A versatile solid phase synthesis of lavendustin A and certain biologically active analogs

Article abstract of DOI:10.1021/jo961719l

Reaction of aminomethylated polystyrene resin (8) with succinic anhydride, followed by esterification of the free carboxylic acid of the product 9 with 2,5-dihydroxybenzaldehyde (4), afforded the resin-linked aldehyde intermediate 10. The key intermediate 10 was converted into the protein-tyrosine kinase inhibitor lavendustin A and certain analogs through reductive amination and reductive alkylation steps, followed by cleavage from the resin. This method enables the preparation of a wide variety of lavendustin A analogs using combinatorial chemistry and parallel synthesis techniques.

Full text of DOI:10.1021/jo961719l

9368
J . Org. Chem. 1996, 61, 9368-9373
A Ver sa tile Solid P h a se Syn th esis of La ven d u stin A a n d Cer ta in
Biologica lly Active An a logs
Rajesh Devraj and Mark Cushman*
Department of Medicinal Chemistry and Molecular Pharmacology, Purdue University,
West Lafayette, Indiana 47907
Received September 6, 1996X
Reaction of aminomethylated polystyrene resin (8) with succinic anhydride, followed by esterification
of the free carboxylic acid of the product 9 with 2,5-dihydroxybenzaldehyde (4), afforded the resin-
linked aldehyde intermediate 10. The key intermediate 10 was converted into the protein-tyrosine
kinase inhibitor lavendustin A and certain analogs through reductive amination and reductive
alkylation steps, followed by cleavage from the resin. This method enables the preparation of a
wide variety of lavendustin A analogs using combinatorial chemistry and parallel synthesis
techniques.
Combinatorial organic synthesis has increasingly be-
abnormal expression or function of PTKs can lead to
several disorders of cell proliferation including cancer.7
Thus there is considerable interest in the development
of PTK inhibitors as anticancer agents.8 Lavendustin A
and its biologically active pharmacophore, benzylamine
2, are potent inhibitors of both receptor and nonreceptor
PTKs.5,9 Structure-activity relationship studies have
revealed that the 2,5-hydroxy substituents on ring A are
important for activity and their deletion and/or their
conversion into O-methyl ethers results in a loss of
inhibitory activity against certain PTKs.10 In addition,
modification of the carboxylic acid on ring B with
hydrophobic aromatic substituents was shown to increase
the inhibitory activity against epidermal growth factor
receptor PTK-stimulated DNA synthesis in a cellular
system.11 We wished to develop a solid phase synthetic
methodology that would incorporate and/or conserve
these structural features in the final products. The
previously reported solid phase synthesis of lavendustin
analogs by Green utilized the carboxylic acid as a handle
to link to the solid support and O-methylated fragments
in ring A, B, and C for the analog synthesis.12 Laven-
dustin A itself was prepared by cleaving the O-methyl
ethers with BBr3, followed by workup with methanol. In
contrast, our strategy utilizes a free phenolic hydroxyl
group for attachment to the resin, which frees the
carboxylic acid for chemical manipulation on solid sup-
port and allows an additional site for generating chemical
diversity. In addition, this synthesis is performed with-
out protection of the free phenolic hydroxyl groups, thus
avoiding the necessity for phenol deprotection.
come a valuable tool for the preparation of structurally
diverse chemical libraries for both lead discovery and lead
development.1 The value of solid phase synthesis in the
generation of large libraries of oligomeric compounds
such as peptides2 and oligonucleotides3 has long been
recognized. Recently, a considerable amount of attention
has been focused on adapting and exploiting the advan-
tages of solid phase synthesis for the production of
libraries of nonoligomeric, small organic molecules for
biological screening.1,4 The development of general, high-
yielding, solid phase reactions is crucial to these efforts.
In addition, new methodologies that avoid extensive
workup, recrystallization, and chromatographic purifica-
tion of the material after cleavage from resin are of
premium value to the chemist involved in multiple,
parallel organic synthesis efforts. We report a solid
phase synthesis of the natural product lavendustin A (1),
a potent protein-tyrosine kinase inhibitor which was first
isolated from Streptomyces griseolavendus.5 The syn-
thetic methodology has been developed to provide access
to lavendustin A combinatorial libraries.
(6) (a) Ullrich, A.; Schlessinger, J . Cell 1990, 61, 203. (b) Yarden,
Y.; Schlessinger, J . Ann. Rev. Biochem. 1988, 57, 443.
(7) (a) Bishop, J . M. Science 1987, 335, 305. (b) Heldin, C. H.;
Betsholtz, C.; Claesson-Welsh, L.; Westermark, B. Biochem. Biophys.
Acta 1987, 907, 219.
(8) (a) Levitzki, A.; Gazit, A. Science 1995, 267, 1782. (b) Burke, T.
R., J r. Drugs Future 1992, 17, 119. (c) Chang, C. J .; Geahlen, R. L. J .
Nat. Prod. 1992, 55, 1529. (d) Dobrusin, E. M.; Fry, D. W. Ann. Rep.
Med. Chem. 1992, 27, 169.
(9) (a) Hsu, C. J .; Persons, P. E.; Spada, A. P.; Bednar, R. A.;
Levitzki, A.; Zilberstein, A. J . Biol. Chem. 1991, 266, 21105. (b) Smyth,
M. S.; Stefanova, I.; Hartmann, F.; Horak, I. D.; Osherov, N.; Levitzki,
A.; Burke, T. R., J r. J . Med. Chem. 1993, 36, 3010.
