Ryanodine action at calcium release channels. 2. Relation to substituents of the cyclohexane ring

Article abstract of DOI:10.1021/jm950712d

Ryanodine (1) and dehydroryanodine (2) are equipotent probes for the ryanodine receptor (ryr) of calcium release channels and differ only in 9(eq)-methyl for 1 and 9,21-methylene for 2. Ryanoids 1 and 2 are used here to prepare novel modifications of the

Full text of DOI:10.1021/jm950712d

J . Med. Chem. 1996, 39, 2339-2346
2339
Rya n od in e Action a t Ca lciu m Relea se Ch a n n els. 2. Rela tion to Su bstitu en ts of
th e Cycloh exa n e Rin g
Phillip R. J efferies,*,† Peter J . Gengo,‡ Michael J . Watson,‡ and J ohn E. Casida*,†
Environmental Chemistry and Toxicology Laboratory, Department of Environmental Science, Policy, and Management,
University of California, Berkeley, California 94720-3112, and Burroughs Wellcome Co., Research Triangle Park,
North Carolina 27709-2700
Received September 27, 1995X
Ryanodine (1) and dehydroryanodine (2) are equipotent probes for the ryanodine receptor (ryr)
of calcium release channels and differ only in 9eq-methyl for 1 and 9,21-methylene for 2.
Ryanoids 1 and 2 are used here to prepare novel modifications of the cyclohexane substituents
to determine their effects on ryr activity and selectivity. 10-Oxo-1 when reacted with carbonyl
and other reagents gave 13 C-10 derivatives including the epi-amine and epi-4-azidobenzoyl
hydrazide as a candidate affinity probe. Four derivatives of 2 included the ∆8-10-hydroxy and
∆8-10-oxo compounds. Defunctionalization of the cyclohexane ring of 2 or its 4,6-ethylboronate
was achieved in part by controlled periodate oxidation of the 9,21-diol to the 21-nor-9-oxo
compounds. These in turn provided access to the 9ax- and 9eq-hydroxy derivatives and to the
21-nor-10-deoxy-9-oxo compound which was converted to 21-nor-10-deoxy-1 and 10-deoxy-2
along with the epimeric 10-deoxy-9-hydroxy compounds. Ryanoids of similar potency to 1 as
inhibitors of [3H]-1 binding in mouse brain, rabbit skeletal muscle, and canine ventricle ryr
preparations and in a rat cardiac contractility assay (inhibition of mechanical response to
electrical stimulation) are epi-1 and the 10-epi-amino, 10-epi-methoxyamino, and 10-epi-
azidobenzoyl hydrazide derivatives and 10-deoxydehydroryanodine. With a few exceptions the
potency of the ryanoids at the cardiac ryr correlates well with their inhibition of cardiac
contractility, indicating that the activity is associated with stabilizing the calcium release
channel in a subconducting state, thereby uncoupling the excitation-contraction process.
In tr od u ction
Natural ryanoids contain a range of substituents at
C-8, C-9, and C-10 of the cyclohexane ring. The most
important ones in amount and biological activity are
ryanodine (1)1 and dehydroryanodine (2)2 which differ
only in the cyclohexane ring with 6ax,10eq-hydroxyls in
both cases and 9eq-methyl for 1 and 9,21-methylene for
2 (Figure 1). The cyclohexane ring in minor and less
active natural ryanoids is modified as follows: 8ax-
hydroxy-10-epi-2 and 8ax-hydroxy-10-(O-methyl)-10-
epi-1 and -2; 9ax-hydroxy-1 and 9ax-hydroxy-10-epi-1;
∆
9,10-8-oxo-10-deoxy-1.3-5 Most of the biological activity
comparisons are based on the inhibition of binding of
[3H]-1 to ryanodine receptors (ryr) in rabbit muscle
sarcoplasmic reticulum preparations.6,7
Chemical modifications in the cyclohexane ring of 1
and 2 generally result in reduced potency as inhibitors
of [3H]-1 binding, e.g. epoxidation of the C-9/C-21 double
bond8 and further modification of the epoxide;4 addition
of thiols at this position;9,10 oxidation or esterification
of the C-10 hydroxyl to 10-oxo-1 and 10-acetyl-1, re-
spectively;8 replacement of the C-10 substituent with a
variety of other moieties mentioned in our review11 and
detailed here. In the C-10eq esters, a basic amino or
guanidino group four or five atoms from the cyclohexane
ring is very favorable for optimal binding whereas a
carboxyl is not suitable.12 Thus, the activity of 1 and 2
at the ryr in calcium release channels is greatly affected
by the nature of the substituents on the cyclohexane
ring.
F igu r e 1. Structures of ryanodine (1) and dehydroryanodine
(2) and partial structure showing the B and C rings viewed
from the opposite side of the molecule.
The present study uses two approaches in further
defining the contribution of substituents in the cyclo-
hexane ring to the action of 1 at the ryr. The first starts
from 10-oxo-1 and 10-oxo-2 to generate a variety of
derivatives including oximes, hydroxyamines, amine,
hydrazone, hydrazine, lactam, aromatic esters, and
products from manipulation of the 9,21-double bond.
The second attempts to defunctionalize the cyclohexane
ring and therefore requires a more extensive series of
reactions. Routes to 10-deoxy-2 and 21-nor-10-deoxy-1
were therefore developed via controlled periodate oxida-
tion of the 9,21-diol, samarium iodide reduction of the
21-nor-9-oxo-10-acetoxy compound, and functionaliza-
† University of California.
‡ Division of Pharmacology, Burroughs Wellcome Co.
X Abstract published in Advance ACS Abstracts, May 15, 1996.
S0022-2623(95)00712-6 CCC: $12.00 © 1996 American Chemical Society

2340 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 12
J efferies et al.
F igu r e 2. Modifications of the 10-hydroxyl substituent starting from 10-keto-1 (3). Conditions: (a) carbonyl reagents; (a1) NH2-
OH; (a2) NH2OMe; (a3) NH2OCH2Ph; (a4) NH2NH2; (b) NaBH4; (c) NaCNBH3; (d) NH4OAc, NaCNBH3; (e) ArC(O)Cl; (f) TsCl,
base.
tion of the 9-oxo group with Tebbe reagent or its removal
by nickel desulfuration of the thioketal.
Str u ctu r a l Mod ifica tion s
Mod ifica tion s of th e 10-Hyd r oxyl Su bstitu en t of
1 (3-16) (F igu r e 2). The 10-oxo compound 3 provided
easy access to derivatives of many types at C-10. We
noted before11 and detail here that borohydride reduc-
tion and recovery from the intermediate borate with
methanol gave largely the 10-epi-hydroxy compound 4.
Carbonyl reagents using standard methods easily con-
verted 10-oxo compound 3 to the corresponding oxime
5, methoxime 6, benzyl oxime 7, and hydrazone 8.
Acidic cyanoborohydride reduction13 then gave the epi-
hydroxyamine 9 (characterized as its benzoyloxy deriva-
tive 10) and the corresponding epi-methoxyamine 11
and epi-(benzyloxy)amine 12, the latter very slowly. The
products, recovered from their borates, were identified
in each case as the 10ax isomers from the small coupling
(∼3 Hz) of H-10. Reductive amination of 3 using
ammonium acetate and cyanoborohydride yielded the
epi-amine 13 characterized as its benzamide 14. Simi-
larly, cyanoborohydride reduction of hydrazone 8 and
methanolysis of the borate gave the hydrazine which
was immediately converted to the benzoyl hydrazide
(not shown) and 4-azidobenzoyl hydrazide (15). Reac-
tion of oxime 5 with tosyl chloride (TsCl) in pyridine
gave an intermediate which was converted gradually
to lactam 16. The oximes are assumed to have the
E-configuration in view of the large interference be-
tween the 12-hydroxyl and the oxime substituent in the
Z isomer. The NMR spectrum of the parent 10-oxo
compound 3 has normal couplings for the protons of the
cyclohexane ring whereas all of the oximes show a
downfield shift (∼0.5 ppm) for the 9-methyl group and
have proton couplings as average values indicating
conformational change which would relieve interaction
of the oxime substituent and the methyl attached to C-9.
Mod ifica tion s of th e Cycloh exa n e Rin g of 2 (17-
20) (F igu r e 3). Protection of 2 as the 4,6-ethylbor-
onate14 [2(EtB)] (prepared with lithium triethylborohy-
dride) allowed direct esterification at C-10 with either
F igu r e 3. Modifications of the cyclohexane ring of 2. Condi-
tions: (a) (1) LiBEt3H, (2) PhCOCl, C5H5N, (3) MeOH, MeNH2;
(b) Pd/C; (c) Swern oxidation.
