J. Nat. Prod. 2005, 68, 1689-1691
1689
N3,5′-Cycloxanthosine, the First Natural Occurrence of a Cyclonucleoside
Robert J. Capon* and Nicholas S. Trotter
Centre for Molecular Biodiversity, Institute for Molecular Bioscience, The University of Queensland, St Lucia,
Queensland, 4072, Australia
Received July 26, 2005
An Eryus sp. of marine sponge from the Great Australian Bight has yielded the first reported natural
occurrence of a cyclonucleoside, N3,5′-cycloxanthosine. The structure of N3,5′-cycloxanthosine was
confirmed by detailed spectroscopic analysis and total synthesis.
During investigations into bioactive metabolites from
Australian marine sponges we had cause to examine an
Eryus sp. (class, Demospongia; order, Astrophorida; family,
Geodiidae) obtained in July 1995 during scientific trawling
operations in the Great Australian Bight. Although the
aqueous EtOH extract of this specimen did not demonstrate
significant biological activity in our target bioassays (an-
timicrobial and antiparasitic), we were intrigued by the
presence of unidentified “aromatic” compounds in the H2O-
soluble partition. In past years we have observed that
sponge extracts can yield novel H2O-soluble metabolites,
such as the unprecedented sulfur sugar 5-thio-D-mannose,1
the nematocidal amino acid analogues echinobetaines A2
and B3, and the acetylcholine mimic esmodil.4 In each of
these cases, the structural proofs were supported by
detailed spectroscopic analysis and total synthesis. Encour-
aged by these experiences and the prospect that the Eryus
sp. may also contain novel metabolites, we embarked on a
detailed chemical analysis.
for 2 (see Table 1) were deceptively simple, indicative of a
nuclear base bearing a single aromatic proton attached to
a sugar moiety, suggestive of a modified nucleoside. Lack
of a measurable 1H NMR coupling or 2D NMR COSY
correlation to the proposed anomeric proton H-1′ (δ 6.15)
was interpreted as evidence of a restricted conformation
featuring dihedral angles that disfavored J1′,2′ coupling.
This analysis was further tempered by the observation of
2D NMR gHMBC correlations from H-1′ to C-2′, C-3′, and
C-4′. Connection of the sugar moiety through N9 was
supported by 2D NMR gHMBC correlations from H-1′ to
C-4 and C-8. The UV and 2D NMR data for 2, while
supportive of a modified nucleoside, did not provide con-
clusive evidence from which to assign unambiguously a
molecular structure. To more fully explore this possibility
and provide an unambiguous assignment of molecular
structure, we embarked on a structure proof by total
synthesis.
An early and plausible candidate structure for 2 was that
of 8,5′-O-cycloinosine; however, although known as a
synthetic compound since 1976,10 the published spectro-
scopic data for 8,5′-O-cycloinosine proved inadequate for
modern comparison purposes. Employing literature pro-
cedures,10-13 we completed a total synthesis of 8,5′-O-
cycloinosine in five steps from adenosine, as outlined in
Scheme 1 (for experimental details see the Supporting
Information). To our disappointment, 8,5′-O-cycloinosine
did not coelute on HPLC with, and the 1H NMR (d6-DMSO
or d4-MeOH) data clearly differed from that of, the natural
product 2. Despite these differences, the data were suf-
ficiently similar for us to conclude that 2 is indeed a
cyclonucleoside, just not 8,5′-O-cycloinosine.
The freeze-dried H2O-soluble partition of an aqueous
EtOH extract of the Eryus sp. was fractionated by elution
through Sephadex G-10 (H2O) followed by C18 HPLC, to
yield the two pure “aromatic” metabolites. The first of these
metabolites was identified as C2-R-D-mannosylpyranosyl-
tryptophan (1). This unusual C-glycosylated tryptophan
was first reported in 1994 by Hofsteenge et al.5 during an
investigation into novel post-translational modification in
human RNases and was subsequently detected by the same
authors in a range of other human proteins including IL-
126 and the erythropoietin receptor.7 Although first recog-
nized as a rare post-translational C-glycosylation event in
human proteins, the free glycosylated amino acid 1 was
subsequently reported in 2000 by Riguera et al.8 from a
marine ascidian. Following this study, in 2001 C2-R-D-
mannosylpyranosyltryptophan (1) was acknowledged by
Yonemura et al.9 as a novel marker of renal function, being
an accurate measure of insulin clearance in patients with
chronic renal failure. Our isolation of C2-R-D-manno-
sylpyranosyltryptophan (1) from an Australian marine
sponge, Eryus sp., suggests that tryptophan C-glycosylation
as a biosynthetic event may be more widespread than
previously documented.
While the reisolation of C2-R-D-mannosylpyranosyltryp-
tophan was noteworthy, of greater interest to us was the
discovery of the co-metabolite, 2. On the basis of high-
resolution ESI(+)MS measurements (M + Na ∆mmu -0.2),
2 was attributed the molecular formula C10H10N4O5, re-
A reappraisal of the data for 2 suggested an alternative
cyclonucleoside structure, namely, N3,5′-cycloxanthosine.
First synthesized in 1963,14 and further characterized by
ORD studies,15 N3,5′-cycloxanthosine remained dormant in
the scientific literature until 2004, at which time it came
to our attention through the publication16 of a convenient
1
quiring eight double-bond equivalents. The H NMR data
1
one-step synthesis from xanthosine. Although the H and
* Corresponding author. Tel: +61-7-33462979. Fax: +61-7-33462101.
E-mail: r.capon@imb.uq.edu.au.
13C NMR (d6-DMSO) data reported for N3,5′-cycloxan-
10.1021/np0502692 CCC: $30.25
© 2005 American Chemical Society and American Society of Pharmacognosy
Published on Web 11/02/2005