Asymmetric catalysis on sequentially-linked columns

Article abstract of DOI:10.1021/ja016556j

We report a catalytic asymmetric reaction process that involves the use of solid-phase reagents and catalysts that constitute the packing of a series of "reaction columns". This process was applied to the catalytic asymmetric synthesis of β-lactams, to yi

Full text of DOI:10.1021/ja016556j

J. Am. Chem. Soc. 2001, 123, 10853-10859
10853
Asymmetric Catalysis on Sequentially-Linked Columns
Ahmed M. Hafez, Andrew E. Taggi, Travis Dudding, and Thomas Lectka*
Contribution from the Department of Chemistry, Johns Hopkins UniVersity, 3400 North Charles Street,
Baltimore, Maryland 21218
ReceiVed July 6, 2001
Abstract: We report a catalytic asymmetric reaction process that involves the use of solid-phase reagents and
catalysts that constitute the packing of a series of “reaction columns”. This process was applied to the catalytic
asymmetric synthesis of â-lactams, to yield pure products with excellent enantio- and diastereoselectivity. We
have identified several advantages to conducting chemical reactions on sequential columns, including ease of
catalyst and reagent recovery and simplified purification steps that preclude the need for chromatography.
Introduction
Scheme 1. Sequential Column Asymmetric Catalysis
Assemblies
Technological advances in synthetic chemistry have given
chemists a large number of choices in the way that they conduct
chemical reactions. In regard to chiral synthesis and drug
discovery, asymmetric catalysis1 and solid-phase chemistry2 are
especially important. Chiral catalysts are usually the most costly
components of asymmetric reactions, and ideally they should
be recoverable from the reaction mixture in a straightforward
fashion.3 In practice, this is rarely the casesthe sensitive catalyst
either decomposes or is difficult to reisolate chromatographi-
cally. Additionally, to be attractive to industry, the process
should avoid the use of column chromatography, and additional
reagents and solvents must be reusable whenever possible. With
these considerations in mind, we have developed a catalytic,
asymmetric reaction process involving the use of columns
packed with solid-phase reagents and catalysts. Substrates are
added to the top of a series of columns, and through gravity or
pressure percolate through the packing, where they react to
produce an enantioenriched product that is eluted at the bottom.
Afterward, solvent-based regeneration of the system prepares
it for another cycle of catalysis. We detail herein what we term
“sequential column asymmetric catalysis” (sequential CAC) and
apply it to the catalytic, asymmetric synthesis of â-lactams,
affording products in good yields, good to excellent diastereo-
selectivity and enantioselectivity (ee). We also discuss how
practical considerations of reagent expense and efficiency led
us to design three assemblies (I, II, and III, Scheme 1) that can
serve as prototypes for other reactions.
however, much of the technology is proprietary. The novelty
of our approach rests on the use of sequentially linked columns
that perform discrete functions in a reaction sequence. Taken
to its logical limit, one can conceive of performing a fairly
complex synthetic sequence all on reaction columns, including
purification steps. Scheme 1 shows three generic assemblies
consisting of jacketed columns and associated linkages. Column
types labeled A, for example, contain stoichiometric reagents
that convert precursors into substrates for the catalytic, asym-
metric reaction. Column type B is packed with the appropriate
asymmetric catalyst, loaded onto a suitable polymeric support.
Column type C contains scavenger resins to remove byproducts
effectively. Variations on this scheme can be imaginedsfor
example, a column can contain both catalysts and reagents
packed together (column type D). Each assembly is designed
to duplicate stages in catalytic reactions, namely substrate
preparation, catalysis, and purification steps.
The concept of employing solid-phase reagents packed into
columns has been relatively unexplored in academic chemistry.
A recent paper details the use of HPLC columns for catalytic,
asymmetric hydrolysis reactions.4 In industry, reaction columns
have been used for large-scale synthesis for quite some time;
(1) (a) For a review, see: Noyori, R. Asymmetric Catalysis in Organic
Synthesis; John Wiley & Sons: New York, 1994. (b) Nugent, W. A.;
RajanBabu, T. V.; Burk, M. J. Science 1993, 259, 479-483. (c) De Camp,
W. H. Chirality 1989, 1, 2-6.
(2) (a) Brown, A. R.; Hermkens, P. H. H.; Ottenheijm, H. C. J.; Rees,
D. C. Synlett 1998, 817-827. (b) Blackburn, C. Biopolymers 1998, 47,
311-351.
(3) For a recent review see: (a) Shuttleworth, S. J.; Allin, S. M.; Sharma,
P. K. Synthesis 1997, 1217-1239. For recent examples see: (b) Kobayashi,
S.; Kusakebe, K.; Ishitani, H. Org. Lett. 2000, 2, 1225-1227. (c) Yu, H.-
B.; Hu, Q.-S.; Pu, L. J. Am. Chem. Soc. 2000, 122, 6500-6501.
(4) Jacobsen, E. N.; Annis, D. A. J. Am. Chem. Soc. 1999, 121, 4147-
4154.
