Halisphingosines A and B, modified sphingoid bases from Haliclona tubifera. Assignment of configuration by circular dichroism and van't Hoff's principle of optical superposition

Article abstract of DOI:10.1021/np300744y

Halisphingosines A (1) and B (2), modified long-chain sphingoid bases, from the marine sponge Haliclona tubifera collected in Brazil, were characterized after conversion to their N-Boc derivatives. The 2R,3R,6R configuration of halisphingosine A, a compound first reported from Haliclona sp. from South Korea, was confirmed using a novel CD approach: deconvolution of exciton coupling from mono- and trinaphthoyl derivatives obtained by derivatization of the natural product. The sensitive CD deconvolution method, applicable to submilligram samples, simultaneously predicted the relative and absolute configuration of three stereocenters in halisphingosine A with precision and accuracy. Halisphingosine B was assigned by correlation to halisphingosine A.

Full text of DOI:10.1021/np300744y

Article
pubs.acs.org/jnp
Halisphingosines A and B, Modified Sphingoid Bases from Haliclona
tubifera. Assignment of Configuration by Circular Dichroism and
van’t Hoff’s Principle of Optical Superposition
Tadeusz F. Molinski,*,†,‡ Renata Biegelmeyer,§ E. Paige Stout,† Xiao Wang,† Mario L. C. Frota, Jr.,⊥
and Amelia T. Henriques§
†Department of Chemistry and Biochemistry and ‡Skaggs School of Pharmacy and Pharmaceutical Sciences, University of California,
San Diego, 9500 Gilman Drive, La Jolla, California 92093-0358, United States
⊥
§Pharmacognosy Laboratory and Department of Biochemistry, Federal University of South Brazil, Porto Alegre, 90610-000, RS,
Brazil
S
* Supporting Information
ABSTRACT: Halisphingosines A (1) and B (2), modified long-chain
sphingoid bases, from the marine sponge Haliclona tubifera collected in
Brazil, were characterized after conversion to their N-Boc derivatives.
The 2R,3R,6R configuration of halisphingosine A, a compound first
reported from Haliclona sp. from South Korea, was confirmed using a
novel CD approach: deconvolution of exciton coupling from mono- and
trinaphthoyl derivatives obtained by derivatization of the natural
product. The sensitive CD deconvolution method, applicable to
submilligram samples, simultaneously predicted the relative and
absolute configuration of three stereocenters in halisphingosine A
with precision and accuracy. Halisphingosine B was assigned by
correlation to halisphingosine A.
iverse variants of the C18 long-chain bases sphingosine
Dand phytosphingosine and the corresponding ceramides
occur throughout the five kingdoms1 including marine phyla2
such as sponges, tunicates, and macroalgae. Modifications
include variable chain lengths, chain branching by methyl
groups, hydroxylation, and polar headgroup modifications.
Linear and branched aminoalkanols from sponges and tunicates
exhibit substantial antifungal activity. “Two-headed’” sphingo-
lipids, constituting C28 or C30 chains functionalized at both
termini as aminoalkanols or aminoalkanediols with four
stereocenters,3 show activity against the fungal pathogens
Candida albicans, C. glabrata, and other species. Heterocyclic
analogues derived from linear sphingoid bases have also been
characterized from sponges and tunicates; these include
azirines4 and piperidines (e.g., pseudodistomins A and B,
from Pseudodistoma sp.5). Other heterocyclic sphingoids
reported from marine invertebrates are the so-called “anhy-
drophytosphingosines” including azetidines (penaresidines A
and B from Penares sp.6) and tetrahydrofurans (pachastriss-
amine from Pachastrissa sp.7synonymous with jaspine B8
and penasins A−E from Penares sp.9).
method for assignment of absolute configuration of sphingoid
bases, the method suffers from the usual limitations: require-
ment of sufficient sample for preparation of S- and R-MTPA
derivatives and NMR analysis and equivocal interpretations for
molecules with multiple contiguous −NH2 and −OH groups.
Other approaches based on degradation have similar draw-
backs, and erroneous assignments have been made in the
past.10 Here, we report the N-Boc derivative 1a of hali-
sphingosine A (1), a C18 sphingoid base first isolated from a
Haliclona sp. collected off Ulleung Island, South Korea,11 and
reisolated in this work from a Brazilian marine sponge,
Haliclona tubifera, collected off Ilha do Arvoredo, Santa
Catarina.12 A new derivative, the N-Boc analogue 2a of
halisphingosine B (2), is also described and characterized by
NMR and MS. Unlike common marine-derived sphingolipids
related to D-ribo-phytosphingosine (i), the new compounds are
L-threo. The configuration of 1a was verfied at the nanomole
scale as 2R,3R,6R [L-threo at the headgroup], by interpretation
of the CD spectra of mono- and tetranaphthoyl derivatives of 1:
specifically, deconvolution of the Cotton effect (CE) arising
from different molecular exciton couplings through application
of van’t Hoff’s principle of optical superposition.13
Although mammalian sphingosines are almost always of the
2S,3R configuration (D-erythro), invertebrate-derived sphingoid
bases exhibit broad stereochemical heterogeneity. Examples of
α,ω-bifunctionalized sphingolipids are known with almost all
possible permutations of the four stereocenters (N = 16).3
While the modified Mosher’s method has been the applied
Special Issue: Special Issue in Honor of Lester A. Mitscher
Received: October 25, 2012
© XXXX American Chemical Society and
American Society of Pharmacognosy
A
dx.doi.org/10.1021/np300744y | J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
turn was coupled to a hydroxymethine signal, H-3 (δ 3.75, m).
