.;
Lebrilla, C. B. J. Nat. Prod. 1999, 62, 1678.
16. For other examples, see: Makarieva, T. N.; Zakharenko, A. M.; Dmitrenok, P. S.;
Guzii, A. G.; Denisenko; Alexandra, V. A.; Savina, A. S.; Dalisay, D. S.; Molinski,
T. F.; Stonik, V. A. Lipids 2009, 44, 1155 and references cited within.
17. Mulder, R. J.; Shafer, C. M.; Dalisay, D. S.; Molinski, T. F. Bioorg. Med. Chem. Lett.
2009, 19, 2928.
10. The in vitro susceptibilities were determined by the broth micro-dilution
method according to the guidelines of the NCCLS: National Committee for
Clinical Laboratory Standards. 2002. Reference method for broth dilution
antifungal susceptibility testing of yeast, 2nd ed. Approved standard M27-A2.
National Committee for Clinical Laboratory Standards, Wayne, PA.
18. Molinski, T. F. J. Nat. Prod. 1993, 56, 1.