Synthesis of tetracyclic heterocompounds as selective estrogen receptor modulators. Part 2. Process improvement for scale-up of 2,5,8-substituted 11,12-dihydro-5H-6,13-dioxabenzo[3,4]cyclohepta-[1,2-a]naphthalene derivatives

Article abstract of DOI:10.1021/op700061x

An improved, reproducible nonchromatographic process for scale-up synthesis of 2,5,8-substituted 11,12-dihydro-5H-6,13-dioxabenzo[3,4]cyclohepta[1,2-a] naphthalene derivatives as selective estrogen receptor modulators (SERMs) is described. The titled comp

Full text of DOI:10.1021/op700061x

Organic Process Research & Development 2007, 11, 731−738
Synthesis of Tetracyclic Heterocompounds as Selective Estrogen Receptor
Modulators. Part 2. Process Improvement for Scale-Up Of 2,5,8-Substituted
11,12-Dihydro-5H-6,13-dioxabenzo[3,4]cyclohepta-[1,2-a]naphthalene
Derivatives
Xun Li,*,† Michael Reuman,† Ronald K. Russell,† Scott Youells,† Sandra Beish,† Zhiyong Hu,† Shawn Branum,†
Nareshkumar Jain,‡ and Zhihua Sui‡
Johnson & Johnson Pharmaceutical Research & DeVelopment, L.L.C., U.S. Research & Early DeVelopment,
1000 Route 202, Raritan, New Jersey 08869, U.S.A.
Abstract:
hundred gram quantities of racemates 11 and 19 were
requested for in vivo biological evaluations. The original
Discovery preparation for 11 was a 9-step, low yielding
(1.1%) synthesis that had four steps with yields in the range
of 35-45%. These low yields were in part due to the need
for silica gel chromatographic purification for intermediates
3a, 4a, 6, and 7 as well as the additional purification of
compound 11 since it was contaminated with tetrabutylam-
monium ion, a byproduct from the tetrabutylammonium
fluoride (TABF) cleavage of the Si-O ether bond in the
last step (Table 1, Discovery route). The above-mentioned
issues required a fresh look at the synthesis before starting
a large-scale production of compounds 11 and 19. It was
believed that the synthetic methods used to prepare com-
pounds I5 could also be used for the production of the seven-
member analogue II. However, the structural difference in
a single methylene group on the C-ring of the compound II
afforded an unexpected number of synthetic challenges.
Herein, we report our results for the improved large-scale
preparation of racemates 11 and 19.
An improved, reproducible nonchromatographic process for
scale-up synthesis of 2,5,8-substituted 11,12-dihydro-5H-6,13-
dioxabenzo[3,4]cyclohepta[1,2-a]naphthalene derivatives as se-
lective estrogen receptor modulators (SERMs) is described. The
titled compounds were prepared in 9-21% overall yield with
high chemical purity (>97%) after nine consecutive synthetic
steps.
Introduction
Tetracyclic heterocompounds I (where R ) H, CH3, or
COC(CH3)3; R1 ) C6H4OCH2CH2N(CH2)5 or other structur-
ally similar substituents) have attracted our interest since they
represent a novel class of compounds with biological
propertiesasselectiveestrogenreceptormodulators(SERMs).1-3
The development of nonchromatographic processes for
Results and Discussion
During our previous scale-up preparation of the six-
member C-ring chromene derivatives I, 3-(2,4-dimethoxy)-
7-methoxycoumarin 3a was identified as the best candidate
and converted to its corresponding 4-bromomethyl analogue
in quantitative yield with high chemical purity under anionic
bromination conditions (LHMDS/NBS); however, 3a was
obtained in only 23% yield after four steps.1,4 This anionic
chemistry could also be applied to help construct the seven-
member C-ring of compounds II by means of trapping a 3a
the scale-up production of 2,5,8-substituted 5,11-dihydro-
chromeno[4,3-c]chromene derivatives (I) (n ) 1) was
reported in our previous work.4,5 Recently the 2,5,8-
substituted 11,12-dihydro-5H-6,13-dioxabenzo 3,4]cyclo-
hepta[1,2-a]naphthalene series II (n ) 2), the C-ring
homolog to compounds I, was also identified as a new class
of highly potent in vitro SERM candidates, and multi-
+
anion with a carbon electrophile (such as a POCH2 , P )
protecting group) to give one carbon hydroxyl-protected
homologue 4-(2-hydroxyethyl)coumarins 4a. The develop-
ment of a one-step synthesis of 3a with higher yield was
necessary to benefit the first non-GMP campaign for race-
mate 11 (g250 g) as well as the second campaign for the
racemate 19 (g250 g). To achieve this immediate goal,
4-methoxy-2-hydroxyacetophenone (1a) was treated with
2,4-dimethoxyphenylacetic acid (2) under Perkin condensa-
tion conditions to afford compound 3a in a slightly improved
yield (36%) after a chromatographic purification.4 Further-
more, when starting material 1a was replaced with the
4-benzyloxy analogue 1b, the desired 7-benzyloxycoumarin
3b was isolated in g60% yield after crystallization of the
crude reaction mixture from 2-propanol (IPA) (step 1 of
Scheme 1).
* Author for correspondence. Telephone: (908) 707-3321. Fax: (908) 526-
6469. E-mail: xli6@prdus.jnj.com.
† High Output Synthesis.
‡ Reproductive Therapeutics.
(1) (a) Chen, N.; Jain, N.; Xu, J.; Reuman, M.; Li, X.; Russell, R. K.; Sui, Z.
Tetrahedron Lett. 2006, 47, 5909. (b) Jain, N.; Xu, J.; Sui, Z. WO2006055694,
2006. (c) Jain, N.; Kanojia, R. M.; Xu, J.; Gou, J.-Z.; Pacia, E.; Lai, M.-T.;
Du, F.; Musto, A.; Allan, G.; Hahn, D.; Lundeen, S.; Sui, Z. J. Med. Chem.
2006, 49, 3056.
(2) Kanojia, R. M.; Jain, N.; Xu, J.; Sui, Z. Tetrahedron Lett. 2004, 45, 5837.
(3) Kanojia, R. M.; Jain, N. F.; Ng, R.; Sui, Z.; Xu, J. WO2003053977, 2003.
(4) (a) Li, X.; Jain, N.; Russell, R. K.; Ma, R.; Branum, S.; Xu, J.; Sui, Z. Org.
Process Res. DeV. 2006, 10, 354. (b) Horva´th, A.; De Smet, K.; Ormerod,
D.; Depre´, D.; Pe´rez-Balado, C.; Govaerts, T.; Van den Heuvel, D.;
Schorpion, I. Org. Process Res. DeV. 2005, 9, 356.
(5) Li, X.; Reuman, M.; Russell, R. K.; Adams, R.; Ma, R.; Beish, S.; Branum,
S.; Youells, S.; Roberts, J.; Jain, N.; Kanojia, R.; Sui, Z. Org. Process Res.
DeV. 2007, 11, 414-421, part 1 of this publication.