(10) (a) Chen, H.; Boiziau, J .; Parker, F.; Maroun, R.; Tocque, B.;
Rouques, B. P.; Garbay-J aureguiberry, C. J . Med. Chem. 1993, 36,
4094. (b) Nussbaumer, P.; Winiski, A. P.; Cammisuli, S.; Hiestand, P.;
Weckbecker, G.; Stu¨tz, A. J . Med. Chem. 1994, 37, 4079.
(11) Chen, H.; Boiziau, J .; Parker, F.; Mailliet, P.; Commercon, A.;
Tocque, B.; Le Pecq, J . P.; Roques, B. P.; Garbay, C. J . Med. Chem.
1994, 37, 845.
Protein-tyrosine kinases (PTKs) play crucial roles in
many signal transduction pathways within cells.6 The
X Abstract published in Advance ACS Abstracts, December 1, 1996.
(1) (a) Patel, D. V.; Gordon, E. M. Drug Design Today 1996, 1, 134.
(b) Thompson, L. A.; Ellman, J . A. Chem. Rev. 1996, 96, 555. (c) Terret,
N. K.; Gardner, M.; Gordon, D. W.; Kobylecki, R. J .; Steele, J .
Tetrahedron 1995, 51, 8135. (d) Lowe, G. Chem. Soc. Rev. 1995, 24,
309.
(2) (a) Gallop, M. A.; Barret, R. W.; Dower, W. J .; Fodor, S. P. A.;
Gordon, E. M. J . Med. Chem. 1994, 37, 1233. (b) Furka, A.; Sebestyen,
F.; Asgedom, M.; Dibo, G. Intl. J . Pept. Protein Res. 1991, 37, 487.
(3) Ellington, A. D.; Szostak, J . W. Nature 1990, 346, 818.
(4) (a) Hermkens, P. H. H.; Ottenheijem, H. C. J .; Ress, D.
Tetrahedron 1996, 52, 4527. (b) Gordon, E. M.; Barret, R. W.; Dower,
W. J .; Fodor, S. P. A.; Gallop, M. A. J . Med. Chem. 1994, 37, 1385.
(5) Onoda, T.; Iinuma, H.; Sasaki, Y.; Hamada, M.; Isshiki, K.;
Naganawa, H.; Takeuchi, T.; Tatsuta, K.; Umezawa, K. J . Nat. Prod.
1989, 52, 1252.
(12) Green, J . J . Org. Chem. 1995, 60, 4287.
S0022-3263(96)01719-7 CCC: $12.00 © 1996 American Chemical Society

Versatile Solid Phase Synthesis of Lavendustin A
J . Org. Chem., Vol. 61, No. 26, 1996 9369
Sch em e 1a
Sch em e 2a
a
Reagents and conditions: (i) Ph3P, DEAD, THF, 24 h; (ii)
NaCNBH3, 1% AcOH in DMAc, 48 h.
Resu lts a n d Discu ssion
Our initial studies were conducted using the TFA-labile
Wang polystyrene resin13 (3) (Scheme 1). A Mitsonobu
coupling14 of the benzyl alcohol of Wang’s resin (3) with
either of the phenolic hydroxyls of 2,5-dihydroxybenzal-
dehyde (4) was expected to provide the acid cleavable
alkylaryl ether linkage.15 Reaction of benzaldehyde 4
with Wang’s resin using the standard Mitsonobu coupling
reagents, triphenylphosphine and DEAD, afforded the
desired support-bound benzaldehyde 5. In this case it
was of no consequence which of the two hydroxyls formed
the ether link with the benzyl alcohol of Wang’s resin,
since both would be susceptible to TFA cleavage. Inde-
pendent to our investigation, another group has recently
reported a similar Mitsonobu coupling with hydroxyben-
zaldehydes.16 Under conditions similar to those reported
above, they found that the Mitsonobu coupling of Wang’s
resin with 2-hydroxybenzaldehyde was unsuccessful. On
the basis of their observation, it seems likely that the
ether formation of the benzyl alcohol of Wang’s resin
occurs with the non-hydrogen-bonded 5-hydroxyl group
of 2,5-dihydroxybenzaldehyde. The efficiency of loading
was determined to be 72% on the basis of the recovery of
pure 2,5-dihydroxybenzaldehyde obtained by cleavage of
resin 5 with CH2Cl2/TFA/Me2S (50/45/5) at room tem-
perature for 1.5 h. The presence of dimethyl sulfide in
the cleavage mixture was essential in order to prevent
decomposition of the cleaved material.
a
Reagents and conditions: (i) pyridine, 24 h; (ii) method A,
DCC, HOBt, pyridine, 2,5-dihydroxybenzaldehyde (4), DMF, 24
h; method B, DCC, DMAP, 4, CH2Cl2, DMF, 24 h; (iii) 5-ami-
nosalicylic acid (6), NaCNBH3, 1% AcOH in DMAc, 48 h; (iv)
MeOH/Et3N/Me2S (75/15/10), 6 × 3 h; (v) method A, o-hydroxy-
benzaldehyde (12), NaCNBH3, 4 Å seives, 1% AcOH in DMAc, 72
h; method B, 12, NaBH(OAc)3, AcOH, 4 Å seives, AcOH, CH2Cl2,
DMF, 72 h.
TFA/Me2S (50/45/5) led to considerable degradation and
unacceptable levels of product purity. In an attempt to
minimize this decomposition, several variables were
explored, including decreasing the concentration of TFA,
increasing the amount of dimethyl sulfide in the cleavage
cocktail, and decreasing the time required for cleavage.
In all instances, the product purity was compromised
significantly due to the formation of highly colored
degradation products. The extensive column chromato-
graphic techniques that would be required to purify the
product after cleavage from the resin and the low yields
made this route highly unattractive and prompted the
investigation of other linker strategies for the synthesis
of the desired compounds.