The C-1−C-3 segment was separated from the C-6 allylic CH-
OH by two CH2 groups, and the remainder of the molecule
was a linear n-alkyl chain. Taken all together, the constitution of
1a was identical to that published,11 except for the presence of
the N-Boc group.
The challenges of assignment of multiple stereocenters in
acyclic natural products can be formidable.14 The original
configurational analysis of 1 was carried out by comparison of
1H NMR spectra of the tetra-Mosher’s ester derivative of 111
with those of the corresponding four synthetic stereoisomers of
sphinganine.15 Although strictly not an accurate comparison (1
contains an additional allylic OH at C-6), the conclusion of the
analysis suggested a likely match for the (2R,3R,6R) isomer. In
order to verify the configuration of 1 and refine methods for
sub-milligram stereoanalysis of complex sphingolipids, we
adapated a variant of Nakanishi and Berova’s method16 for
phytosphingosines based on circular dichroism (CD) of
naphthimide−naphthoate derivatives obtained by two-step
sequential derivatization. Nakanishi’s method introduces a C-
2 naphthimide group selectively by conversion of the primary
NH2 in the first step by condensation with naphthalene-2,3-
dicarboxylic acid, followed by per-O-naphthoylation of the
remaining OH groups. Degeneracy is avoided in the super-
position of pairwise exciton couplings because of the different
UV λmax of naphthimide−naphthoate chromophore pairs. This
choice provides “fingerprint” CEs that sufficiently discriminate
all possible diastereomers and their corresponding enantiomers.
We favored, instead, a single-step exhaustive naphthoylation of
1a that we expected would provide the same discrimination of
erythro and threo diastereomers by pairwise interactions of
naphthamide−naphthoate pairs, but overlaid with an additional
CE from the distal allylic 2-naphthoate at C-6. Application of
the latter CE for assignment of allylic alcoholsarising from
helicity of the exciton coupled 1Lb transition of the 2-
naphthoate chromophore with the ene π−π* transition (Figure
2)was first introduced by Nakanishi and co-workers17 for
Figure 1. Sphingoid bases from Haliclona tubifera and erythro-D-
ribosphingosine.
RESULTS AND DISCUSSION
■
Ethyl acetate-soluble extracts of the sponge H. tubifera were
fractionated by flash chromatography (NH3-saturated MeOH/
CH2Cl2) to obtain a polar ninhydrin-positive fraction
containing a mixture of long-chain bases. In order to simplify
the purification, the long-chain base mixture was converted to
the corresponding N-Boc derivatives (di-tert-butylcarbonate,
NaHCO3, THF(aq)) and then separated by multiple rounds of
reversed-phase HPLC (C18-bonded silica), leading to pure N-
Boc derivatives 1a and 2a. Characterization of 1a by MS and
NMR (Table 1) demonstrated that it was the N-Boc derivative
of 1 (2-aminooctadec-7-ene-1,3,6-triol), a compound first
reported by Mansoor and co-workers in 2007.11 Analysis of
the COSY spectrum (CD3OD, 500 MHz) assigned downfield
signals to a secondary allylic alcohol (H-6, δ 4.37, dt, J = 9.0, 5.7
Hz; H-7, δ 5.31, dddd, J = 9.3, 7.8 Hz, 1.5, 1.5; H-8, δ 5.45, dt, J
= 9.3, 6.4 Hz) and a contiguous spin system associated with the
terminal headgroup, CH2OH (H-1a, δ 3.55, m; H1b, δ 3.60,
m), adjacent to C-2 CH-NH-(t-Boc) (H-2, δ 3.54, m), which in
Table 1. H and 13C NMR Data of Halisphingosine A (1a, CD3OD, 500 and 125 MHz) and Per-acetyl Halisphingosine B (8,
1
CDCl3, 600 MHz)
1a
8
b
a
b
a
position
δC, type
δH (J in Hz)
3.55 m, 3.60 m
δC, type
δH (J in Hz)
1
2
62.8, CH2
56.5, CH
63.1, CH2
49.8, CH
4.03 ddd (15.9, 11.2, 6.2)
4.41 m
3.54 m
2-NH
5.63 br t (5.8)
5.06 m
3
70.7, CH
34.7, CH2
38.5, CH2
67.9, CH
133.7, CH
131.8, CH
28.4, CH2
3.75 m
71.9, CH
30.5, CH2
34.0, CH2
73.8, CH
30.5, CH2
4
1.45 m
1.54 m
5
1.38 m, 1.56 m
4.37 dt (9.0, 5.7)
1.50 m
6
4.84 p (12.5, 6.0)
1.54 m
7
5.31 dddd (9.3, 7.8, 1.5, 1.5)
5.45 dt (9.3, 6.4)
2.09 m
8
22.6−31.5, CH2
1.20−1.32 m
9
10−17
18
23.5−32.5, CH2
14.1, CH3
1.29−1.41 m
0.91 t (6.7)
14.2, CH3
0.88 t (6.7)
N-Boc CH3
28.5, CH3
1.45 s
c
Ac
23.0, CH3
21.0, CH3
21.0, CH3
23.0, CH3
2.08 s
2.06 s
2.04 s
2.03 s
c
Ac
c
Ac
c
Ac
a
b
br = broad; s = singlet; d = doublet; dt = doublet of triplets; t = triplet; m = multiplet; p = pentet. Multiplicities of 13C NMR signals established by
c
edited HSQC and DEPT. Acetyl H and 13C signals cross-correlated by HSQC but not assigned.
1
B
dx.doi.org/10.1021/np300744y | J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
stereoclusters. In other words, the ensemble should be
interpretable as the superposition of each set of independent
exciton coupled systems according to van’t Hoff’s principle of
optical superposition.13 This was shown to be the case, and the
configuration of 1 was verified as follows.
