A Novel Cleavage Technique to Generate Small Molecule Compounds and Libraries via a Two-Resin System

Article abstract of DOI:10.1021/jo971347w

Application of organic synthesis to solid supports has led to the successful implementation of combinatorial chemistry in the drug discovery process. This paper describes a novel use of the Hofmann elimination of tetrasubstituted amine salts on solid-phas

Full text of DOI:10.1021/jo971347w

J . Org. Chem. 1998, 63, 1027-1032
1027
A Novel Clea va ge Tech n iqu e To Gen er a te Sm a ll Molecu le
Com p ou n d s a n d Libr a r ies via a Tw o-Resin System
Xiaohu Ouyang, Robert W. Armstrong, and Martin M. Murphy*
Department of Small Molecule Drug Discovery, Amgen Inc., 1840 DeHavilland Drive,
Thousand Oaks, California 91320-1789
Received J uly 22, 1997
Application of organic synthesis to solid supports has led to the successful implementation of
combinatorial chemistry in the drug discovery process. This paper describes a novel use of the
Hofmann elimination of tetrasubstituted amine salts on solid-phase resin to generate diverse
combinatorial libraries of trisubstituted amines. Highly pure compounds were isolated without
further purification by the addition of a second resin as the source reagent to promote the required
elimination. The use of mixed resin systems to generate compounds is a novel application of bead-
based technologies.
In tr od u ction
laboratories, in which we have expanded the scope of the
inputs to attain an acceptable number of inputs to
generate large libraries (>10 000) of trisubstituted amines.
Figure 2 is an example of our early general synthesis
methodology involving the use of monosubstituted amines.
In addition, we will discuss our efforts in which we
developed new methodology to bypass the problem of
extraction but still maintain high yields and acceptable
purity standards.
In an effort to define the generality of the Michael
addition a large variety of amines were employed. The
diversity of disubstituted amines, cyclic and noncyclic,
was explored first as only a second step of alkylation was
needed prior to elimination. These inputs were found to
proceed smoothly. In addition, a number of monosub-
stitiuted amines were employed successfully. This was
immediately followed by the reductive amination of
aldehyde and ketone inputs which allowed for greater
diversity. To examine the products of the Michael
addition step, a select but diverse number of alkylating
reagents were used to study the methodology necessary
to generate the desired compounds. Table 1 contains a
partial list of amines, aldehydes were applicable, and
alkylating agents employed in this study.
As reported by Morphy et al.,4a we have also demon-
strated that the Hofmann elimination could be utilized
to generate the desired trisubstituted amines by the use
of an excess of triethylamine (TEA). Unfortunately, after
Hofmann elimination only the use of time-consuming
aqueous extraction effectively provided for the full re-
moval of the unwanted triethylamine hydrobromide or
iodide salts formed in the elimination step. A recently
published follow-up paper by Morphy,4b cited an improved
extraction method by the use of solid-phase biphasic
columns. However both of these techniques create a
serious problem if one is making large libraries (>1000
compounds) or one uses automation equipment. An
alternative strategy studied was the application of heat7
which produced compounds that contained no TEA salts,
The emergence of combinatorial chemistry as a major
contributor to the drug discovery process1 has led to the
very rapid development of novel synthetic methodologies2
for the organic synthesis of nonpeptidyl small molecules
on solid support resins. In addition, recent advances in
the use of new types of linker strategies2a,3 have expanded
the diversity of compounds generated upon cleavage from
the resin. In particular, a recent report4 described the
use of a linker strategy in which substituted amines were
added in a Michael fashion5 to an acrylate bound resin
as seen in Figure 1.
Subsequent alkylation to quaternize the trisubstituted
amine followed by the addition of a suitable base allowed
for the Hofmann elimination of the desired product and
the regeneration of the starting acrylate resin. A limita-
tion in this effort was noted by the lack of diverse
monosubstituted amines implemented and the cited need
to extract each compound to ensure an acceptable level
of purity;4a this is especially troublesome in the case of
large library production6 which might employ automation
equipment. We report on concurrent efforts in our
* Phone: (805) 447-1923. Fax: (805) 499-5714. E-mail: mmurphy@
amgen.com.
(1) (a) Gallop, M. A.; Barrett, R. W.; Dower, W. J .; Fodor, S. P. A.;
Gordon, E. M. J . Med. Chem. 1994, 37, 1233-1251. (b) Gallop, M. A.;
Barrett, R. W.; Dower, W. J .; Fodor, S. P. A.; Gordon, E. M. J . Med.
Chem. 1994, 37, 1384-1401.
(2) For recent reviews on solid-phase organic synthesis, see (a)
Thompson, M. A.; Ellman, J . A. Chem. Rev. (Washington, D.C.) 1996,
96, 555-600. (b) Terrett, N. K.; Gardner, M.; Gordon, D. W.; Kobylecki,
R. J .; Steele, J . Tetrahedron 1995, 51, 8135-8173. (c) Fruchtel, J . S.;
J ung, G. Angew. Chem., Int. Ed. Engl. 1996, 35, 17-42. (d) Balkenhohl,
F.; von dem Bussche-Hunnefeld, C.; Lansky, A.; Zechel, C. Angew.
Chem., Int. Ed. Engl. 1996, 35, 2288-2337.
(3) Most recently (a) Han, Y.; Walker, S. D.; Young, R. N. Tetrahe-
dron Lett. 1996, 37, 2703-2706. (b) Beaver, K. A.; Seigmund, A. C.;
Spear, K. L. Tetrahedron Lett. 1996, 37, 1145-1148. (c) Boehm, T. L.;
Hollis Showalter, H. D. J . Org. Chem. 1996, 61, 6498-6499. (d) Ngu,
K.; Patel, D. V. Tetrahedron Lett. 1997, 38, 973-976. (e) Routledge,
A.; Abell, C.; Balasuramanian. Tetrahedron Lett. 1997, 38, 1227-1230.
(f) See refs 2a and 2b for a list of other linkers and strategies.
(4) (a) Morphy, J . R.; Rankovic, Z.; Rees, D. C. Tetrahedron Lett.
1996, 37, 3209-3212. (b) Brown, A. R.; Rees, D. C.; Rankovic, Z.;
Morphy, J . R. J . Am. Chem. Soc. 1997, 119, 3288-3295.
(6) 1H NMR spectra of the compounds were found to contain a large
excess of triethylamine. An example in the Supporting Information is
provided.
(7) The resin with a quaternary ammonium compound in DMF was
simply heated at 60-80 °C.
(5) (a) Ley, S. V.; Mynett, D. M.; Koot, W.-J . Synlett. 1995, 60, 6006-
6007. (b) Cody, D. R.; DeWitt, S. H. H.; Hodges, J . C.; Kiely, J . S.;
Moos, W. H.; Pavia, M. R.; Roth, B. D.; Schroeder, M. C.; Stankovic,
C. J . PCT Int. App. WO 9408711; Chem Abstr. 1995, 122, 106536.
S0022-3263(97)01347-9 CCC: $15.00 © 1998 American Chemical Society
Published on Web 01/28/1998

