Synthesis of a stable mimic for the ring closed form of gallein, featuring a novel one pot boron tribromide mediated intramolecular cyclisation process

Article abstract of DOI:10.1055/s-1999-2762

Treatment of 9-{2-[1-(1,3-dioxalan-2-yl)ethyl]phenyl}-3,4,5,6-tetrabenzyloxy-9H-xanthen-9-o l 14 with excess boron tribromide at -78°C, and warming to room temperature, provides a facile route to 3,4,5,6-tetrahydroxy-spiro(1H-indene-1,9-[9H]-xanthene-3-(2

Full text of DOI:10.1055/s-1999-2762

LETTER
1133
Synthesis of a Stable Mimic for the Ring Closed Form of Gallein, Featuring a
Novel One Pot Boron Tribromide Mediated Intramolecular Cyclisation
Process
Andrew P. A. Crew, Amanda J. Lyons, Nicholas P. Camp*
Department of Chemistry, Proteus Molecular Design Limited, Beechfield House, Lyme Green Business Park, Macclesfield, Cheshire,
SK11 0JL, UK
Fax: +44 (1625) 500666; E-mail N.P.Camp@Proteus.co.uk
Received 13 April 1999
Abstract: Treatment of 9-{2-[1-(1,3-dioxalan-2-yl)ethyl]phenyl}-
3,4,5,6-tetrabenzyloxy-9H-xanthen-9-ol 14 with excess boron tri-
bromide at -78 °C, and warming to room temperature, provides a
facile route to 3,4,5,6-tetrahydroxy-spiro(1H-indene-1,9-[9H]-xan-
thene-3-(2H)-one 2, a stable mimic for the ring closed form of gal-
lein.
Key words: gallein, indanone, triphenylcarbinol, boron tribromide,
intramolecular cyclisation
Scheme 2
Using a pharmacophore search based upon known inhibi-
Our retrosynthetic analysis of 2 was sparked by consider-
tors of HIV integrase, we identified gallein 1 (Scheme 1)
ation of the ring-opened form 1b of gallein (Scheme 3).
as a potential inhibitor for this enzyme. Although, gallein
We felt that replacement of the acid moiety in 1b with a
possessed an IC50 value of 170nM1 against HIV integrase
methyl ketone could allow an irreversible cyclisation to
for 3’-processing in vitro, we were concerned by potential
occur, via its enol or enolate form 4, to produce 2.
selectivity problems due to indiscriminate reactivity of
both the ring closed 1a and ring opened 1b forms towards
other enzyme nucleophiles. For instance, irreversible
binding could be achieved by nucleophilic attack at the
lactone carbonyl in 1a or by Michael type reactions at sev-
eral sites on the semi-quinone intermediates of 1b.
Scheme 1
In our quest to unravel the biological mode of action of
gallein towards HIV integrase and to determine which
structural features of the molecule were important for en-
zyme binding, we wished to prepare a stable mimic of the
Where R = Li or H and R′ = alkyl
ring-closed form 1a. For instance, a carbocyclic analogue
2 (Scheme 2) would be resistant to both intramolecular
and intermolecular ring opening reactions and so should
suffer less from the aforementioned reactivity problems,
but would still possess many of the hydrogen bonding ca-
pabilities of gallein.
Scheme 3
We predicted that the formation of a quinone type inter-
mediate such as 4 could be achieved by treatment of the
triphenylcarbinol derivative 5 under acidic or Lewis acid-
Synlett 1999, No. 07, 1133–1135 ISSN 0936-5214 © Thieme Stuttgart · New York