The N-Boc derivative 1a (ca. 1 mg) was converted to the
acetonide 3 (Scheme 1) under standard conditions. Analysis of
1
the H NMR (500 MHz, CDCl3) spectrum of 3, in particular
the vicinal coupling constants, verified a chair conformation of
the 1,3-dioxane ring, possessing an axial C-2 N-Boc group and
an equatorial C-3 substituent. It follows from the small vicinal
coupling observed between H-2 and H-3 (J = 1.9 Hz) that 3
has the threo relative configuration at C-2−C-3 using arguments
similar to those applied by Braekman to the acetonide of a
related threo sphingolipid from Haliclona vansoesti.19 Attempted
naphthoylation of 3 was sluggish and incomplete under
standard conditions (2-naphthoyl chloride, pyridine, 100 °C,
24 h); however, smooth conversion to mononaphthoate 4 was
achieved with the alternative acylating reagent, N-(2′-
naphthoyl)imidazole (DBU, CH3CN, 80 °C, ∼4 h).20 The
CD spectrum of 4 (CH3CN, Figure 4b) revealed the long-
wavelength half of the expected split CE as a simple broad
negative band [λ 236 (Δε −21.9)] rising to a weak, positive CE
(Δε ∼8, λ = 190 nm, truncated). Thus the C-6 configuration of
4 and 1b is verified as R.
Naphthoylation of a fraction enriched in 1 gave, after HPLC
purification, the tetra(2′-naphthoyl) derivative 5. In contrast to
4, the CD spectrum (CH3CN) of 5 (Figure 4c) exhibited
expectedly more complex CEs [λ 243 (Δε −20.1), 235 (+17.4),
224 (−12.5)] generated by superposed exciton coupling
contributions of the stereoclusters at C-1−C-3 and C-6−C-
7−C-8.
For the purposes of CD stereoanalysis of the headgroup in 5,
model compounds of known configuration were required. The
erythro and threo tri(2′-naphthoyl) diastereomers 6b and 7b
were prepared (Scheme 2) from L-serine using a variation of the
method reported earlier.3c The CD spectrum of threo-7b
(Figure 3) showed a strong negative split CE [λ 237 (Δε
−19.3), 221 (+7.2), peak to trough, A = 26.5], while the CD of
erythro-6b (Figure 3) was distinctly different: although the sign
of the split CE was still negative the magnitude was far larger [λ
226 (Δε +56), 242 (−68.6), A = 124]. Both erythro-6b and
Figure 2. Newman diagrams of major contributing conformers and
predicted signs of exciton coupled CD spectra in (a, b) S, and R
secondary allylic 2-naphthoate esters,16,18 (c, d) N,O,O-tris-naphthoyl
derivatives of 2-aminoalkane-1,3-diols, (R,R)-i, and (R,S)-ii, and (e, f)
(S,S)-i, and (S,R)-ii (contributions from 1-naphthoate−2-naphthamide
pairs are not shown, but their Cotton effects have the same signs with
amplified magnitude A). Np = 2-naphthoyl. Blue bar = 2-naphthoate;
red bar = 2-naphthamide.
secondary alcohols from their allylic benzoates and successfully
validated by others18 to allylic 2-naphthoate esters. Because the
exciton coupled chromophores at the headgroup C-1−C-3 and
the C-6−C-8 are separated by two CH2 groups in an acylic
chain backbone, little interaction is expected between the two
Scheme 1. Synthesis of Acetonide 3 and Mono- and Trinaphthoyl Derivatives 4 and 5 from 1
C
dx.doi.org/10.1021/np300744y | J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
HPLC (tR = 34.5 min, RP C18, 7:3 CH3CN/H2O; one peak by
1
co-injection), HRMS, and H NMR (CDCl3, 500 MHz; see
Supporting Information), in particular, by comparison of the
chemical shifts of the four narrow-line CH3(CO) signals and
H-1−H-3 (Table 1). Thus, the configurations of 1a and 2a are
the same, and given that the two natural products co-occur in
H. tubifera, it is highly likely the absolute configurations are also
identical.
The application of CD combined with deconvolution of
superposed CEs is exceedingly sensitive; in the case of 5 we
estimate the limits of detection to be approximately 30 pmol.
The method should be applicable to the structures of other
sphingoid base natural products containing similar remotely
separated stereoelements.
The pathogenic yeast Candida albicans is modestly inhibited
by sphingosine or phytosphingosine; however, many modified
sphingolipids possess significantly more potent antifungal
activity.21 Unfortunately, homogeneous samples of 1 for testing
could not be purified from the limited amounts of sponge
material available or by deprotection of pure 1a,22 and
insufficient amounts of 2a were available for assay. Efforts in
our laboratory to procure additional samples of 1 by total
synthesis are under way.
In conclusion, a new sphingolipid, halisphingosine B, was
characterized as its per-acetyl derivative 8 along with 1a, the N-
Boc derivative of halisphingosine A (1). A variant of
Nakanishi’s CD protocol16 for assignment of sphingolipids,
combined with van’t Hoff’s principle of optical superposition,
was applied to verify the absolute configuration of the known
compound 1. These tools should find general application for
configurational analysis of sphingoid-based natural products.
Figure 3. Measured circular dichroism spectra (CH3CN, 24 °C) of
per-naphthoyl derivatives prepared from L-serine (see Scheme 2). (a)
erythro-6b ((2S,3R), dashed line, and (b) threo-7b (2S,3S), solid line.