A reductive amination5,17 of support-bound benzalde-
hyde 5 with 5-aminosalicylic acid (6), followed by cleavage
from the resin, would afford the lavendustin A active
pharmacophore (2). The reductive amination reaction
was conducted in 1% AcOH in N,N-dimethylacetamide
(DMAc) using a large excess of both the amine (6-7
equiv) and sodium cyanoborohydride (18-20 equiv) as
the reducing agent to afford the support-bound benzy-
lamine 7. Unfortunately, the attempted cleavage of the
desired product 2 from the solid support with CH2Cl2/
(13) Wang, S. J . Am. Chem. Soc. 1973, 95, 1328.
(14) Mitsonobu, O. Synthesis 1981, 1.
(15) (a) Richter, L. S.; Gadek, T. R. Tetrahedron Lett. 1994, 35, 4705.
(b) Rano, T. A.; Chapman, K. T. Tetrahedron Lett. 1995, 36, 3789. (c)
Krchnak, V.; Flegelova, Z.; Weichsel, A. S.; Lebl, M. Tetrahedron Lett.
1995, 36, 6193.
(16) Hamper, B. C.; Dukesherer, D. R.; South, M. S. Tetrahedron
Lett. 1996, 37, 3671.
(17) (a) Coy, D. H.; Hocart, S. J .; Sasaki, Y. Tetrahedron 1988, 44,
835. (b) Bray, A. M.; Chiefari, D. S.; Valerio, R. M.; Maeji, N. J .
Tetrahedron Lett. 1995, 36, 5081. (c) Gordon, D. W.; Steele, J . Bioorg.
Med. Chem. Lett. 1995, 5, 47.
Commercially available aminomethyl polystyrene resin
(8) was acylated with succinic anhydride (3 equiv) in
pyridine for 24 h at room temperature (Scheme 2). After
the resin was subjected to a second round of reaction and
a negative Kaiser test was ensured, the IR spectrum
(KBr) of the derivatized resin 9 revealed two new
absorptions at 1721 and 1655 cm-1, indicating that the
desired substitution had occurred. Two different reaction
conditions were explored for the esterification of the

9370 J . Org. Chem., Vol. 61, No. 26, 1996
Devraj and Cushman
support-bound acid with 2,5-dihydroxybenzaldehyde.
Coupling of 9 with DCC (3 equiv), hydroxybenzotriazole
(3 equiv), and 2,5-dihydroxybenzaldehyde (4) (3 equiv)
in DMF/pyridine (9/1) as the solvent afforded the desired
support-bound benzaldehyde ester (10). In a similar
manner, a modified Steglich esterification18 of resin 9
with DCC (3 equiv), DMAP (3.5 equiv), and 4 (3 equiv)
in DMF/CH2Cl2 (4/1) also afforded 10. It was found that
a double-coupling protocol consistently provided much
higher loading of the aldehyde onto the solid support.
Reductive amination of support-bound benzaldehyde 10
with 5-aminosalicylic acid (6) under the conditions re-
ported above (room temperature, 48 h) afforded the
benzylaniline-linked solid support 11. Several basic
cleavage conditions were then explored to release the
desired product 2 from the solid support in high yield
and purity. Attempted hydrolysis of the ester with
varying concentrations of NH4OH in dioxane resulted in
degradation, and none of the desired product could be
isolated. Nonhydrolytic transesterification protocols were
next attempted with ca. 0.25 equiv of sodium methoxide
employed and the secondary aromatic nitrogen undergo-
ing the reaction. Reductive alkylation using the condi-
tions reported above failed to afford any product, and the
starting material was recovered intact. Longer reaction
times and addition of 4 Å molecular sieves to the reaction
mixture led to a better conversion of the starting mate-
rial. The desired support-bound lavendustin A (13) was
thus obtained by driving the reaction to completion by
subjecting the resin to the reductive alkylation conditions
three times. Cleavage of the loaded resin 13 under the
standard conditions reported above afforded the natural
product lavendustin A (1) in 82% purity and 72% yield.
We also investigated the use of sodium triacetoxyboro-
hydride [NaBH(OAc)3] as a reducing agent in the reduc-
tive alkylation.17c,21 Reaction of support-bound benzyla-
niline 11 with o-hydroxybenzaldehyde (6 equiv) and
NaBH(OAc)3 (8 equiv) in CH2Cl2/DMAc (4/1) at room
temperature for 72 h did not afford any of the desired
product on cleavage, and only unreacted starting material
was observed in the cleavage mixture. However, addition
of AcOH (10 equiv) and 4 Å molecular sieves to the
reaction mixture did increase the rate of reaction and
after three rounds of reductive amination afforded the
desired support-bound lavendustin A (13). In comparison
to the NaCNBH3 method of reductive amination, the
NaBH(OAc)3 method provided for a modest increase in
the yield (77%) of lavendustin A (1).
in methanol19 and 10% triethylamine in methanol.20
A
qualitative (TLC) estimation of the progress of transes-
terification indicated that although the desired product
2 was formed, it was accompanied by highly colored, polar
degradation products, with better results being obtained
with 10% triethylamine in methanol.
Optimization of the cleavage conditions with Et3N in
methanol proved to be challenging. Subjection of the
loaded resin 11 to cleavage for 30 min with 10% Et3N in
MeOH at room temperature afforded the desired product
2 in good purity (>95%). However, less than 10% yield
of the product was obtained. Longer reaction times
inevitably led to the formation of highly colored degrada-
tion products which compromised the purity of the
product. Heating the reaction mixture to 50 °C in order
to accelerate the rate of cleavage resulted in the forma-
tion of a deep red solution and complete degradation of
the product. These results suggested that compounds
containing a 2,5-dihydroxy substitution pattern are
readily susceptible to degradation under the basic condi-
tions of cleavage and probably oxidize to quinones
through free radicals and/or other mechanisms. In an
attempt to circumvent this problem, dimethyl sulfide
(Me2S) was incorporated in the cleavage cocktail, and the
desired product was cleaved from the resin in small
portions using repeated reactions, thus minimizing the
contact of the cleaved material with the cleavage cocktail.
