and combine multiple substrates in a systematic fashion. The
reaction sequence is orchestrated via a regulated flow system,
eventually yielding the desired product and allowing for easy
regeneration and reuse of the resin-bound reagents; in
essence, a “synthesis machine.” We are actively pursuing
this as the ultimate goal of our sequentially linked column
assembly methodology.
Several criteria guided the choice of our target molecule.
We wanted it to be a structure with a degree of complexity
that could be assembled from readily available starting
materials in four or more steps on the columns. We also
wanted one of the steps to be a new use of an asymmetric
reaction that we had developed in our own labs. In other
steps, we wished to make new uses of solid-phase reagents.
Finally, the target had to have notable medicinal activity,
be a currently investigated drug candidate, or be a natural
product of importance. Along those lines, the metallopro-
teinase inhibitor BMS-275291 (1), which is now in stage III
clinical trials as a treatment for cancer, seemed an attractive
candidate.6
Figure 1. Components of the solid-phase “synthesis machine”.
stoichiometric base for dehydrohalogenation and as the
catalyst for the chlorination (the beads can be regenerated
by a simple flush cycle with Hu¨nig’s base, followed by THF).
The reaction mixture then flows into column B (Figure 1),
containing piperazino resin 4, to remove any remaining acid
chloride.9 Next, the reaction mixture encounters column E,
where it is mixed with the peptide that is synthesized on
columns C and D (the second branch of the system). Two
simple amino acid derivatives are coupled on a column
containing the versatile carbodiimide-based resin 3 to yield
peptide 10 using THF as the flow solvent. Removal of the
Fmoc group from product 10 is accomplished on column D,
which contains tris-(2-aminoethyl)amine resin 5.10 Both of
these reactions are accomplished in relatively short periods
of time, making them ideal for a column based flow system.
This balance between the residence time of the substrate on
the column and the flow rate of the mobile phase is key to
the success of such a system. Rapid elution does not allow
for complete conversion, while excessive residence times
often promote undesired side reactions.11 It is also important
to note that in all cases, the results appear to be better when
a column with a smaller inner diameter is employed (e.g.,
13 mm was found to be optimal for many steps); this allows
for maximum interaction with the reactant beads as the
substrate percolates through the solid phase.
To assemble 1 on a sequentially linked system, we
envisaged an apparatus containing columns linked both in
parallel and in series, which reflects the convergent nature
of the proposed synthesis (Figure 1). While the concept is
intriguing, there are several limitations to this approach: most
notable is the necessity for a “base solvent” that is introduced
at the top of the assembly in which all subsequent reactions
must be conducted (although quantities of different solvents
can be introduced along the way, producing admixtures). For
our purposes, we were lucky to discover that THF was a
nearly ideal solvent/cosolvent for all of our transformations.
In one branch of the assembly, acid halide precursor 8
undergoes asymmetric chlorination7 to yield the active ester
9 in 88% ee8 in 3 h. This is the first instance in which our
cinchona alkaloid-catalyzed asymmetric halogenation reac-
tion has been performed on a solid-phase system. Quinine-
loaded Wang resin-based beads 2 were packed into a jacketed
addition funnel, saturated with THF, and then cooled to 0
°C. Acid chloride 8 and chlorinating agent 7 are then added
in sequence to the top of the system and allowed to drip
through the column. The cinchona alkaloid serves as both a
N-Deprotected peptide 11 in THF then flows to meet
R-chloroester 9 in “junction” column E, which is loaded with
a simple diluant (Celite) that acts as a benign reaction
medium. A longer residence time at this step is necessary to
(6) (a) Heath, E. I.; Grochow, L. B. Drugs 2000, 59, 1043-1055. (b)
Marshall, J. L.; Baidas, S.; Bhargava, P.; Rizvi, N. IDrugs 2000, 3, 518-
524.
(7) (a) Wack, H.; Taggi, A. E.; Hafez, A. M.; Drury, W. J., III; Lectka,
T. J. Am. Chem. Soc. 2001, 123, 1531-1532. (b) France, S.; Wack, H.;
Taggi, A. E.; Hafez, A. M.; Wagerle, T. R.; Shah, M. H.; Dusich, C. L.;
Lectka, T. J. Am. Chem. Soc. 2004, 126, 4245-4255.
(8) Enantiomeric excess was determined by HPLC. An aliquot of the
reaction mixture (which ordinarily would flow to the next column) was
taken and analyzed.
(9) Since secondary amines do not react with our R-halo esters, the acid
chloride scavenging step can be performed very rapidly (<1 h).
(10) Tris-(2-ethylamino)amine acts as an effective deblocking agent and
scavenger for dibenzofulvene while suppressing formation of precipitates
or emulsions: Carpino, L. A.; Sadat-Aalaee, D.; Beyermann, M. J. Org.
Chem. 1990, 55, 1673-1675.
(11) Weber, A. P. Chem. Eng. 1969, 76, 79-80.
3010
Org. Lett., Vol. 7, No. 14, 2005