TsCl or benzoyl chloride to obtain 17. Oxidation of the
∆8-10-hydroxy compound 18 gave the ∆8-10-oxo deriva-
tive 19, as previously noted in our review,11 and
analogously of 2 to the methylene ketone 20, which
unfortunately proved to be unstable.
Defu n ction a liza tion of th e Cycloh exa n e Rin g of
2 (21-33) (F igu r e 4). The initial goal was to deoxy-
genate 2 at C-10, i.e. to prepare 10-deoxy-2 (30). The
synthesis started with the 9,21-diol (produced from 2
by osmium tetroxide oxidation)8 initially protected as
the ethylboronate [2(EtB)] to block the 4,12-diol from
periodate oxidation and to reduce water solubility of the
intermediates. Oxidation with 1 molar equiv of sodium
periodate then gave the 21-nor-9-oxo derivative [21(EtB)]
which was reduced with superhydride to the 21-nor-9ax-
hydroxy compound [22(EtB)] [assigned from the small
coupling (∼4 Hz) of H-10], from which the parent (24)
was recovered by methanolysis with methanol/methyl-
amine. Treatment of 22 with samarium iodide15 gave
the 21-nor-10-deoxy-9-oxo derivative (27) as its ethyl-
boronate in 30% yield. Mild acetylation of 21(EtB)
yielded exclusively the 10-acetate which on treatment
with samarium iodide gave the 21-nor-10-deoxy-9-oxo
compound (27) as its ethylboronate in high yield.
Reduction of the latter ethylboronate with sodium
borohydride in methanol and removal of the boronate

Ryanodine Cyclohexane Substituents
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 12 2341
F igu r e 4. Defunctionalization of the cyclohexane ring of 2. (a) OsO4/C5H5N; (b) NaIO4; (c) LiEt3BH or LAH; (d) Ac2O/C5H5N; (e)
SmI2; (f) LAH; (g) Tebbe reagent; (h) (1) (CH2SH)2/BF3‚Et2O, (2) MeOH/MeNH2, (3) Ni; (i) POCl3/C5H5N.
group gave the 21-nor-10-deoxy-9ax-hydroxy compound
28 with a smaller amount of the 9eq epimer 29, and
these were assigned from the line shapes of the NMR
signals for the C-9 protons.
the 21-nor-10-deoxy-∆8/∆9 compounds 33. The main
and less polar product showed NMR data characteristic
of the 1,4-oxygen-bridged cyclohexane ring of the 21-
nor-10-deoxy-6,9-oxido compound 32 which would arise
by internal SN2 reaction with the intermediate chloro-
phosphate or derived ion pair.
An alternative method was developed for conversion
of 2 to the 21-nor-9-oxo compound 23 without use of the
boronate group. Osmylation of 2 in pyridine and
purification as for the boronate gave the known 9,21-
diol.8 Although the 4,12-diol group is oxidized rapidly
by periodic acid,16 reactions with periodate are slower
at pH >4,17 and in practice the more exposed 9,21-diol
oxidized much faster at neutral pH. Reduction of the
21-nor-9-oxo compound 23 with samarium iodide gave
its 10-deoxy derivative 27 in poor yield, not improved
by inclusion of pivalic acid,18 the major byproduct being
the 21-nor-9eq-hydroxy derivative 25 assigned from the
large coupling (9.8 Hz) for the diaxial protons at C-9
and C-10. Brief reduction of 23 with lithium aluminum
hydride gave a mixture of the 21-nor-9ax-hydroxy com-
pound with its 9eq epimer (24 and 25). Mild acetylation
of 23 gave in high yield the 21-nor-9-oxo-10-acetoxy
compound 26 which was efficiently reduced with sa-
marium iodide to the 10-deoxy derivative 27. Brief
reduction of 27 with lithium aluminum hydride gave
both epimeric alcohols (28 and 29). Reaction of oxo
compound 27 with the Wittig reagent in dimethyl
sulfoxide (DMSO)19 was unsuccessful. On the other
hand methylenation with Tebbe reagent19 gave a good
yield of 10-deoxy-2 (30).
Str u ctu r e-Activity Rela tion sh ip s
Rabbit skeletal muscle, mouse brain, and canine left
ventricular membranes were used for assays of [3H]-1
binding and inhibition thereof by the test compounds.14
A pharmacologic response mediated by the cardiac ryr
was also examined involving contraction on electrical
stimulation of cardiac muscle using rat right ventricular
muscle strips.14 Conditions specifically considered in
the companion paper14 are not detailed here. The data
in Tables 1 and 2 are generally for compounds with
>10% of the activity of 1 in at least one of the assays.
Compounds of lower activity or those closely related to
tabulated compounds (oxime, hydrazone, and substi-
tuted hydroxyamines) are given in the Supporting
Information.
Mod ifica tion s of th e 10-Hyd r oxyl Su bstitu en t of
1 (3-16) (Ta ble 1, F igu r e 2). This series derived from
10-oxo-1 (3) generally retains moderate to high activity
at ryrs with fair correspondence for the different tissues
where such data are available. The epi-hydroxy com-
pound 4 is 34-97% as active as 1 in the 4 assays, i.e.
the normal and epi stereochemistry are equally favor-
able. Oximes 5-7 and hydrazone 8 are similar in
potency to oxo compound 3. The hydroxyamines and
amine approximate the activity of 1 when small sub-
stituents are present (9, 11, and 13) but are considerably
less potent with aryl moieties (10 and 12). Although
the epi-benzamide 14 is less active than the epi-amine
13, the azidobenzoyl hydrazide 15 is 67% as active as 1
at the skeletal muscle ryr. In this candidate photoprobe,
in contrast with earlier candidates,22 the photoreactive
substituent lies near the cage structure. Lactam 16
Another goal was to obtain 21-nor-10-deoxy-1 (31).
Although this is an expected product of the 9-tosylhy-
drazone and borohydride in acetic acid,20 no deoxygen-
ation was observed. However, 31 was obtained from 27
by conversion to the ethylene thioketal21 using boron
trifluoride etherate, methanolysis of the borate ester
and then brief desulfurization with nickel. Deoxygen-
ation of the 9-position was also achieved by treatment
of the 9ax-hydroxy compound 28 with phosphoryl chloride/
pyridine which gave a small alkene fraction containing

2342 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 12
J efferies et al.
Ta ble 1. Modifications of the 10-Hydroxyl Substituent of
Ryanodine and the Cyclohexane Ring of Dehydroryanodine in
Relation to Action at Calcium Release Channels
24 and 28 show high potency in the functional assay
relative to their 9eq isomers 25 and 29, but this isomer
difference is not evident in the ryr assay where they are
relatively less effective. Deoxygenation at C-10 affects
the activity differently depending on the other substit-
uents and the bioassay system, i.e. greatly reduces ryr
activity in the 21-nor-9-oxo series (27 versus 23) and
confers similar potency in the 21-nor-9ax- and -9eq-
hydroxy series (28 and 29 versus 24 and 25). On the
same basis, 25 and 29 show similar low values in all
assays whereas 24 and 28 appear to differ significantly
in the canine ventricle compared with the ryr assays.
There is little potency loss in any of the assay systems
on removing the 10-hydroxyl substituent (30 versus 2),
somewhat more for the 21-nor-10-deoxy compound 31,
and the deoxy-∆8/∆9 compound 33 is of only weak
activity as is the 6,9-oxido compound 32.
activity relative to ryanodinea
ryanodine receptor
10-substituent or
modification of
cyclohexane ring
rat
rabbit mouse canine ventricular
muscle brain ventricle strip assay
no.
1 ryanodine
100
100
100
100
Selected Modifications of the
10-Hydroxy Substituent of 1b
30
3 oxo
4 10-epi-hydroxy
5 oxime
6 methoxime
9 epi-hydroxyamine
11 epi-methoxyamine
13 epi-amine
97
26
13
32
68
68
69
78
17
36
34
34
31
13
65
14 epi-benzamide
9.5
15 epi-4-azidobenzoylhydrazide 67
16 lactam
42
7.4
11
Cor r ela tion betw een Rela tive P oten cies of Ry-
a n od in e Der iva tives a s In h ibitor s of [3H]-1 Sp e-
cific Bin d in g in Th r ee r yr P r ep a r a tion s a n d of
Ca r d ia c Con tr a ctility of Ra t Mu scle Str ip s (F igu r e
5). The data in this study and a companion paper14
indicate little if any structural specificity of 1 and its
analogs in differentiating the [3H]-1 binding site of
cardiac ryr versus either rabbit skeletal muscle ryr (r
) 0.88, n ) 29) or mouse brain ryr (r ) 0.93, n ) 23).