Treatment of resin 11 six times with MeOH/Et3N/Me2S
(75/15/10) for 3 h at room temperature under an argon
atmosphere each time afforded the desired lavendustin
A active pharmacophore 2 in 86% yield (based on
incorporation of 2,5-dihydroxybenzaldehyde on the resin)
and excellent purity (98%) after filtration through silica
gel. This general method involving repeated cleavage
reactions, to minimize contact time with the cleavage
cocktail, was used in subsequent steps and consistently
provided materials in good yields and purity.
This solid phase synthetic methodology was extended
to the synthesis of a few lavendustin A analogs. A double
reductive amination of benzaldehyde-linked solid support
10 with 3-aminobenzoic acid (14) and NaCNBH3 in 1%
AcOH in DMAc afforded the desired benzylaniline-linked
solid support 15 (Scheme 3). Cleavage of resin 15 under
the standard conditions afforded the deshydroxy laven-
dustin A active pharmacophore 16 in 88% yield and 92%
purity. A triple reductive amination of resin 15 with
benzaldehyde (6 equiv), NaCNBH3 (18 equiv), and 4 Å
molecular sieves in 1% AcOH in DMAc afforded the
desired support-bound dibenzylaniline 17. Cleavage of
this resin under the standard conditions afforded the
lavendustin A analog 18 in 73% yield.
The free carboxylic acid functionality in resin 15
presented an additional site for chemical modification.
As mentioned before, modification at this site with
hydrophobic side chains has resulted in analogs with
superior activity against epidermal growth factor receptor
PTK. Amide bond formation was investigated with
phenethylamine in the presence of BOP-Cl, DIC, or EDCI
as the coupling agent. Excellent results were obtained
when resin 15 was reacted with 5 equiv of phenethy-
lamine in the presence of EDCI (4 equiv), hydroxyben-
zotriazole (4 equiv), and iPr2NEt (5 equiv) in a DMF/
CH2Cl2 (1/1) mixture (Scheme 4). The support-bound
phenethyl amide 19 was obtained following a double-
coupling protocol to ensure complete conversion of all the
acid sites to the amide. After the resin 19 was subjected
to the standard cleavage conditions, the desired analog
21 was obtained in good yield (85%) and purity (93%)
following filtration through silica gel. In a similar
manner, the support-bound benzyl amide 22 was ob-
tained by cleavage of resin 20 formed from the reaction
of resin 15 with benzylamine. The next step was to
introduce a benzyl group on the secondary nitrogen atom
present in the support-bound amide 19 in an attempt to
The synthesis of lavendustin A (1) on solid support
required a reductive alkylation of resin 11 with 2-hy-
droxybenzaldehyde (12). This reaction proved to be very
sluggish due the unreactivity of both the aldehyde
(18) Neises, B.; Steglich, W. Angew. Chem., Int. Ed. Engl. 1978, 17,
522.
(19) Frenette, R.; Friesen, R. W. Tetrahedron Lett. 1994, 35, 9177.
(20) Lyle, T.; Brady, S.; Ciccarone, T.; Colton, C.; Paleveda, W.;
Veber, D.; Nutt, R. J . Org. Chem. 1987, 52, 3752.
(21) Abdel-Magid, A. F.; Maryanoff, C. A.; Carson, K. G. Tetrahedron
Lett. 1990, 31, 5595.

Versatile Solid Phase Synthesis of Lavendustin A
J . Org. Chem., Vol. 61, No. 26, 1996 9371
Sch em e 3a
Sch em e 4a
a
Reagents and conditions: (i) 3-aminobenzoic acid (14), NaC-
NBH3, 1% AcOH in DMAc, 48 h; (ii) MeOH/Et3N/Me2S (75/15/10),
6 × 3 h; (iii) benzaldehyde, NaCNBH3, 4 Å seives, 1% AcOH in
DMAc, 72 h.
synthesize a fully functionalized lavendustin A analog.
A nucleophilic displacement strategy was chosen in favor
of the rather sluggish reductive amination protocol
utilized previously for resins 11 and 15. Reaction of
support-bound phenethyl amide 19 twice with benzyl
bromide (4 equiv) and proton sponge (1.2 equiv) as the
base in anhydrous DMSO at room temperature for 8 h
afforded the support-bound dibenzylaniline amide 23.
Cleavage of resin 23 under the standard conditions
afforded the desired product 24 in very good purity (94%)
and yield (80%) after filtration through silica gel.
In conclusion, we have demonstrated a convenient solid
phase synthesis of the potent PTK inhibitor lavendustin
A and representative analogs via a base-labile linker
approach utilizing commercially available starting ma-
terials and avoiding protection/deprotection steps. In
addition, the careful optimization of the cleavage condi-
tions furnished the rather sensitive para-hydroxylated
phenolic target molecules in good yields and superior
purities. This methodology, together with the ready
commercial availability of substituted amines, benzal-
dehydes, and benzyl halides offers the potential for the
synthesis of a large library of structurally related analogs
as PTK inhibitors.
a
Reagents and conditions: (i) phenethylamine (for 19) or
benzylamine (for 20), EDCl, HOBt, iPr2NEt, DMF, CH2Cl2, 6 h;
(ii) MeOH/Et3N/Me2S (75/15/10), 6 × 3 h; (iii) benzyl bromide,
proton sponge, DMSO, 8 h.