We have published a number of preliminary reports concern-
ing â-lactam synthesis using chiral nucleophiles as catalysts with
ketenes and imines as the reacting substrates.5 We realized
advantages of conducting this type of reaction on a column,
including obviating the need to isolate and/or manipulate highly
reactive ketenes; separating the different solid-phase components
10.1021/ja016556j CCC: $20.00 © 2001 American Chemical Society
Published on Web 10/11/2001

10854 J. Am. Chem. Soc., Vol. 123, No. 44, 2001
Hafez et al.
Scheme 3. Preparation of Polymer-Supported Quinine
Scheme 2. Catalytic, Asymmetric Synthesis of â-Lactams
with Solid Phase Reagents and Catalyst
Derivatives
easily, and recycling the polymers for additional catalytic
reactions; and finally, avoiding strong agitation that can degrade
resin beads.6 Scheme 2 illustrates the chemical steps in the
catalytic, asymmetric synthesis of â-lactams with solid-phase
reagents and catalyst, including a ketene generation step (SP
base), the catalytic step (SP catalyst), and a purification step
(SP scavenger).
The well-precedented7 in situ generation of ketenes with
triethylamine (or other tertiary amines such as Hu¨nig’s base) is
not applicable to our chemistry as the amine itself catalyzes
the cycloaddition of ketene 2 and imino ester 3a.8 The presence
of the byproduct hydrochloride salts compromises the reaction
by lowering chemical yield and eroding diastereoselectivity.5b
These facts prompted us to employ resin-bound dehydrohalo-
genation reagents to produce contaminant-free ketene solutions
under inert atmosphere at reduced temperature. In our initial
attempts, solid-phase bases such as Amberlite IRA-67,9 a tertiary
amine-based polymer, and resin-bound peralkylated guanidines10
failed to promote ketene formation when phenylacetyl chloride
1a in THF was passed through a packing of resin beads in a
jacketed column at -78 °C. To our satisfaction, we found that
the extremely basic resin BEMP 5,11 containing a triamino-
phosphoramide imine bound to a polymeric support,12 produces
ketenes rapidly and in high yield. By slowly passing a THF
solution of phenylacetyl chloride 1a through a jacketed addition
funnel containing the polymer 5 (1.1 basic equivalent) at -78
°C, a straw-colored solution of phenylketene 2a was eluted. We
found that we could also form highly reactive ketenes such as
ethylketene, phenoxyketene, benzyloxyketene, acetoxyketene,
and phthalamidoketene, and believe that this approach should
be amenable to the formation of pure solutions of many other
reactive ketenes. An advantage to placing BEMP resin 5 in a
column is that in a reaction flask the resin was generally found
to interfere with subsequent chemical reactions (it tended to
destroy the imino ester and epimerize the â-lactam product).13
Unfortunately, we found that the quality of BEMP seemed to
be dependent on the supplier and lot number.14 Less active lots
of BEMP resin would sometimes let unreacted acid chloride
through the column. We could compensate for lower quality
BEMP by adding a larger excess of the resin to the column
and by agitating the resin beads through magnetic stirring on
the column. Usually, however, this was not necessary.
Results and Discussion
Preparation and Optimization of the Polymer-Supported
Quinine Derivatives. In a preliminary report, we discovered
that when a cold solution of phenylketene 2a (-78 °C) is treated
with 1 equiv of imino ester 3a and 10 mol % benzoylquinine
(BQ), â-lactam 4a is formed in high ee and dr after purification
by column chromatography.5b Under these conditions the
catalyst is difficult to recover in a pure form from the reaction
mixture, so that a solid-phase-based system was immediately
deemed desirable. Moreover, the column asymmetric catalysis
(CAC) system eliminated the shortcomings of performing this
reaction with the solid-phase components in a conventional
reaction flask. In the next step, we attached quinine units to a
solid support (Scheme 3) to synthesize the chiral packing of
the catalytic reaction columns.15
Preliminary reactions were performed with quinine attached
to carboxypolystyrene resin 7,16 using assembly I. After obtain-
ing a relatively poor diastereomeric ratio (dr) of 3:1 (cis:trans)
(Table 1, entry 1), we chose to derivatize the inexpensive and
very high-loading Wang resin.17 We felt that higher loading
beads would be more effective as a catalyst. The other important
(5) (a) Wack, H.; Drury, W. J., III; Taggi, A. E.; Ferraris, D.; Lectka, T.
Org. Lett. 1999, 1, 1985-1988. (b) Taggi, A. E.; Hafez, A. M.; Wack, H.;
Young, B.; Drury, W. J., III; Lectka, T. J. Am. Chem. Soc. 2000, 122, 7831-
7832. (c) Hafez, A. M.; Taggi, A. E.; Wack, H.; Drury, W. J., III; Lectka,
T. Org. Lett. 2000, 2, 3963-3965.
(6) For other advantages over homogeneous systems see: (a) Kamahori,
K.; Ito, K.; Itsuno, S. J. Org. Chem. 1996, 61, 8321-8324. (b) Itsuno, S.;
Ito, K.; Maruyama, T.; Kanda, N.; Hirao, A.; Nakahama, S. Bull. Chem.