These systems contain mixed receptor types: skeletal
muscle, two isoforms of ryr-1; brain, ryr-2 as the main
component with ryr-1 in parts of this organ; cardiac,
mainly ryr-2 making this organ source perhaps the most
homogeneous.7 Bound 1 is retained with a 76 kDa
fragment of peptide bearing the carboxyl terminus after
hydrolysis both with and without photoaffinity attach-
ment.22,23 Sequence studies show that this part of the
protein is largely conserved across the various receptors
so that it is unlikely that very large differences would
be found in the selectivity of ryanodine analogs acting
on ryr preparations from different organs.
The cardiac contractility assay14 is a physiological
function test with intact tissue and is therefore depend-
ent on both penetration to the site and potency at the
ryr. With the exception of five numbered compounds
and the 4,12-oxygen-bridged derivatives (see Figure
legend) which are at least 5-fold more potent on a
relative basis in the cardiac contractility than in the
cardiac ryr assay, there was a good correlation between
these two assays (r ) 0.85, n ) 15), consistent with
interaction of the ryanoids at the high-affinity 1 receptor
on sarcoplasmic reticulum calcium release channels
rendering them unable to enter normal open or closed
gating configurations. The negative inotropic effect was
slow in onset and progression consistent with the
intrinsic activity being associated with stabilizing the
calcium release channel in a subconducting state24 and
uncoupling the excitation-contraction process. Also
consistent with this mode of action was the ability of
these derivatives to increase internal [Ca2+] in electri-
cally stimulated cardiomyocytes isolated from adult
rats.25
Selected Modifications of the
Cyclohexane Ring of 2b
17 10-benzoyloxy
18 ∆8-10-hydroxy
19 ∆8-10-oxo
3.9
26
12
17
27
11
a
Values for 1 are as follows: IC50s of 3.5 and 3.3 nM for rabbit
muscle and mouse brain; Ki of 1.8 nM for cardiac ventricle
preparation; and IC50 of 34 nM for rat ventricular muscle strip
assay. Standard error values as percent of the mean of three to
five experiments (each in duplicate for receptor assays) were
independent of the compound and averaged 10, 12, 4, and 6 for
the ryr assays (rabbit muscle, mouse brain, and canine ventricle)
b
and ventricular strip assay, respectively. Data for 7, 8, 10, 12,
and 20 are given in the Supporting Information.
Ta ble 2. Defunctionalization of the Cyclohexane Ring of
Dehydroryanodine in Relation to Action at Calcium Release
Channels
activity relative to ryanodineb ) 100
ryanodine receptor
selected
modifications of
cyclohexane ring
rat
ventricular
rabbit mouse
canine
no.a
muscle brain ventricle strip assay
21 21-nor-9-oxo (EtB)
23 21-nor-9-oxo
13
10
8.3
7.1
6.5
2.7
0.42
5.9
26
28
24 21-nor-9ax-hydroxy
28 21-nor-10-deoxy-9ax
hydroxy
0.62
0.20
0.50
0.40
-
30 10-deoxydehydro
31 21-nor-10-deoxy
37
6.5
77
18
45
121
a
Data for 22, 25-27, 29, 32, and 33 are given in the Supporting
Information. Standard values for 1 are given in Table 1.
b
appears to be less active on cardiac preparations relative
to skeletal muscle ryr than the other 10-hydroxy vari-
ants.
Mod ifica tion s of th e Cycloh exa n e Rin g of 2 (17-
20) (Ta ble 1, F igu r e 3). The 10-benzoyloxy compound
17 is 4-17% as active as 1, falling in the same range
as 10-acetoxy-1.8 Introduction of unsaturation at C-8
reduces potency to about one-third, i.e. alcohols 18
versus 1 and ketones 19 versus 3.
Defu n ction a liza tion of th e Cycloh exa n e Rin g of
2 (21-33) (Ta ble 2, F igu r e 4). The target compounds
10-deoxy-2 (30) and 21-nor-10-deoxy-1 (31) were com-
pared with intermediates in their preparation from 2.
Replacing the 9-methylene of 2 with 9-oxo as in 21 and
23 reduces potency to about 10% and the epimeric
alcohols 24 and 25 from reduction of 23 are even less
active (except for 24 in the functional assay). 9ax-
Hydroxy-1 and 9ax-hydroxy-10-epi-1 occur naturally and
have low activity3,4 similar to that of compound 24,
which is a 21-nor analog. The 9ax-hydroxy compounds
Exp er im en ta l Section
Ch em istr y. Gen er al. The companion paper14 gives sources
of 1 and 2, the general methods of product isolation, and the
conditions for 1H and 13C NMR and HRMS. Reactions were
carried out on a semimicroscale with product isolation by
preparative TLC (0.25- or 0.5-mm silica gel F254) or rotary

Ryanodine Cyclohexane Substituents
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 12 2343
F igu r e 5. Correlation between relative potencies of ryanodine derivatives (1 is rated as 100) as inhibitors of [3H]-1 specific
binding at cardiac ryr versus either skeletal muscle and brain ryr (left) or cardiac contractility of rat muscle strips (right). All
data involving discrete potency values in this study and a companion paper14 are plotted with the exception of the following: 24
and 28 in this study; 18, 23, 27, and the 4,12-oxygen-bridged derivatives in the companion paper. Compounds without discrete
potency values had little or no activity in both the cardiac ryr and contractility assays. Correlation coefficients and n are as
follows: canine left ventricle cardiac ryr versus rabbit skeletal muscle ryr (left figure, upper correlation line) r ) 0.88 and n ) 29;
cardiac ryr versus mouse brain ryr (left figure, lower correlation line) r ) 0.93 and n ) 23; cardiac ryr versus rat strip contractility
assay (right figure) r ) 0.85 and n ) 15.
chromatography (1-mm silica gel GF254) as crystalline solids
or resins. HPLC (C-18 reverse phase silica column developed
with discriminating MeOH-water mixtures) was used to
verify the absence of 1 or 2 in the final samples. Sample
purities were estimated to be >98% based on NMR, TLC, and
HPLC. NMR data which differ little from the fully assigned
reference spectra26,27 are given only in abbreviated form.28
Mod ifica tion s of th e 10-Hyd r oxyl Su bstitu en t of 1 (3-
16). 10-Oxo-1 (3). Ryanodine (1, 30 mg, 0.061 mmol) in
DMSO (0.2 mL) and CH2Cl2 (2 mL) was added gradually at
-60 °C to a solution from addition of DMSO (0.2 mL, 2.6 mmol)
to CH2Cl2 (2 mL) and oxalyl chloride (25 mg, 0.20 mmol) in
CH2Cl2 (0.2 mL). After 10 min Et3N (0.2 mL, 1.45 mmol) was
added, and after a further 10 min the solution was allowed to
warm to ambient temperature. Saturated NaHCO3 aqueous
solution was added and the product isolated with ethyl acetate
and purified by rotary chromatography in CHCl3/MeOH (49:
1) to give 38 (20 mg, 67%) and 1 (7 mg, 23%) with CHCl3/MeOH
(9:1): 1H NMR (CD3OD) δ 7.04 (m, H-26), 6.86 (m, H-24), 6.24
(m, H-25), 5.69 (s, H-3), 2.62 (d, J ) 13.9 Hz) and 2.08 (d, J )
13.9 Hz) AB H2-14, 2.21 (m, H-13), 2.99 (m, H-9), 2.43 (dt, J )
13, 13, and 6 Hz, H-7ax), 1.44 (dd, J ) 13 and 6 Hz, H-7eq),
1.95 (m, H-8eq), 1.75 (dq, J ) 13, 13, 13, and 5 Hz, H-8ax), 1.25
(s, H3-17), 1.08 (d, J ) 6.8 Hz) and 0.74 (d, J ) 7.0 Hz) (H3-18
and -19), 0.93 (s, H3-20), 1.00 (d, J ) 7.0 Hz, H3-21); 13C NMR
(CD3OD) δ 66.0 (C-1), 82.9 (C-2), 90.8 (C-3), 91.5 (C-4), 49.5
(C-5), 84.0 (C-6), 26.7 (C-7), 31.9 (C-8), 41.4 (C-9), 215.0 (C-
10), 89.7 (C-11), 97.0 (C-12), 30.8 (C-13), 42.6 (C-14), 103.6 (C-
15), 11.0 (C-17), 18.7 and 19.3 (C-18 and -19), 12.5 (C-20), 13.7
(C-21), 161.0 (C-22), 123.2 (C-23), 116.9 (C-24), 110.9 (C-25),
125.6 (C-26).
