Synthesis of ring B unsaturated estriols. Confirming the structure of a diagnostic analyte for Smith-Lemli-Opitz syndrome

Article abstract of DOI:10.1021/ol016224j

(Equation presented) Brief partial syntheses are described for ring B unsaturated estriols, which are candidate metabolites diagnostic for Smith-Lemli-Opitz syndrome prenatally. These steroids are also likely metabolites of the Premarin preparation used in estrogen replacement therapy. Equilin (8) was converted in three steps to 7-dehydroestriol, which was isomerized to 8-dehydroestriol. The simplicity of the transformations belies the lability of these previously inaccessible metabolites and their synthetic precursors.

Full text of DOI:10.1021/ol016224j

ORGANIC
LETTERS
2001
Vol. 3, No. 16
2547-2550
Synthesis of Ring B Unsaturated
Estriols. Confirming the Structure of a
Diagnostic Analyte for
Smith−Lemli−Opitz Syndrome
,‡
Li-Wei Guo,† Cedric H. L. Shackleton,* and William K. Wilson†
Department of Biochemistry and Cell Biology, Rice UniVersity, Houston, Texas 77005,
and Childrens Hospital Oakland Research Institute, 5700 Martin Luther King Jr. Way,
Oakland, California 94609
cshackleton@chori.org
Received June 4, 2001
ABSTRACT
Brief partial syntheses are described for ring B unsaturated estriols, which are candidate metabolites diagnostic for Smith−Lemli−Opitz syndrome
prenatally. These steroids are also likely metabolites of the Premarin preparation used in estrogen replacement therapy. Equilin (8) was
converted in three steps to 7-dehydroestriol, which was isomerized to 8-dehydroestriol. The simplicity of the transformations belies the lability
of these previously inaccessible metabolites and their synthetic precursors.
The Smith-Lemli-Opitz syndrome1 (SLOS) is an autosomal
recessive disorder of cholesterol biosynthesis caused by a
defect in the enzymatic conversion of 7-dehydrocholesterol
to cholesterol (Scheme 1). This defect results in an abnormal
accumulation of 7- and 8-dehydrocholesterol (1 and 2), which
can be detected by GC/MS analysis of blood from affected
individuals.2 Prenatal detection is also important because
SLOS is a serious birth defect characterized by mental
retardation, multiple developmental malformations, and a
high carrier frequency.3 Prenatal diagnosis presently neces-
sitates amniocentesis or chorionic villus sampling.4 We are
developing less invasive diagnostic methods that target
aberrant metabolites recently found in urine and serum during
pregnancy.5 Validation of these methods requires authentic
standards of estrogen metabolites unsaturated in ring B (e.g.,
3 and 4). These dehydroestriols are also of interest as
candidate metabolites of equine steroids contained in Pre-
marin, which is widely used in estrogen replacement therapy.6
Equilin (8) and estriol (6) have been known since the
seminal isolation studies of estrogens in the 1930s,7 but
definitive syntheses of dehydroestriols have not been re-
ported. Although an early patent describes the Birch reduc-
† Department of Biochemistry and Cell Biology, Rice University.
‡ Childrens Hospital Oakland Research Institute.
(4) Kratz, L. E.; Kelley, R. I. Am. J. Med. Genet. 1999, 82, 376-381.
(5) (a) Shackleton, C. H. L.; Roitman, E.; Kratz, L. E.; Kelley, R. Prenat.
Diagn. 2001, 21, 207-212. (b) Shackleton, C. H. L.; Roitman, E.; Kratz,
L. E.; Kelley, R. I. J. Clin. Endocrinol. Metab. 1999, 84, 1157-1159. (c)
Shackleton, C. H. L.; Roitman, E.; Kratz, L. E.; Kelley, R. I. Steroids 1999,
64, 446-452.
(1) (a) Kelley, R. I.; Hennekam. R. C. M. J. Med. Genet. 2000, 37, 321-
335. (b) Waterham, H. R.; Wanders, R. Biochim. Biophys. Acta 2000, 1529,
340-356. (c) Porter, F. D. Mol. Genet. Metab. 2000, 71, 163-174. (d)
Battaile, K. P.; Steiner, R. D. Mol. Genet. Metab. 2000, 71, 154-162.
(2) (a) Tint, G. S.; Irons, M.; Elias, M. R.; Batta, A. K.; Frieden, R.;
Chen, T. S.; Salen, G. New Engl. J. Med. 1994, 330, 107-113. (b) Kelley,
R. I. Clin. Chim. Acta 1995, 236, 45-58.
(6) (a) Bhavnani, B. R. Proc. Soc. Exp. Biol. Med. 1998, 217, 6-16. (b)
Bhavnani, B. R.; Cecutti, A. J. Soc. Gynecol. InVestig. 1995, 2, 424 (Abstract
No. 415).
(3) Battaile, K. P.; Battaile, B. C.; Merkens, L. S.; Maslen, C. L.; Steiner,
R. D. Mol. Genet. Metab. 2001, 72, 67-71 and references therein.
(7) Fieser, L. F.; Fieser, M. Steroids; Reinhold Publishing Co.: New
York, 1959; pp 452-465.
10.1021/ol016224j CCC: $20.00 © 2001 American Chemical Society
Published on Web 00/00/0000