365 nm. Column chromatography was carried out using
Merck silica gel (230-400 mesh). Analytical reverse phase
high-performance liquid chromatography was done with a
diode-array detector or a UV/vis detector at 254 nm. p-
(Benzyloxy)benzyl alcohol (Wang’s) resin (substitution 0.96
mmol/g) and aminomethyl polystyrene resin (substitution 0.81
mmol/g) were purchased from Calbiochem-Novabiochem In-
ternational. Commercially available starting materials and
reagents were purchased from Aldrich Chemical Company,
Inc. All solvents were dried over 4 Å molecular sieves before
use. DMF was distilled from calcium hydride. Pyridine was
stored over KOH pellets.
Mitson obu Cou p lin g of 2,5-Dih yd r oxyben za ld eh yd e
(4) to Wa n g’s Resin (3) To Affor d Loa d ed Resin 5.
A
solution of 4 (0.28 g, 2 mmol) and triphenylphosphine (0.53 g,
2 mmol) in anhydrous THF (6 mL) was added to a suspension
of washed (THF) Wang’s resin (0.96 mmol/g, 1.04 g, 1 mmol)
preswelled in THF (10 mL). This was followed by the slow
addition of DEAD (0.35 g, 2 mmol) in THF (3 mL). The
mixture was gently stirred for a period of 24 h at room
temperature and filtered. The resin was washed with THF
(3 × 15 mL), DMF (3 × 15 mL), MeOH (3 × 15 mL), THF (3
× 15 mL), and CH2Cl2 (3 × 15 mL) and dried under reduced
pressure. The entire procedure was repeated again to afford
Exp er im en ta l Section
Analytical thin-layer chromatography was performed on
Whatman silica gel 60 K6F glass-coated plates with fluorescent
indicator, and spots were visualized with UV light at 254 or

9372 J . Org. Chem., Vol. 61, No. 26, 1996
Devraj and Cushman
support-bound benzaldehyde. The loading on solid support
was determined to be 0.62 mmol/g (72%) by cleavage of loaded
resin 5 twice with CH2Cl2/TFA/Me2S (50/45/5) for 1.5 h and
isolation of pure 2,5-dihydroxybenzaldehyde.
was filtered, washed well with 1% AcOH in DMAc, H2O,
DMAc, EtOH, and CH2Cl2 (3 × 10 mL each), and dried, and
the reductive amination procedure was repeated to afford the
benzylaniline-linked solid support 11.
Red u ctive Am in a tion of 5 w ith 5-Am in osa licylic Acid
To F or m Loa d ed Resin 7. Derivatized resin 5 (0.62 mmol/
g, 0.32 g, 0.2 mmol) was washed with 1% AcOH in DMAc (2 ×
5 mL). 5-Aminosalicylic acid (0.19 g, 1.2 mmol) was added to
a preswelled suspension of the resin 5 in 1% AcOH in DMAc
(7 mL). Solid NaCNBH3 (0.23 g, 3.6 mmol) was added in
portions over a period of 3 h and the suspension gently stirred
for 48 h at room temperature. The resin was filtered, washed
well with 1% AcOH in DMAc, DMAc, CH2Cl2, and EtOH (3 ×
7 mL, each), and dried under reduced pressure to afford the
benzylaniline-linked solid support 7. Attempted cleavage of
the desired product 2 from this resin resulted in considerable
degradation.
Clea va ge of La ven d u stin A Active P h a r m a cop h or e 2
fr om Solid Su p p or t 11. Resin 11 (100 mg) was subjected to
cleavage (using freshly prepared cleavage mixture) in
a
manner similar to that reported above for the cleavage of 4
from resin 10. The residue was filtered through a syringe
containing silica gel (1.5 × 4 cm) using CHCl3/MeOH (4/1) as
the solvent. The filtrate was evaporated under a stream of
argon, reevaporated from anhydrous EtOH (2×), and then
further dried under reduced pressure over CaSO4 to afford 13.8
mg (86%, based on the loading of 4 on solid support) of
lavendustin A active pharmacophore (2). HPLC analysis (25-
75% MeOH in 0.1% aqueous TFA over 30 min, flow rate 1 mL/
min, Phenomenex C-8 column) indicated this material (reten-
tion time ) 9.04 min) to be 98% pure: 1H NMR (300 MHz,
CD3OD) δ 7.89 (s, 1 H), 7.58 (d, 1 H, J ) 2.7 Hz), 7.11 (dd, 1
H, J ) 2.7, 8.7 Hz), 6.82 (d, 1 H, J ) 8.7 Hz), 6.70 (d, 1 H, J
) 8.7 Hz), 6.67 (d, 1 H, J ) 3 Hz), 6.62 (dd, 1 H, J ) 3, 8.7
Hz), 4.29 (s, 2H); low-resolution FABMS m/ z 276 (MH+); high-
resolution FABMS calculated MH+ 276.0872, found 276.0866.
Red u ctive Alk yla tion of Resin 11 w ith o-Hyd r oxyben -
za ld eh yd e To Affor d 13. Meth od A. Support-bound ben-
zylaniline 11 (0.16 g, 0.1 mmol assuming 100% conversions)
was suspended in 1% AcOH in DMAc (5 mL). o-Hydroxyben-
zaldehyde (12, 0.09 g, 0.7 mmol) was added to this suspension
followed by 4 Å molecular sieves (0.08 g). After 1 h, solid
NaCNBH3 (0.13 g, 2.1 mmol) was added in portions over 2 h
to this suspension and the reaction continued for 72 h at room
temperature. The resin was filtered, washed with 1% AcOH
in DMAc, H2O, DMAc, EtOH, and CH2Cl2 (3 × 8 mL each),
and dried under reduced pressure. This procedure of reductive
alkylation was repeated twice on the dried resin to afford
lavendustin A-linked solid support 13.