1028 J . Org. Chem., Vol. 63, No. 4, 1998
Ouyang et al.
F igu r e 1. Synthesis of trisubstituted amines.
F igu r e 2. Combinatorial synthesis of trisubstituted amines using three building blocks.
Ta ble 1. Sa m p le of Bu ild in g Block s Used
In an attempt to solve the paradigm of attaining high
purity by an alternate means, we first attempted cleavage
of the quaternary amine salts from resin by the use of 1
1
equiv of TEA. This proved to be helpful as the H NMR
spectra of the cleaved compounds contained less TEA
salts.11 However, there was still an undesired excess of
the base still present.
At the same time a series of articles had described the
use of resin bound scavengers which were able to
sequester excess reagents and leave the desired pure
products in the solution. Most notably, ion-exchange
resins had been demonstrated to be useful as reagents
and scavengers in a variety of solution-phase parallel
syntheses so as to remove unwanted reaction byproducts
and eliminate the need for aqueous workups.9 More
recently, applications involving specific functional groups
have been introduced with outstanding results.10
Resu lts a n d Discu ssion
It was postulated that the use of a basic ion-exchange
resin might act as a regenerating agent for a solution
(9) (a) Parlow, J . J . Tetrahedron Lett. 1996, 37, 5257-5260. (b) Gayo,
L. M.; Suto, M. J . Tetrahedron Lett. 1997, 38, 513-516.
(10) (a) Kaldor, S. W.; Siegel, M. G.; Fritz, J . E.; Dressman, B. A.;
Hahn, P. J . Tetrahedron Lett. 1996, 37, 7193-7196. (b) Kaldor, S. W.;
Fritz, J . E.; Tang, J .; McKinney, E. R. Bio. Med. Chem. Lett. 1996, 6,
3041-3044. (c) Flynn, D. L.; Crich, J . Z.; Devraj, R. V.; Hockerman, S.
L.; Parlow, J . J .; South, M. S.; Woodard, S. J . Am. Chem. Soc. 1997,
119, 4874-4881. (d) Booth, R. J .; Hodges, J . C. J . Am. Chem. Soc. 1997,
119, 4882-4886.
but it was observed that the products were contaminated
with other unidentified impurities.8 Neither of these
methods was found to give satisfactory levels of depend-
able yield and/or purity of the desired products.
(11) 1H NMR spectra of various compounds were seen to contain
TEA in a range of 5-60% of an equivalence. Recently Brown et al.
utilized powdered K2CO3 to effectively deprotonate the diisopropyl-
ethylamine salt in situ. See ref 4b.
(8) 1H NMR spectra of the compounds were found to contain a
unidentified polymeric contamination.

