6
812
J.-T. Fan et al. / Tetrahedron Letters 51 (2010) 6810–6813
6
*
2
3
4
1
due ) of 1 were optimized at the B3LYP/6-31G level using GAUSSIAN
Ala , N-Me Tyr , and Ala . The absolute configurations of Ala ,
1
6
2
4
0
3 program package. Our calculation results indicated that the
L
-
Ala , and Ala were also identified as D(R), L(S), and L(S), respec-
5
6
Tyr /
correlations (see Supplementary data). For the absolute stereo-
chemistry at C-6db, C-6 b, C-6f, firstly, H-6bb was suggested to
be b-oriented based on the coupling constant of H-6bb/H-6 (dd,
J = 15.6, 11.2 Hz). Then considering the NOE correlations of H-
bb/H-6db and H-6 b/H-6db and the rigidity of oxygen-containing
five-membered ring, H-6db and H-6 b were placed at b orienta-
tion. However, the ROESY spectrum in methanol-d could not iden-
tify the absolute configuration at C-6f. Fortunately, NOE
correlation of H-6f with H-6db was clearly observed in DMSO-d
which implied that H-6f was also b-oriented. So the absolute
configurations of C-6db, C-6 b, and C-6f were assigned as R, S,
L
-Residue isomer was in accordance with the observed NOE
tively, by application of the Marfey’s method. Furthermore, it pos-
sessed two 1,2,4-trisubstituted benzene rings and
additional signals arising from a glucopyranosyl moiety.
HMBC cross-peaks of Tyr -NMe/Ala -CO and Tyr -NMe/Tyr -
CO revealed the sequence of the two tyrosines. Then crucial
a
set of
e
5
4
6
5
a
6
e
HMBC correlations of H-5da/C-6
with NOE correlations of H-5da/H-6db indicated the linkage of
C-5 a and C-6 b. As for the glucopyranosyl moiety of 2, HMBC
correlation of the anomeric proton (d 5.02, d, J = 7.0 Hz) (b-form)
eb and H-6db/C-5ea, together
e
4
e
e
H
6
,
with C-6f indicated that the sugar group was ligated to f-position
of Tyr-6. Enzymatic hydrolysis of 2 with b-glucosidase yielded
e
D-glucose, which was detected by TLC and optical rotation,
1
9
and R, respectively. Therefore, through the above-mentioned
NMR and DFT analysis, the structure of 1 was determined as shown
in Figure 2.
½
aꢁ
+103.7 (c 0.10 H
2
O).
In the ROESY spectrum, NOE correlation that was apparently
observed between H-5 and H-6 suggested that the N-methyl
peptide bond between Tyr and Tyr was cis. Similarly, two geom-
D
a
a
Rubiyunnanin B (2)17 was obtained as an amorphous solid and
5
6
5
6
5
6
possessed the molecular formula C46
41.3917, [M+Na] , calcd 941.3908), with 21 degrees of unsatura-
tion. The H and C NMR spectra of 2 in methanol-d
showed some characteristics of RAs and demonstrated the pres-
ence of two conformers in a ratio of 84:16. An extensive compari-
H
58
N
6
O
14 by HRESIMS (m/z
etries (
L-Tyr /L-Tyr and D-Tyr /D-Tyr ) of 2 were optimized, and
+
5
6
9
the calculation results supported the L-Tyr /L-Tyr structure (see
Supplementary data). Moreover, H-5da exhibited NOE correlation
with Tyr -NMe, which indicated that the benzene ring of Tyr
was at the left angle of the benzene ring of Tyr . Consequently,
based on the above-mentioned NMR and DFT analysis, the struc-
ture of 2 was determined as shown in Figure 2.
1
13
4
(Table 1) also
6
5
6
son of 1D and 2D NMR data of 2 with those of 1 and RA-XII (3) also
established the presence of the 18-membered ring, that is,
1
D
-Ala ,
Comparing with all known 30 RAs, both 1 and 2 have unique
skeletons. Except for the 18-membered ring, compound 1 has extra
5
6
two rings formed between Tyr and Residue via a phenolic oxygen
linkage and a new carbon–carbon bond. In addition, compound 2 is
discovered for the first time that it lacks the typical phenolic oxy-
gen linkage, but rather has a carbon–carbon bond at the ortho-posi-
5
6
tions of the hydroxyl groups of Tyr and Tyr .
The biosynthetic mechanism for plant cyclopeptides has not
1
8
been well established, except for the cyclotides. The hexapep-
tide precursor of RAs could be synthesized via a non-ribosomal
4
,19
peptide biosynthetic mechanism.
To form a mature RA, sev-
eral modification steps are expected to take place, among which
is the phenolic coupling step that leads to the most characteris-
tic structural feature of RAs. It is well established that the cou-
pling is via a free radical mechanism, typically at ortho–ortho,
5
ortho–para, or para–para position. The hydroxyl of Tyr and the
6
ortho-carbon of Tyr can generate free radicals. An oxygen–car-
bon coupling would lead to the 14-membered cycloisodityrosine
moiety that is found in all RAs. For compound 2, the coupling is
between two ortho-carbons. To form the very unusual three-ring
6
system of compound 1, the aromatic ring of Tyr of a RA
precursor needs to be reduced first. This sets a stage for an
electrophilic aromatic substitution-like reaction, using the
6
meta-carbon of Tyr as the electrophile to add to the ortho-
5
carbon of Tyr . This step could be facilitated by a protonation
6
on Tyr ring and the lone-pair electrons on the oxygen bridge
between the two tyrosine residues. An elimination of the proton
5
on Tyr ring driven by the restoration of the aromatic ring of
5
Tyr would lead to the formation of the new carbon–carbon
bond of 1. Possible biosynthetic pathways for RAs, 1 and 2 are
shown in Figure 3.
Cytotoxicities of 1 and 2 against 11 cancer cell lines (HepG2,
BEL-7402, SMMC-7721, MDA-MB-231, DU-145, PC-3, A549, BGC-
8
23, Hela, U251, and B16) were measured by SRB assay with the
2
0
maximum concentration of 50 lg/mL. Only compound 2 exhib-
ited moderate cytotoxicities with IC50 values of 31.3, 13.2, 33.7,
.3, 16.3, 21.9, 4.8, 5.6, 21.3, 6.5, and 3.6 g/mL, respectively. Fur-
thermore, the inhibitory effects of 1 and 2 on NO production in LPS
and IFN- -induced RAW 264.7 murine macrophages were exam-
ined. Only compound 2 was found to inhibit NO production with
IC50 value of 10.7 g/mL.
7
l
c
Figure 2. Selected 2D NMR correlations of 1 and 2.
l