10-epi-Hyd r oxy-1 (4). 10-Oxo-1 (3, 13 mg, 0.026 mmol)
was reduced with NaBH4 (20 mg, 0.48 mmol) in MeOH at -18
°C during 45 min. Acidification (HCl), extraction with ethyl
acetate, and methanolysis of the borate by distillation with
MeOH during 3 h gave a mixture, separated by rotary
chromatography with CHCl3/MeOH (19:1) to give 10-epi-1 (4,
10 mg, 77%): 1H NMR (CD3OD/CDCl3) δ 6.99 (m, H-26), 6.85
(m, H-24), 6.20 (m, H-25), 5.54 (s, H-3), 3.81 (d, J ) 3.0 Hz,
H-10), 2.0 (m, H-9), 2.47 (d, J ) 14.0 Hz) and 1.85 (d, J ) 14
Hz) AB H2-14, 2.17 (m, H-13), 2.04 (m, H-7ax), 2.50 (m, H-8ax),
2.35 (m, H-8eq), 1.29 (s, H3-17), 1.03 (d, J ) 6.8 Hz) and 0.73
(d, J ) 7.0 Hz) (H3-18 and -19), 0.85 (s, H3-20), 0.91 (d, J )
6.9 Hz, H3-21); 13C NMR (CD3OD) δ 65.2 (C-1), 83.4 (C-2), 89.7
(C-3), 91.3 (C-4), 49.5 (C-5), 85.2 (C-6), 26.6 and 23.6 (C-7 and
-8), 31.7 (C-9), 74.7 (C-10), 80.9 (C-11), 95.1 (C-12), 29.9 (C-
13), 40.2 (C-14), 102.0 (C-15), 10.4 (C-17), 18.8 and 18.2 (C-18
and -19), 11.5 (C-20), 17.1 (C-21), 161.5 (C-22), 121.9 (C-23),
116.5 (C-24), 110.4 (C-25), 124.9 (C-26). A later fraction of 1
(1 mg, 8%) was eluted with CHCl3/MeOH (17:3).
Oxim e a n d Hyd r a zon e Der iva tives of 10-Oxo-1 (5-8).
The oximes (5-7) were prepared by heating 3 in pyridine
solution with the hydroxyamine hydrochloride at 50 °C for 1-2
h and purification by rotary chromatography in CHCl3/MeOH
mixtures. Oxime 5 had Rf 0.50 (CHCl3/MeOH, 9:1): NMR
(CD3OD) as for 3 except 1H δ 2.15 (m), 1.90 (m), 1.66 (m) and
1.46 (m) (H2-7 and H2-8), 2.93 (sext, J ) 6.5 Hz, H-9), 1.31 (s,
H3-17), 0.93 (s, H3-20), 1.42 (d, J ) 7.2 Hz, H3-21); 13C δ 84.1
and 85.6 (C-6 and -11), 28.9 (C-8), 33.9 (C-9), 161.6 (C-10), 85.6
(C-11), 19.1 (C-21). The methoxime (6) had Rf 0.37 (CHCl3/
MeOH, 19:1): NMR spectra as for oxime 5 and 1H δ 1.36 (d,
J ) 7 Hz, H3-21), 3.84 (s, OMe); 13C NMR (CD3OD) δ 163.4
(C-10), 62.6 (OMe); HRMS (FAB) m/z C27H36N2O9H+ 521.2499,
found 521.2506. The benzyl oxime 7 had Rf 0.39 in CHCl3/
MeOH (19:1): NMR spectra as for oxime 5 except 1H δ 5.04
(d, J ) 12.5 Hz) and 5.08 (d, J ) 12.5 Hz) AB (CH2Ar), 1.39
(d, J ) 6.9 Hz, H3-21), 1.16 (s, H3-17), 0.73 (phenyl); 13C NMR
(CD3OD) δ 77.5, 129.0, 129.3, 129.5, and 138.7 (Ar). The
hydrazone 8 prepared with hydrazine at pH 8 was purified by
TLC using a silanized plate with CHCl3/MeOH (19:1): NMR
(CD3OD) as for oxime 5 except 13C δ 30.6 (C-9), 156.3 (C-10),
16.1 (C-21).
epi-Hyd r oxya m in es 9-12. Oxime 5 (28 mg, 0.055 mmol)
in MeOH (2 mL) was acidified (0.1% HCl) and kept at pH 3-4
(pH paper) while NaCNBH3 (30 mg, 0.48 mmol) was added
during 1 h. After 12 h, aqueous HCl was added and the
product isolated by extraction into ethyl acetate and distilla-
tion with MeOH (3 h). Hydroxyamine 9 (8.7 mg, 31%) was
isolated by TLC with CHCl3/MeOH/MeNH2 (83:14:3): Rf 0.28
1
in CHCl3/MeOH (9:1); NMR (CD3OD) as for 4 except H δ 2.97
(d, J ) 2.5 Hz, H-10), 2.20 (m, H-9), 1.03 (d, J ) 7.0 Hz, H3-
21); 13C δ 25.8 and 27.6 (C-7 and -8), 69.0 (C-10), 84.8 (C-11),
19.4 (C-21); HRMS (FAB) m/z C25H36N2O9Na+ 531.2319, found
531.2308. (Benzoyloxy)amine 10 was prepared in high yield
by reaction either with benzoyl chloride in THF/Et3N or with
N-(benzoyloxy)succinimide: HRMS (FAB) m/z C32H40N2O10Na+
635.2581, found 635.2590. Methoxyamine 11 (30%), Rf 0.31
in CHCl3/MeOH (19:1), was obtained in the same way from
the methoxime: NMR (CD3OD) as for hydroxyamine 9 except
1H δ 3.51 (s, OMe) 3.34 (d, J ) 3.9 Hz, H-10); 13C δ 66.6 (C-
10), 61.5 (OMe); HRMS (FAB) m/z C26H38N2O9H+ 523.2656,
found 523.2660. (Benzyloxy)amine 12 (Rf 0.27), prepared as
above (15%), was separated from the benzyl oxime (Rf 0.22)
by TLC with CHCl3/MeOH/MeNH2 (94:5:1): NMR (CD3OD)
as for hydroxyamine 9 except 1H δ 4.63 (d, J ) 14 Hz) and
4.70 (d, J ) 14 Hz) CH2Ar, 3.40 (d, J ) 3.7 Hz, H-10), 2.39 (d,
J ) 13.8 Hz) and 1.79 (d, J ) 13.8 Hz) AB H2-14, 7.28 (Ar);
HRMS (FAB) m/z C32H42N2O9H+ 599.2969, found 599.2979.
10-epi-Am in e 13 a n d Its Ben za m id e 14. 10-Oxo-1 (3, 20
mg, 0.041 mmol) and ammonium acetate (0.17 g, 2.2 mmol)

2344 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 12
J efferies et al.
in MeOH (3 mL) were set aside with NaCNBH3 (12 mg, 0.19
mmol) for 24 h. Excess borohydride was destroyed by acidi-
fication (HCl), and the amine borate was recovered from the
neutral solution with ethyl acetate, distilled with methanol
for 4 h, and purified by rotary chromatography with CHCl3/
MeOH/MeNH2 (83:14:3) to give 13 (10 mg, 53%): NMR (CD3-
143.4 (C-9), 66.4 (C-10), 41.9, 39.5, and 37.1 (C-14, 2 MeS),
108.4 (C-15), 110.9 (C-21); HRMS (FAB) m/z C29H40NO13BS2-
Na+ 708.1959, found 708.1944.
∆8-10-Hyd r oxy-1 (18) a n d ∆8-10-Oxo-1 (19). Dehydro-
ryanodine (2) was isomerized by heating at 50 °C in MeOH
with Pd/C for 5 h and the product29 (18, 74%) separated by
rotary chromatography with CHCl3/MeOH (18:1-17:3). Oxi-
dation of 18 (20 mg, 0.041 mmol) with oxalyl chloride (8 mg,
0.067 mmol) as for 1 and TLC with CHCl3/MeOH (19:1) gave
the ketone (19, 12 mg, 60%): NMR (CD3OD) as for 3 except
1H δ 2.72 (d, J ) 14 Hz) and 2.07 (d, J ) 14 Hz) AB H2-14,
2.95 (dq, J ) 18 and 2.5 Hz, H-7ax), 2.17 (obsc, H-7eq), 6.69 (m,
H-8), 1.31 (s, H3-17), 1.78 (m, H3-21); 13C NMR (CD3OD) δ 84.0
and 86.6 (C-6 and -11), 32.9 (C-7), 148.0 (C-8), 135.4 (C-9),
199.0 (C-10), 15.4 (C-21).