EXPERIMENTAL SECTION
■
General Experimental Procedures. General experimental
procedures are described elsewhere.23 CD spectra were recorded on
a Jasco J810 spectropolarimeter in 1 mm path length under the
following parameters: slit 2 nm, scan speed 100 nm/min, digital
resolution 1 nm, 8−16 accumulations. 1H NMR spectra were recorded
on a dual-channel Jeol ECA 500 MHz spectrometer (1H NMR,
500.1599 MHz), an Agilent 500 MHz spectrometer equipped with an
XSens 13C{1H} cryoprobe, or a dual-channel Bruker 600 MHz
spectrometer (1H NMR, 599.5560 MHz) equipped with an Avance III
dual-channel console and a 1.7 mm 1H{13C,15N} microcryoprobe.
Spectra are referenced to residual solvent signals for CD3OD or
CDCl3 (1H, δ 3.31 or 7.26 ppm; 13C, δ 49.0 or 77.0 ppm, respectively).
Accurate mass measurements were carried out on an Agilent 6230 ESI-
TOF mass spectrometer, equipped with an Agilent 1200 capillary
HPLC system. Solvents for sub-milligram preparative HPLC were
distilled from commercial HPLC-grade stock.
threo-7b displayed biphasic CEs that were higher in magnitude
than their corresponding tribenzoyl derivatives3c and, as with
the latter compounds, easily discriminated the four possible
stereoisomers with the added advantage of approximately 5- to
10-fold higher sensitivity.
The CD spectrum of 5 was compared with “hybrid CD”
spectra generated by simple linear combinations of the CD
spectra of 7b and 4 (Figure 4) and their mirror images (not
shown; see Supporting Information for hybrid CD of 4 and
erythro-6b). The only combination (Figure 3c) that matched
the observed CEs of 5 corresponds with the 2R,3R,6R
configuration, thereby confirming the assignment for 1
proposed by Mansoor and co-workers.11
Animal Material. Specimens of Haliclona tubifera were collected
by hand with scuba (−15 to −20 m) in 2010 at Ilha do Arvoredo,
Santa Catarina, Brazil. The samples were freeze-dried and stored at
Halisphingosine B (2a) was isolated in low yield (∼0.4 mg)
from an earlier-eluting fraction that yielded 1a. The formula
C23H47NO5Na established from ESI HRMS suggested one
−18 °C for 4 months before use. The sponge was identified by Joao
̃
1
double-bond equivalent less than 1a. Inspection of the H
Luis de Fraga Carraro, and specimens are deposited in the Museu de
NMR spectrum of 2a (Table 1) showed the absence of the
signals of the disubstituted olefin in 1a but preservation of the
spin system corresponding to H-1−H-6, suggesting the new
compound was 7,8-dihydrohalisphingosine A. Halisphingosine
derivatives 1a and 2a were correlated as follows (Scheme 3).
Compound 1a was hydrogenated (H2, Pd−C, MeOH) and the
product converted to the per-acetyl derivative 8 by removal of
the Boc group (TFA, CH2Cl2) and acetylation (Ac2O, pyridine,
rt) followed by purification by HPLC. Compound 8 was also
obtained by two-step conversion of 2a (∼0.1 mg, TFA,
CH2Cl2; Ac2O, pyridine). Both samples of 8 were identical by
Cien
Zoobotan
̂
cias Naturais−Porifera (MCNPOR) collection of the Fundaca̧ o
̃
̂
ica do Rio Grande do Sul, Brazil.
Purification of Halisphingosines A and B: N-Boc Derivatives
1a and 2a. A sample of H. tubifera (56.1g wet wt) was extracted with
MeOH (3 × 500 mL). The resulting extract was adjusted to 9:1
MeOH/H2O and partitioned with hexanes (3 × 300 mL). The
aqueous MeOH layer was removed, and the H2O layer was extracted
with EtOAc (3 × 300 mL). The EtOAc-soluble fraction was partially
(500 mg) purified by flash chromatography on a silica column using a
gradient of MeOH (NH3 satd)/CH2Cl2 (5:95−10:90, v/v). The
ninhydrin positive bands were combined, resulting in seven fractions.
The sixth fraction (46.8 mg) was subjected to a reversed-phase (C18)
D
dx.doi.org/10.1021/np300744y | J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Figure 4. Measured CD spectra (CH3CN, 24 °C) and hybrid CD spectra. (a) N,O,O′,O″-Tetranaphthoyl compound 5 prepared from natural
product 1. (b) Mononaphthoate 4 prepared from 3 (see Scheme 1) and N,O,O′-trinaphthoyl compound threo-7b (see Scheme 2). (c) CD of 5 (solid
line) overlaid with hybrid CD spectrum (dashed) = [CD(4) − CD(7b), dashed]. (d) CD of 2 (solid line) overlaid with hybrid CD spectrum
(dashed) = [CD(4) + CD(7b)]. See Supporting Information for comparisons of the CD of 5 with hybrid spectra derived from 4 and erythro-6b.
cartridge with a solvent gradient of MeOH/H2O (50:50−100:0, v/v).