Soc. Jpn. 1986, 59, 3329-3331.
(7) (a) Palomo, C.; Aizpurua, J. M.; Inaki, G.; Oiarbide, M. Eur. J. Org.
Chem. 1999, 3223-3235. (b) Tidwell, T. T. Ketenes; John Wiley & Sons:
New York, 1995. (c) Lynch, J. E.; Riseman, S. M.; Laswell, W. L.; Tschaen,
D. M.; Volante, R. P.; Smith, G. B.; Shinkai, I. J. Org. Chem. 1989, 54,
3792-3796.
(8) We have used R-imino ester 3a in the catalytic, asymmetric synthesis
of R-amino acid derivatives: (a) Drury, W. J., III; Ferraris, D.; Cox, C.;
Young, B.; Lectka, T. J. Am. Chem. Soc. 1998, 120, 11006-11007. (b)
Ferraris, D.; Young, B.; Dudding, T.; Lectka, T. J. Am. Chem. Soc. 1998,
120, 4548-4549. Imine 3a was first used in work by: (c) Tschaen, D. H.;
Turos, E.; Weinreb, S. M. J. Org. Chem. 1984, 49, 5058-5064.
(9) Available from Aldrich (1.6 mequiv/mL exchange capacity).
(10) Polymer-supported base of the hindered tertiary base 7-methyl-1,5,7-
triazabicyclo[4.4.0]dec-5-ene available from Novabiochem (TBD-methyl
polystyrene, 200-400 mesh, 2.0-3.0 mmol/g).
(11) The pKa of the conjugate acid of BEMP in DMSO is 16.2. See:
O’Donnell, M. J.; Delgado, F.; Hostettler, C.; Schwesinger, R. Tetrahedron
Lett. 1998, 39, 8775-8778.
(12) Schwesinger, R.; Willaredt, J.; Schlemper, H.; Keller, M.; Schmitt,
D.; Fritz, H. Chem. Ber. 1994, 127, 2435-2454.
(13) BEMP resin beads often stick to the walls of the reaction vessel
decreasing the efficiency of the dehydrohalogenation.
(14) BEMP is available from both Fluka and Aldrich. The BEMP from
Fluka was of higher quality.
(15) Other examples of polymer supported asymmetric catalysts utilizing
cinchona alkaloids: Bolm, C.; Gerlach, A. Eur. J. Org. Chem. 1998, 21-27.
(16) Carboxypolystyrene HL (100-200 mesh, 1.0-1.6 mmol/g) is
available from Novabiochem.

Asymmetric Catalysis on Sequentially-Linked Columns
J. Am. Chem. Soc., Vol. 123, No. 44, 2001 10855
Table 1. Polymer-Supported Quinine-Catalyzed Asymmetric
Synthesis of â-Lactam 4a from Phenylacetyl Chloride 1a and Imino
Ester 3aa
the dr to 2:1 with 87% ee (entry 4). These results confirmed
our hypothesis that for whatever reason, a long, rigid linker is
beneficial to reaction selectivity.
polymer-supported
quinine derivative
We found it necessary to run each newly synthesized resin a
small number of times before we received consistent results.
We noticed “bleeding” of quinine and other materials in all of
our initial runs with each catalyst resin even after multiple
washings during their preparations.19 Once the resins have aged
through several uses (5-10 runs), they have been employed
multiple times with no significant loss in yield or selectivity.
One resin batch has been used over 60 times and still maintains
its efficacy to date.
entry
drb
% eec
% yieldd
1
2
3
4
7
3:1
10:1
13:1
2:1
90
93
90
87
52
65
62
64
6a
6b
6c
a The reaction was carried out with phenylacetyl chloride 1a (0.13
mmol) and imino ester 3a (0.13 mmol), using assembly I. b Determined
by 1H NMR of crude residue (cis:trans). c Determined by HPLC analysis
with a Regis Technologies (R,R)-Whelk-01 Chiral HPLC column.
d Isolated yield after crystallization of concentrated residue.
Application of the Supported Quinine Derivatives in the
Asymmetric Synthesis of â-Lactams Using Assembly I. In
detail, assembly I consists of two jacketed columns linked
together by a ground glass joint; the top column is packed with
a polymer-supported dehydrohalogenating agent that produces
analytically pure, extremely reactive ketenes from inexpensive
and widely available acid chlorides. The middle column is
packed with a nucleophile-based solid-phase asymmetric cata-
lyst. Between the two columns, the imine is added to the system.
An additional column is optional (although in many cases highly
desirable) and packed with a scavenger resin to remove any
unreacted ketene or imine from the eluent. Assembly of the
system began by loading two fritted, jacketed columns (each 2
cm wide) under nitrogen, the top column with the BEMP resin
5, and the middle column with catalyst-loaded beads 6a (3 cm).