1
OD/CDCl3) as for hydroxyamine 9 except H δ 3.21 (d, J ) 3.7
Hz, H-10), 2.49 (d, J ) 13.7 Hz) and 1.84 (d, J ) 13.7 Hz) AB
H2-14, 0.94 (d, J ) 7.2 Hz, H3-21); 13C δ 30.8 (C-9), 57.5 (C-
10), 19.3 (C-21). Benzamide 14 prepared in THF with benzoyl
chloride/Et3N was purified by TLC with CHCl3/MeOH (19:1):
NMR (CD3OD) as for hydroxyamine 9 except 1H δ 7.71, 7.47
and 7.37 (Ar), 4.38 (d, J ) 3.70 Hz, H-10), 1.35 (s, H3-17), 0.82
(d, J ) 6.8 Hz, H3-21); 13C δ 29.6 (C-9), 53.3 (C-10), 131.8,
130.0, 128.7 and 126.9 (Ar); HRMS (FAB) m/z C32H40N2O9H+
597.2812, found 597.2827.
10-Oxo-2 (20). Dehydroryanodine (2, 30 mg, 0.061 mmol)
was oxidized in the same way as 1 using oxalyl chloride (7
mg, 0.056 mmol). The product was purified by TLC with
CHCl3/MeOH (19:1) containing 1% quinol and was stored
similarly in MeOH at -20 °C: NMR (CD3OD) as for 3 except
1H δ 5.30 (br s) and 5.85 (br s) H2-21, 2.75 (m), 2.35 (m), 2.15
(m), and 1.55 (m); 13C NMR (CD3OD) δ 84.0 and 83.3 (C-6 and
-11), 24.9 and 26.7 (C-7 and -8), 144.3 (C-9), 201.5 (C-10), 123.2
(C-21).
10-ep i-H yd r a zin e a n d It s 4-Azid ob en zoyl h yd r a zid e
(15). 10-Oxo-1 (3, 15 mg, 0.031 mmol) was added to hydrazine
(50 mg, 1.8 mmol) in MeOH at pH 7. After formation of the
hydrazone was complete (2 h), NaCNBH3 (25 mg, 0.40 mmol)
was added and the solution maintained at pH 5-6 for 12 h.
The solution was acidified (pH 1-2) and left for 3 h, basified
(NaHCO3), extracted with ethyl acetate, and deborated with
methanol as above to give crude hydrazine (Rf 0.49 in CHCl3/
MeOH/MeNH2, 78:19:3): NMR (CD3OD) as for hydroxyamine
9 except 1H δ 2.97 (d, J ) 3.8 Hz, H-10), 1.03 (d, J ) 7.0 Hz,
H3-21); 13C δ 67.2 (C-10). The benzoyl hydrazide, prepared as
for the benzamide 14, was purified by TLC with CHCl3/MeOH
(19:1), Rf 0.18: NMR (CD3OD/CDCl3) as for benzamide 14
except 1H δ 3.11 (d, J ) 3.5 Hz, H-10), 1.12 (d, J ) 7.2 Hz,
H3-21); HRMS (FAB) m/z C32H41N3O9H+ 612.2921, found
612.2948. The 4-azidobenzoyl hydrazide 15 was prepared in
the same way from the acid chloride and was purified twice
by TLC in CHCl3/MeOH (92:8). NMR as for the benzoylhy-
drazide except 1H 7.70, 7.46, and 7.41 (phenyl).
La cta m 16. The oxime 5 (10 mg, 0.02 mmol), TsCl (30 mg,
0.16 mmol), and pyridine (30 mg, 0.38 mmol) were dissolved
in THF (2 mL). An early faster spot on TLC (Rf 0.88 CHCl3/
MeOH (9:1)) was replaced with the lactam (Rf 0.45) overnight.
Isolation by addition of water, acidification, extraction with
ethyl acetate, and TLC in CHCl3/MeOH (9:1) gave the lactam
16 (5.7 mg, 57%): NMR (CD2OD) as for 3 except 1H δ 3.02
(ddq, J ) 11.5, 2.0, and 7.0 Hz, H-9), 1.58 (dt, J ) 13, 13, and
6 Hz, H-7ax), 1.80 (br q), 1.75 (dq, J ) 13, 13, 13, and 5 Hz,
H-8ax), 0.92 (s, H3-20), 1.09 (d, J ) 6.7 Hz, H3-21); 13C δ 30.0
(C-8), 38.7 (C-9), 181.2 (C-10), 18.2 (C-21).
Defu n ction a liza tion of th e Cycloh exa n e Rin g of 2 (21-
33). 21-Nor -9-oxo-1 (23) a n d Its Eth ylbor on a te (21). The
boronate (80 mg, 0.15 mmol) in pyridine (2 mL) was added to
OsO4 (60 mg, 0.24 mmol) in pyridine (1 mL). After 12 h
saturated aqueous solutions of NaHCO3 (2 mL) and NaHSO3
(2 mL) were added, and the solution was stirred for 1 h.
Acidification (HCl), extraction with ethyl acetate, and chro-
matography on silica (4 g) gave a trace of 1 (2 mg) with CHCl3/
MeOH (20:1), and then the diol(EtB) (38 mg, 45%) was
recovered with CHCl3/MeOH (93:7). A small amount of
deborated diol (5 mg, 6%) was eluted with CHCl3/MeOH (85:
15). The diol(EtB) above (32 mg, 0.057 mmol) in water (4 mL)
and THF (1 mL) were treated dropwise with stirring with 0.1
M NaIO4 (0.6 mL, 0.06 mmol). The product isolated with ethyl
acetate was purified by rotary chromatography using CHCl3/
MeOH mixtures to give ketol boronate 22 (26 mg, 94%): TLC
Rf 0.60 for CHCl3/MeOH, 9:1; 1H NMR (CD3OD) as for 2(EtB)
except δ 4.65 (s, H-10), 2.54 (d, J ) 13.9 Hz) and 1.97 (d, J )
13.9 Hz) AB H2-14, 2.75 (m, H-8ax), 2.40 (m, H-8eq), 2.20 (m,
H-7ax), 1.66 (ddd, J ) 1.0, 5.9, and 13.0 Hz, H-7eq), 0.87 (m)
and 0.75 (m) (EtBO2); 13C NMR (CD3OD) δ 44.9 (C-5), 85.2
(C-6), 25.6 (C-7), 34.3 (C-8), 209.7 (C-9), 72.5 (C-10), 90.0 (C-
11), 95.8 (C-12), 7.7 and 5.9 (br) (EtBO2).
Hydroxylation of 2 (50 mg, 0.096 mmol) as described above
gave the known diol8 with the modification that the pyridine
was removed from the aqueous layer and the latter extracted
five times with an equal volume of ethyl acetate. The organic
layers were combined, evaporated, and chromatographed on
silica using CHCl3/MeOH (93:7) to remove traces of 1 and then
CHCl3/MeOH (85:15) to give the diol (45 mg, 72%). Oxidation
with 1 molar equiv of 0.05 M periodate at 10 °C and TLC with
CHCl3/MeOH (80:20) gave the ketol 23 (Rf 0.60, 36 mg, 77%):
1H NMR (CD3OD) δ 7.05 (m, H-26), 6.88 (m, H-24), 6.24 (m,
H-25), 5.66 (s, H-3), 4.85 (s, H-10), 2.59 (d, J ) 14.0 Hz) and
1.94 (d, J ) 14.0 Hz) AB H2-14, 2.70 (m, H-8ax), 2.33 (m, H-8eq),
2.30 (m, H-7ax), 1.65 (dd, J ) 7.2 and 11.9 Hz, H-7eq), 2.25 (m,
H-13), 1.37 (s, H3-17), 1.09 (d, J ) 6.7 Hz) and 0.75 (d, J ) 6.3
Hz) (H3-18 and -19), 0.96 (s, H3-20); 13C NMR (CD3OD) 65.5
(C-1), 83.0 (C-2), 91.0 (C-3), 91.2 (C-4), 49.8 (C-5), 85.0 (C-6),
27.5 (C-7), 35.6 (C-8), 211.1 (C-9), 74.2 (C-10), 92.3 (C-11), 96.0
(C-12), 30.9 (C-13), 41.7 (C-14), 103.5 (C-15), 9.9 and 12.8 (C-
17 and -20), 18.8 and 19.4 (C-18 and -19), 161.5 (C-22), 123.2
Mod ifica tion s of th e Cycloh exa n e Rin g of 2 (17-20).
Deh yd r or ya n od in e 4,6-Di-O-et h ylb or on a t e [2(E t B)].
