Two useful photolabile surfaces for solid-phase synthesis

Article abstract of DOI:10.1021/ol0056224

formula presented o-Nitrobenzyl-based photolabile surfaces 3 and 4 have been synthesized from 2-phenylcyclohexanone and aminopropylsiloxane-grafted controlled pore glass. The procedures are simple and inexpensive and generate materials whose minimally fun

Full text of DOI:10.1021/ol0056224

ORGANIC
LETTERS
2000
Vol. 2, No. 6
851-853
Two Useful Photolabile Surfaces for
Solid-Phase Synthesis
Troy D. Ryba and Patrick G. Harran*
Department of Biochemistry, The UniVersity of Texas Southwestern Medical Center at
Dallas, Dallas, Texas 75390-9038
pharra@biochem.swmed.edu
Received February 4, 2000
ABSTRACT
o-Nitrobenzyl-based photolabile surfaces 3 and 4 have been synthesized from 2-phenylcyclohexanone and aminopropylsiloxane-grafted controlled
pore glass. The procedures are simple and inexpensive and generate materials whose minimally functionalized reactive chromophore should
tolerate a range of subsequent chemistries. Time- and solvent-dependent analysis of photodegradation demonstrates a performance comparable
to related photolabile systems.
We are attempting to develop an artificial analogy to small
molecule, processive biosynthesis: one that operates through
synthetic restructuring of functionalized polymers. In these
experiments, surface-bound oligopeptides are arrayed as
starting materials for synthesis rather than being individual
end points. Therefore, a format which enables efficient
assembly of immobilized, unprotected peptides and also
provides for versatility in their manipulation is vital. Because
radicals, carbenoids, and active oxidants will play a role in
this chemistry, commonly used organic polymer matrices are
not attractive solid supports. Moreover, the system is
designed to interface with whole-cell and receptor-based
activity screens. Synthetic linkers degradable with acid and/
or base can complicate this objective which, in other contexts,
has been accomplished using photolabile inserts and product
release using light.1 The same tactic seemed appropriate for
the current application, and we first employed a material
formed from aminopropylsiloxane-grafted controlled pore
glass (1, amino-CPG) and o-nitrobenzylamine 2 (Figure 1)
for solid-phase peptide synthesis. This combination performs
well; however, our synthesis of 22 is time-consuming and
inefficient. Herein, we describe syntheses and preliminary
performance of related photolabile reagents 3 and 4: materi-
als which retain a reactive chromophore of relatively high
(1) The preparation and utility of photolabile linkers has been re-
viewed: (a) Gordon, K.; Balasubramanian, S. J. Chem. Technol. Biotechnol.
1999, 74, 835-851. (b) Backes, B. J.; Ellman, J. A. Curr. Opin. Chem.
Biol. 1997, 1, 86-93. (c) Lloyd-Williams, P.; Albericio, F.; Giralt, E.
Tetrahedron 1993, 49, 11065-11133. (d) Pallai, V. N. R. Synthesis 1980,
1-26.
Figure 1.
10.1021/ol0056224 CCC: $19.00 © 2000 American Chemical Society
Published on Web 03/02/2000