respectively) were treated with excess (Boc)2O and K2CO3 in THF/
The fractions eluted with 80% and 100% of MeOH (29.3 and 10 mg,
H2O (4:1), and the mixture was stirred for 2 h. After filtration of the
E
dx.doi.org/10.1021/np300744y | J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
the volatiles were removed under a stream of N2, and the residue was
purified by HPLC (Luna, 3 μm silica, gradient 65:35 hexanes/EtOAc
to 100:0 over 60 min, 0.5 mL min−1) to give the mononaphthoate
ester 4 (0.6 mg, 76%) as a colorless oil: UV (CH3CN) λmax (log ε) 290
(3.67), 280 (3.83), 270 (3.76), 237 (4.76) nm; CD, see Figure 4 and
Supporting Information; 1H NMR (CDCl3) 8.58 (s, H-1′), 8.05 (d, J =
8.6 Hz, H-3′), 7.95 (1H, d, J = 8.1 Hz, H-8′*), 7.880 (1H, d, J = 8.6
Hz, H-4′), 7.875 (1H, d, J = 8.1 Hz, H-5′*), 5.83 (m, H-6), 5.61 (m,
H-8), 5.47 (1H, bd, J = 9.9 Hz, H-7), 5.28 (1H, d, J, 10.0 Hz, NH),
4.02 (1H, d, J = 12 Hz, H1a) 3.87 (1H, bt, J = 6.7 Hz, H-3), 3.73 (1H,
d, J = 12 Hz, H-1b), 3.47 (1H, bd, J = 10.4 Hz, H-2), 2.23 (H-9, m),
1.85 (m), 1.65 (m), 1.43 (s, t-BuO), 1.42 (s, CCH3), 1.38 (s, CCH3),
1.45−1.23 (m), 0.87 (3H, t, J = 6.9 Hz, H-18), *interchangeable;
HRMS m/z 632.3919 [M + Na]+ (calcd for C37H55NO6Na, 632.3922).
Preparation of N,O′,O″,O′″-Tetranaphthoyl Derivative 5. A
partially purified fraction (0.5 mg, ∼1.2 μmol) containing largely 1 in a
solution of DBU (5.4 μL, 36 μmol) and excess N-(2′-naphthoyl)-
imidazole (2.7 mg, 12.0 μmol) in CH3CN (0.2 mL) was heated in a
sealed vial at 70 °C overnight. After cooling, the volatiles were
removed under a stream of N2 and the residue was purified by
analytical HPLC (silica, Luna 3 μm, hexanes to EtOAc over 60 min,
0.5 mL min−1) to give the tetranaphthoyl derivative 5 (0.3 mg, 27%):
UV (CH3CN) λmax (log ε) 290 (4.10), 280 (4.26), 270 (4.20), 233
Scheme 2. Synthesis of Model Compounds erythro-6b and
threo-7b
Scheme 3. Conversion of 1a and 2a to Peracetyl Compound
8
1
(5.14) nm; CD, see Figure 4 and Supporting Information; H NMR
(CDCl3) 8.56 (1H, s, H-1′), 8.56 (H-1′, s), 8.51 (1H, s, H-1′), 8.27
(H-1′, s), 8.2 − 7.2 (m), 6.90 (1H, d, J = 9.5 Hz, NH), 5.78 (1H, m),
5.67 (1H, m), 5.55 (1H, m), 5.42 (1H, m), 5.03 (1H, m), 4.71 (1H, s,
m, H-1a), 4.63 (1H, s, m, H-1b), 2.18 (2H, m), 1.95 (2H, m), 1.81
(2H, m), 1.25−1.65 (m), 0.83 (3H, t, J = 6.7 Hz, H-18); HRMS m/z
932.4517 [M + H]+ (calcd 932.4521 for C62H62NO7, 932.4521).
Preparation of L-erythro- and L-threo-Trinaphthoyl Model
Compounds 6a and 7a. The starting aminodiols L-erythro-6a and L-
threo-7a were prepared from 8 derived from L-serine as described
previously3c and used without further purification. A solution of crude
6a (5.0 mg, 0.0257 mmol), DBU (117 mg, 0.771 mmol), and N-(2′-
naphthoyl)imidazole (57 mg, 0.257 mmol) in CH3CN (1.0 mL) was
sealed in a vial and heated at 70 °C overnight. After removal of the
volatiles, the residue was first purified by TLC (4:1 hexanes/EtOAc)
and then semipreparative HPLC (RP C18 Luna column, gradient
elution with 60:40 to 100:0 CH3CN/H2O over 15 min, 4.0 mL min−1)
to give erythro-6b as a white powder (0.7 mg, 5%). Conversion of 7a
into threo-7b was achieved in a similar fashion (2.9 mg, 14%).
L-erythro-6a: UV (CH3CN) λmax (log ε) 233 (5.14), 280 (4.52)
nm; [α]23D −67.0 (c 0.060, CHCl3); CD, see Figure 3 and Supporting
solids and concentration, the entire filtrate was subjected to RP-HPLC
(Luna C18, 10 × 250 mm, 5 μm; gradient elution, 65:45 to 100:0
CH3CN/H2O over 35 min, ELSD detection) to give N-Boc derivative
1a (8.0 mg). The entire fourth N-Boc fraction was subjected to RP-
HPLC (Luna C18, 10 × 250 mm, 5 μm; 7:3 CH3CN/H2O, 2.5 mL
min−1) to give N-Boc-halisphingosine B (2a, 0.4 mg).
N-Boc-halisphingosine A (1a): colorless oil; [α]D +12.7 (c, 0.20,
CHCl3); 1H NMR (CD3OD, 500 MHz), see Table 1; 13C NMR
(CDCl3) 156.4 (NH-(CO)), 132.5* (C-7), 132.3* (C-8), 79.6 (O-
C(CH3)3, 73.2 (C-3), 67.7 (C-6), 65.6 (C-1), 54.2 (C-2), 37.4 (C-4§),
34.2 (C-5§), 31.7 (C-16), 29.7 (C-9), 29.41 (C-15§), 29.38 (C-10§),
1
§
28.94 (C-11§), 28.35 (3×C, OC(CH3)3 ), 27.7 (C-12§), 25.41 (C-
Information; H NMR (600 MHz, CDCl3) δ 8.58 (1H, s), 8.49 (1H,
13§), 25.23 (C-14§), 22.6 (C-17), 14.1 (C-18) (*,§ interchangeable);
HRMS m/z 438.3192 [M + Na]+ (calcd for C23H415NO5Na, 438.3195).