10.1021/op700061x CCC: $37.00 © 2007 American Chemical Society
Published on Web 06/21/2007
Vol. 11, No. 4, 2007 / Organic Process Research & Development
•
731

Table 1. Summary of the reaction conditions and yields for the preparation of compounds 11 and 19
route
scale-up
scale-up
scale-up
step
1
discovery routea
(TBDMS route)b
(benzyloxy route) c
(methoxy route for 19)d
1a, Ac2O, Et3N,
148 °C, 48 h; 36% of 3a
1b, Ac2O, Et3N, 148 °C,
48 h; 60% of 3b
2
3a, (TMS)2NLi, SEM-Cl,
THF, -20 °C, 1 h; 45% of 4a
3b, (TMS)2NLi, MOM-Br,
THF, -20 °C, 1 h; 95% of 4b
3
4a, BBr3, CH2Cl2,
20 °C, 36 h
4b, BBr3, CH2Cl2,
36-38 °C, 24 h;
>95% of 5
90-99% of 5
(workup by addition
of water to the reaction
mixture at -76 °C)
(workup by addition
of the boron complex
into EtOH at 0 °C,
or to water at 20 °C)
4
5
5, Ph3P, DIAD THF,
20 °C,18 h; 42% of 6
5, Ph3P, DIAD THF, 20 °C,
18 h; 97% of 6
6, TBDMS-Cl, Et3N,
CH2Cl2, 0-20 °C,
24 h; 35% of 7
6, TBDMS-Cl, Et3N, CH2Cl2,
0-20 °C, 24 h; 60% of 7
6, BnBr, K2CO3, CH2Cl2,
38-40 °C, 24 h; 92% of 7a
6, MeI, K2CO3, DMF,
20 °C, 4 h; 85% of 14
6
7
7, DIBALH, CH2Cl2,
-20 °C, 2 h; 95% of 8
8, 12, nBuLi, THF,
-78 °C, 1 h
7, DIBALH, CH2Cl2,
-20 °C, 2 h; 99% of 8
8, 12, nBuLi, THF, -78 °C, 1 h
7a, DIBALH, CH2Cl2,
-40 °C 1 h; 99% of 8a
8a, 12, nBuLi, THF, -78 °C, 1 h
14, DIBALH, CH2Cl2,
-76 °C, 6 h; 73% of 15
15, 12, nBuLi,
THF, -78 °C, 8 h;
95% of 18
8
9
9, HCl, CH2Cl2,
0-20 °C, 1 h
9, HCl, CH2Cl2, 0-20 °C, 1 h
9a, HCl, toluene 20 °C, 1 h
18, HCl, CH2Cl2,
10-15 °C, 45 min;
56% of 19 (2 steps)
10, TBAF, THF, 20 °C,
18 h; 50% of 11 (3 steps)
10, TBAF, THF, 20 °C,
18 h, 72% of 11 (3 steps)
10a, Pd/C, H2, 3.06 atm,
EtOAc/CH3OH, 20 °C,
20 h; 41% of 11 (3 steps)
a Five chromatographic purifications were required (steps 1, 2, 4, 5, and 9). b One chromatographic purification was required (step 9). c Zero chromatographic
purification was required. d Zero chromatographic purification was required.
Scheme 1
Originally, the one carbon homologation on the 4-methyl
group was accomplished by treating 3a anion with SEM-
Cl, which resulted in a moderate isolated yield (45%) of
compound 4a after chromatography. When SEM-Cl was
replaced with the more active alkylating reagent, MOM-
Br,1a,6 a doubled isolated yield (95%) of 4b was obtained in
high chemical purity (96%, HPLC area %) without chro-
matographic separation (scaled-up TBDMS route, step 2 of
732
•
Vol. 11, No. 4, 2007 / Organic Process Research & Development

Scheme 2
Scheme 1). During pilot production, the unstable and
carcinogenic MOM-Br was replaced with the stable and less
harmful pivaloyoxymethyl chloride (POM-Cl), which also
produced the POM-homologated compound in high yield
(93%) and good chemical purity (85.4%).4b
The Discovery deprotection of 4a using BBr3 (6 equiv)
in CH2Cl2 was complete after 36 h at room temperature,7
and then addition of water (g12 mL/g of 4a) to the reaction
mixture at -76 °C afforded >95% isolated yield of crude
tetraol 5. The scaled-up deprotection of 4b with reduced
amounts of BBr3 (4 equiv) in CH2Cl2 was complete in 24 h
at 36-38 °C; however, less than 10% of crude 5 was isolated
after acidic aqueous-organic solvent extraction, when a large
volume of water (g80 mL/g of 4b) was used in the workup.
The tetraol 5 was identified as a major component in the
aqueous phase (pH e 3), which was very difficult to extract
into any organic solvent (such as CH2Cl2, EtOAc, and n-butyl
alcohol) even when the aqueous phase pH was e1 or
saturated with solid NaCl. This problem was resolved by
the isolation of the crude boron complex of 5 as a solid and
then careful portionwise addition of this crude complex to
EtOH (200 proof) at 0 °C. Ethanol removal in vacuo and
repeated chasing of the resulting material with MeOH (×4)
produced tetraol 5 in quantitative yield with 86% chemical
purity (HPLC area %). Another improvement was realized
during the preparation of racemate 19, where the slurry of
crude boron complex 5 was directly quenched into a limited
volume of water (g25 mL/g but e57 mL/g of 4b), and the
resulting aqueous phase was allowed to stand at least 20 h;
the desired tetraol 5 was isolated as a crystalline solid in
72-86% yield with good purity (>80%, HPLC area %).
The tetraol 5 was converted to 2,8-bisphenol 6 using
modified Mitsunobu cyclization conditions (DIAD and Ph3P
in THF)8 to afford cyclic 2,8-bisphenol 6 in 97% isolated
yield with good quality (82%, HPLC area %). The protection
of the 2,8-bisphenol groups of 6 was best accomplished with
2.3 equiv of TBDMSCl in an acceptable yield (60%) of 2,8-
bis-silyl lactone 7 as a yellowish solid without chromato-
graphic purification (scaled-up TBDMS route, steps 4 and
5 of Scheme 1).9
developed conditions to give a 138% recovery yield of crude
diol 9, which contained about 30% of 1-(2-phenoxyethyl)-
1
piperidine (12a) as determined by H NMR and HPLC
analyses (the isolated yield >100% of theory was due to
the crude material containing solvent residues and other
uncharacterized impurities. The yields were calculated based
on pure products). The crude diol 9 was cyclized to 2,8-bis-
silyloxy benzopyran derivative 10 in quantitative yield, and
at this point, 12a was removed as its HCl salt during workup.
Following the Discovery procedure, the silyl protecting
groups of 10 were cleaved using TBAF (2 equiv) in THF to
afford crude 11 (158% isolated yield) after workup, which
contained 84% (HPLC area %) of 11.10 However, an
approximate 8% of tetrabutylammonium group (TBA+) was
observed in the 1H NMR, an impurity that was not reported
in the original preparation. It was found impossible to remove
this impurity from compound 11 by either organic acid-
base aqueous solvent partition or crystallization. This result
was due to the reaction of the phenolic groups of 11 (pKa of
2-OH ) 10.19 and pKa of 8-OH ) 9.88, unpublished internal
communication) with the tetrabutylammonium group to
create a salt that is soluble in both organic and aqueous
solvents. Instead of using TBAF, several other reagents (such
as KF/HBr/DMF, CBr4/CH3OH, and/or HCl/CH3OH) were
investigated for this step, but either an incomplete reaction
or a complicated mixture was observed. This problem was
resolved by converting crude 11 to its 2,8-dibenzyloxy 10a
(R2 ) Bn) or 2,8-bis-acetoxy 10b (R2 ) Ac) derivatives,
followed by organic/aqueous partitioning, and then cleavage
of the protecting groups to give a pure 11. This protection/
deprotection process was not used on scale due to limited
time and other results (vide supra). It was realized that when
the reaction was run with less than 1 equiv of TBAF (0.6
equiv) at a longer reaction time the tetrabutylammonium
group contamination of crude 11 was minimized to (2%,
which was removed by chromatographic purification to
afford 11 (85% isolated yield; 94%, HPLC area %) on 250-g
scale. After solving this issue, monobromo compound 13
was identified as another major impurity (3.1%) that
presented in the above-obtained 11 (Scheme 3) by 1H NMR
and LC/MS spectroscopic analyses. The batch of 12 used
for this step was checked and found to contain ∼3% of the
undesired dibromo compound 12b. The side chain 12 was
simply redistilled to remove the dibromo side product 12b.