Scheme 1. Metabolism of Fetal ∆5,7 and ∆5,8 Sterols to
Scheme 2
Estriols
tion of 6,8-didehydroestriol 15 to 3 (identified based on its
UV spectrum),8 reductions of similar equilenins generally
give mixtures of olefins.9 Other synthetic efforts have been
thwarted by the facile aromatization of equilins.10 This Letter
describes the first synthesis of homogeneous samples of
dehydroestriols.
In our hands, bromination of equilin with CuBr2 in
MeOH10 also gave complex mixtures. 1D and 2D NMR
analysis of standards and crude reaction mixtures led to
identification of 10 bromosteroids, which, together with
equilenin and equilin, accounted for >95% of the steroids
observed in most reactions. Knowledge of reporter signals
for the numerous bromosteroids facilitated byproduct iden-
tification and optimization of reaction conditions. Thus, use
of 3 equiv of CuBr2 in methanol13 led mainly to bromination
in ring A, dehydrogenation to equilenins, and rapid conver-
sion of the desired 16-bromosteroids 9a and 9b to dibromides
and equilenins (Table 1, entries 1 and 2). Shorter reaction
times and a smaller excess of CuBr2 resulted in large amounts
of unbrominated steroids (8 and 10) and low conversion to
the desired products (Table 1, entries 3-5). In THF, the yield
of 16-bromoequilins doubled to 21%, but the product still
consisted mainly of equilenins and ring A brominated
equilins (Table 1, entry 6). However, reaction in CHCl3-
EtOAc17 gave >70% conversion to 16-bromoequilins (Table
1, entry 7), and these conditions were sufficiently reproduc-
ible to afford gram quantities of the desired products as a
2:1 mixture of 16R- and 16â-bromo epimers. Under specific
reaction conditions,13a this mixture was cleanly hydrolyzed
to 16R-hydroxyequilin (7) without formation of 16-keto
byproducts.18 Reduction of 7 with NaBH4 led to the target
7-dehydroestriol (3).19
With the intention of synthesizing 8-dehydroestriol (4) by
a parallel bromination-hydrolysis-reduction scheme, we
prepared 8-dehydroestrone (12) by isomerizing equilin with
LiNHCH2CH2NH2 in ethylenediamine (Scheme 3).20 How-
ever, refluxing 12 with CuBr2 in CHCl3-EtOAc resulted in
virtually no bromination at C-16 or in ring A, the product
consisting of a 1:1:2 mixture of 10, 12, and 9(11)-dehy-
droestrone. An alternative attempt to prepare 13 by LiNHCH2-
CH2NH2 isomerization of 7 gave none of the expected
products.21 However, isomerization of triol 3, which lacks
the potentially labile13 16,17-ketol functionality of 7, was
Our primary target was the preparation of compounds 3
and 4. Considering the modest amounts of material required
for bioanalytical purposes, we focused on partial synthesis
from available estrogens, such as equilin, equilenin (10), and
the Torgov diene.11 In a standard synthetic approach to
estriols, the 16-hydroxyl is introduced by acid hydrolysis of
a 16R,17R-epoxide formed from the enol acetate of estrone.12
However, application of this method to equilin resulted in
aromatization to 16R-hydroxyequilenin.10 Another approach
to 16-hydroxylation entails 16-bromination of estrone, fol-
lowed by hydrolysis in DMF to the ketol.13 Although many
unsaturated 17-ketosteroids can be selectively brominated
at C-16 with CuBr2 in refluxing methanol,13,14 this reaction
was reported to give a complex mixture for equilin.10 Aiming
to overcome these problems, we set out to devise conditions
for implementing the simple ring D manipulations shown in
the retrosynthetic analysis (Scheme 2) without triggering the
indicated side reactions, namely, aromatization of ring B,15
epimerization at C-14,7 double-bond isomerization,7,16 and
ketol rearrangement.13a
(8) Marshall, D. J. U.S. Patent 3 470 159, 1969.
(9) Marshall, D. J.; Deghenghi, R. Can. J. Chem. 1969, 47, 3127-3131.
(10) Ikegawa, S.; Kurosawa, T.; Tohma, M. Steroids 1990, 55, 250-
255.
(11) 3-Methoxyestra-1,3,5(10),8,14-pentaen-17-one. For leading refer-
ences, see: Tanaka, K.; Nakashima, H.; Taniguchi, T.; Ogasawara, K. Org.
Lett. 2000, 2, 1915-1917.
(12) Leeds, N. S.; Fukushima, D. K.; Gallagher, T. F. J. Am. Chem.
Soc. 1954, 76, 2943-2948. Estrones can also by hydroxylated at C-16
microbiologically: Pan, S. C.; Principe, P. A.; Junta, B. U.S. Patent 3,-
431,174, 1969.
(13) (a) Numazawa, M.; Nagaoka, M. J. Org. Chem. 1982, 47, 4024-
4029. (b) Numazawa, M.; Kimura, K.; Nagaoka, M. Steroids 1981, 38,
557-565.
(14) Glazier, E. R. J. Org. Chem. 1962, 27, 4397-4399.
(15) Junghans, K. Chem. Ber. 1975, 108, 2824-2826.
(16) Jacquesy, J. C.; July, G.; Gesson, J. P. C. R. Acad. Sci., Ser. C
1972, 274, 969-971 and references therein.
(17) King, L. C.; Ostrum, G. K. J. Org. Chem. 1964, 29, 3459-3461.
Org. Lett., Vol. 3, No. 16, 2001
2548