Small Molecule Compounds via a Two-Resin System
J . Org. Chem., Vol. 63, No. 4, 1998 1029
F igu r e 3. One-pot Hofmann elimination and extraction.
F igu r e 4. Possible mechanism for Amberlite-promoted Hofmann elimination.
amine catalyst which could promote the Hofmann elimi-
nation reaction. In this process, one could theoretically
accomplish the cleavage step of the product amine and
at the same time remove undesired the salt with an
excess of a second resin bound amine. Evaporation of
the filtrate would afford one pure product free of the
undesired amine salts as indicated by Figure 3.
in the same vessel.12 The remarkable aspect of this
reaction is that the presence of two polymeric reagents is
the only requirement to effect the transformation. Two
possible explanations for this effect can be reasoned.
First, the resin containing the quaternary ammonium
compounds undergoes a thermal elimination of a small
amount of product. The newly created amine hydrobro-
mide salt acts upon the ion-exchange Amberlite, or the
rink resin, losing the salt and transforming to the salt
free amine in a catalytic like cycle. This derived amine
can then act as the base, promoting Hofmann elimination
to provide more product as shown in Figure 4. An
alternative explanation could be that there remains a
trace amount of residual base from previous handling in
the matrix of the resins which could initiate the elimina-
tion reaction.
In an effort to evaluate different Hofmann elimination
conditions, we picked representative examples from
several trisubstituted amine libraries and conducted
Hofmann elimination reactions using TEA, Amberlite
IRA-95 basic ion-exchange resin, and deprotected rink
amide polystyrene resin. It must be noted that the yields
are based upon the reported loading of hydroxymethyl
resin (a single batch was used for this experiment). As
shown Table 2, compounds 1-6 were synthesized by the
Michael addition of a secondary amine to the acrylate
resin, followed by quarternization employing various
alkylating agents (alkyl, allyl, propargyl, and benzyl
Upon implementation of this protocol, this strategy was
found to be quite successful. Simple agitation overnight
of the resin bound quaternary ammonium compound, a
trace amount of TEA and the weakly basic ion-exchange
resin Amberlite in DMF, produced the highly pure
desired product with good yields after filtration and
evaporation of the filtrate.
Further experimentation with Amberlite weakly basic
ion-exchange resin in DMF and the resin bound quater-
nary ammonium compounds, in the absence of any base,
proved to be successful. Highly pure products were
isolated after routine filtration and evaporation of the
excess filtrate. It was found that yield was a function of
time with the optimal reaction length 18 h for library
production. Lastly, we performed the same experiment
(no base) using deprotected rink amide resin. To our
surprise we obtained identical results to those of highly
purified products. Less efficient yields were observed
though an extended reaction time to give comparable
results.
A quick review of the literature revealed only one other
example on the use of more than one polymeric reagent
(12) Parlow, J . J . Tetrahedron Lett. 1995, 36, 1395-1396 and
references therein.

