6
0
W.-C. Lai et al. / Steroids 89 (2014) 56–62
compound contains one more ring than the parent 20E. To facili-
tate the understanding of the different rearrangements of 20E dur-
ing its photo-irradiation resulting structure 5, we applied the
specific atomic numbering of the parental compound to compound
one-dimensional NOESY experiments were performed, irradiating
H-9, H -18, H -19 and H -21, respectively in the latter cases. The
3
3
3
NOESY results were in accordance with structure 5 and ruled out
5a. The schematic stereo-structure of 5 with the atomic numbering
is depicted in Fig. 2A; the arrows indicate the detected characteris-
tic spatial proximities, whereas the structure of Fig. 2B shows the
refined stereochemistry obtained by PM3 semi-empirical calcula-
tion [20]. For a better visualization of the steric arrangement of
compound 5, the C(23)–C(27) side chain is replaced with R.
The HRMS measurement of 6 yielded molecular formula
1
13
5
(see Fig. 1). For the H and C signal assignment, we first iden-
1
tified the five singlets in the H NMR spectrum. The identification
of the geminal H C-26 (1.16/29.0) and H C-27 (1.17/29.9) groups
3
3
is straightforward owing to their mutual HMBC correlation. The
HMBC responses of these methyl groups assigned the quaternary
C-25 (71.4) and also the H
the 3.71/42.0, 3.71/26.8 and 3.71/20.3 cross-peaks revealed the
assignment of HC-22 (3.71/77.6), H C-23 (1.71; 1.42/26.8) and
the third angular methyl H C-21 (1.37/20.3). By utilizing the HMBC
correlations of H -21, the quaternary C-20 (93.1) and HC-17 (3.13/
1.9) were assigned, and the H-17/H C-18 cross-peak (3.13/24.5)
revealed the assignment of the fourth angular methyl group. The
HMBC correlations of H -19 0.95/49.6 and 0.95/45.8, respectively,
2
C-24 methylene (42.0). In addition,
27 44 8
C H O . Therefore, one oxygen atom is incorporated into the
1
13
2
parental 20E. The H and C NMR chemical shifts corresponded
to those previously published by Harmatha et al. for 14 -hydro-
3
a
3
peroxy-20-hydroxyecdysone [11], a known photoproduct of 20E.
The presence of the peroxy group in position 14 of compound 6
4
3
is straightforward considering the characteristic (d
paramagnetic shift of C-14 to d 96.6 ppm.
D
ꢂ 12 ppm)
3
assigned the HC-5 and HC-9 methines. For their unambiguous dif-
ferentiation selective one-dimensional TOCSY experiments were
utilized. Irradiation of the well-separated H-3 signal (3.98q,
J ꢂ 2.5 Hz) revealed the 7-membered spin system of the A-ring,
whereas excitation of the signal at 3.27 gave the 5-membered spin
Despite the unchanged molecular formula, C27
pound 7 as compared to that of 20E, significant differences could
H O of com-
44 7
1
3
1
be observed in the C and H NMR spectra. A comparison of the
chemical shifts with those of 20E strongly suggested the shift of
7
,8
8,14
the
D
double bond to the
D
position. The observed 10 ppm
system of H-9, H
ation of H-17 (3.13dd, J: 8.0; 4.7 Hz) assigned the 5-membered spin
system of H-17, H -16 and H -15 hydrogen atoms. To achieve an
unequivocal assignment of C-14, the HMBC correlations of the
-15 and the H -16 hydrogens were utilized. Their responses to
2
-11 and H
2
-12. A similar experiment, the irradi-
paramagnetic shift of d C(6)@O to 216.1, together with the absence
of the characteristic signals of @CH in position 7, revealed that this
3
1
2
2
carbon atom evolved to sp CH (d H-7 3.53s; d C-7 59.2). These H
1
3
and C chemical shifts also indicated the presence of one –OH
group attached to C-7. Considering the strong NOE contacts of
H
2
2
the signal at d 175.5 ppm revealed that the original OH substituted
quaternary sp C atom turned into a lactone O@CAO carbon atom.
3
the protons H -19–H-7 and also to H-5 (2.42dd) hydrogen atoms
detected in the selective one-dimensional NOESY spectrum, the b
position of these atoms is straightforward, i.e. the 7-OH group is
3
Despite the absence of the characteristic H-7 olefin signal, the d
1
99.5 ppm chemical shift of C-6 strongly indicated the presence of
in a position and the cis A/B ring-junction remained unchanged.