oxidation potential (in comparison to commercial veratryl-
based systems) and are produced concisely from a common
precursor.
Commercial 2-phenylcyclohexanone is nitrated with 85%
nitronium tetrafluoroborate to afford an excess of 2′ isomer
5 (Scheme 1).3 Purified 5 is then either converted to
tert-butyl dicarbonate and the formed imide (8) treated with
LiOH to generate ω-carbamoyl carboxylic acid 9. Amidation
of 9 with amino-CPG (TBTU, DIPEA, HOBt) followed by
exposing the resin to anhydrous CF3CO2H gives target
material 4. On a larger (10-15 g) scale, the conversion of 5
to 9 is carried out without intermediate purifications.
The density of reactive sites presented by batches of 3
(50-60 µmol/g)7 and 4 (120-125 µmol/g) is determined
by acylation with N-Fmoc-Gly-OH and quantifying the
dibenzofulvene released upon DBU treatment (2% volume
in DMF).8 Photolyses are performed as stirred resin suspen-
sions (borosilicate vials) in degassed solvent (vide infra)
using a 450-W medium-pressure mercury vapor lamp
positioned 5-8 cm from the reaction vessel. No attempt is
made to attenuate power levels9 although a long-pass filter
(cut-on 348 nm, Oriel Corp.) is routinely utilized.
Scheme 1a
A cited advantage of veratryl-based photolabile inserts is
an increased photocleavage rate, due to higher molar
absorptivity, at wavelengths (>350 nm) desirable for com-
binatorial chemistry applications.10 Interestingly, photophysi-
cal measurements in solution indicate that this increase in
absorptivity can, in certain instances, be offset by a marked
decrease in quantum yield.11 Immobilization introduces
additional variables. For example, Holmes has demonstrated
that the photodecomposition of 11 (Hg(Xe) arc) occurs with
t1/2 ) 0.66 min in pH 7.4 phosphate-buffered saline (Figure
2).10b In contrast, a solid-phase analogy of the process (i.e.,
a (a) NO2BF4, CH3NO2, 10 °C, 15 min (46% + 12% 4′ isomer);
(b) m-CPBA, Na2HPO4, CH2Cl2, rt, 3 h (99%); (c) H2SO4, NaN3,
PhH, rt, 12 h (66%); (d) (Boc)2O, DMAP, THF, rt (91%); (e) LiOH,
THF/H2O (82% for 9, 93% for 10).
caprolactone 6 via Baeyer-Villiger oxidation (m-CPBA) or
to caprolactam 7 via Schmidt rearrangement (H2SO4, NaN3).
In both instances, ring expansion proceeds regioselectively
with migration of the more substituted cyclohexanone
R-carbon (C2).4 Lactone 6 is saponified and the resultant
hydroxy acid (10) transferred onto the surface of amino-
CPG (Biosearch Technologies) through amidation (TBTU,5
DIPEA, HOBt). The composite formed (3) is functionally
identical to a known polystyrene-based resin (six steps from
o-nitrobenzaldehyde)6 although the current preparation is
advantageous in that sensitive organometallic reagents and
protecting group manipulations are not required. To complete
the amine congener of 3, lactam 7 is N-acylated with di-
Figure 2.
photoinduced release of 13 from 12) requires 3 h to reach
95% conversion. For comparison, resin 4 loaded with
N-Fmoc-Gly-OH evolves N-Fmoc-Gly-NH2 on photolysis (p-
dioxane, Hg vapor lamp, g348 nm) wherein yield peaks at
75% after roughly 90 min (Figure 3).12 In p-dioxane, the
(2) Compound 2 is prepared from known 2,2-dimethyl-3-(tert-butoxy-
carbonyl)amino-3-(2′-nitrophenyl)propionic acid. Sternson, S. M.; Schreiber,
S. L. Tetrahedron Lett. 1998, 39, 7451-7454.
(3) The use of 95% NO2BF4 (Aldrich) results in lower regioselectivity
(ortho:para ) 1.4:1). The 85% reagent (Aldrich) is contaminated primarily
with NOBF4. Reaction mixtures using 95% NO2BF4 can be doped with 10
mol % of NOBF4 to reconstitute, in part, a more selective nitration mixture.
Nonselective nitration of 2-phenylcyclohexanone has been described:
Prager, R. H.; Tippett, J. M.; Ward, A. D. Aust. J. Chem. 1978, 31, 1989-
2001.
(4) The conversion of 5 to 7 generates trace byproducts, none of which
account for more than 5% of unrecovered mass. It is possible that one of
these materials is a regioisomeric lactam.
(5) TBTU ) O-benzotriazol-1-yl-N,N,N′,N′-tetramethyluronium tetrafluo-
roborate, DIPEA ) N,N-diisopropylethylamine, HOBt ) N-hydroxyben-
zotriazole.
(6) Rodebaugh, R.; Fraser-Reid, B.; Geysen, H. M. Tetrahedron Lett.
1997, 38, 7653-7656.
(7) The discrepancy in available sites for CPG derivatized with 9 and
10 is unexplained. In solution, the condensation of 10 with n-BuNH2 (TBTU,
DIPEA, HOBt, DMF) affords the corresponding hydroxy amide in 82%
isolated yield. Self-condensation and/or re-lactonized products are not
detected. Moreover, there is no significant difference in immobilized yields
using 9 or 10 on organic polymer supports (Tentagel, Argopore).
(8) Newcomb, W. S.; Deegan, T. L.; Miller, W.; Porco, J. A., Jr.
Technical Bulletin 012; Argonaut Technologies Inc.: San Carlos, CA. This
information can be downloaded from www.argotech.com.
(9) 983 mW/cm2 at the 366 nm line (50 cm distance from the lamp).
(10) (a) Holmes, C. P.; Jones, D. G. J. Org. Chem. 1995, 60, 2318-
2319. (b) Holmes, C. P. J. Org. Chem. 1997, 62, 2370-2380.
(11) Krafft, G. A.; Randall Sutton, W.; Cummings, R. T. J. Am. Chem.
Soc. 1988, 110, 301-303.
852
Org. Lett., Vol. 2, No. 6, 2000