The scavenger resin20 10 was loaded into a column and attached
to the bottom of the apparatus. All columns were flushed with
THF under nitrogen. The BEMP column was cooled to -78
°C with a dry ice/acetone mixture, and the catalyst loaded
Figure 1. Linkers for polymer supported quinine derivatives.
consideration concerned the linker length and shape. Our
hypothesis was that the linkers would in all probability affect
the reaction chemistry in terms of rate and selectivity. One can
argue that long, rigid linkers should project the catalytic quinine
moiety more effectively away from the Wang resin bulk, thus
mimicking solution chemistry more effectively. Short or “floppy”
linkers would arguably allow close proximity of the catalytic
moiety with the resin and thus lower the rate and affect
selectivity in an unpredictable way. We derivatized the Wang
resin with three different linkers to probe different possibilities
(Figure 1). One linker is long and “floppy” (azelaic acid 8c),
and the others consist of a series of phenyl groups of varying
length (8a and 8b).18 The Wang resin was treated with a large
excess of the diacid chloride of each of the linkers to ensure
predominant monoesterification. The quinine units were then
attached to the derivatized resin by another simple esterification
reaction to form solid-phase catalysts 6a-c. Incorporation of
quinine onto the resin averages 60%. Indeed we found that the
length of the catalyst-solid-phase linker is crucial to the success
of the reactionsshort linkers give inferior results. Two factors
could be in play here: presumably the steric encumbrance of
the polymeric support has a negative effect (Table 1), and cross-
linking of the resin could have an effect (although experimental
evidence suggests that the degree of cross-linking is similar
[within 10%] for resins 6a-c).
column was cooled to -43 °C by dry ice/acetonitrile.21
A
solution of phenylacetyl chloride 1a in THF was added to the
top column and allowed to percolate by gravity through the
BEMP resin and onto the lower catalyst-loaded resin of the
middle column. Imino ester 3a was then added through a port
onto the middle column. The reaction was initiated by allowing
a slow drip of THF from the catalyst column, enough to allow
complete elution of the column contents dropwise over the
course of 2 h. After passing through the scavenger resin column
(which traps unreacted ketene and imine byproducts), the eluted
reaction mixture was concentrated to afford â-lactam 4a in 93%
ee (Table 2, entry 1). Crystallization of the residue affords
analytically pure material in 65% yield (>99% ee, 98/2 dr).
To ready the apparatus for another catalytic cycle, the
columns were separated and regenerated. The catalyst-loaded
resin column was washed with methanol, methylene chloride,
and diethyl ether and dried under high vacuum. The BEMP resin
was regenerated by rinsing with phosphazene base P4-t-Bu in
THF/MeCN (1:1), until the eluent was free from Cl-,12 and then
dried under high vacuum at 120 °C. The scavenger resin was
washed with a triethylamine solution in THF, methylene
chloride, and diethyl ether and dried under high vacuum.22 This
reaction has been successfully run through the catalyst column
60 times with no significant loss in selectivity or yield (90% ee
and 62% yield for run 60). The resin beads can either be stored
in the column itself or removed to serve another purpose. When
quinidine is similarly attached to the Wang resin and used in
The quinine-supported Wang resin 6a (with the terephthaloyl
linker 8a) afforded â-lactam 4a in 65% yield with 10:1 cis:trans
dr. The enantioselectivity (ee) has been improved from 91% to
93% since our initial published results (entry 2).5c The slightly
longer, more rigid quinine-supported Wang resin 6b with the
biphenyl linker 8b improved the dr to 13:1 in comparable yield
(62%) and with comparable ee (90%) (entry 3). The longer,
floppy quinine supported Wang resin 6c with the C9 linker 8c
afforded â-lactam 4a in 64% yield and dramatically dropped
(19) See Experimental Section for detailed instructions on washing and
drying the catalyst resins.
(17) Wang resin (100-200 mesh, 1.6-3.0 mmol/g) is available from
Novabiochem.
(18) Other examples of linker length’s effect on selectivities: (a) Kim,
B. M.; Sharpless, K. B. Tetrahedron Lett. 1990, 31, 3003-3006. (b) Pini,
D.; Petri, A.; Nardi, A.; Rosini, C.; Salvadori, P. Tetrahedron Lett. 1992,
32, 5175-5178.
(20) Aminomethylated polystyrene (EHL, 200-400 mesh, 2.0-3.0
mmol/g) from NovaBiochem was employed as the scavenger resin.
(21) The optimal temperature for column asymmetric catalysis of
phenylketene 2a and imino ester 3a was found to be -43 °C.
(22) Recycling of the scavenger resin is limited to 15-20 cycles due to
the generation of polymer-bound amides.

10856 J. Am. Chem. Soc., Vol. 123, No. 44, 2001
Hafez et al.