7
,8
2
a conjugated
D
-6-one moiety in the B ring. The HMBC correla-
Both of the two quaternary sp carbon signals at d 124.6 and d
153.1 exhibited strong responses to the H-9 and H-7 signals in
the HMBC spectrum, respectively, and their differentiation pro-
tions of the H-5 hydrogen to the O@C-6 and to the quaternary
C-7 atoms (d 142.8 ppm), in addition with the H-9/C-7 (3.27/
42.8) and H-9/C-8 (3.27/145.0) cross-peaks, completed the signal
assignment of the B ring. The C-7 signal exhibits d 20.6 ppm para-
magnetic, whereas C-8 d 23.1 ppm diamagnetic shift in compar-
@
1
vided the H
3
-18/C-14 (0.90/153.1) cross-peak. The detailed NMR
1
1
D
analysis combining two-dimensional H, H-COSY, -NOESY,
-
D
ROESY, edited gs-HSQC, gs-HMBC and one-dimensional sel-ROESY
and DEPTQ measurements allowed the complete structural assign-
ment of all hydrogen and carbon signals (see Table 1), and con-
firmed the suggested structure for compound 7. It should also be
noted, that the 7-deoxy analog of 7 has been reported by Canonica
et al. [9], and their NMR chemical shifts correlate well with our
data.
Unsurprisingly, our results show fundamental differences as
compared to the previous studies on the photo-transformation of
20E due to highly different experimental conditions. First of all,
the powerful laser impacted the solution at much higher energy
than previously used [9–11], and the wavelength was also differ-
ent: in our case an optically pure, 266 nm UV–light was used, in
contrast with the Pyrex-filtered (>300 nm) UV–light provided by
Hg lamps. Moreover, in our case, methanol was used as a solvent,
which is a better solvent for 20E than water and allows to reach
higher concentrations of starting material (and products). Last
but not least, our experimental set-up made it a challenge to
remove oxygen from the system and to provide an inert atmo-
sphere, therefore our conditions must have been more oxidative
than those of the previous studies.
Compound 1 was formed by an oxidative side chain cleavage
between C-20 and C-22, which, due to the large distance from
the B-ring chromophore, should be the result of rather inter- than
intra-molecular processes. At the employed wavelength and
energy, direct photolysis of methanol (which could result in a large
amount of strong oxidizing species) is unlikely [21]. Oxygen or
other radicals (i.e. OH radical originated from C-14, see below) pro-
vided by neighboring 20E molecules can however give a reasonable
explanation for the formation of 1. In the case of all other
ison with the corresponding values measured for the parental 20E,
revealing that the substituent on the @C-7 atom should be oxygen.
3
Surprisingly, a strong correlation between the H -18 methyl
hydrogens and the @C-8 carbon atom was also detected, providing
evidence of a contraction in the C-ring. In the course of the photo-
chemical transformation of 20E, a similar rearrangement was
reported by Canonica et al. [9,10]. They explained the contraction
by assuming that the initially produced O-6, C-8 diradical would
undergo a homolytic cleavage of the 13–14 bond, and that the
recombination of C-13 and C-8 radicals led to the five-member
C-ring. In our case (due to the high distances between possible
members of the new lactone ring), the homolytic cleavage of the
8
–14 bond also had to be taken into account. Following such a
cleavage, C-14 can reasonably form two possible lactone rings:
either with the 20-OH resulting a six-member lactone, or with
2
tures are feasible; in the case of compound 5, an H-22/C-7 correla-
tion; and in the case of 5a, an H-22/C-14 HMBC correlation should
be expected. However, probably due to spatial reasons, neither of
these correlations could be detected even in the case where the
HMBC measurement was optimized to 5 Hz long-range couplings
instead of 10 Hz. In the course of building-up the fourth and fifth
rings, the configuration around the C-20 and C-22 chirality centers
2-OH forming a seven-member lactone ring. A priori both struc-
should remain unchanged. Therefore, in structure 5, both H
and H-22 necessarily exist in the position, whereas in case of
a they are in a b position. To choose between the two alternative
structures and to elucidate the configuration of the stereogenic
centers, and to determine the or b position of the methylene
hydrogen atoms, two-dimensional and set of selective
3
C-20
a
5
a
a
a