2
(50 mg, 0.10 mmol) in THF (2 mL) was treated with 1 M
lithium triethyl borohydride in THF (1 mL, 1.0 mmol) at 0 °C
for 10 min. The product isolated by addition of water,
acidification (HCl), and ethyl acetate extraction was set aside
in MeOH for 15 min to decompose complex boronates. TLC
with CHCl3/MeOH (9:1) (Rf 0.55) gave 2(EtB) (45 mg, 85%):
1
NMR (CD3OD/CDCl3) as for the methylboronate14 except H δ
0.91 (m) and 0.72 (m) (EtB); 13C NMR δ 8.2 and 7.1 (br) (EtB);
HRMS C27H36O9NBNa+ 552.2381, found 552.2389. The bor-
onate was hydrolyzed by slow distillation with MeOH during
3 h. Alcoholysis could be avoided by evaporation of hydroxylic
solvents at <30 °C.
Ester s fr om 2(EtB) In clu d in g th e 10-Ben zoa te 17.
2(EtB) (10 mg, 0.019 mmol) was treated with benzoyl chloride
(100 mg, 0.71 mmol) in pyridine (1 mL) for 24 h. The product
was deborated by slow distillation with MeOH (4 h) and
purified by TLC in CHCl3/MeOH (92:8) to give 17 (6 mg,
53%): NMR (CD3OD/CDCl3) as for 4 except 1H δ 5.49 (H-10),
4.71 and 4.81 (H2-21), 8.14, 7.56, and 7.45 (Ar). The 10-
tosylate (EtB) was prepared similarly during 4 days and
separated by TLC using CHCl3/MeOH (9:1): NMR as for
2(EtB) except 1H δ 5.49 (H-10), 4.88 and 4.87 (H2-21), 1.07 (s,
H3-17), 7.76, 7.27, and 2.16 (ArCH2); 13C δ 145.5 (C-9), 68.6
(C-10), 110.9 (C-21), 141.6, 133.7, 129.0, 127.7, 21.1 (ArCH2).
Mesylation under these conditions gave the 10,15-bis(mesy-
loxy) compound (45%): NMR as for 2(EtB) except 1H δ 5.38
(H-10), 5.18 and 5.04 (H2-21), 2.87 (d) and 2.63 (d) AB H2-14,
1.35 (s, H3-17), 3.14 (s) and 3.16 (s) 2 MeS; 13C δ 60.8 (C-1),
(C-23), 117.1 (C-24), 111.0 (C-25), 125.6 (C-26); HRMS C24H31
-
NO10Na+ 516.1846, found 516.1841.
21-Nor -9a x-h yd r oxy-1 (24) a n d Its Eth ylbor on a te (22).
The ketol boronate 21 (10 mg, 0.019 mmol) in THF (2 mL)
was treated at 0 °C with 1 M LiEt3BH in THF (0.3 mL, 0.3
mmol). After 10 min the solution was diluted with water,
acidified (HCl), extracted with ethyl acetate, and purified by
rotary chromatography with CHCl3/MeOH mixtures to give
the diol 22 (4 mg, 39%): NMR (CD3OD) as for 2(EtB) except
1H δ 3.99 (d, J ) 4 Hz, H-10), 3.96 (m, H-9), 2.65 (d, J ) 13.9

Ryanodine Cyclohexane Substituents
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 12 2345
Hz) and 1.97 (d, J ) 13.9 Hz) AB H2-14; 13C NMR (CD3OD) δ
21.1 and 28.4 (C-7 and -8), 66.8 (C-9), 72.7 (C-10), 87.0 (C-11),
8.0 and 6.7 (br) (EtBO2). Methanolysis of the boronate was
carried out by slow distillation with methanol containing
initially 5% MeNH2 (40%) during 4h. Chromatography on
silica (0.5 g) and elution with CHCl3/MeOH (9:1) containing
1% NH3 gave diol 24 (8 mg): 1H NMR (CD3OD) as for ketol
23 except δ 4.19 (d, J ) 3.9 Hz, H-10), 3.92 (m, H-9), 2.15 (dt,
J ) 13, 13, and 4.2 Hz, H-7ax), 1.27 (ddd, J ) 13.5, 3.0, and
1.0 Hz, H-7eq), 2.01 (m, H-8ax), 1.78 (m, H-8eq), 1.42 (s, H3-17);
13C NMR (CD3OD) 88.8 and 85.9 (C-6 and -11), 25.4 and 28.8
(C-7 and -8), 67.1 (C-9), 73.0 (C-10), 97.2 (C-12); HRMS (FAB)
m/z C24H33NO10Na+ 518.2002, found 518.1988.
ethylboronate with NaBH4 in MeOH at -20 °C and then
deboronation with MeOH/MeNH2 gave a similar mixture of
28 and 29.
10-Deoxy-2 (30). Ketone 27 (10 mg, 0.021 mmol) in THF
(0.2 mL) and toluene (0.5 mL) was treated with Tebbe reagent
(0.5 M in toluene) (0.2 mL, 0.10 mmol). After 15 min 0.25 M
NaOH (1 mL) and ether were added. After 15 min the pH
was adjusted to 10, and the ether extract was washed with
saturated NaHCO3, evaporated, and left for 12 h. The product
was separated from inorganic material by extraction into ether
and purified twice by TLC (2 × 0.5 mm plates) with CHCl3/
MeOH (9:1) solution to give 30 (6.4 mg, 63%): 1H NMR (CD3-
OD/CDCl3) δ 6.99 (m, H-26), 6.87 (m, H-24), 6.21 (m, H-25),
5.46 (s, H-3), 4.71 (q, J ) 1.9 Hz) and 4.64 (q, J ) 1.9 Hz)
(H2-21), 2.71 (br d, J ) 12.4 Hz, H-10ax), 2.43 (d, J ) 13.8 Hz)
and 1.82 (d, J ) 13.8 Hz) AB H2-14, 2.40 (br t, H-8ax), 2.20 (m,
H-13), 2.13 (m, H-8eq), 2.10 (br d, J ) 12.5 Hz, H-10eq), 2.05
(m, obsc, H-7ax), 1.36 (br dd, J ) 12.2 and 4.0 Hz, H-7eq), 1.29
(s, H3-17), 1.03 (d, J ) 6.7 Hz) and 0.75 (d, J ) 6.3 Hz) (H3-18
and -19), 0.85 (s, H3-20); 13C NMR (CD3OD) δ 64.9 (C-1), 83.0
(C-2), 91.3 (C-3), 91.4 (C-4), 48.3 (C-5), 86.0 (C-6), 27.5 and
29.6 (C-7 and -8), 149.5 (C-9), 35.2 (C-10), 83.8 (C-11), 94.4
(C-12), 29.8 (C-13), 40.8 (C-14), 101.5 (C-15), 10.5 (C-17), 18.3
and 18.8 (C-18 and -19), 12.3 (C-20), 110.7 and 110.6 (C-21
and -22), 122.0 (C-23), 116.8 (C-24), 110.6 (C-25), 125.1 (C-
26); HRMS (FAB) m/z C25H33NO8Na+ 498.2104, found 498.2120.