N-Boc-halisphingosine B (2a): colorless oil; H NMR (CD3OD,
600 MHz) δ 3.75 (1H, t, J = 5.6 Hz), 3.60 (1H, m), 3.50−3.54 (3H,
m), 1.45 (9H, s), 1.27−1.38 (30H, m), 0.90 (3H, t, J = 6.7 Hz);
HRMS m/z 440.3343 [M + Na]+ (calcd for C23H47NO5Na, 440.3352).
Preparation of Acetonide 3. A solution of N-Boc halisphingosine
A (1a) (2.0 mg) in acetone and 2,2-dimethoxypropane (12 μL),
CH2Cl2 (0.2 mL), and pyridinium p-toluenesulfonate (1.2 mg, 4.8
μmol) was stirred at room temperature overnight. The mixture was
treated with solid K2CO3, filtered, and concentrated under a stream of
N2. The residue was resuspended, and the soluble portion separated by
column chromatography (oversized pipet, silica, 4:1 hexanes/EtOAc)
s), 8.35 (1H, s), 8.03 (1H, dd, J = 8.5, 1.7 Hz), 7.94 (1H, dd, J = 8.6,
1.7 Hz), 7.92−7.85 (4H, m), 7.85−7.76 (5H, m), 7.73 (1H, d, J = 8.6
Hz), 7.59−7.42 (5H, m), 7.39 (1H, d, J = 8.6 Hz), 5.55 (1H, dt, J =
8.6, 4.3 Hz), 5.01 (1H, ddt, J = 8.6, 6.0, 4.3 Hz), 4.84 (1H, dd, J = 11.8,
6.0 Hz), 4.76 (1H, dd, J = 11.8, 4.3 Hz), 2.15−2.05 (1H, m), 2.00−
1.90 (1H, m), 1.69−1.60 (2H, m), 1.02 (3H, t, J = 7.4 Hz); 13C NMR
(126 MHz, CDCl3) δ 167.39, 167.35, 167.1, 135.8, 135.7, 135.0, 132.8,
132.51, 132.47, 131.6, 131.51, 131.50, 129.51, 129.50, 129.2, 128.7,
128.6, 128.5, 128.3, 127.93, 127.87, 127.85, 127.84, 127.80, 126.92,
126.87, 126.84, 126.7, 125.3, 125.2, 123.7, 76.0, 63.3, 52.2, 34.6, 19.2,
14.0; HRMS m/z 618.2250 [M + Na]+ (calcd for C39H33NO5Na,
618.2251).
L-threo-7b: UV (CH3CN) λmax (log ε) 231 (5.22), 280 (4.61) nm;
[α]23 +29.6 (c 0.071, CHCl3); CD, see Figure 3 and Supporting
1
to obtain 3 as a pale yellow oil (1.3, 60%): H NMR (CDCl3, 500
D
1
Information; H NMR (500 MHz, CDCl3) δ 8.58 (1H, s), 8.53 (1H,
MHz) δ 5.48 (1H, dt, J = 9.3, 6.4 Hz), 5.36 (1H, dddd, J = 9.3, 7.8, 1.5,
1.5 Hz), 5.29 (1H, d, J = 10.0 Hz), 4.41 (1H, dt, J = 7.6, 6.0 Hz), 4.05
(1H, dd, J = 10.2, 1.7 Hz), 3.89 (1H, m), 3.75 (1H, dd, J = 10.2, 1.7
Hz), 3.5 (1H, dddd, J = 10.0, 1.9, 1.7, 1.7 Hz), 2.08 (2H, m), 1.45 (6H,
s, C(CH3)2), 1.40 (2H, t-BuO, s), 1.35−1.24 (21H, m), 0.88 (3H, t, J
= 7.0 Hz); ESI HRMS m/z 478.3504 [M + Na]+ (calcd for
C26H49NO5Na, 478.3503).
Preparation of Mononaphthoate Ester 4. A mixture of
acetonide 3 (0.6 mg), DBU (3.0 μL, 20 μmol), and N-(2′-
naphthoyl)imidazole20 (1.4 mg, 6.5 μmol) was dissolved in CH3CN
(0.2 mL), sealed in a vial, and heated at 60 °C overnight. After cooling,
s), 8.27 (1H, s), 8.02 (1H, dd, J = 8.8, 1.6 Hz), 7.98 (1H, dd, J = 8.5,
1.6 Hz), 7.92−7.73 (10H, m), 7.62−7.49 (5H, m), 7.46 (1H, dd, J =
8.4, 6.7 Hz), 6.92 (N-H, 1H, d, J = 9.2 Hz), 5.71 (1H, dt, J = 8.9, 4.8
Hz), 5.03 (1H, ddt, J = 8.9, 5.8, 4.8 Hz), 4.72 (1H, dd, J = 11.6, 5.8
Hz), 4.65 (1H, dd, J = 11.6, 4.8 Hz), 2.05−1.95 (1H, m), 1.95−1.86
(1H, m), 1.64 − 1.49 (2H, m), 1.00 (3H, t, J = 7.3 Hz); 13C NMR
(126 MHz, CDCl3) δ 167.8, 166.9, 166.8, 135.8, 135.7, 134.9, 132.7,
132.55, 132.49, 131.54, 131.50, 131.47, 129.52, 129.50, 129.1, 128.8,
128.6, 128.51, 128.49, 128.36, 127.90, 127.89, 127.87, 127.82, 127.78,
127.02, 126.95, 126.92, 126.86, 126.7, 125.21, 125.16, 123.5, 73.8,
F
dx.doi.org/10.1021/np300744y | J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
64.9, 51.9, 34.1, 18.9, 14.0; HRMS m/z 618.2254 [M + Na]+ (calcd for
C39H33NO5Na, 618.2251).