Scheme 3a
Table 1. Bromination of Equilin (8) with CuBr2: Effects of
Reaction Conditions on Product Distributiona
a (a) CuBr2, CHCl3-EtOAc 1:1, reflux; (b) NaOH, DMF-H2O
3:1, rt, 1.5 h; (c) LiNHCH2CH2NH2, NH2CH2CH2NH2, 30 °C, 1 h;
(d) NaBH4, MeOH, 0 °C, 2 h; (e) LiNHCH2CH2NH2, NH2CH2-
CH2NH2, 40 °C, 48 h.
1
a Product distributions were determined by H NMR. Desired products
phase HPLC afforded 4, 14, and 15, which were character-
ized by 2D NMR and NOE difference spectroscopy to
confirm the regio- and stereochemical structure assignments.
With availability of authentic samples of the dehydroestri-
ols, we compared their GC mobilities and mass spectral
fragmentation with those of the SLOS urinary metabolites24
(9a and 9b) are highlighted. b Chloroform-ethyl acetate 1:1. c Molar ratio
of CuBr2 to 8.
more promising. Despite the poor solubility of 3 and its
sluggish rate of isomerization, reaction conditions were found
to give 4 as the major product.22,23 Semipreparative reverse-
indicated a 9:5:3:3 mixture of 4, 3, 14, and 15. Preparative HPLC (250 ×
21.2 mm C18 column, MeOH-H2O 35:65, 7 mL/min) gave 14 (tR 146 min),
15 (tR 159 min), 3 (tR 165 min), and 4 (tR 172 min). NMR (CD3OD, 25
°C): 3 δ 0.636 (s), 3.578 (d, 5.4 Hz), 4.043 (ddd, 9.1, 5.4, 2.0 Hz), 5.358
(br d, 3.4 Hz); 4 δ 0.757 (s), 3.549 (d, 5.3 Hz), 4.119 (ddd, 9.0, 5.3, 1.8
Hz); 14 δ 0.946 (s), 3.560 (d, 6.9 Hz), 3.999 (ddd, 8.8, 7.9, 6.9 Hz); 15 δ
0.667 (s), 3.668 (d, 5.6 Hz), 4.228 (ddd, 9.1, 5.6, 2.1 Hz).
(23) The forcing isomerization conditions resulted in partial epimerization
of 4 to 14 and prompted a thorough structure of determination of all products
by NMR. In contrast, byproducts were negligible in the preparation of 12
(3% 3, 1% 6-dehydroestrone, and 1-2% 14â steroids) and are frequently
absent in base-catalyzed olefin isomerizations: Pines, H.; Stalick, W. M.
Base-catalyzed reactions of hydrocarbons and related compounds; Aca-
demic Press: New York, 1977; Chapter 2.
(18) Equilin (101 mg) and freshly ground CuBr2 (166 mg) were heated
in CHCl3-EtOAc (25 mL each) for 2 h under vigorous reflux (to remove
HBr). The crude product (140 mg; 48% 9a, 24% 9b) was stirred for 1.5 h
at room temperature in DMF-water (3:1, total 10 mL) containing 2 equiv
of NaOH. MPLC on silica gel (EtOAc-hexane 3:7) gave 7 (78 mg,
containing some ∆6,8 material). Attempts to purify 7 by reverse-phase HPLC
(MeOH-H2O 35:65) gave a 19:1 mixture of 7 and its ∆6,8 analogue.
Consequently, the equilenins formed during bromination and hydrolysis were
removed after reduction of 7 to 3. NMR (CDCl3, 25 °C): 9a δ 0.814 (s),
4.589 (d, 7.3 Hz), 5.449 (m); 9b δ 0.995 (s), 4.263 (t, 8 Hz), 5.500 (m); 7
δ 0.865 (s), 4.420 (d, 8.3 Hz), 5.509 (m).
(19) Ketol 7 (446 mg) was reduced with NaBH4 (47 mg) in MeOH (25
mL) for 2 h at 0 °C. Methanol was removed at <20 °C in a stream of N2
(higher temperatures resulted in formation of 15). Addition of cold saturated
NH4Cl (10 mL) followed by extraction with EtOAc gave 3 (452 mg, 81%
purity). HPLC purification (250 × 21.2 mm C18 column, MeOH-H2O
45:55) of a 50-mg sample gave 3 (35 mg, 99% purity).
(24) Isolation of dehydroestriols: Urine from a pregnant woman
carrying an SLOS fetus was processed by our standard methods for
analyzing urinary steroids: Shackleton, C. H. L. J. Steroid Biochem. Mol.
Biol. 1993, 45, 127-140. Briefly, steroid sulfates and glucuronides from a
C18 solid-phase extraction (SPE) of the urine sample were hydrolyzed with
(20) Raijmakers, P. H. U.S. Patent 5,739,363, 1998.
Helix pomatia (Roman snail) digestive juice (Sigma-Aldrich). The resulting
unconjugated steroids were reextracted by SPE and fractionated on Sephadex
LH-20 (100 × 10 mm column; cyclohexane-ethanol 4:1) as described:
Setchell, K. D.; Shackleton, C. H. L. Clin. Chim. Acta 1973, 47, 381-388.
GC/MS analysis of individual 5-mL fractions (as TMS ethers) revealed
that dehydroestriol was eluted between 140 and 165 mL. The only two
steroids found in this fraction were dehydroestriol and didehydroestriol 15.
(21) Reaction of 7 (50 mg) with 0.3 M Li in ethylenediamine (3 mL)
gave mainly unreacted 7, whereas 1 M Li led to a complex mixture.
(22) Dehydroestriol 3 (100 mg) was heated at 40 °C for 48 h in
ethylenediamine (3.4 mL) containing LiNHCH2CH2NH2 (prepared by
adding 13.4 mmol of MeLi-LiBr in ether to ethylenediamine, followed by
evaporation of the ether at 55 °C). NMR of the crude product (104 mg)
Org. Lett., Vol. 3, No. 16, 2001
2549