30 was also obtained in 25% yield by using CH2I2/Zn/TiCl4.30
30 was not available by treatment of 27 with the Wittig
reagent in DMSO.18
21-Nor -9eq-h yd r oxy-1 (25), 21-Nor -9-oxo-10-a cetoxy-1
(26), a n d 21-Nor -10-d eoxy-9-oxo-1 (27). The ketol 23 (58
mg, 0.12 mmol) was dissolved in acetic anhydride (0.5 mL, 0.53
mmol) and pyridine (0.05 mL). Monoacetylation was complete
in 70 min at ambient temperature. Excess acetic anhydride
was hydrolyzed by gradual addition of MeOH (1.5 mL) at 25
°C. The product isolated after acidification (HCl) and extrac-
tion with ethyl acetate was washed with saturated NaHCO3
solution and chromatographed on silica (5.0 g). Elution with
CHCl3/MeOH (97:3) gave the crystalline acetate 26 (54 mg,
84%): NMR (CD3OD) as for the ketol 23 except 1H δ 5.42 (s,
H-3), 5.85 (br s, H-10), 2.67 (d, J ) 14.1 Hz) and 1.89 (d, J )
14.5 Hz) AB H2-14, 2.70 (m, H-8ax), 2.35 (m, H-8eq), 1.49 (s,
H3-17), 0.90 (s, H3-20); 13C NMR (CD3OD) δ 47.7 (C-5), 85.0
and 88.9 (C-6 and -11), 26.2 and 35.2 (C-7 and -8), 205.0 (C-
9), 74.0 (C-10), 20.7 and 170.8 (acetate). The acetate 26 (16
mg, 0.030 mmol) in THF/MeOH (4:1, 5 mL) was treated
gradually at -60 °C under argon with 0.1 M SmI2 in THF (1
ml, 0.1 mmol). Reduction was fast, and the reaction was
repeated twice. Workup with saturated K2CO3 (2 mL) and
NaHSO3 (2 mL) solutions, extraction with ethyl acetate,
evaporation of the solvents, and separation from salts with
ethyl acetate gave the crude ketone. Purification was effected
by chromatography on silica (3.5 g). Elution with CHCl3/
MeOH (9:1) gave the ketone 27 (38 mg, 89%): NMR (CD3OD)
21-Nor -10-d eoxy-1 (31). Ketone 27 (3.8 mg, 0.0079 mmol)
in ether (2 mL) was treated with 1,2-ethanedithiol (0.075 mL,
0.80 mmol) and BF3‚Et2O (0.1 mL, 0.77 mmol). After 15 min
water was added, and the product was extracted with ether
and washed with saturated NaHCO3. The ether extract was
deborated by slow distillation (1 h) with MeOH/MeNH2 as
described above for 24 and stirred with Ni (∼100 mg) in EtOH
(2 mL) for 2 h. Addition of water, acidification (HCl), extrac-
tion with ethyl acetate, separation from Ni, and purification
by TLC with CHCl3/MeOH (9:1) gave 31 (2.4 mg, 66%): 1H
NMR (CD3OD/CDCl3) δ 6.97 (m, H-26), 6.85 (m, H-24), 6.21
(m, H-25), 5.46 (s, H-3), 2.43 (d, J ) 13.8 Hz) and 1.86 (d, J )
13.5 Hz) AB H2-14, 2.20 (m, H-13), 1.28 (s, H3-17), 1.02 (d, J
) 6.6 Hz) and 0.75 (d, J ) 6.3 Hz) (H3-18 and -19), 0.82 (s,
H3-20); 13C NMR (CD3OD/CDCl3) as for 30 except 25.1 and
26.7 (C-7 and -10), 19.3 and 20.6 (C-8 and -9), 83.6 (C-11), 94.7
(C-12); HRMS (FAB) m/z C24H33NO8Na+ 486.2104, found
486.2095. The sequence of these steps is important since the
borate complexes with Ni. Longer exposure to Ni gives a poor
yield of 31.
1
as for diol 24 except H δ 2.98 (d, J ) 15.3 Hz) and 2.23 (obsc)
H-10, 2.47 (d, J ) 14.1 Hz) and 1.90 (d, J ) 14.5 Hz) AB H2-
14, 2.33 (m, H-7ax), 1.89 (m, H-7eq), 2.55 (m, H-8ax), 2.35 (m,
H-7ax), 1.49 (s, H3-17); 13C NMR (CD3OD) δ 29.7 (C-8), 212.6
(C-9), 43.0 (C-10), 87.6 (C-11), 93.9 (C-12). Reduction of the
ketol (23, 48 mg, 0.097 mmol) as described above and TLC
using first CHCl3/MeOH (85:15) and then 8:2 gave the ketone
27 (10.5 mg, 23%), unreacted ketol (12 mg, 25%), and diol 25
(7 mg, 15%): NMR (CD3OD) as for ketol 23 except 1H δ 4.00
(d, J ) 8.9 Hz, Heq-10), 3.80 (m, Heq-9), 2.09 (dt, J ) 12.5, 12.5,
and 4.5 Hz, H-7ax), 0.89 (s, H3-20); 13C NMR (CD3OD) 86.0 and
88.0 (C-6 and -11), 25.6 and 26.8 (C-7 and -8), 72.5 and 73.0
(C-9 and -10); HRMS (FAB) m/z C24H33NO10Na+ 518.2002,
found 518.1988. Similarly acetylation of 21-nor-9-oxo-1(EtB)
(21) and reduction as above for 26 gave 21-nor-10-deoxy-9-
oxo-1(EtB) in 80% yield (reduced to 30% without the acetyla-
tion step).
21-Nor -10-d eoxy-6,9-oxid o-1 (32) a n d 21-Nor -10-d eoxy-
∆8/∆9-1 (33). The axial alcohol 28 (16 mg, 0.033 mmol) in
pyridine (2.0 mL) was treated dropwise with phosphoryl
chloride (0.2 mL, 2.1 mmol) at -20 °C. The solution was
warmed to ambient temperature (15 min) and the crude
product isolated by addition to water, acidification (HCl),
extraction with ethyl acetate and TLC with CHCl3/MeOH (19:
1) to give the ether 32 (5.8 mg, 38%): 1H NMR (CD3OD) as
for 30 except δ 4.50 (t, J ) 4.5 Hz, H-9), 2.44 (ddd, J ) 13.1,
4.0, and 2.0 Hz, H-10exo), 1.28 (m, H-10endo), 2.54 (d, J ) 13.8
Hz) and 1.77 (d, J ) 13.8 Hz) AB H2-14, 2.25 (obsc, H-7endo),
1.53 (ddd, J ) 12, 8, and 4 Hz, H-7exo), 1.62-1.75 (obsc, H2-8),
1.24 (s, H3-17); 13C NMR (CD3OD) δ 46.3 (C-5), 94.6 (C-6), 22.4
(C-7), 31.3 and 34.3 (C-8 and -10), 88.6 (C-11); HRMS (FAB)
m/z C24H31NO8Na+ 484.1947, found 484.1956. A less mobile
fraction (2.7 mg) appeared to be a mixture of 8- and 9-enes
(4:1) (33). Major component: 1H NMR (CD3OD) δ 5.64 (s, H-3),
5.55 (br m) and 5.65 (br m) H-8 and -9, 2.65 (d, J ) 13.8 Hz)
and 2.01 (d, J ) 13.5 Hz) AB H2-14, 2.69 (br m) H2-7 and H2-
10, 2.23 (m, H-13), 1.30 (s, H3-17), 1.08 (d, J ) 6.8 Hz) and
0.74 (d, J ) 7.0 Hz) H3-18 and -19, 0.94 (s, H3-20). Minor
component: 1H NMR (CD3OD) δ 6.0 (br m), 5.85 (br m), 5.63
(s), 1.32 (s), 0.96 (s); 13C NMR (CD3OD) 86.2 and 89.4 (C-6
and -11), 28.6 and 29.7 (C-7 and -10), 124.8 and 123.6 (C-8
and -9), 97.0 (C-12); HRMS (FAB) m/z C24H31NO8Na+ 484.1947,
found 484.1950.
21-Nor -10-d eoxy-9a x- a n d -9eq-h yd r oxy-1 (28 a n d 29).
Ketone 27 (28 mg, 0.058 mmol) in THF (2 mL) was treated
with ∼1 M LiAlH4 in ether (1.5 mL, ∼1.5 mmol). After 5 min
the product was isolated by addition of water, acidification
(HCl), and separation by TLC with CHCl3/MeOH (85:15) to
give the less polar 28 (Rf 0.43, 16 mg, 55%): NMR (CD3OD)
as for ketol 23 except 1H δ 4.0 (br m, W1/2h ) 9 Hz, H-9), 2.23
(m, H-13), 2.25 (dt, J ) 13, 13, and 5.0 Hz, H-7ax), 1.30 (m,
H-7eq), 1.95 (m, H-8ax), 1.70 (m, H-8eq), 1.32 (s, H3-17); 13C NMR
(CD3OD) 83.8 and 87.2 (C-6 and -11), 23.6 and 29.1 (C-7 and
-8), 68.2 (C-9), 32.3 (C-10), 11.2 (C-17); HRMS (FAB) m/z
C
24H33NO9Na+ 502.2053, found 502.2054. The more polar
fraction was the equatorial alcohol 29 (Rf 0.18, 7 mg, 15%):
1H NMR as for ketol 23 except δ 4.0 (tt, J ) 5.4 and 10.7 Hz,
H-9), 2.11 (dt, H-7ax), 1.33 (br d, J ) 12 Hz, H-7eq), 1.95-1.65
(m, H2-10 and H2-8), 0.89 (s, H3-20); 13C NMR (CD3OD) 92.5
(C-4), 83.3 and 87.4 (C-6 and -11), 26.1 and 29.8 (C-7 and -8),
68.3 (C-9), 35.2 (C-10), 11.2 (C-17); HRMS (FAB) m/z C24H33
-
NO9Na+ 502.2053, found 502.2064. Reduction of 27 as the

2346 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 12
J efferies et al.
(11) J efferies, P. R.; Casida, J . E. Ryanoid chemistry and action. In
Natural and Engineered Pest Management Agents; Hedin, P. A.,
Menn, J . J ., Hollingworth, R. M., Eds.; American Chemical
Society Symposium Series No. 551, American Chemical Soci-
ety: Washington, DC, pp 130-144.