Scheme 4. Column Asymmetric Catalysis Assembly I
Table 2. Polymer-Supported Quinine-Catalyzed Asymmetric
Synthesis of â-Lactams 4 from Acid Chlorides 1 and Imino Ester
3a with Quinine Derivative 6aa
a The reaction was carried out with the appropriate acid chloride 1
(0.13 mmol) and imino ester 3a (0.13 mmol), using assembly I with
catalyst 6a. b Determined by 1H NMR of crude residue (cis/trans).
c Determined by HPLC analysis with a Regis Technologies (R,R)-
Whelk-01 Chiral HPLC column. d Isolated yield after crystallization
of concentrated residue. e Polymer-supported quinidine on Wang resin
used as catalyst.
place of the catalyst-loaded beads 6a (entry 2), the other
enantiomer 4c of â-lactam 4a is produced in 61% yield, 10:1
dr (cis:trans), and 95% ee (crystallization affords optically pure
material >99% ee). Furthermore, these different catalyst
columns can be used and stored as needed.
Several other â-lactams were produced by varying the acid
chloride employed in the reaction. Butyryl chloride 1d, when
mixed with 3a using assembly I, afforded â-lactam 4d (entry
3) in modest yield (53%) and good selectivities (13:1 dr, 91%
ee). Similarly, â-lactam 4e (entry 4) was produced in 63% yield,
12:1 dr, and 94% ee. Acetoxyacetyl chloride 1f was used with
3a to afford 4f in 60% yield (entry 5) with good selectivity
(14:1 dr, 91% ee).
Application of the Catalytic, Resin-Bound Quinine in the
Asymmetric Synthesis of â-Lactams Using Assembly II. As
an illustration of the potential flexibility of our system, we have
implemented our sequential CAC technique to conduct a
reaction with four discrete steps: (1) formation of reactive
ketenes in one column; (2) formation of imines in situ from
corresponding R-chloroamines in another parallel column; (3)
catalysis of the condensation of the ketene and imine to form a
â-lactam product in the third column; and (4) removal of
unwanted byproducts from the reaction stream with a scavenger
resin. Ketenes 2 and imines 3 cannot be made simultaneously
in a single column due to the different temperatures needed to
effect reaction. To maximize utility, we modified our apparatus
to accommodate this more elaborate application (Scheme 5).
Chloroglycine derivatives 923 are very easily synthesized,
handled, and stored, and serve as convenient and potentially
widely applicable precursors to reactive imines such as 3,
which are more difficult to make and handle through our other
standard procedures. The flexibility in the choice of R group
on 9 demonstrates its versatility; as an illustration of utility we
have investigated, along with sulfonyl groups (Ts 9a), the
corresponding acyl groups (COPh 9b). Acyl-substituted â-
lactams are inhibitors of prostate specific antigen (PSA) and
cytomegalovirus protease, for example.24 Imines 3a,b can be
made at room temperature from chloroglycines 9a,b by using
a variety of dehydrohalogenation bases. We have utilized several
polymer-supported bases, including BEMP 5, TBD-methyl
polystyrene,10 and piperidinomethyl polystyrene,25 to make
imine 3a,b from 9a,b by means of mechanical stirring in THF.
The solid-phase bases were filtered from the imine solution and
the solution was added to assembly I, much the same way as
the preformed imine 3a, to afford â-lactam 4a (Table 3),
although this procedure is obviously more tedious. The applica-
tion of these polymer-supported bases in a continuous flow
system without any external agitation of the solution failed to
generate the imine to any appreciable extent. These bases were
presumably not basic enough (as positive charge builds up on
spent resin, the basicity drops precipitously), and our column
(24) For PSA, see: (a) Adlington, R. M.; Baldwin, J. E.; Becker, G.
W.; Chen, B.; Chen, L.; Cooper, S. L.; Hermann, R. B.; Howe, T. J.;
McCoull, W.; McNulty, A. M.; Neubauer, B. L.; Pritchard, G. J. J. Med.
Chem. 2000, 44, 1491-1508. (b) Adlington, R. M.; Baldwin, J. E.; Becker,
G. W.; Chen, B.; Cooper, S. L.; Hermann, R. B.; Howe, T. J.; McCoull,
W.; McNulty, A. M.; Neubauer, B. L.; Pritchard, G. J. Bioorg. Med. Chem.
Lett. 1997, 7, 1689-1694. For cytomegalovirus protease, see: (c) Bonneau,
P. R.; Hasani, F.; Plouffe, C.; Malenfant, E.; LaPlante, S. R.; Guse, I.;
Ogilvie, W. W.; Plante, R.; Davidson, W. C.; Hopkins, J. L.; Morelock,
M. M.; Cordingley, M. C.; De´ziel, R. J. Am. Chem. Soc. 1999, 121, 2965-
2973. (d) Yoakim, C.; Ogilvie, W. W.; Cameron, D. R.; Chabot, C.; Guse,
I.; Hache´, B.; Naud, J.; O’Meara, J. A.; Plante, R.; De´ziel, R. J. Med. Chem.
1998, 41, 2882-2891.