(2) (a) Carter, G. T.; Rinehart, K. L. J. Am. Chem. Soc. 1978, 100,
7441−2. (b) Cuardos, R.; Montejo de Garcini, E.; Wandosell, F.;
Conversion of 1a and 2a to Per-acetyl Derivative 8. A sample
of 1a (0.5 mg) and Pd/C (1.0 mg) was taken up in MeOH (0.4 mL)
and purged with H2 for 2 min. The heterogeneous mixture was
allowed to stir at rt under H2 (1 atm) for 4 h. Complete hydrogenation
was verified by LCMS. The mixture was then filtered (0.45 μm syringe
filter) and then concentrated under a stream of N2. The resultant oil
(0.5 mg) was then reconstituted in CH2Cl2/TFA (3:1, 0.2 mL) and
stirred at rt for 30 min. The solvent was then removed under a stream
of N2, and the oil was reconstituted in pyridine (0.1 mL) and acetic
anhydride (0.1 mL). The reaction was allowed to stir at 35 °C under
N2 for 3 h. The volatile reagents were removed, and the crude oil was
purified by RP-HPLC (Luna C18, 4.6 × 250 mm, 5 μm; 7:3 CH3CN/
H2O, 0.85 mL min−1) to give 7,8-dihydroperacetate-halisphingosine A
(8, 0.4 mg). Conversion of 2a using the conditions of the last two of
the three preceding reactions followed by purification gave 8, identical
(HRMS, 1H NMR, LC) with the product obtained from 1a. 1H NMR
́
Faircloth, G.; Fernandez-Sousa, J. M.; Avila, J. Cancer Lett. 2000, 152,
23−29.
For a review of bioactive amino alcohols from marine
organisms, see (c) Molinski, T. Curr. Med. Chem.: Anti-Infect. Agents
2004, 3, 197−220.
(3) (a) Makarieva, T. N.; Denisenko, V. A.; Stonik, V. A. Tetrahedron
Lett. 1989, 30, 6581−6584. (b) Molinski, T. F.; Makarieva, T. N.;
Stonik, V. A. Angew. Chem., Int. Ed. 2000, 39, 4076−4079. (c) Nicholas,
G. M.; Molinski, T. F. J. Am. Chem. Soc. 2000, 122, 4011−4019.
(d) Makarieva, T. N.; Zakharenko, A. M.; Dmitrenok, P. S.; Guzii, A.
G.; Denisenko, V. A.; Savina, A. S.; Dalisay, D. S.; Molinski, T. F.;
Stonik, V. A. Lipids 2009, 44, 1155−1162. (e) Makarieva, T. N.; Guzii,
A.; Denisenko, V. A.; Dmitrenok, P. S.; Santalova, E. A.; Pokanevich, E.
V.; Molinski, T. F.; Stonik, V. A. J. Nat. Prod. 2005, 68, 255−257.
(f) Makarieva, T. N.; Dmitrenok, P. S.; Zakarenko, A. M.; Denisenko,
V. A.; Guzzi, A. G.; Li, R.; Skepper, C. K.; Molinski, T. F.; Stonik, V. A.
J. Nat. Prod. 2007, 70, 1991−1998. (g) Zhou, B.-N.; Mattern, M. P.;
Johnson, R. K.; Kingston, D. G. I. Tetrahedron 2001, 57, 9549−9554.
(4) (a) Molinski, T. F.; Ireland, C. M. J. Org. Chem. 1988, 53, 2103−
5. (b) Salomon, C. E.; Williams, D. H.; Faulkner, D. J. J. Nat. Prod.
1995, 58, 1463−1466. (c) Skepper, C. K.; Molinski, T. F. J. Org. Chem.
2008, 73, 2592−2597.
1
(CDCl3, 600 MHz) see Table 1; H NMR (C6D6, 600 MHz) δ 5.16
(1H, m), 5.08 (2H, m), 4.61 (1H, dddd J = 13, 9.7, 6.6, 3.1 Hz), 4.05
(1H, ddd, J = 11, 6.5, 1.7 Hz), 3.94 (1H, ddd, J = 11, 6.5, 1.7 Hz), 1.79
(3H, s), 1.70 (3H, s), 1.69 (3H, s), 1.55 (2H, m), 1.51 (3H, s), 1.15−
1.47 (24H, m), 0.91 (3H, t, J = 6.7 Hz); HRMS, m/z 508.3244 [M +
Na]+ (calcd for C26H47NO7Na, 508.3245). Separate co-injections of 8
derived from 1a or 2a gave a single peak (LCMS, rt 13.38 min, linear
gradient from 60% to 100% CH3CN(aq)/0.1% formic acid, over 15
min; Kinetex C18, 4.6 × 150 mm, 2.7 μm).
(5) (a) Kobayashi, J.; Naitoh, K.; Doi, Y.; Deki, K.; Ishibashi, M. J.
Org. Chem. 1995, 60, 6941−6945. (b) Kiguchi, T.; Yuumoto, Y.;
Ninomiya, I.; Naito, T. Chem. Pharm. Bull. 1997, 45, 1212−1215 , and
key references within.
(6) Kobayashi, J. i.; Cheng, J.-F.; Ishibashi, M.; Walchli, M. R.;
Yamamura, S.; Ohizumi, Y. J. Chem. Soc., Perkin Trans. 1 1991, 1135−
1137.
(7) Kuroda, I.; Musman, M.; Ohtani, I. I.; Ichiba, T.; Tanaka, J.;
Gravalos, D. G.; Higa, T. J. Nat. Prod. 2002, 65, 1505−1506.
(8) Ledroit, V.; Debitus, C.; Lavaud, C.; Massiot, G. Tetrahedron Lett.
2003, 44, 225−258.