1030 J . Org. Chem., Vol. 63, No. 4, 1998
Ta ble 2. Com p a r in g Differ en t Hofm a n n Elim in a tion Con d ition s
Ouyang et al.
a
Yield based upon use of 1.00 g of hydroxymethyl resin with a labeled loading of 1.2 mmol/g. The Hofmann elimination for all three
experiments was carried out for 18 h. Evaporation of solvents was conducted at the same length of time for accurate comparison of three
b
experiments. Note: increased reaction times for Amberlite and rink amide results in substantially higher yields. Purity was estimated
d
from the crude 1H NMR ((2.5%). c Reaction sequence: 1,2,3,4-tetrahydroisoquinoline and 4-fluorobenzyl bromide. Reaction sequence:
f
3-methylpiperidine and bromoacetophenone. e Reaction sequence: butylamine, benzaldehyde, and allyl bromide. Reaction sequence:
g
h
butylamine, butyraldehyde, and benzyl bromide. Reaction sequence: benzylamine, cyclopentanone, and methyl iodide. Reaction
i
sequence: phenethylamine, cyclohexanone, and methyl iodide. Reaction sequence: piperazine, benzoyl chloride, and methyl iodide.
Reaction sequence: piperazine, phenyl isocyanate, and benzyl bromide.
j
bromides and R-bromo ketones) and subsequent Hofmann
elimination. The Hofmann precursor for compounds
7-10 was synthesized by Michael addition of primary
amines, benzylamine and phenethylamine, followed by
reductive alkylation with aliphatic, aromatic aldehydes
and cyclic ketones. Quarternization and subsequent
Hofmann elimination provided the desired compounds.
Compounds 11-14 were generated by the simply use of
piperazine as a linker, followed by acylation or urea
formation; alkylation, and elimination thus provided the
unsymmetrical piperazine derivatives.
extraction method. The rink amide method displayed
slightly lower yields in some of the examples, which
indicated lower efficiency of ion exchange with the HBr
salt, which could due to the lower loading of the amine
functionality on rink amide than Amberlite or the basic-
ity of the resin.
To compare and contrast the effects of different Hof-
mann elimination conditions on resin, compound 2 was
1
studied under five different conditions. The H NMR of
the Hofmann elimination products in deuterated Metha-
nol is displayed in Figure 5.
In general, the Amberlite method provided slightly
higher yields than that involving the use of the TEA/
Spectrum B (yield 65.2%) demonstrates that aqueous
extraction effectively remove the undesired TEA salts