(23) Steglich, W.; Jager, M.; Jaroch, S.; Zistler, P. Pure Appl. Chem.
1994, 66, 2167-2170.
(25) Available from Novabiochem (HL, 200-400 mesh, 3.3-3.8 mmol/
g).

Asymmetric Catalysis on Sequentially-Linked Columns
J. Am. Chem. Soc., Vol. 123, No. 44, 2001 10857
Scheme 5. Column Asymmetric Catalysis Assembly II
Scheme 6. Dual Ketene and Imine Generation from
Powdered and Solid-Phase Bases in the Asymmetric
Synthesis of â-Lactams
the reagent columns feed, and a scavenger resin column (type
C) below the catalytic column. One of the top columns was
packed with BEMP resin 5, the other top column with the NaH/
Celite mix (6:1 NaH:Celite w/w ratio, 3 cm). The catalytic
column was loaded with catalyst beads 6a (3 cm). The two top
columns were connected to the catalyst column by a Claisen
adapter. After all the columns were flushed with THF, the
BEMP column was cooled to -78 °C with dry ice/acetone
mixture. The NaH/Celite column was kept at room temperature,
and the catalyst loaded column was cooled to -43 °C with a
dry ice/acetonitrile mixture. A solution of phenylacetyl chloride
1 in THF was added to the top of the column and allowed to
percolate by gravity through the BEMP resin and onto the
catalyst-loaded resin of the middle column. Concurrently, the
R-chloroamine 9a, in a solution of THF, was added to the NaH/
Celite column and allowed to drip by gravity through solid bases
onto the catalyst-loaded resin of the middle column. The reaction
was initiated by allowing a slow drip of THF from the bottom
of the column to allow complete elution of the column contents
over the course of 2 h. Surprisingly, little experimentation with
regard to flow rates was undertaken before this fairly optimum
“drip rate” was determined. After passing through the scavenger
resin column the eluted reaction mixture was concentrated to
afford fairly pure â-lactam 4a in 90% ee and 10:1 dr (cis:trans).
Simple crystallization of the residue affords optically and
analytically pure material (>99% ee, 98/2 cis-trans dr) in 62%
yield. Starting with chloroglycine 9b and implementing a similar
procedure affords â-lactam 4b in 33% yield, 13/1 dr, and 92%
ee.27 There is reason to believe that a whole range of ketenes
and acyl-substituted chloroglycines can be used in this procedure
to produce a spectrum of â-lactam products.
Table 3. Polymer-Supported Quinine-Catalyzed Asymmetric
Synthesis of â-Lactam 4a from Phenylacetyl Chloride 1a and
R-chloroamine 9a with Quinine Derivative 6aa
entry
polymer-supported base
BEMP
TBD-methyl polystyrene
drb % eec % eed % yielde
1
2
3
8:1
88
91
91
>99
>99
>99
55
61
59
10:1
piperidinomethyl polysyrene 10:1
a The reaction was carried out with phenylacetyl chloride 1a (0.13
mmol) and chloroglycine 9a (0.13 mmol), using assembly I with catalyst
6a. b Determined by 1H NMR of crude residue (cis/trans). c Determined
by HPLC analysis with a Regis Technologies (R,R)-Whelk-01 Chiral
HPLC column. d Ee after crystallization of concentrated residue.
e Isolated yield after crystallization of concentrated residue.
was not long enough to form the imine sufficiently. An
alternative to making an impracticably long column was to find
a strong, inexpensive base that could be packed into a shorter
column. After considerable experimentation, we found that
mixing Celite26 with NaH in the column serves as the best
packing to form the imine. NaH is the actual stoichiometric
base that dehydrohalogenates the R-chloroamine 9a to form
imine 3a, while the Celite acts as a diluant, slowing down the
flow rate of the solution and thus allowing more time for the
formation of the imine.
Application of the Supported Quinine in the Asymmetric
Synthesis of â-Lactams Using Assembly III. The expense of
solid-phase-based BEMP prompted us to seek more economical
ways of ketene generation. We recently reported a procedure
for the synthesis of reactive ketenes by the use of an amine
dehydrohalogenating catalyst and fine-mesh powdered K2CO3
as the stoichiometric base.28 Our concept was to generate ketenes
with columns of powdered K2CO3 to which chiral catalyst
loaded beads have been added. This type of system satisfies
the potential need for a shuttle base when using powdered
In detail, assembly II consists of three fritted, jacketed
columns (each 2 cm wide), including two top columns (type
A) for reagent synthesis, a catalytic column (type B) into which
(27) Scavenger column was not used due to the propensity of primary
amines opening N-acyl â-lactams.
(28) Hafez, A. M.; Taggi, A. T.; Wack, H.; Esterbrook, J.; Lectka, T.
Org. Lett. 2001, 3, 2049-2051.
(26) Celite 545 (diatomaceous earth) was employed in the reaction.

10858 J. Am. Chem. Soc., Vol. 123, No. 44, 2001
Hafez et al.