Finally, the impurity 13 could be cleanly removed by
debromination of the mixture under a catalytic hydrogenation
The addition of freshly prepared Grignard reagent of 12
to lactone 7 was unsuccessful (cf. the six-member C-ring
series I5), probably due to the high oxidation level on
2-carbonyl and/or carbonyl conjugation with aromatic A- and
D-rings that deactivates its electrophilicity. Lactone 7 was
therefore reduced to lactol 8, a compound with low oxidation
level at C-2 where nucleophilic acceptability is also in-
creased, with DIBALH (1.2 equiv) in a quantitative yield
with high chemical purity (96%, HPLC area %). Similar to
the 6-member C-ring process,5 an excess of lithium reagent
12c (Scheme 2) was added to lactol 8 under the well-
(6) (a) Fieser, M. Fieser and Fieser’s Reagents for Organic Synthesis; John
Wiley & Sons: New York, 1982; Vol. 10, pp 56-57. (b) Williams, R. M.;
Im, M.-N.; Cao, J. J. Am. Chem. Soc. 1991, 113, 6976.
(7) Suzuki, A.; Hara, S. In Encyclopedia of Reagents for Organic Synthesis;
Paquette, L. A. Ed.; John Wiley & Sons: New York, 1995; Vol. 1, pp
645-648.
(8) Mitsunobu, O.; Wada, M.; Sano, T. J. Am. Chem. Soc. 1972, 94, 679.
(9) Greene, T. W.; Wuts, P. G. M. ProtectiVe Groups in Organic Synthesis,
3rd ed.; John Wiley & Sons: New York, 1999; pp 273-275.
(10) Li, H.-Y. In Encyclopedia of Reagents for Organic Synthesis; Paquette, L.
A. Ed.; John Wiley & Sons: New York, 1995; Vol. 7, pp 4728-4732.
Vol. 11, No. 4, 2007 / Organic Process Research & Development
•
733

Scheme 3
of them provided a clean reduction. Optimized DIBALH
reduction conditions were developed where the reaction
temperature was rigorously controlled below -70 °C to
achieve a high conversion (95%) of 14 after 6 h to the desired
lactol 15 in 73% (HPLC area %), while the formation of
side products 16 and 17 were minimized to 17% and 5%,
respectively (Scheme 4). This crude 15 was used in next
step without further purification.
Scheme 4
to afford highly pure (>97%, HPLC area %) of the desired
11 in quantitative isolated yield (Scheme 3).
Since the 2,8-dibenzyloxy compound 10a could be
converted to 11 by hydrogenolysis without over-reduction,
a non-TABF process was developed to prepare compound
11 (Scheme 1).11 2,8-Bisphenol 6 was converted to 2,8-
dibenzyloxy derivative 7a in quantitative isolated yield, after
treatment with benzyl bromide and K2CO3 in CH2Cl2.
DIBALH reduction of 7a resulted in quantitative preparation
of the desired lactol 8a, which was further treated with the
lithium reagent 12c to produce the crude diol 9a. Acid-
catalyzed ring-closure of 9a afforded compound 10a, which
after removal of the benzyl protecting groups under catalytic
hydrogenation conditions (5% Pd/C in MeOH)11,12 and
crystallization from EtOAc and IPA (9:1) afforded compound
11 in 41% (over three steps starting from compound 6) with
>97% chemical purity. Since this benzyloxy route was more
cost effective and also provided more chemically robust
intermediates (7a, 8a, 9a, and 10a), it was recommended to
the pilot plant for the GMP campaign of compound 11.
The second campaign was changed to prepare more than
250 g of 2,8-dimethoxy racemic 19, due to the in vitro
screening results which concluded that compound 19 was
the preferred candidate for the advanced biological studies
in vivo. To achieve this goal, cyclic 2,8-bisphenol 6 was
prepared as before from tetraol 5 under Mitsunobu cycliza-
tion conditions in quantitative yield with 88% chemical purity
(HPLC area %). Since the tetraol 5 used in this campaign
was prepared by the crystallization from water, it was
necessary to determine the contents of boronic acid and/or
water in 5 by combustion analysis before use in this reaction,
because the success of this Mitsunobu cyclization reaction
was water- and B(OH)3 dependent. If tetraol 5 was contami-
nated with a high residue of either boron (>2-3%, 1.0 equiv
of B(OH)3) or water (>6.8%, 1.3 equiv of water) this
cyclization failed. The 2,8-bisphenol 6 was converted to its
2,8-bis-methoxy lactone 14, which was isolated as a solid
by filtration in 85% yield and high chemical purity (96%,
HPLC area %) (Scheme 1).
Addition of the side chain to the crude lactol 15 was also
accomplished using the lithium reagent 12c. This addition
required at least 2.0 equiv of the lithium reagent and was
conducted below -70 °C in THF. After the workup, it was
difficult to remove the excess des-bromo 12a from diol 18.
A method was developed where the crude product mixture
was partitioned between CH3CN and heptane; des-bromo 12a
was extracted into the heptane phase, and the more polar
diol product 18 remained in the CH3CN phase. This method
afforded crude diol 18 (>70% chemical purity, plus 17% of
16, 5% of 17, and 8% of unidentified unknowns; HPLC area
%) as a thick amber oil that was used in the next step without
further purification.
Acid-catalyzed B-ring closure of crude diol 18 was
accomplished with concentrated HCl in CH2Cl2, which
resulted in near a quantitative isolation of crude 19 that was
with 80% chemical purity. Crude 19 was purified by the same
partition method as applied to diol 18 to afford the desired
product 19 in 97% chemical purity (HPLC area %) and a
moderate yield (67%), since some of 19 was extracted into
the heptane layer. This 97% pure material was further
partitioned between heptane and CH3CN to afford the final
product 19 (99.5%) in 56% isolated yield over two steps.
Conclusions
A reproducible process was developed for the scale-up
synthesis of compound 11 in 21.4% isolated yield over nine
steps with >97% chemical purity and required only one
chromatographic purification to remove tetrabutylammonium
impurity from the crude product 11. The desired compound
19 was also prepared by the nonchromatographic “methoxy
route” process in reproducible 41% isolated yield and
>99.5% chemical purity. And finally, the benzyl-protected
series provided the best impurity-free, nonchromatographic
route, which was used for large-scale GMP preparation of
compound 11 after some additional improvements were made
during the pilot production.
When the reduction of lactone 14 to lactol 15 was done
under conditions (DIBALH/CH2Cl2 at -20 °C) similar to
those used to prepare lactol 8 in the first campaign, the
desired lactol 15 was obtained only in 18% yield along with
two unexpected major side products 16 (19%) and 17 (52%).