(12) Gerzon, K.; Humerickhouse, R. A.; Besch, H. R., J r.; Bidasee,
K. R.; Emmick, J . T.; Roeske, R. W.; Tian, Z.; Ruest, L.; Sutko,
J . L. Amino- and Guanidinoacylryanodines: Basic Ryanodine
Esters with Enhanced Affinity for the Sarcoplasmic Reticulum
Ca2+-Release Channel. J . Med. Chem. 1993, 36, 1319-1323.
(13) Hutchins, R. O.; Natale, N. R. Cyanoborohydride. Utility and
Applications in Organic Synthesis. A Review. Org. Prep. Proc.
Int. 1979, 11, 201-244.
(14) J efferies, P. R.; Blumenkopf, T. A.; Gengo, P. J .; Cole, L. C.;
Casida, J . E. Ryanodine Action at Calcium Release Channels.
I. Importance of Hydroxyl Substituents. J . Med. Chem. 1996,
39, 2331-2338.
(15) Molander, G. A.; Hahn, G. Lanthanides in Organic Synthesis.
2. Reduction of R Hetero Substituted Ketones. J . Org. Chem.
1980, 51, 1135-1138.
(16) Kelly, R. B.; Whittingham, D. J .; Wiesner, K. The Structure of
Ryanodine. I. Can. J . Chem. 1951, 29, 905-910.
(17) Buist, G. J .; Bunton, C. A. The Mechanism of Oxidation of
R-Glycols by Periodic Acid. Part I. Ethylene Glycol. J . Chem.
Soc. 1954, 1406-1413.
(18) Kusuda, K.; Inanaga, J .; Yamaguchi, M. A Highly Efficient
Deoxygenation of R-Oxygenated Esters Via SmI2-Induced Elec-
tron Transfer Process. Tetrahedron Lett. 1989, 30, 2945-2948.
(19) Pine, S. H.; Shen, G. S.; Huang, H. Ketone Methylenation Using
the Tebbe and Witting ReagentssA Comparison. Synthesis
1991, 165-167.
(20) Hutchins, R. O.; Natale, N. R. Sodium Borohydride in Acetic
Acid. A Convenient System for the Reductive Deoxygenation
of Carbonyl Tosylhydrazones. J . Org. Chem. 1978, 43, 2299-
2301.
(21) Fieser, L. F. Preparation of Ethylenethioketals. J . Am. Chem.
Soc. 1954, 76, 1945-1947.
(22) Witcher, D. R.; McPherson, P. S.; Kahl, S. D.; Lewis, T.; Bentley,
P.; Mullinnix, M. J .; Windass, J . D.; Campbell, K. P. Photoaf-
finity Labeling of the Ryanodine Receptor/Ca2+ Release Channel
with an Azido Derivative of Ryanodine. J . Biol. Chem. 1994,
269, 13076-13079.
(23) Callaway, C.; Seryshev, A.; Wang, J .-P.; Slavik, K. J .; Needle-
man, D. H.; Cantu, C. III; Wu, Y.; J ayaraman, T.; Marks, A. B.;
Hamilton, S. L. Localization of the High and Low Affinity [3H]-
Ryanodine Binding Sites on the Skeletal Muscle Ca2+ Release
Channel. J . Biol. Chem. 1994, 269, 15876-15884.
(24) Sitsapesan, R.; Williams, A. J . Mechanisms of Caffeine Activa-
tion of Single Calcium-Release Channels of Sheep Cardiac
Sarcoplasmic Reticulum. J . Physiol. 1990, 423, 425-439.
(25) Watson, M. J .; Gengo, P. J . Effects of Ryanodine on Calcium
Homeostasis in Single Adult Rat Cardiomyocytes. Biophys. J .
1995, 68, A180.
(26) Waterhouse, A. L.; Holden, I.; Casida, J . E. Ryanoid Insecti-
cides: Structural Examination by Fully Coupled Two-dimen-
sional 1H-13C Shift Correlation Nuclear Magnetic Resonance
Spectroscopy. J . Chem. Soc., Perkin Trans. 1985, 1011-1016.
(27) Krishnamurthy, V. V.; Casida, J . E. COLOC-S: a Modified
COLOC Sequence for Selective Long-Range X-H Correlation 2D
NMR Spectroscopy. Magn. Reson. Chem. 1987, 25, 837-842.
(28) 1H and 13C signals that differ <0.03 and <1 ppm, respectively,
from the analogous signals of the reference compound are not
given.
Biology. Previously described procedures were used to
assay [3H]-1 binding in the ryr of rabbit skeletal muscle,3
mouse brain,9 and canine ventricle14 membranes. The struc-
ture-activity findings for ryr recognition were compared with
those for mediation of a pharmacological response measured
as inhibition of mechanical response to electrical stimulation
of cardiac muscle.14
Ack n ow led gm en t. We thank our University of
California laboratory colleagues Gary Quistad, J oyce
J ames and Motohiro Tomizawa for helpful discussions
and Elisabeth Lehmberg and Loretta Cole for receptor
assays with skeletal muscle and brain. The research
was greatly facilitated by the advice of Todd Blumen-
kopf of the Division of Chemistry of Burroughs Wellcome
Co. (currently at Pfizer Central Research, Groton CT).
This work was supported in part by National Institute
of Environmental Health Sciences Grant PO1 ES00049
and by a grant from Burroughs Wellcome Co., Research
Triangle Park, NC.
Su p p or tin g In for m a tion Ava ila ble: Extended Tables 1
and 2 with potency values for weakly active compounds (2
pages). Ordering information is given on any current mast-
head page.
Refer en ces
(1) Wiesner, K. The Structure of Ryanodine. Adv. Org. Chem. 1972,
295-316.
(2) Waterhouse, A. L.; Holden, I.; Casida, J . E. 9,21-Didehydro-
ryanodine: a New Principal Toxic Constituent of the Botanical
Insecticide Ryania. J . Chem. Soc., Chem. Commun. 1984, 1265-
1266.
(3) J efferies, P. R.; Toia, R. F.; Brannigan, B.; Pessah, I. N.; Casida,
J . E. Ryania Insecticide: Analysis and Biological Activity of 10
Natural Ryanoids. J . Agric. Food Chem. 1992, 40, 142-146.
(4) Welch, W.; Ahmad, S.; Airey, J . A.; Gerzon, K.; Humerickhouse,
R. A.; Besch, H. R., J r.; Ruest, L.; Deslongchamps, P.; Sutko, J .
L. Structural Determinants of High-Affinity Binding of Ryanoids
to the Vertebrate Skeletal Muscle Ryanodine Receptor:
A
Comparative Molecular Field Analysis. Biochemistry 1994, 33,
6074-6085.
(5) Humerickhouse, R. A.; Besch, H. R., J r.; Gerzon, K.; Ruest, L.;
Sutko, J . L.; Emmick, J . T. Differential Activating and Deacti-
vating Effects of Natural Ryanodine Congeners on the Calcium
Release Channel of Sarcoplasmic Reticulum: Evidence for
Separation of Effects at Functionally Distinct Sites. Mol.
Pharmacol. 1993, 44, 412-421.
(6) Pessah, I. N.; Waterhouse, A. L.; Casida, J . E. The Calcium-
Ryanodine Receptor Complex of Skeletal and Cardiac Muscle.
Biochem. Biophys. Res. Commun. 1985, 128, 449-456.
(7) Coronado, R.; Morrissette, J .; Sukhareva, M.; Vaughan, D. M.
Structure and Function of Ryanodine Receptors. Am. J . Physiol.
1994, 266, C1485-C1504.
(8) Waterhouse, A. L.; Pessah, I. N.; Francini, A. O.; Casida, J . E.
Structural Aspects of Ryanodine Action and Selectivity. J . Med.
Chem. 1987, 30, 710-716.
(9) J efferies, P. R.; Lehmberg, E.; Lam, W.-W.; Casida, J . E.
Bioactive Ryanoids from Nucleophilic Additions to 4,12-Seco-4,
12-dioxoryanodine. J . Med. Chem. 1993, 36, 1128-1135.
(10) Kahl, S. D.; McPherson, P. S.; Lewis, T.; Bentley, P.; Mullinnix,
M. J .; Windass, J . D.; Campbell, K. P. Radioimmunoassay for
the Calcium Release Channel Agonist Ryanodine. Anal. Bio-
chem. 1994, 218, 55-62.
(29) Ruest, L.; Taylor, D. R.; Deslongchamps, P. Investigation of the
constituents of Ryania speciosa. Can. J . Chem. 1985, 63, 2840-
2843.
(30) Hibino, J .; Okazoe, T.; Takai, K.; Nozaki, H. Carbonyl Methyl-
enation of Easily Enolizable Ketones. Tetrahedron Lett. 1985,
26, 5579-5580.
J M950712D