Table 4. Polymer-Supported Quinine-Catalyzed Asymmetric
Synthesis of â-Lactam 4a from Phenylacetyl Chloride 1a and Imino
Ester 3aa
In a control experiment, by utilizing only the quinine-
supported resin 6a as both a dehydrohalogenating agent and
catalyst (column type A), the â-lactam can be isolated in 61%
yield with a 7:1 dr and an ee of 91% (entry 2). In the event, a
solution of phenylacetyl chloride 1a in THF was added to the
top of the quinine-loaded column. Imino ester 3a in THF was
then added. The reaction was initiated by allowing a slow drip
of THF from the bottom of the column to allow complete elution
of the column contents over the course of 2 h. After passing
though the scavenger resin the eluted reaction mixture was
concentrated to afford crude â-lactam 4a in 91% ee. Crystal-
lization of the residue affords optically and analytically pure
material in 61% yield.
entry
1
column packing
drb
% eec
% yieldd
6a/K2CO3 (1/5)
6a
3:1
7:1
88
91
58
61
a The reaction was carried out with phenylacetyl chloride 1a (0.13
mmol) and imino ester 3a (0.13 mmol), using assembly III with catalyst
6a. b Determined by 1H NMR of crude residue (cis/trans). c Determined
by HPLC analysis with a Regis Technologies (R,R)-Whelk-01 Chiral
HPLC column. d Isolated yield after crystallization of concentrated
residue.
Scheme 7. Column Asymmetric Catalysis Assembly III
Conclusions
One can envision the synthesis of considerably more complex
enantioenriched compounds by using the sequential CAC
methodology. Reactants can be passed through any number of
discrete columns containing recyclable reagents connected in
parallel or in series. It should be noted that the use of solid-
supported catalysts and reagents, including scavenger resin
columns, greatly simplifies the purification of the reaction
products. In the future, we will apply the concept of serial CAC
to the synthesis of more complex molecules in multistep and/
or multicatalytic procedures.
Experimental Section
General Procedure for the Synthesis of Polymer-Supported
Quinine Derivatives. Wang resin (2.0 g, 5.06 mmol, 2.53 mmol/g)
was added via addition funnel to a solution of terephthaloyl chloride
(2.57 g, 12.7 mmol) and triethylamine (1.74 mL, 12.7 mmol) in 30
mL of THF at 0 °C over 2 h. The reaction was allowed to warm to
room temperature over 10 h. The resin was washed on a course glass
frit with 50 mL of THF and 50 mL of diethyl ether to remove excess
reagents and reaction byproducts. The derivatized resin was then dried
under vacuum for 3 h. To the resin at 0 °C was added a solution of
quinine (3.28 g, 10.1 mmol) and triethylamine (1.13 g, 11.1 mmol) in
60 mL of THF. The reaction was allowed to warm to room temperature
over 10 h. Methanol (40 mL) was then added and the reaction stirred
for 1 h. The resin was filtered off and washed with an additional 40
mL of methanol. The resin 6a was then placed in a Soxhlet extractor
and refluxed with acetone for 24 h. 6a was removed from the extractor
and dried under vacuum for 24 h. Incorporation of quinine catalyst
was determined to be 60% based on the recovered quinine (2.3 g, 7.1
mmol). Anal. Found for 7: C, 56.70; H, 4.91; N, 1.55. Found for 6a:
C, 81.73; H, 7.22; N, 2.51. Found for 6b: C, 77.12; H, 6.58; N, 2.68.
Found for 6c: C, 81.32; H, 7.28; N, 2.87.
General Procedure for Column Catalysis Using Assembly I. Two
fritted, jacketed columns (each 2 cm wide) were loaded under nitrogen,
the top column with BEMP resin 5 (1.1 basic equivalents) and the
bottom with catalyst-loaded beads 6a (3 cm). The scavenger resin 10
was loaded into a column and attached to the bottom of the apparatus.
All columns were flushed with THF under nitrogen. The BEMP column
was cooled to -78 °C with dry ice/acetone mixture. The catalyst-loaded
column was cooled to -43 °C with a dry ice/acetonitrile mixture. A
solution of phenylacetyl chloride 1a (20 mg, 0.13 mmol) in THF (1
mL) was added to the top of the column and allowed to drip by gravity
through the BEMP resin and onto the catalyst-loaded resin of the middle
column. Imino ester 3a (33 mg, 0.13 mmol in 0.5 mL of THF) was
then added through a port onto the middle column. The reaction was
initiated by allowing a slow drip of THF from the bottom of the column
to allow complete elution of the column contents over the course of
2 h. After passing though the scavenger resin the eluted reaction
mixture was concentrated to afford crude â-lactam 4a in 93% ee and
10:1 cis:trans dr. Crystallization of the residue affords optically and
analytically pure material in 65% yield. The standard reaction for
assembly I can be scaled up by a factor of 10 or more (ca. 330 mg of
acid chloride), using the same amount of catalyst resin with comparable
results.
carbonates. The polymer-supported quinine beads effect dehy-
drohalogenation, and then presumably transfer their protons to
the solid carbonate next door. As to the intriguing question of
whether this transfer would involve bead-to-particle contact or
is conducted through the flow medium was addressed by a
control experiment. We attempted to produce solutions of ketene
simply by filtering a solution of acid chloride through a cold
column (-78 °C) containing powdered K2CO3, but were un-
successful. The catalyst beads are in fact needed to produce
ketene on the column. The implications of this experiment on
the questions of packing, flow, and morphology remain to be
sorted out.