1134
A. P. A. Crew et al.
LETTER
ic conditions. A further disconnection of 5 provides a pro-
tected tetrahydroxyxanthone precursor 6 and a dilithiated
species 7 (R = Li) as simple starting materials for the syn-
thesis.
Herein we wish to describe our novel one pot reaction
for the conversion of 9-{2-[1-(1,3-dioxalan-2-yl)-eth-
yl]phenyl}-3,4,5,6-tetrabenzyloxy-9H-xanthen-9-ol 14
into 3,4,5,6-tetrahydroxy-spiro(1H-indene-1,9-[9H]-xan-
thene-3-(2H)-one 2. In our initial investigation of the
chemistry required for the construction of the indanone
ring system, we used 3,3’,4,4’-tetramethoxybenzophenone
82, as a more reactive electrophile than the corresponding
xanthone, in the synthesis of a model analogue 3. Our at-
tempts to react a dilithio species such as 73 with the benz-
ophenone model 8 proved unsuccessful in the short term,
Reagents: (i) 10 equiv. BBr3 1M in CH2Cl2, -
providing numerous side products, and no desired cyclisa- 78 °C to room temp.
tion material. We simplified proceedings by preparing a
masked ketone in the form of dioxolane 10, previously de-
scribed by Rees,4 which undergoes rapid metal-halogen
exchange upon treatment with n-butyllithium in THF at
-78 °C. Thus, treatment of 3,3',4,4'-tetramethoxybenz-
ophenone 8 with 1.2 equivalents of this metallated species
at -78 °C in THF, and warming to room temperature, pro-
vided the triphenylcarbinol 11 in 67% yield after
purification5 (Scheme 4).
Scheme 5
We then turned our attention to the synthesis of the gallein
analogue 2. We synthesised the 3,4,5,6-tetramethoxyxan-
thone precursor 6 using the method of Lin7 but found a
marked decrease in both its solubility and reactivity com-
pared to the benzophenone precursor 8. To overcome this,
3,4,5,6-tetrabenzyloxyxanthone (13) was prepared8 and
treated with excess lithiated dioxolane, (derived from 10)
via cannulation at the higher temperature of 0 °C in THF.
This provided the crude triphenylcarbinol derivative 14 as
a red gum which was used directly in the boron tribromide
reaction, producing the desired indanone 2 in 54% yield
over both steps (Scheme 6).9 Analogue 2 possessed an
IC50 value of 1.8mM against HIV integrase for 3'-process-
ing.
Reagents: (i) 1 equiv. 9, 2 equiv. ethylene glycol, 0.2 equiv. pTSA, to-
luene, 150 °C Dean Stark (90%); (ii) 1.2 equiv. 10, 1.2 equiv. n-BuLi,
THF, -78 °C, 20 min; (iii) 1 equiv. 3,3',4,4'-tetramethoxybenzopheno-
ne (8), THF, -78 °C to room temp., (67%).
Scheme 4
Having failed to achieve the desired cyclisation under
standard ketal deprotection conditions, we attempted a
ketal deprotection under conditions that would also allow
demethylation of the phenols. In this way, we hoped to
generate a reactive quinone type intermediate such as 4
and improve our chances of cyclisation. Indeed, reaction
of 11 under anhydrous conditions with ten equivalents of
boron tribromide in dichloromethane at -78 °C, and
warming to room temperature, furnished the desired in-
danone 3 and the vinyl bromide 126 as the only two prod-
ucts in an approximate 1:1 ratio (Scheme 5).
Reagents: (i) 2.5 equiv. 10, 2.5 equiv. n-BuLi, THF, -78 °C, 20 min;
(ii) 1 equiv. tetrabenzyloxyxanthone 13, THF, 0 °C to room temp; (iii)
7 equiv. BBr3 1M in CH2Cl2, -78 °C to room temp., (54% over both
steps).
Reducing the amount of boron tribromide to 7 equivalents
and quenching the reaction mixture after 10 minutes at
room temperature improved the yield of 3 to 82% and re-
duced the formation of the vinyl bromide 12 to 18%.
Scheme 6
Synlett 1999, No. 07, 1133–1135 ISSN 0936-5214 © Thieme Stuttgart · New York