Several solvent and temperature combinations, as well as
nonconventional reducing agents were investigated, but none
(11) Greene, T. W.; Wuts, P. G. M. ProtectiVe Groups in Organic Synthesis,
3rd ed.; John Wiley & Sons: New York, 1999; pp 76-86.
(12) (a) Heathcock, C. H.; Ratcliffe, R. J. Am. Chem. Soc. 1971, 93, 1746. (b)
Schmidhammer, H.; Brossi, A. J. Org. Chem. 1983, 48, 1469.
Experimental Section
Starting materials, reagents, and solvents were obtained
from commercial suppliers and were used without further
734
•
Vol. 11, No. 4, 2007 / Organic Process Research & Development

purification, except for 4-bromophenol that was used to
quenched with saturated NH4Cl (2.0 L) and extracted with
EtOAc (2.0 L). The organic phase was condensed in vacuo
at 60 °C to give 395.0 g of crude product, which was
crystallized in IPA/EtOAc (1400 mL/200 mL). The solid was
collected by filtration and dried in an oven under house
vacuum (∼120 mmHg) at 60 °C for 18 h. There was obtained
341.0 g (89% isolated yield; 96% HPLC area %) of coumarin
4b as a slightly tan solid. Recrystallization of the material
from the mother liquor (∼160.0 g) again in IPA/EtOAc (1000
mL/200 mL) afforded an additional 24.4 g (6.4%) of pure
compound 4b. 1H NMR (300 MHz, CDCl3) δ 2.76-3.0 (m,
2 H), 3.19 (s, 3 H), 3.46 (t, J ) 6.8, 2 H), 3.74 (s, 3 H), 3.86
(s, 3 H), 5.15 (s, 2 H), 6.55 (s, 1 H), 6.57 (dd, J ) 0.8, 8.1,
1 H), 6.88-6.98 (m, 2 H), 7.08 (d, J ) 8.4, 1 H), 7.32-
7.56 (m, 5 H), 7.61 (d, J ) 8.3, 1 H). LC/MS m/z 447 (MH+),
469 (MNa+).
prepare side-chain 12. All the melting points are uncorrected
1
and determined on a MEL-TEMP 3.0 apparatus. H NMR
spectra were recorded at 300 MHz on a Bruker Avance-300
instrument, and mass spectra were recorded on an Agilent
Series 180 LC/MS instrument (positive/negative modes). The
chemical purity was determined on an Agilent Series 1100
system at UVmax ) 254 and 340 nm, using a ZORBAX
Eclipse XDB-Phenyl column (4.6 mm ID × 5 cm, 3.5 µ) at
40 °C with flow rate of 1.0 mL/min and run time of 10.0
min. Solvent system: A 80% H2O + 0.1% TFA; B 20%
CH3CN. Gradient: B 20%/0.0 min, B 20%/1.0 min, B 90%/
6.0 min, B 90%/8.0 min, B 55%/9.0 min, B 20%/10.0 min.
Rochelle’s solution (40%, wt/vol) was prepared by dissolving
potassium sodium tartrate tetrahydrate (×400 g) in deion-
ization water (D.I. water, × 1 L) at 20 °C.
All reactions were carried out in a four-neck round-bottom
flask (RBF, 1-22 L), equipped with a thermocouple control-
ler, an overhead mechanical stirrer, a condenser, and a
pressure-equalization addition funnel and nitrogen inlet/outlet
whenever they were required.
3-(2,4-Dihydroxyphenyl)-7-hydroxy-4-(2-hydroxyethyl)-
2H-chromen-2-one (5). A 12-L RBF was charged with CH2-
Cl2 (4.8 L) and compound 4b (250.0 g, 0.56 mol). The
solution was stirred at 20 °C under N2, BBr3 (99+%, 561.2
g, 2.24 mol) was added via a double-tipped needle under
very mild N2 pressure over a 20-min period (this was a mildly
exothermic process; the final internal temperature was 34
°C). This mixture was gradually heated to 38 °C and gently
refluxed for 20 h. The reaction was monitored by LC/MS
and HPLC. The reaction was cooled to 20 °C, and the solid
was filtered under N2 atmosphere (This operation must be
done under N2 atmosphere in a hood with good Ventilation,
due to the presence of excess BBr3 that will decompose to
HBr when it meets with atmospheric moisture in the
atmosphere. The filtrate must be cooled to 0 °C and, with
stirring under N2, quenched with IPA (>500 mL) dropwise.
This quenched solution is then stirred at 20 °C for 18 h,
before disposal.), washed with CH2Cl2 (1.0 L × 2), and dried
under N2 for 20 min. Another 12-L RBF was charged with
EtOH (200 proof, 2.0 L) and cooled to 0 °C in an ice bath.
The above solid was added portionwise over a 10-min period
with fast agitation (this was an exothermic process; the final
internal temperature was ∼26 °C), and the cherry solution
was stirred at 20 °C for 18 h. The solvent was concentrated
at 65 °C under house vacuum and then under high vacuum
(10 mmHg); the resulting material was redissolved in MeOH
(×4) and concentrated until the solid reached a constant
weight. There was obtained 182.5 g (104% isolated yield)
of crude tetraol 5 with 86% chemical purity (HPLC, area
%) as a foamy solid. This crude material was used in the
7-(Benzyloxy)-3-(2,4-dimethoxyphenyl)-4-methyl-2H-
chromen-2-one (3b). A 5-L Morton flask was charged with
2,4-dimethoxyphenylacetic acid (2) (254.0 g, 1.29 mol, 98%),
o-xylene (1.0 L), Et3N (310.0 mL, 2.18 mol), and Ac2O
(145.0 mL, 1.41 mol) and was stirred at 20 °C for 30 min.
2-Hydroxy-4-benzyloxyacetophenone (1b) (314.0 g, 1.29
mol, 99%) was added to the reaction mixture and brought
to reflux (150 °C) for 50 h, while Et3N (1033 mL, 5.0 equiv)
and Ac2O (655 mL, 5.0 equiv) were added continuously via
syringe pumps in a ratio of ∼1:2 (vol/vol) over the same
time period (50 h); meanwhile, ∼50 mL of distillate was
removed via a Dean Stark trap at intervals every 60 min
during the daytime. The progress of the reaction was
monitored by LC/MS and HPLC. After completion, the
mixture was cooled to 20 °C, the solvents were removed by
distillation under high vacuum (2-4 mmHg), and the
resulting crude semisolid was crystallized from IPA (3.0 L)
to afford 307.2 g (60%) of 7-benzyloxycoumarin 3b as a
brown solid with 99% (HPLC, area %) chemical purity,
which was used in the next step without further purification.
1H NMR (300 MHz, CDCl3) δ 2.20 (s, 3 H), 3.76 (s, 3 H),
3.84 (s, 3 H), 5.16 (s, 2 H), 6.56 (s, 1 H), 6.58 (dd, J ) 1.0,
8.1, 1 H), 6.88-6.98 (m, 2 H), 7.09 (d, J ) 8.4, 1 H), 7.32-
7.54 (m, 5 H), 7.56 (d, J ) 8.4, 1 H). LC/MS m/z 403 (MH+),
425 (MNa+).