Meth od B. o-Hydroxybenzaldehyde (12, 0.73 g, 0.6 mmol)
and 4 Å molecular sieves (0.08 g) were added to a suspension
of the loaded resin 11 (0.16 g, 0.1 mmol) in CH2Cl2/DMF (4/1,
5 mL). Glacial AcOH (57 µL, 1 mmol) was then added and
the reaction continued for 1 h. Solid NaBH(OAc)3 (0.21 g, 1
mmol) was added in portions to this suspension over a period
of 2 h and the reaction continued for 72 h at room temperature.
The resin was filtered, washed with DMF, H2O, EtOH, and
CH2Cl2 (3 × 8 mL each), dried, and resubjected to the same
procedure twice more to afford the desired solid support 13.
Clea va ge of La ven d u stin A (1) fr om Solid Su p p or t 13.
The loaded resin 13 (100 mg) obtained by method A was
subjected to cleavage, using freshly prepared cleavage mixture,
in a manner similar to that reported above for the cleavage of
4 from loaded resin 10. The residue was filtered through a
syringe containing silica gel (1.5 × 4 cm) using CHCl3/MeOH
(4/1) as the solvent. The filtrate was evaporated under a
stream of argon and re-evaporated from anhydrous EtOH (2×)
to afford the desired product 1 as a light brownish-gray solid.
HPLC analysis (35-90% MeOH in 0.1% aqueous TFA over
30 min, flow rate 1 mL/min, Phenomenex C-8 column) of this
material indicated it (retention time ) 10.72 min) to be 82%
pure. An analytical sample was obtained by further purifying
this residue on silica gel (1.5 × 5 cm), eluting with CHCl3/
MeOH (4/1). Evaporation under a stream of argon afforded
14.9 mg (72%, based on loading of 4 on solid support) of
lavendustin A (1): 1H NMR (CD3OD, 300 MHz) δ 7.68 (d, 1
H, J ) 2.7 Hz), 7.25 (dd, 1 H, J ) 2.7, 8.8 Hz), 7.14 (dd, 1 H,
J ) 7.5, 8.0 Hz), 7.04 (d, 1 H, J ) 7 Hz), 6.81 (d, 1 H, J ) 8.0
Hz), 6.73 (2 overlapping d, 2 H, J ) 7, 8.8 Hz), 6.66 (d, 1 H, J
) 8.7 Hz), 6.58 (dd, 1 H, J ) 3, 8.7 Hz), 6.52 (d, 1 H, J ) 3
Hz), 4.66 (s, 2 H), 4.60 (s, 2 H); low-resolution FABMS m/ z
382 (MH+); high-resolution FABMS calculated MH+ 382.1291,
found 382.1291.
Der ivatization of Am in om eth ylated P olystyr en e Resin
(8) w ith Su ccin ic An h yd r id e To Yield 9. Commercially
available aminomethylated polystyrene resin 8 (manufactur-
er’s substitution 0.81 mmol/g, 2.47 g, 2 mmol) was preswelled
in pyridine (35 mL). A solution of succinic anhydride (0.6 g, 6
mmol) in pyridine (12 mL) was added to this suspension and
the mixture gently stirred for 24 h at room temperature. The
resin was then filtered and washed well with pyridine, DMF,
CH2Cl2, and EtOH (3 × 20 mL each). The resin was then dried
under reduced pressure over KOH pellets and subjected to
another round of reaction to afford the desired succinic acid-
derivatized resin 9: IR (KBr) 1721, 1655 cm-1
.
Ester ifica tion of Resin 9 w ith 2,5-Dih yd r oxyben za l-
d eh yd e To Give 10. Meth od A. Solid DCC (1.24 g, 6 mmol)
and HOBt (0.81 g, 6 mmol) were added to a suspension of resin
9 in DMF/pyridine (9/1, 30 mL). Solid 2,5-dihydroxybenzal-
dehyde (4, 0.83 g, 1.2 mmol) was added next and the mixture
gently stirred at room temperature for 24 h. The resin was
filtered, washed with pyridine (1 × 25 mL) and DMF, CH2-
Cl2, and EtOH (3 × 25 mL each), and dried under reduced
pressure over KOH pellets. This procedure was repeated again
on this resin to afford the benzaldehyde-linked solid support
10.
Meth od B. To a suspension of dried resin 9 in DMF/CH2-
Cl2 (4/1, 30 mL) were added DCC (1.24 g, 6 mmol) and DMAP
(0.86 g, 7 mmol) followed by 2,5-dihydroxybenzaldehyde (0.83
g, 1.2 mmol). The suspension was gently stirred for 24 h at
room temperature and filtered. The resin was washed with
DMF, CH2Cl2, and EtOH (3 × 25 mL each) and then dried
under reduced pressure. This coupling protocol was repeated
again to afford the support-bound benzaldehyde 10.
Cleavage of 2,5-Dih ydr oxyben zaldeh yde (4) fr om Solid
Su p p or t 10. The cleavage cocktail was prepared by degassing
a solution of MeOH (7.5 mL) and Et3N (1.5 mL) with argon.