Figure 1. Comparison of GC/MS spectra of tris-TMS derivatives of authentic 7-dehydroestriol (A), 8-dehydroestriol (B), and the
dehydroestriol isolated from urine (C). The molecular ions are at m/z 502, and major fragments are formed by losses of trimethylsilanol
(-90) and methyl groups (-15). GC retention times for A, B, and C were 18.77 ( 0.03 min.
(Figure 1). Although the TMS ethers of 7- and 8-dehy-
droestriols coeluted on the nonpolar column used and shared
the same parent and fragment ions, the isomers could be
distinguished by the relative abundance of these ions. The
steroid isolated from urine had the abundance profile of
8-dehydroestriol. This finding does not completely exclude
production and excretion of 7-dehydroestriol by SLOS
patients since only a few affected individuals have so far
been studied. In addition, 7-dehydroestriol is less stable and
may undergo aromatization to the didehydroestriol (15) found
in urine.
will facilitate the establishment of routine noninvasive
prenatal diagnosis for SLOS.
Acknowledgment. This work was supported by Grants
HL-49122 and R03HD-39707 from the National Institutes
of Health. We are grateful to Dr. Richard Kelley for
providing the urine samples and to Ms. Esther Roitman for
valuable technical assistance.
Supporting Information Available: Tables of NMR
signal assignments, NOE difference results, HR-MS data,
1
conformational analysis of 4 and 14, and H NMR spectra
of 3, 4, 14, and 15. This material is available free of charge
via the Internet at http://pubs.acs.org.
In conclusion, we have developed simple and efficient
methods for preparing estrogen metabolites unsaturated in
ring B. The availability of these reference dehydroestriols
OL016224J
2550
Org. Lett., Vol. 3, No. 16, 2001