7-(Benzyloxy)-3-(2,4-dimethoxyphenyl)-4-(2-methoxy-
ethyl)-2H-chromen-2-one (4b). A 5-L RBF was cooled to
-20 °C (dry ice/acetone) and charged with lithium bis-
(trimethylsilyl)amide ((TMS)2NLi, 1.12 L, 1.12 mol, 1.0 M
in THF). Compound 3b (345.0 g, 0.858 mol) in anhydrous
THF (1.5 L) was added to the reaction mixture over a 10-
min period and stirred at -20 °C for 45 min. Bromomethyl
methyl ether (MOMBr, 90%, 125.0 mL, 1.12 mol) was added
to the reaction mixture over a 10-min period (there was a
mild exotherm, the final internal temperature was 0 °C), and
stirring was continued at -20 °C for 1 h (the reaction was
monitored by LC/MS and HPLC). The reaction mixture was
1
next step without further purification. H NMR (300 MHz,
CD3OD) δ 2.84-3.06 (m, 2 H), 3.56-3.78 (m, 2 H), 4.88
(br s, 4 H), 6.41 (dd, J ) 0.9, 8.2, 1 H), 6.43 (s, 1 H), 6.77
(d, J ) 1.0, 1 H), 6.82-6.93 (m, 2 H), 7.74 (d, J ) 8.3, 1
H). LC/MS m/z 315 (MH+), 337 (MNa+).
An Alternative Workup Process. The completed reac-
tion mixture of compound 4b (355.1 g, 0.795 mol) with BBr3
(1074.6 g, 4.29 mol) in CH2Cl2 (8 L) was quenched by
cautiously transferring the material to a 22-L separatory flask
containing water (8 L) via a 1/4-in. Teflon tube. After the
transfer was complete, the CH2Cl2 layer was removed, and
the water layer was filtered into a 19-L filter flask. The
Vol. 11, No. 4, 2007 / Organic Process Research & Development
•
735

separatory flask was rinsed with hot water (2 × 500 mL);
these additional extracts were added to the 19-L flask, and
the combined mixture was allowed to crystallize overnight.
The solid was collected by filtration and dried under house
vacuum at 45 °C to give 178.7 g (72% isolated yield) of the
tetraol 5 as a pale-yellow solid with 72% chemical purity
(HPLC, area %).
at 60 °C for 18 h. There was obtained 230.0 g (39% isolated
yield; 96% HPLC area %) of the desired 2,8-bissilyl lactone
7 as a yellowish fine powder. The mother liquor was
concentrated to give a 610.0 of thick oil which contained
1
∼25-30% of lactone 7 as determined by H NMR. Chro-
matographic purification (SiO2, 2.0 kg, eluted with EtOAc/
hexane (0/100 to 20/80)) of this oil gave crude 7 (321.0 g),
which was slurried in pentane (260 mL) to give an additional
2,8-Dihydroxy-11,12-dihydro-6,13-dioxabenzo[3,4]-
cyclohepta[1,2-a]naphthalene-5-one (6). A 12-L RBF was
charged with tetraol 5 (251.7 g, 0.802 mol) and anhydrous
THF (4.0 L). This suspension was cooled to -5-0 °C,
diisopropyl azodicarboxylate (DIAD, 664.5 mL, 3.206 mol)
was added over a 35-min period, and the mixture was stirred
at -5 °C for 30 min followed by the addition of triphen-
ylphosphine (841.0 g, 3.206 mol) solution in THF (1.6 L)
over a 30-min period. After the addition, the mixture was
warmed to 20 °C and stirred for 18 h, and the progress of
the reaction was monitored by LC/MS and HPLC. The
solvent was concentrated in vacuo at 60 °C, and the resulting
material was dissolved in CH2Cl2 (6.0 L) and washed with
2 N NaOH solution (4 L, 2 L, 1 L). The aqueous phases
were combined and back-extracted with CH2Cl2 (1.6 L). This
alkaline phase was cooled to 0 °C, acidified to pH ∼1-2
with concentrated HCl solution (37%, ∼1.8 L) (the final
internal temperature was 16 °C), and the resulting slurry was
stirred at 10 °C for 1 h. The solid was collected by filtration,
and the filter cake was washed with D.I. H2O (500 mL × 5,
or until pH ∼6-7). This solid was dried by air-suction for
18 h and then placed in a vacuum oven at 70 °C under house
vacuum for 72 h until a constant weight was achieved. There
was afforded 230.7 g (97.2% isolated yield, 82% by HPLC
area %) of 2,8-bisphenol lactone 6 as a yellow-green solid.
1H NMR (300 MHz, DMSO-d6) δ 2.94 (t, J ) 5.9, 2 H),
4.56 (t, J ) 5.8, 2 H), 6.52 (d, J ) 0.9, 1 H), 6.68 (dd, J )
0.8, 8.2, 1 H), 6.78 (d, J ) 1.0, 1 H), 6.86 (dd, J ) 0.9, 8.8,
1 H), 7.44 (d, J ) 8.3, 1 H), 7.81 (d, J ) 8.4, 1 H), 9.7 (s,
1 H), 10.6 (s, 1 H). LC/MS m/z 297 (MH+), 319 (MNa+).
2,8-Bis(tert-butyldimethylsilyloxy)-11,12-dihydro-6,13-
dioxabenzo[3,4]-cyclohepta[1,2-a]naphthalene-5-one (7).
A 12-L RBF was charged with CH2Cl2 (2730 mL) and 2,8-
bisphenol lactone 6 (82%, 411.2 g, 1.389 mol). This
suspension was stirred at 20 °C for 5 min, and then Et3N
(449.2 mL, 3.223 mol) was added over a 5-min period. tert-
Butyldimethylsilyl chloride (TBDMS-Cl, 97%, 466.3 g 3.001
mol) was added portionwise (40-g portions at 5-min inter-
vals) over a 60-min period while the reaction temperature
was maintained between 23 and 30 °C. After the addition,
the mixture was stirred at 20 °C for 20 h. The reaction was
1
124.9 g (21.2%) of pure 7. H NMR (300 MHz, CDCl3) δ
0.24 (s, 6 H), 0.27 (s, 6 H), 1.0 (s, 18 H), 2.93 (t, J ) 6.0,
2 H), 4.68 (t, J ) 6.1, 2 H), 6.66 (d, J ) 1.1, 1 H), 6.78 (dd,
J ) 0.9, 8.1, 1 H), 6.82 (dd, J ) 0.9, 8.2, 1 H), 6.88 (s, 1
H), 7.52 (d, J ) 8.4, 1 H), 7.69 (d, J ) 8.4, 1 H). LC/MS
m/z 525 (MH+), 547 (MNa+).
2,8-Dibenzyloxy-11,12-dihydro-6,13-dioxabenzo[3,4]-
cyclohepta[1,2-a]naphthalene-5-one (7a). Compound 7a
(8.9 g, 92% isolated yield; 90% HPLC, area %) was prepared
in a similar manner as 7 from 2,8-bisphenol 6 (6.0 g, 20.3
mmol; 89%) with benzyl bromide (2.0 equiv) and K2CO3
1
(4.4 equiv) in CH2Cl2 at 38 °C for 24 h to afford 7a. H
NMR (300 MHz, CDCl3) δ 2.96 (t, J ) 5.8, 2 H), 4.68 (t,
J ) 5.9, 2 H), 5.10 (s, 2 H), 5.15 (s, 2 H), 6.78 (d, J ) 0.8,
1 H), 6.91 (dd, J ) 0.9, 8.3, 1 H), 6.92-7.01 (m, 2 H),
7.29-7.50 (m, 10 H), 7.56 (d, J ) 8.4, 1 H), 7.72 (d, J )
8.3, 1 H). LC/MS m/z 477 (MH+), 499 (MNa+).