Dimethyl sulfide (1 mL) was then added to this mixture, and
the cleavage cocktail was used immediately. This solution (4
mL) was added to dried resin 10 (100 mg) obtained from
method A, and the mixture was gently stirred under argon
for 3 h at room temperature. The resin was filtered and
washed with MeOH (2 mL) and the filtrate evaporated at once
under a stream of argon. The resin was subjected to this
procedure a total of six times. The combined residue was
filtered through a syringe packed with silica gel (1.5 × 4 cm)
using 4/1 CHCl3/MeOH as the solvent. The filtrate was
evaporated under a stream of argon to afford 8.8 mg (92%) of
the desired product 4 as a yellow solid: mp 99-100 °C; 1H
NMR (300 MHz, CDCl3) δ 10.61 (s, 1 H), 9.82 (s, 1 H), 7.08
(dd, 1 H, J ) 3, 9 Hz), 7.01 (d, 1 H, J ) 3 Hz), 6.91 (d, 1 H, J
) 9 Hz), 4.95 (s, 1 H); low-resolution FABMS m/ z 139 (MH+).
In a similar manner the cleavage of resin 10 (100 mg)
obtained by method B afforded 8 mg (84%) of pure 2,5-
dihydroxybenzaldehyde (4).
Red u ctive Am in a tion of Resin 10 w ith 5-Am in osa li-
cylic Acid To Affor d 11. 5-Aminosalicylic acid (6) (0.28 g,
1.8 mmol) was added to a suspension of resin 10 (0.43 g, 0.3
mmol assuming 100% reaction over the previous steps) in 1%
AcOH in DMAc (8 mL). Solid NaCNBH3 (0.34 g, 5.4 mmol)
was added in portions over 3 h to the suspension and the
reaction continued for 48 h at room temperature. The resin
In a similar manner cleavage of resin 13 (100 mg), obtained
by method B, afforded after purification on silica gel (as
mentioned above) 15.9 mg (77%) of lavendustin A identical in
all respects to the sample obtained above.
Red u ctive Am in a tion of Resin 10 w ith 3-Am in oben -
zoic Acid To Give 15. The reductive amination was per-

Versatile Solid Phase Synthesis of Lavendustin A
J . Org. Chem., Vol. 61, No. 26, 1996 9373
formed as described for the synthesis of resin 11 using 6 equiv
of 3-aminobenzoic acid (14) and 18 equiv of NaCNBH3. After
a second round of reductive amination, the resin was dried
under reduced pressure to afford the desired support-bound
benzylaniline 15.
resin 19 (100 mg) was cleaved to afford an oil, which was
filtered through a syringe containing silica gel (1.5 × 4 cm)
using CHCl3/MeOH (8/1). The filtrate was evaporated under
a stream of argon, re-evaporated from anhydrous EtOH (2×),
and dried under reduced pressure to afford 17 mg (85%, based
on the loading of 4 on solid support) of lavendustin A analog
21 as a dark yellow solid. HPLC analysis (35-90% MeOH in
0.1% aqueous TFA over 30 min, flow rate 1 mL/min, Phe-
neomenex C-8 column) indicated that the solid (retention time
) 15.40 min) was 93% pure: 1H NMR (CD3OD, 300 MHz) δ
7.1-7.3 (m, 6 H), 7.02 (m, 1 H), 6.94 (d, 1 H, J ) 7.5 Hz), 6.76
(dd, 1 H, J ) 8 Hz), 6.71 (d, 1 H, J ) 3 Hz), 6.63 (d, 1 H, J )
8.4 Hz), 6.50 (dd, 1 H, J ) 3, 8.4 Hz), 4.26 (s, 2 H), 3.54 (t, 2
H, J ) 7.5 Hz), 2.87 (t, 2 H, J ) 7.5 Hz); low-resolution FABMS
m/ z 363 (MH+); high-resolution FABMS calculated MH+
363.1709, found 363.1717.
Clea va ge of An a log 22 fr om Resin 20. In a manner
similar to that reported above for the cleavage resin 19, resin
20 (100 mg) was cleaved to afford 15.7 mg (82%) of the
lavendustin A analog 22 as a yellowish-orange solid. HPLC
analysis (20-95% CH3CN containing 0.05% H3PO4 in 0.05%
aqueous H3PO4 over 35 min, flow rate 1 mL/min, Vydac C-18
column) indicated that this material (retention time ) 17.65
min) was 82% pure: 1H NMR (CD3OD, 300 MHz) δ 7.02-7.35
(m, 7 H), 6.77 (m, 1 H), 6.71 (d, 1 H, J ) 3 Hz), 6.62 (d, 1 H,
J ) 9 Hz), 6.50 (dd, 1 H, J ) 3, 9 Hz), 4.53 (s, 2 H), 4.26 (s, 2
H); low-resolution FABMS m/ z 349 (MH+); high-resolution
FABMS calculated MH+ 349.1476, found 349.1468.
Alk yla tion of Resin 19 w ith Ben zyl Br om id e To Affor d
23. Benzyl bromide (48 µL, 0.4 mmol) and proton sponge (20
mg, 0.12 mmol) were added to a suspension of the dried resin
19 (0.17 g, 0.1 mmol) in anhydrous DMSO (3 mL). This
suspension was gently stirred at room temperature for 8 h.
The resin was filtered and washed well with DMSO, EtOH,
and CH2Cl2 (3 × 5 mL each). The resin was dried and the
alkylation procedure repeated again to afford the benzylated
solid support 23.