ASSOCIATED CONTENT
* Supporting Information
Additional experimental details and H NMR NMR data for
compounds 1b, 2b, 6a, 7a, 12, and 13 are available free of
charge via the Internet at http://pubs.acs.org.
■
S
1
(9) Ando, H.; Ueoka, R.; Okada, S.; Fujita, T.; Iwashita, T.; Imai, T.;
Yokoyama, T.; Matsumoto, Y.; van Soest, R. W. M.; Matsunaga, S. J.
Nat. Prod. 2010, 73, 1947−1950.
(10) The absolute configurations of diastereomeric erythro and threo
2-aminotetradec-5,7-dien-3-ols from a Papua New Guinea sponge,
Haliclona sp., were first misassigned as (2S,3R) and (2S,3S),
respectively, (a) Gulavita, N. K.; Scheuer, P. J. J. Org. Chem. 1989,
54, 366−369 and revised upon total synthesis. (b) Mori, K.; Matsuda,
H. Liebigs Ann. Chem. 1992, 2, 131−137.
(11) Mansoor, T. A.; Park, T.; Luo, X.; Hong, J.; Lee, C.-O.; Jung, J.
H. Nat. Prod. Sci. 2007, 13, 247−250.
(12) Compound 1 (ref 11) ((2R,3R,6R,Z)-2-aminooctadec-7-ene-
1,3,6-triol) had not been assigned a trivial name. Here, the name
“halisphingosine A” is coined for 1 for convenience of referencing and
relational properties to the co-occurring new compound, hali-
sphingosine B (2).
(13) van’t Hoff, J. H. Die Lagerung der Atome im Raume; Vieweg:
Braunschweig, 1908; Chapter 8, pp 95−97.
(14) For a contemporary review of integrated approaches to
assignment of configuration in marine natural products, see:
Morinaka, B. I.; Molinski, T. F. Tetrahedron 2012, 68, 9307−9343.
(15) Li, S.; Wilson, W. K.; Schroepfer, G. J., Jr. J. Lipid Res. 1999, 40,
117−1125.
AUTHOR INFORMATION
Corresponding Author
*Tel: +1 (858) 534-7115. Fax: +1 (858) 822-0386. E-mail:
tmolinski@ucsd.edu.
Notes
■
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Y. Su and A. Mrse (UCSD) for assistance with
HRMS and NMR measurements, respectively, and J. L. de
Fraga Carraro for collection of the sponge and the photograph
used in the graphic abstract. Purchases of the Agilent TOF mass
spectrometer and the Jeol 500 MHz NMR spectrometer were
made possible with funds from the NIH Shared Instrument
Grant program (S10RR025636) and the NSF Chemical
Research Instrument Fund (CHE0741968), respectively.
While at UCSD, R.B. was supported by a travel fellowship
(PDSE-CAPES 8301/11-5, Brazil) and E.P.S was supported by
Ruth L. Kirschstein National Research Service Awards (NIH
T32 CA009523). We are grateful for funding for this work from
the NIH (AI100776 to T.F.M).
(16) (a) Shirota, O.; Nakanishi, K.; Berova, N. Tetrahedron 1999, 55,
13643−13658. (b) Kawamura, A.; Berova, N.; Dirsch, V.; Mangoni, A.;
Nakanishi, K.; Schwartz, G.; Bielawska, A.; Hannun, Y.; Kitagawa, I.
Bioorg. Med. Chem. Lett. 1996, 4, 1035−1043.
(17) Gonnella, N. C.; Nakanishi, K.; Martin, V. S.; Sharpless, K. B. J.
Am. Chem. Soc. 1982, 104, 3775−1376.
(18) (a) Schneider, C.; Schreier, P.; Humpf, H. U. Chirality 1997, 9,
563−567. (b) Molinski, T. F.; Brzezinski, L. J.; Leahy, J. W.
Tetrahedron: Asymmetry 2002, 13, 1013−1016.
DEDICATION
■
Dedicated to Dr. Lester A. Mitscher, of the University of
Kansas, for his pioneering work in the discovery of bioactive
natural products and their derivatives.
REFERENCES
■
(19) Devijver, C.; Salmoun, M.; Daloze, D.; Braekman, J. C.; De
Weerdt, W. H.; De Kluijver, M. J.; Gomez, R. J. Nat. Prod. 2000, 63,
978−980.
(1) Pruett, S. T.; Bushnev, A.; Hagedorn, K.; Adiga, M.; Haynes, C.
A.; Sullards, M. C.; Liotta, D. C.; Merrill, A. H., Jr. J. Lipid Res. 2008,
49, 1621−1639.
G
dx.doi.org/10.1021/np300744y | J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
(20) Ikemoto, N.; Lo, L.-C.; Nakanishi, K. Angew. Chem., Int. Ed.
1992, 31, 890−891.
(21) Nicholas, G. N.; Li, R.; MacMillan, J. B.; Molinski, T. F. Bioorg.
Med. Chem. Lett. 2002, 12, 2159−2162.
(22) Attempted removal of the N-Boc group in 1a using the standard
protocol (50% CF3COOH in CH2Cl2, rt) or milder conditions (10
equiv, −20 °C) resulted in decomposition, most likely through SN1
solvolysis of the allylic OH. More benign conditions for deprotection
of N-Boc-alkanamines (H2O, 100 °C, overnight, reported by Wang, J.;
Liang, Y.-L.; Qu, J. J. Chem. Soc., Chem. Commun. 2009, 5144−5146 ),
when applied to 1a, returned only unreacted starting material.
(23) Dalisay, D. S.; Rogers, E. W.; Edison, A.; Molinski, T. F. J. Nat.
Prod. 2009, 72, 732−738.
H
dx.doi.org/10.1021/np300744y | J. Nat. Prod. XXXX, XXX, XXX−XXX