2,8-Dimethoxy-11,12-dihydro-6,13-dioxabenzo[3,4]-
cyclohepta[1,2-a]naphthalene-5-one (14). Compound 14
(72.8 g, 85% isolated yield; 96%, HPLC area %) was
prepared in a similar manner as 7 from 2,8-bisphenol 6 (78.0
g, 0.263 mol) with iodomethane (3.0 equiv) and K2CO3 (4.0
equiv) in DMF at 20 °C for 4 h to afford 14. 1H NMR (300
MHz, DMSO-d6) δ 2.98 (t, J ) 5.9, 2 H), 3.81 (s, 3 H),
3.90 (s, 3 H), 4.64 (t, J ) 6.0, 2 H), 6.75 (d, J ) 0.9, 1 H),
6.85 (dd, J ) 0.9, 8.0, 1 H), 7.01 (dd, J ) 0.9, 8.1, 1 H),
7.08 (d, J ) 0.8, 1 H), 7.59 (d, J ) 8.3, 1 H), 7.93 (d, J )
8.4, 1 H). LC/MS m/z 325 (MH+), 347 (MNa+).
2,8-Bis(tert-butyldimethylsilyloxy)-11,12-dihydro-5H-
6,13-dioxabenzo[3,4]-cyclohepta[1,2-a]naphthalene-5-ol (8).
A 12-L RBF was charged with CH2Cl2 (4.0 L) and lactone
7 (360.0 g, 0.687 mol); the solution was stirred under
nitrogen and cooled to -20 °C in a dry ice/acetone bath.
Diisobutylaluminum hydride (DIBALH, 1.0 M in CH2Cl2,
823.0 mL, 0.823 mol) was added dropwise over a 45-min
period, while the reaction temperature was maintained at -20
°C. After the addition, the reaction was stirred at -20 °C
for 1.5 h, and the reaction was monitored by LC/MS, TLC
1
(EtOAc/hexane, 2/8), and H NMR analyses. The reaction
was quenched with Rochelle’s solution (2.4 L, 40% of
potassium sodium tartrate tetrahydrate in D.I. water) and
transferred to a 22-L three-neck separatory flask. Additional
Rochelle’s solution (13.0 L) was added, and the mixture was
agitated at room temperature for 2 h. After phase separation,
the aqueous phase was extracted with CH2Cl2 (1.6 L × 2),
and the combined organic phases were washed with Roch-
elle’s solution (3.0 L) and brine (3.0 L). The organic phase
was concentrated in vacuo at 30 °C, and the resulting material
was placed in a vacuum oven (under 10 mmHg) at 60 °C
for 18 h. There was obtained 368.0 g (101% isolated yield,
1
monitored by LC/MS and H NMR. The reaction mixture
was transferred to a 12-L three-neck separatory flask, washed
with 0.1 N HCl (1.36 L × 2), saturated NaHCO3 (1.36 L),
and brine (1.36 L). The solvent was concentrated in vacuo
and kept under high vacuum (10 mmHg) at 60 °C for 1 h.
The resulting semisolid material (791.0 g, 108% recovery
yield) was triturated with pentane (780 mL) and stirred at
20 °C for 20 min and then at -10 °C for 30 min. The solid
was collected by filtration, washed with pentane (20 mL ×
3), and then placed in a vacuum oven under house vacuum
736
•
Vol. 11, No. 4, 2007 / Organic Process Research & Development

96% HPLC area %) of 2,8-bissilyl lactol 8 as a slightly
yellowish solid, which was used in the next step without
further purification. 1H NMR (300 MHz, CDCl3) δ 0.24 (s,
12 H), 1.0 (s, 18 H), 2.73-2.99 (m, 2 H), 3.01 (d, J ) 7.6,
1 H), 4.46-4.60 (m, 2 H), 6.09 (d, J ) 7.8, 1 H), 6.56 (dd,
J ) 1.1, 8.4, 1 H), 6.57-6.63 (m, 2 H), 6.68 (dd, J ) 1.2,
8.3, 1 H), 7.2 (d, J ) 8.3, 1 H), 7.42 (d, J ) 8.4, 1 H).
LC/MS m/z 527 (MH+), 549 (MNa+).
2 H), 4.15 (t, J ) 6.3, 2 H), 4.56 (m, 1 H), 5.68 (br s, 1 H),
641-6.48 (m, 2 H), 6.58 (d, J ) 1.0, 1 H), 6.78 (d, J ) 8.9,
2 H), 6.90-7.0 (m, 3 H), 7.06 (d, J ) 8.4, 1 H), 7.18 (dd,
J ) 1.0, 8.3, 1 H), 7.24 (d, J ) 8.3, 1 H). LC/MS m/z 732
(MH+), 754 (MNa+), and 206 (MH+ of 12a).
(Z)-2-(5-(Hydroxy(4-(2-(piperidin-1-yl)ethoxy)phenyl)-
methyl)-(8-benzyoxy-2,3-dihydrobenzo[b]oxepin-4-yl)-5-
benzyoxyphenol (9a). Compound 9a was prepared in a
similar fashion as 9 from compound 8a (90%, 4.0 g, 8.36
mmol) with the side-chain 12 (2.3 equiv) and n-BuLi (2.3
equiv) in THF at -78 °C for 1 h to afford 9a (8.37 g crude
product, 146% isolated yield; 68%, HPLC area %). 1H NMR
(300 MHz, CDCl3) δ 1.44 (m, 2 H), 1.63 (m, 4 H), 2.51 (m,
4 H), 2.64 (m, 2 H), 2.80 (t, J ) 6.0, 2 H), 4.08 (m, 2 H),
4.12 (t, J ) 6.1, 2 H), 4.98 (m, 1 H), 5.10 (s, 2 H), 5.12 (s,
2 H), 6.50-7.02 (m, 9 H), 7.10-7.51 (m 12 H). LC/MS
m/z 684 (MH+), 706 (MNa+).
2,8-Dibenzyloxy-11,12-dihydro-5H-6,13-dioxabenzo-
[3,4]-cyclohepta[1,2-a]naphthalene-5-ol (8a). Compound 8a
was prepared in a similar fashion as 8 from compound 7a
(4.0 g, 8.4 mmol, 90%) with DIBALH (1.2 equiv) in CH2-
Cl2 at below -40 °C for 1 h to afford 8a (4.1 g, 102%
1
isolated yield, 90% HPLC area %). H NMR (300 MHz,
CDCl3) δ 2.91 (m, 2 H), 4.54 (m, 2 H), 5.06 (s, 2 H), 5.08
(s, 2 H), 6.08 (d, J ) 8.9, 1 H), 6.65 (dd, J ) 0.9, 8.3, 1 H),
6.66-6.76 (m, 2 H), 6.80 (dd, J ) 1.0, 8.1, 1 H), 7.20-
7.50 (m, 13 H). LC/MS m/z 479 (MH+), 501 (MNa+).