Clea va ge of An a log 24 fr om Resin 23. Utilizing the
standard conditions reported above, resin 23 (100 mg) was
cleaved to afford an oil which was filtered through silica gel
(1.5 × 4 cm) using CHCl3/MeOH (8/1). The filtrate was
evaporated under a stream of argon, re-evaporated twice from
anhydrous EtOH, and dried to afford 18.9 mg (80%) of
lavendustin A analog 24 as a dark yellow solid. HPLC
analysis (35-90% MeOH in 0.1% aqueous TFA over 30 min,
flow rate 1 mL/min, Phenomenex C-8 column) indicated that
this material (retention time ) 26.07 min) was 94% pure: 1H
NMR (CD3OD, 300 MHz) δ 7.12-7.33 (m, 12 H) 6.96 (d, 1 H,
J ) 7.5 Hz), 6.81 (dd, 1 H, J ) 3, 7.5 Hz), 6.65 (d, 1 H, J ) 8.1
Hz), 6.51 (overlapping peaks, 2 H), 4.69 (s, 2 H), 4.59 (s, 2 H),
3.51 (t, 2 H, J ) 7.2 Hz), 2.84 (t, 1 H, J ) 7.2 Hz);
low-resolution FABMS m/ z 453 (MH+); high-resolution FABMS
calculated MH+ 453.2178, found 453.2160.
Clea va ge of An a log 16 fr om Resin 15. Resin 15 (100
mg) was subjected to cleavage (using freshly prepared cleavage
mixture) in a manner similar to that reported above for the
cleavage of compound 4 from resin 10. The residue was
filtered through a syringe containing silica gel (1.5 × 4 cm)
using CHCl3/MeOH (4/1) as the solvent. The filtrate was
evaporated under a stream of argon, evaporated twice again
after addition of anhydrous EtOH each time, and then further
dried under reduced pressure to afford 13.4 mg (88%, based
on the loading of 4 on solid support) of lavendustin A active
pharmacophore analog 16. HPLC analysis (25-75% MeOH
in 0.1% aqueous TFA over 30 min, flow rate 1 mL/min,
Phenomenex C-8 column) indicated that this material was 92%
pure (retention time ) 10.45 min): 1H NMR (CD3OD, 300
MHz) δ 7.31 (d, 1 H, J ) 3 Hz), 7.26 (d, 1 H, J ) 7.8 Hz), 7.15
(dd, 1 H, J ) 7.5, 8.1 Hz), 6.83 (dd, 1 H, J ) 3, 7.5 Hz), 6.71
(d, 1 H, J ) 3 Hz), 6.63 (d, 1 H, J ) 8.8 Hz), 6.50 (dd, 1 H, J
) 3, 8.8 Hz), 4.27 (s, 2 H); low-resolution FABMS m/ z 259
(M+); high-resolution FABMS calculated M+ 259.0845, found
259.0842.
Red u ctive Alk yla tion of Resin 15 w ith Ben za ld eh yd e
To P r ovid e Loa d ed Resin 17. A triple reductive amination
of resin 15 with benzaldehyde (6 equiv) and NaCNBH3 (18
equiv) was performed as described above for the synthesis of
resin 13 to give the dibenzylaniline-linked polystyrene support
17.
Clea va ge of An a log 18 fr om Resin 17. Cleavage of resin
17 (100 mg) using the standard procedure reported above,
followed by filtration through silica gel, afforded 18 as a yellow
solid. HPLC analysis (35-90% MeOH in 0.1% aqueous TFA
over 30 min, flow rate 1 mL/min, Phenomenex C-8 column)
indicated that this material (retention time ) 22.06 min) was
78% pure. This sample was further purified on silica gel (1.5
× 5 cm), eluting with CHCl3/MeOH (4/1). Evaporation of the
solvent under a stream of argon afforded 14 mg (73%, based
on the substitution level of resin 10) of lavendustin A analog
18 as a pale yellow solid: 1H NMR (CD3OD, 300 MHz) δ 8.0
(d, 1 H, J ) 8.4 Hz), 7.15-7.58 (m, 7 H), 6.88 (dd, 1 H, J ) 3,
8.4 Hz), 6.64 (d, 1 H, J ) 8.4 Hz), 6.52 (overlapping d and dd,
2 H, J ) 3 Hz and J ) 3, 8.4 Hz), 4.70 (s, 2 H), 4.60 (s, 2 H);
low-resolution FABMS m/ z 349 (M+); high-resolution FABMS
calculated M+ 349.1314, found 349.1307.
Am id a tion of Resin 15 w ith P h en eth yla m in e To Af-
for d Loa d ed Resin 19. Resin 15 (0.31 g, 0.2 mmol, assuming
100% conversion) was suspended in DMF/CH2Cl2 (1/1, 6 mL),
and diisopropylethylamine (176 µL, 1 mmol) was added to this
suspension. Solid EDCI‚HCl (0.15 g, 0.8 mmol) and HOBt
(0.11 g, 0.8 mmol) were added to the mixture, followed
immediately by the addition of phenethylamine (125 µL, 1
mmol). The reaction was continued for 6 h at room temper-
ature. The resin was then filtered and washed with DMF,
EtOH, and CH2Cl2 (3 × 8 mL each). The resin was dried under
reduced pressure, and the coupling protocol was repeated
again on this resin to afford the phenethylamide-linked solid
support 19.
Ack n ow led gm en t. This research was made possible
by contract NO1-CM-17512, awarded by the National
Cancer Institute.
Su p p or tin g In for m a tion Ava ila ble: 1H NMR spectra
and HPLC traces of 1, 2, 16, 18, 21, 22, and 24 (14 pages).
This material is contained in libraries on microfiche, im-
mediately follows this article in the microfilm version of the
journal, and can be ordered from the ACS; see any current
masthead page for ordering information.
Am id a tion of Resin 15 w ith Ben zyla m in e To Yield 20.
In a manner similar to that described above for the synthesis
of solid support 19, benzylamine was coupled with resin 15 to
afford benzylamide-linked solid support 20.
Clea va ge of An a log 21 fr om th e Loa d ed Resin 19.
Utilizing the standard conditions reported above, the loaded
J O961719L