LETTER
Synthesis of a Stable Mimic for the Ring Closed Form of Gallein
1135
In conclusion, we have developed a novel, one pot reac-
tion for the construction of a stable mimic of gallein. This
boron tribromide mediated reaction was developed using
a model benzophenone derived triphenylcarbinol and fea-
tures a tandem sequence of ketal deprotection, phenol
deprotection and an intramolecular cyclisation. This sta-
ble analogue 2 retains significant activity against HIV in-
tegrase in vitro, although somewhat lower than the parent
molecule gallein. This suggests that both the reactive lac-
tone moiety and the catecholic functionalities are impli-
cated in gallein’s potency against this enzyme.
Acknowledgement
We thank our collaborator Dr Samson Chow at UCLA for providing
the HIV integrase inhibition data.
References an d Notes
(1) The full biological data associated with gallein and its
analogues will be disclosed in a future publication.
(2) The 3,3',4,4'-tetramethoxybenzophenone 8 was prepared from
veratrole and 3,4-dimethoxybenzoyl chloride under Friedel-
Crafts conditions.
(3) The dilithio species 7 was prepared from 2-
bromoacetophenone 9 by treatment with 1 equiv. of lithium
hexamethyldisilazide, followed by 2 equiv. of n-BuLi at
-78 °C in THF.
Typical experimental procedure: Triphenylcarbinol 11.
To a solution of dioxolane 10 (730 mg, 3 mmol) in dry THF (15 ml
was added n-BuLi (1.2 ml of a 2.5 M solution in hexanes) at -78 °C
under an atmosphere of argon. After 20 min at -78 °C, the resultant
white suspension was treated with 3,3',4,4'-tetramethoxybenzophe-
none 8 (755 mg, 2.5 mmol) in dry THF (5 ml) and the reaction mix-
ture warmed slowly to rt. The solution was quenched with sat. aq.
NH4Cl (5 ml), extracted with ethyl acetate (2 x 50 ml) and the or-
ganic layers washed with water (10 ml), brine (10 ml) then dried
(MgSO4) and concentrated in vacuo. The residue was triturated with
Et2O at 0 °C to yield the triphenylcarbinol 11 as a white solid (780
(4) Hickey, D. M. B.; MacKenzie, R.; Moody, C. J.; Rees, C. W.
J. Chem. Soc. Perkin Trans. I. 1987, 921.
(5) The predicted acid lability of the tertiary alcohol in 11 was
first noted when preparing the 1H NMR solution in CDCl3
where a pink solution was observed. Indeed, treatment of a
methanolic solution of 11 with dil. HCl provided a purple
solution, possibly indicating the generation of a resonance
stabilised intermediate.
(6) Selected data for 12:1H NMR (300 MHz, DMSO-d6) 8.90 (br,
4H), 7.40 (m, 4H), 7.12 (s, 1H), 6.65 (m, 4H), 6.45 (dd,
J = 7.7, 2.0 Hz, 2H); 13C NMR (75.47 MHz, CDCl3) 149.84,
145.36, 145.00, 144.75, 141.11, 133.29, 127.78, 127.22,
125.35, 120.87, 118.65, 118.00, 115.71, 115.61, 65.62;
HRMS: calculated for (MH+) C21H16O4Br :411.0232; found
411.0232. The halogenative deoxygenation of ketones by
cleavage of 1,3-benzodioxoles to give vinyl bromides using
boron tribromide has been published; Napolitano, E.; Fiaschi,
R.; Mastrorilli, E.; Synthesis 1986, 122.
(7) Lin, C. N.; Liou, S. S.; Ko, F. N., Teng., C. M.; J. Pharm. Sci.
1992, 81, 1109.
(8) The 3,4,5,6-tetrabenzyloxyxanthone 13 was prepared by
demethylation of 6 using boron tribromide and treatment of
the resulting 3,4,5,6-tetrahydroxyxanthone with benzyl
bromide.
1
mg, 67%). H NMR (300 MHz, CDCl3) 7.62 (dd, J = 7.9, 1.2 Hz,
1H), 7.12-7.20 (m, 1H), 7.0 (dt, J = 8.0, 1.4 Hz, 1H), 6.85 (d, J = 1.9
Hz, 2H), 6.58-6.65 (m, 3H), 6.35 (dd, J = 8.4, 2.1 Hz, 2H), 3.80 (s,
6H), 3.70 (s, 6H), 3.60 (m, 2H), 2.82 (m, 2H), 1.85 (s, 3H) 1.50 (br.,
1H); 13C NMR (75 MHz, CDCl3) 148.96, 148.19, 145.14, 142.29,
141.50, 132.72, 128.81, 127.53, 127.02, 120.83, 111.60, 111.13,
109.97, 83.48, 64.81, 64.38, 56.22, 56.20, 30.07. HRMS: calculated
for C27H30O7:466.1992; found 466.1987.
Indanone 3. To 11 (70 mg, 0.15 mmol) was added BBr3 (1.05 ml
of a 1.0 M solution in CH2Cl2, 1.05 mmol) dropwise at -78 °C under
an argon atmosphere. After 1h. at -78 °C, the reaction mixture was
warmed slowly to rt. After 10 min at rt, the reaction mixture was
poured into a 25 ml beaker containing a mixture of ice and 5% aq.
HCl. and stirred for 30 min. The reaction mixture was then extracted
with ethyl acetate (2 x 25 ml) and the organic layers washed with
water (10 ml), brine (10 ml) then dried (MgSO4) and concentrated
in vacuo. The residue was purified by flash chromatography on sil-
ica gel (6% MeOH, CH2Cl2) to yield the indanone 3 as a white solid
(43 mg, 82%). 1H NMR δ (300 MHz, DMSO-d6) 8.79 (br. s, 4H),
7.67 (t, J = 7.3Hz, 1H), 7.64 (d, J = 6.8Hz, 1H), 7.45 (t, J = 7.1Hz,
1H), 7.35 (d, J = 7.7Hz, 1H), 6.60 (d, J = 8.3Hz, 2H), 6.5 (d,
(9) The corresponding vinyl bromide could also be isolated (67%)
when the cyclisation was carried out using 10 equiv. of BBr3
(at -78 °C) and left to stir for 16 h at 20 °C. This could be
successfully converted to the ketone 2 by treatment with 15%
HOAc in c.HCl at 130 °C.
J = 2.3Hz, 2H), 6.34 (dd, J = 8.3 & 2.3Hz, 2H) and 3.2 (s, 2H); 13
C
NMR (75 MHz, CDCl3) 205.31, 161.35, 145.70, 144.57, 138.89,
135.80, 128.64, 123.59, 119.22, 116.49, 115.98, 56.40, 55.39;
HRMS: calculated for C21H16O5:348.0998; found 348.0998.
Article Identifier:
1437-2096,E;1999,0,07,1133,1135,ftx,en;L04499ST.pdf
Synlett 1999, No. 07, 1133–1135 ISSN 0936-5214 © Thieme Stuttgart · New York