2,8-Dimethoxy-11,12-dihydro-5H-6,13-dioxabenzo[3,4]-
cyclohepta[1,2-a]naphthalene-5-ol (15). Compound 15 was
prepared in a similar fashion as 8 from compound 14 (32.4
g, 0.10 mol) with DIBALH (1.5 equiv) in CH2Cl2 at - 70
°C for 6 h to afford 15 (30.0 g, 92% isolated yield). 1H NMR
(300 MHz, DMSO-d6) δ 2.72-2.98 (m, 2 H), 3.21 (d, J )
5.4, 1 H), 3.78 (s, 3 H), 3.80 (s, 3 H), 4.40-4.57 (m, 2 H),
5.99 (d, J ) 5.6, 1 H), 6.60 (d, J ) 0.8, 1 H), 6.61-6.66
(m, 2 H), 6.79 (dd, J ) 0.9, 8.2, 1 H), 7.18 (d, J ) 7.9, 1
H), 7.46 (d, J ) 8.3, 1 H). LC/MS m/z 327 (MH+), 349
(MNa+).
(Z)-2-(5-(Hydroxy(4-(2-(piperidin-1-yl)ethoxy)phenyl)-
methyl)-(8-methoxy-2,3-dihydrobenzo[b]oxepin-4-yl)-5-
methoxyphenol (18). Compound 18 was prepared in a
similar fashion as 9 from compound 15 (87.0 g, 0.27 mol)
with the side-chain 12 (2.44 equiv) and n-BuLi (2.31 equiv)
in THF at -78 °C for 8 h to afford 18 (145.0 g crude product
1
(102% isolated yield; 80%, HPLC area %). H NMR (300
MHz, CDCl3) δ 1.43 (m, 2 H), 1.62 (m, 4 H), 2.52 (m, 4
H), 2.75 (m, 2 H), 2.80 (t, J ) 5.6, 2 H), 3.73 (s, 3 H), 3.78
(s, 3 H), 4.04 (m, 2 H), 4.11 (t, J ) 5.5, 2 H), 4.58 (m, 1 H),
5.69 (s, 1 H), 6.41-6.56 (m, 3 H), 6.62 (d, J ) 0.8, 1 H),
6.76 (d, J ) 9.0, 2 H), 6.91 (d, J ) 8.1, 1 H), 6.96 (d, J )
7.9, 1 H), 7.08 (d, J ) 8.4, 1 H), 7.16 (d, J ) 8.9, 2 H).
LC/MS m/z 532 (MH+), 554 (MNa+).
(Z)-2-(5-(Hydroxy(4-(2-(piperidin-1-yl)ethoxy)phenyl)-
methyl)-(8-tert-butyldimethylsilyloxy-2,3-dihydrobenzo-
[b]oxepin-4-yl)-5-(tert-butyldimethylsilyloxy)phenol (9). A
12-L RBF was charged with 1-[2-(4-bromophenoxy)ethyl]-
piperidine (12) (233.6 g, 0.822 mol) and anhydrous THF
(1.8 L) under N2. The solution was stirred and cooled to
-76 °C; n-butyllithium (n-BuLi, 2.5 M in hexane, 329.0 mL,
0.822 mol) was added dropwise over a 60-min period, while
the internal reaction temperature was maintained between
-76 to -73 °C. The mixture was stirred for 20 min, and a
solution of lactol 8 (188.0 g, 0.357 mol) in anhydrous THF
(1.8 L) was added dropwise over an 80-min period at < -73
°C. The reaction was stirred for an additional hour, and the
progress of the reaction was monitored by LC/MS, TLC
(CH2Cl2/CH3OH, 9/1), and 1H NMR. After completion, the
reaction was quenched with saturated NH4Cl solution (740
mL) at -78 °C and the mixture was allowed to warm to 20
°C with stirring. The solvent was concentrated in vacuo at
60 °C to give an oily material, which was redissolved in
EtOAc (2.8 L) and washed with D.I. H2O (2.8 L × 2). The
aqueous phase was extracted with EtOAc (1.8 L × 2), and
the combined organic phases were washed with brine (2.5
L). The solvent was concentrated in vacuo at 65 °C to give
1-(2-{4-[2,8-Bis(tert-butyldimethylsilyloxy)-11,12-dihy-
dro-5H-6,13-dioxabenzo[3,4]cyclohepta[1,2-a]naphthalene-
5-yl]phenoxy}ethyl)piperidine (10). A 12-L RBF was
charged with CH2Cl2 (3.6 L) and the crude diol 9 (360.0 g,
0.3574 mol) under N2, and the mixture was cooled to 10
°C. A solution of HCl (37%, 152.8 mL, 1.43 mol) was added
dropwise over a 15-min period with fast agitation. After the
addition, the reaction was stirred for 30 min at 20 °C, and
the reaction was monitored by LC/MS, TLC (CH2Cl2/CH3-
1
OH, 9/1), and H NMR. The mixture was transferred to a
12-L three-neck separatory flask, diluted with CH2Cl2 (1.23
L), and washed with D.I. H2O (4.0 L × 2), saturated NaHCO3
solution (3.0 L), and brine (3.0 L). The organic phase was
dried over Na2SO4 (1.0 kg) and then concentrated in vacuo
at 50 °C. The resulting cherry, syrupy material was placed
under high vacuum at 60 °C for 18 h to afford 252.0 g
(98.9%) of crude compound 10, which was used in the next
step without further purification. 1H NMR (300 MHz, CDCl3)
δ 0.18 (s, 6 H), 0.22 (s, 6 H), 0.97 (s, 9 H), 0.99 (s, 9 H),
1.44 (m, 2 H), 1.61 (m, 4 H), 2.50 (m, 4 H), 2.66 (t, J )
6.0, 2 H), 2.89 (m, 2 H), 4.06 (t, J ) 6.1, 2 H), 4.69 (m, 2
H), 6.04 (s, 1 H), 6.32 (d, J ) 0.9, 1 H), 6.39 (dd, J ) 1.0,
8.2, 1 H), 6.53 (dd, J ) 0.9, 8.3, 1 H), 6.59 (d, J ) 1.0, 1
H), 6.68 (d, J ) 9.0, 2 H), 6.98 (d, J ) 8.3, 1 H), 7.11 (d,
J ) 8.4, 1 H), 7.35 (d, J ) 9.0, 2 H). LC/MS m/z 714 (MH+),
736 (MNa+).
1
the crude diol 9 (359.8 g, 138%; of which the H NMR
spectra indicated ∼30% of des-bromo 12a), which was used
1
in the next step without further purification. H NMR (300
MHz, CDCl3) δ 0.20 (s, 6 H), 0.23 (s, 6 H), 0.96 (s, 9 H),
0.99 (s, 9 H), 1.46 (m, 2 H), 1.61 (m, 4 H), 2.53 (m, 4 H),
2.73-2.81 (m, 2 H), 2.80 (t, J ) 6.2, 2 H), 3.98-4.14 (m,
Vol. 11, No. 4, 2007 / Organic Process Research & Development
•
737

1-(2-{4-[2,8-(Dibenzyloxy)-11,12-dihydro-5H-6,13-
dioxabenzo[3,4]cyclohepta[1,2-a]naphthalene-5-yl]-
phenoxy}ethyl)piperidine (10a). Compound 10a was pre-
pared in a similar fashion as 10 from compound 9a (8.37 g,
68%, HPLC area %) with concentrated HCl (4.9 mL) in
toluene at 20 °C for 1 h to afford 10a (5.46 g, 98% isolated
yield; 96%, HPLC area %) (the excess of des-bromo 12a
was removed as HCl salt during aqueous workup). 1H NMR
(300 MHz, CDCl3) δ 1.42 (m, 2 H), 1.62 (m, 4 H), 2.49 (m,
4 H), 2.74 (m, 2 H), 2.81 (t, J ) 5.7, 2 H), 4.06 (m, 2 H),
4.13 (t, J ) 5.8, 2 H), 5.0 (s, 2 H), 5.08 (s, 2 H), 6.05 (s, 1
H), 6.46 (d, J ) 0.8, 1 H), 6.53 (dd, J ) 0.9, 8.2, 1 H), 6.68
(dd, J ) 0.9, 8.2, 1 H), 6.73 (d, J ) 0.9, 1 H), 6.78 (d, J )
8.9, 2 H), 6.92 (m, 1 H), 7.01 (d, J ) 8.3, 1 H), 7.18 (d, J
) 9.0, 2 H), 7.20-7.51 (m, 10 H). LC/MS m/z 666 (MH+),
688 (MNa+).