In detail, assembly III consists of one fritted, jacketed column
(2 cm wide, type D) that was loaded with a mixture of catalyst-
loaded beads 6a and powdered carbonate in a 5/1 w/w ratio.
The scavenger resin 10 was loaded into a column and attached
to the bottom of the apparatus. After flushing the columns with
THF, the catalyst-carbonate loaded column was cooled to -43
°C with a dry ice/acetonitrile mixture. In an initial experiment,
we found that when phenylacetyl chloride 1a and imine 3a were
added to the quinine-supported resin 6a and K2CO3 mix in a
jacketed column at -43 °C, the reaction afforded â-lactam 4a
in 58% yield (Table 4, entry 1). The dr values of the product,
however, seemed to be eroded by the presence of the K2CO3
(3:1 dr and 88% ee). Crystallization of the reaction material,
however, provided optically pure material as an 98/2 mixture
of diastereomers. The special appeal of reactions with assembly
III is that they can be scaled up easily. For example, we made
1 g of pure lactam 4a simply by increasing the amount of
carbonate proportionally while employing the same loading of
catalyst beads (10/1 ratio of carbonate to catalyst).

Asymmetric Catalysis on Sequentially-Linked Columns
J. Am. Chem. Soc., Vol. 123, No. 44, 2001 10859
General Procedure for Column Catalysis Using Assembly II.
Three fritted, jacketed columns (each 2 cm wide) were loaded under
nitrogen, one of the top columns with BEMP resin 5 (1.1 basic
equivalents) and the other top column with a NaH/Celite mix (3 cm),
while the bottom was loaded with catalyst beads 6a (3 cm). The two
top columns were connected to the catalyst column by a Claisen adapter.
The scavenger resin 10 was loaded into a column and attached to the
bottom of the apparatus. All columns were flushed with THF under
nitrogen. The BEMP column was cooled to -78 °C with a dry ice/
acetone mixture. The NaH/Celite column was kept at room temperature.
The catalyst-loaded column was cooled to -43 °C with a dry
ice/acetonitrile mixture. A solution of phenylacetyl chloride 1a (20 mg,
0.13 mmol) in THF (1 mL) was added to the top of the column and
allowed to drip by gravity through the BEMP resin and onto the catalyst-
loaded resin of the middle column. The R-chloroamine 9a (33 mg, 0.13
mmol in 0.5 mL of THF) was then added to the NaH/Celite column
and allowed to drip by gravity through solid bases and onto the catalyst-
loaded resin of the middle column. The reaction was initiated by
allowing a slow drip of THF from the bottom of the column to allow
complete elution of the column contents over the course of 2 h. After
passing though the scavenger resin the eluted reaction mixture was
concentrated to afford crude â-lactam 4a in 90% ee and 10:1 cis:trans
dr. Crystallization of the residue affords optically and analytically pure
material in 62% yield.
with catalyst-loaded beads 6a (0.30 mmol) and K2CO3 in a 1/5 w/w
ratio (9 cm). The scavenger resin 10 was loaded into a column and
attached to the bottom of the apparatus. All columns were flushed with
THF under nitrogen. The catalyst/K2CO3 loaded column was cooled
to -43 °C with a dry ice/acetonitrile mixture. A solution of phenylacetyl
chloride 1a (471 mg, 3.00 mmol) in THF (10 mL) was added to the
top of the column and onto the catalyst-loaded resin of the middle
column. Imino ester 3a (774 mg, 3.00 mmol in 10 mL THF) was then
added through a port onto the catalyst-loaded resin column. The reaction
was initiated by allowing a slow drip of THF from the bottom of the
column to allow complete elution of the column contents over the course
of 2 h. After passing though the scavenger resin the eluted reaction
mixture was concentrated to afford crude â-lactam 4a in 89% ee and
3:1 cis:trans dr. Crystallization of the residue affords optically and
analytically pure material in 58% yield.
Acknowledgment. T.L. thanks the NIH, DuPont, Eli Lilly,
and the NSF Career Program for general lab support, the Dreyfus
Foundation for a Teacher-Scholar Award, and the Alfred P.
Sloan Foundation for a Fellowship.
Supporting Information Available: Experimental details
(PDF). This material is available free of charge via the Internet
at http://pubs.acs.org.
General Procedure for Column Catalysis Using Assembly III.
One fritted, jacketed column (2 cm wide) was loaded under nitrogen
JA016556J