photodegrade acylated 10 immobilized on grafted polystyrene
(THF/H2O).6
Solid-phase peptide synthesis on resin 4 is facile and
efficient. Using Fmoc-protected monomers and TBTU-
mediated couplings, semiautomated syntheses of four- and
five-residue oligomers accommodate the incorporation of all
proteinogenic and designed residues thus far examined. The
glass matrix tolerates a corrosive reagent mixture (95%
CF3CO2H, 5% thioanisole, 5% 1:1 HS(CH2)2SH/H2O) used
for side-chain deprotections. Protected Leu-enkephalin-amide
(YGGFL-NH2) is synthesized on 4 and, following exposure
to the above mixture (3h, rt), photoreleased in 40% yield
(93% per step average) and 88% purity (HPLC, Vydac C18).
The combination of an inorganic matrix and a minimally
functionalized linking group has advantages in our experi-
ments that should also be useful in other applications. The
resins load at low density, there are no issues of solvent-
dependent swelling, nor is there a need to protect 3/4 from
normal working light. The extent of their chemical and
mechanical stability remains to be determined although
strong aqueous base (NaOH, KOH) should be avoided.
Finally, the use of 9 and 10 is not limited to immobilization
on glass. The simplicity and low cost of synthesizing each
from 2-phenylcyclohexanone will hopefully facilitate their
use with other solid supports in the range of solid-phase
chemistry now being explored by the community at large.
Figure 3. Time and solvent-dependent photorelease of Fmoc-Gly-
NH2 from 4 acylated with Fmoc-Gly-OH (450W Hg vapor lamp,
rt, g348 nm). Aliquots (5 µL) of reaction solution (10 mg of resin
suspended in 0.5 mL of degassed p-dioxane) removed by syringe
and analyzed by HPLC (4.6 × 250 mm Vydac C18, linear gradiant
40 f 60% 1:1 i-PrOH/CH3CN in H2O at 0.5 mL min-1) using
Fmoc-Gly-OH (Bachem) as external standard. Inset: The same
photochemical experiment with varying solvents and product
analysis after 2 h of irradiation: lane 1, p-dioxane; lane 2, 4:1 H2O/
MeOH; lane 3, CH3CN; lane 4, 1:1 THF/H2O; lane 5, THF.
Acknowledgment. Financial support provided by the
Advanced Research Program of the Texas Higher Education
Coordinating Board, The Robert A. Welch Foundation, and
junior faculty awards administered through the Howard
Hughes Medical Institute and the University of Texas
Southwestern Medical Center are gratefully acknowledged.
P.G.H. is the recipient of an NSF-CAREER award.
purity of isolated product is >95% (1H NMR and HPLC),
although both yield and purity do vary with reaction medium.
Of the solvents examined, p-dioxane has proven most
generally effective (Figure 3). Hydroxyl-presenting particles
3 loaded with N-Fmoc-Gly-OH release unchanged acid in
77% yield (85% purity) after 2.0 h of photolysis in p-dioxane.
This contrasts with the 12-24 h required to similarly
Supporting Information Available: Experimental pro-
1
cedures, characterization data, and H NMR spectra for
5-10. This material is available free of charge via the
(12) It is common to detect residual dibenzofulvene, accounting for
∼15% of the original loading capacity, when photolyzed glass particles
(dark red after 2 h of irradiation) are re-subjected to DBU treatment.
Internet at http://pubs.acs.org.
OL0056224
Org. Lett., Vol. 2, No. 6, 2000
853