1-(2-{4-[2,8-(Dimethoxy)-11,12-dihydro-5H-6,13-di-
oxabenzo[3,4]cyclohepta[1,2-a]naphthalene-5-yl]phenoxy}-
ethyl)piperidine (19). Compound 19 was prepared in a
similar fashion as 10 from compound 18 (132.0 g, 0.248
mol) with concentrated HCl (61 mL) in CH2Cl2 at 10-15
°C for 45 min to afford 19 (84.0 g, 56% isolated yield;
99.5%, HPLC area %). 1H NMR (300 MHz, CDCl3) δ 1.41
(m, 2 H), 1.58 (m, 4 H), 2.46 (m, 4 H), 2.71 (t, J ) 6.0, 2
H), 2.89 (t, J ) 5.8, 2 H), 3.73 (s, 3 H), 3.79 (s, 3 H), 4.03
(t, J ) 6.1, 2 H), 4.69 (t, J ) 5.5, 2 H), 6.06 (s, 1 H), 6.36
(d, J ) 1.0, 2 H), 6.48 (dd, J ) 0.9, 8.3, 1 H), 6.58 (dd, J
) 0.9, 8.2, 1 H), 6.68 (d, J ) 9.1, 2 H), 7.02 (d, J ) 8.4, 1
H), 7.17 (d, J ) 8.3, 1 H), 7.36 (d, J ) 9.0, 2 H). LC/MS
m/z 514 (MH+), 536 (MNa+).
5-[4-(2-Piperidin-1-yl-ethoxy)phenyl]-11,12-dihydro-
5H-6,13-dioxabenzo[3,4]cyclohepta[1,2-a]naphthalene-
2,8-diol (11). A 12-L RBF was charged with compound 10
(253.0 g, 0.356 mol) in anhydrous THF (5.0 L) with agitation
at 20 °C under nitrogen. Tetrabutylammonium fluoride
(TBAF, 1.0 M in THF, 213 mL, 0.213 mol) was added
dropwise over a 1-h period with fast agitation, and the
reaction was then stirred for 18 h; the reaction was monitored
by LC/MS, TLC (CH2Cl2/CH3OH, 9/1), and HPLC ((reten-
tion time/area %): unknown (2.36 min, 1.9%), 11 (3.90 min,
94%), 13 (4.18 min, 3.1%), mono-TBDMS-11 (5.11 min,
0.1), and 10 (6.08 min, 0.0%)). After the reaction was
completed, EtOAc (2.5 L) and brine (2.5 L) were added to
the reaction and stirred for 10 min. The mixture was
transferred to a 12-L separatory flask, and the aqueous phase
was separated; the organic phase was washed with brine (2.5
L, 1.0 L × 6). The organic phase was dried over Na2SO4
(1.0 kg), and then concentrated in vacuo at 50 °C. The
resulting cherry-colored material was placed under high
vacuum at 60 °C for 2 h to afford 273.8 g (158%, 84% of
11, which contained residual EtOAc; HPLC area %) of crude
11. Chromatographic purification (SiO2 × 2.4 kg; eluted with
CH2Cl2/CH3OH (97/3 to 93/7)) of the above crude product
afforded 139.3 g (81% isolated yield, 94% of 11 and 3.1%
of 13; HPLC area %) of tetrabutylammonium impurity
contaminated 11. LC/MS m/z 486 (MH+), 243 (MH+ of
Bu4N+).
In addition, compound 11 was also prepared from the
hydrogenolysis of 10a (3.73 g, 5.61 mmol; 96%) using Pd/C
(5 mol %) in EtOAc/MeOH (60/60, mL/mL) under hydrogen
pressure (3.06 atm) at 20 °C for 20 h to afford the crude
product 11 (3.68 g, 91% isolated yield). Crystallization of
crude 11 in EtOAc/IPA (9:1) afforded pure 11 (2.28 g, 62%
isolated yield; 98% HPLC area %), which compared exactly
1
to the H NMR and LC/MS spectroscopic data with the
compound 11 prepared from the TBDMS-protected route.
Debromination of 13. A sample of monobromo 13
contaminated 11 (139.3 g, 0.287 mol) was dissolved in CH3-
OH (600 mL), and the solution was transferred to a 2-L Parr
glass hydrogenation bottle followed by the addition of the
catalyst (5% palladium on carbon, 30.5 g) under N2
atmosphere. The mixture was flushed with hydrogen (∼1.36
atm × 3) and then the shaker bottle was refilled with
hydrogen to 3.06 atm and shaken for 16 h. The reaction was
monitored by LC/MS and HPLC. After completion, the
reaction vessel was vented with N2 (∼1.36 atm × 3), and
the suspension was filtered through a pad of Celite. After
the filter cake was washed with CH3OH (100 mL × 3), the
combined filtrate was concentrated in vacuo at 50 °C, and
the product was further dried for 2 h under high vacuum (2
mmHg) at 65 °C and then at 50 °C for 18 h. There was
obtained 132.1 g (95% isolated yield) of pure racemate 11
1
as a pink, foamy solid (a partial HBr salt). H NMR (300
MHz, CD3OD) δ 1.46 (m, 2 H), 1.61 (m, 4 H), 2.53 (m, 4
H), 2.75 (t, J ) 6.0, 2 H), 2.85 (m, 2 H), 4.07 (t, J ) 6.1,
2 H), 4.61 (m, 2 H), 4.88 (s, 2 H), 6.02 (s, 1 H), 6.18 (d, J
) 0.8, 1 H), 6.36 (dd, J ) 0.7, 8.1, 1 H), 6.50 (dd, J ) 0.6,
8.0, 1 H), 6.52 (d, J ) 0.6, 1 H), 6.79 (d, J ) 9.1, 2 H), 7.0
(d, J ) 8.2, 1 H), 7.16 (d, J ) 8.1, 1 H), 7.34 (d, J ) 9.0,
2 H). LC/MS m/z 486 (MH+), 508 (MNa+). Elementary
analysis Calcd for C30H31N1O5‚1.0 H2O (MW ) 503.60):
C, 71.55; H, 6.60; N, 2.78; %H2O, 1.0. Found: C, 71.37;
H, 6.52; N, 2.70; %H2O, 1.52.
Received for review March 12, 2007.
OP700061X
738
•
Vol. 11, No. 4, 2007 / Organic Process Research & Development