Small Molecule Compounds via a Two-Resin System
J . Org. Chem., Vol. 63, No. 4, 1998 1031
discovered a novel application of solid-solid reactions by
employing the ion-exchange resin Amberlite. The rink
amide resin is also effective, albeit at lower efficiency.
The Amberlite method has proven to be most successful
for our library production. Amberlite-promoted Hofmann
eliminations in solution-phase organic syntheses have
proven to be successful, as several examples have shown.
Results on this application and other novel solid-solid
interactions will be forthcoming.
Exp er im en ta l Section
Ma ter ia ls. All solid-phase reactions were carried out
at room temperature. Reagents were purchased from
Aldrich and Acros and used without further purification.
Amberlite IRA-95 weakly basic ion exchange was pur-
chased from Sigma and was washed before use. (Hy-
droxymethyl and rink amide) polystyrene resins were
purchased from Novabiochem with a loading of 1.20 and
0.68 mmol/g, respectively.
Gen er a l Meth od s. All reactions were carried out in
peptide synthesis vessels and agitated on an orbit shaker.
Concentration of solutions after workup was performed
by reduced pressure rotary evaporation. NMR spectra
were obtained on a 500 MHz instrument with CDCl3 as
the solvent unless noted. Infrared spectra were obtained
on a Perkin-Elmer IR spectrometer Spectra 2000. Data
were reported in % transmittance. High-resolution mass
spectra (HRMS) were obtained from electron spray mass
spectrometer. Gas chromatography was performed on a
Hewlett-Packard HP 6890 series GC system with a HP
G1099 column and a HP 5973 mass selective detector.
Gen er a l P r oced u r e for P r ep a r in g Acr yla te Resin
on Solid P h a se. A typical experimental procedure is
as follows: hydroxymethyl polystyrene resin (35 g, 42
mmol) [1.2 mmol/g] was added to a 1 L reaction vessel.
After the addition of dichloromethane (250 mL) and
diisopropylethylamine (73 mL, 420 mmol), followed by
four equal portion additions of acryloyl chloride (34 mL,
420 mmol). The mixture was slowly stirred for 4 h. The
resin was then washed with dichloromethane (500 mL)
and methanol (500 mL) for four cycles. The procedure
was repeated, and the acylated resin was air-dried.
Gen er a l P r oced u r e for P r ep a r in g Tr isu bstitu ted
Am in e Syn th esis on Solid P h a se. All trisubstituted
amines gave clean 400 MHz 1H NMR, 13C NMR, and GC-
MS. A portion of the acrylate resin (5 g, 6 mmol) was
mixed with 50 mL of DMF and butylamine (11.9 mL, 120
mmol) and agitated over 18 h. The resin was washed
with DMF (1 × 50 mL), dichloromethane (1 × 50 mL),
methanol (2 × 50 mL), and ether (1 × 50 mL) and dried
in vacuo. To the resin were added 80 mL of DMF,
benzaldehyde (12.2 mL, 120 mmol), and glacial acetic acid
(200 µL), and the mixture was shaken for 0.5 h. NaBH-
(OAc)3 (25 g, 120 mmol) was added to the mixture and
left to agitate for 18 h. The resin was washed with
methanol (2 × 50 mL), DMF (2 × 50 mL), and methanol
(2 × 50 mL) and subsequently dried. The dry resin was
then suspended in a solution of allyl bromide (10.4 mL,
120 mmol) in DMF (50 mL) and agitated over another
18 h. Filtration was followed by rinsing with DMF (1 ×
50 mL), dichloromethane (2 × 50 mL), and methanol (2
× 50 mL) and dried in vacuo. To three peptide vessels
were added exactly 1.00 g of the dry resin with tetra-
substituted amine salts. To the first vessel were added
167 µL of TEA (1.2 mmol) and 40 mL of DMF, to the
F igu r e 5. Comparison of methods to promote Hofmann
elimination.
that are observed in the crude product A (yield 108.7%).
[Note: spectrum A contains an unusually low amount of
the base triethylamine; examples of 1H NMR spectra
containing large amounts of TEA in the crude products
are found in Supporting Information.] The products
exposed to a basic resin, C (yield 68.6%) and E (yield
37.8%), can be seen to be identical in quality with respect
to B. Spectrum D (yield 82.9%) indicates a highly pure
product, but a slight shift of some proton signals is noted.
This shift effect was seen in other products under similar
conditions.
Con clu sion
In summary, we have developed a route to generate
large diverse libraries of trisubstituted amines under
very mild conditions which gives the desired products in
high yields and eliminates the need for additional time-
consuming purification steps. In addition, we have

1032 J . Org. Chem., Vol. 63, No. 4, 1998
Ouyang et al.
second vessel were added 3.0 g of pre-washed Amberlite
IRA-95 ion-exchange resin (4.7 mequiv/g) and 40 mL of
DMF, and to the third vessel were added 3.0 g of
deprotected rink amide resin [Novabiochem, 0.68 mmol/
g] and 40 mL of DMF. All three vessels were agitated
over 18 h. Vessel one was filtered and evaporated to give
a white solid which was partitioned between 10 mL of
DCM and 10 mL of 5% aqueous sodium bicarbonate. The
organic layer was dried over MgSO4, passed through a
frit filter, and rotary evaporated. A colorless thick oil
was obtained (79 mg, 32% overall yield). Vessels two and
three were both filtered and evaporated. The Amberlite
method yielded 91 mg (37% yield, 99% purity by GC/MS)
of material. The rink amide method yielded 40 mg (16%
yield, 99% purity by GC/MS) of material. All compounds
were characterized by proton and carbon NMR, GC-MS,
and HR-MS.
Ack n ow led gm en t. We thank Alex Virgillio and
Alex Kiselyov for key suggestions on ion-exchange
resins. In addition we thank Sherri Lee, Wade Russu,
Marcus Strawn, Tom Huang, and Paul J ohnston for
implementation of library synthesis and Ying Luo for
assistance on NMR.
Su p p or tin g In for m a tion Ava ila ble: Characterization
1
data for all compounds, including 400 and 500 MHz H NMR
and 13C NMR data and spectra, FT-IR, GC-MS, and HR-MS
data, compounds 1-14 (66 pages). This material is contained
in libraries on microfiche, immediately 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.
J O971347W