Kinetics of self-immolation: Faster signal relay over a longer linear distance?

Article abstract of DOI:10.1021/ol900433g

A counterintuitive observation of a faster signal relay over a longer linear distance prompted detailed kinetic studies of self-immolation reactions. With appropriate conformational bias, trigger-to-reporter signal transduction can take an efficient "shor

Full text of DOI:10.1021/ol900433g

ORGANIC
LETTERS
2009
Vol. 11, No. 10
2065-2068
Kinetics of Self-Immolation: Faster
Signal Relay over a Longer Linear
Distance?
Ho Yong Lee, Xuan Jiang, and Dongwhan Lee*
Department of Chemistry, Indiana UniVersity, 800 East Kirkwood AVenue,
Bloomington, Indiana 47405
dongwhan@indiana.edu
Received March 2, 2009
ABSTRACT
A counterintuitive observation of a faster signal relay over a longer linear distance prompted detailed kinetic studies of self-immolation reactions.
With appropriate conformational bias, trigger-to-reporter signal transduction can take an efficient “shortcut” that outperforms conventional
pathways involving repetitive quinone methide rearrangements and elimination.
Chemical structures that can amplify and transmit externally
triggered structural changes over a long distance are critical
Scheme 1. Mechanism of Self-Immolation and Representative
functional components in signal transduction.1 Controlled chain
fragmentation of self-immolative molecules is one drastic
chemical mode of operation that satisfies such criteria. Here,
repetitive bond rearrangement reactions following a single
triggering event propagate through the molecular backbone to
release remotely located signaling units (Scheme 1).2-5 The
self-immolative nature of one such motif, quinone methide
(QM), has been extensively exploited within this context.
Since the seminal independent studies that appeared in the
literature in 2003,3a,4a,b,5 numerous prototypes of signal-
amplifying self-immolative sensors and drug delivery ve-
Structural Motifs Undergoing Sequential Bond
Cleavage Reactions
(1) Kay, E. R.; Leigh, D. A.; Zerbetto, F. Angew. Chem., Int. Ed. 2007,
46, 72–191.
(2) (a) Wang, W.; Alexander, C. Angew. Chem., Int. Ed. 2008, 47, 7804–
7806. (b) Shabat, D. J. Polym. Sci., Part A: Polym. Chem. 2006, 44, 1569–
1578
.
10.1021/ol900433g CCC: $40.75
Published on Web 04/16/2009
 2009 American Chemical Society

hicles have been reported.3-5 However, surprisingly little is
known about the structure-reactivity relationships underpin-
ning the mechanism of signal relay, including the rearrange-
ment kinetics of QM itself (Scheme 2).6 In this paper, we
however, the solution turned yellow with the appearance of
strong (ε ) 23 000 M-1 cm-1) absorption at λmax,abs ) 407
nm and development of blue emission at λmax,em ) 455 nm,
indicating quantitative release of umbelliferone as the final
product of the reaction (Figure 1a and Scheme 3). Under
similar conditions, 2 reacted similarly with F-, but its
Scheme 2
.
Self-Fragmentation through 1,4- (top) and
1,6-Rearrangement (bottom)
Figure 1. (a) UV-vis and emission spectra of 1 prior to (dotted
lines; λexc ) 300 nm) and after (solid lines; λexc ) 380 nm) reaction
with F- (100 equiv) in THF. (b) Time-dependent changes in the
fluorescence intensity of 1 and 2 monitored at λ ) 460 nm (λexc
)
describe mechanistic studies on well-defined trigger-
relay-reporter conjugates that undergo externally triggered
chain fragmentation. Our kinetic evidence points toward an
alternative mechanism of self-immolation, by which a signal
travels to distant positions through a reaction pathway that
does not necessarily involve repetitive bond rearrangements.
In addition to dictating the kinetics of trigger-to-reporter
signal transduction, the chemical structure of the linker units
profoundly influenced the mechanism of chain fragmentation,
which has significant implications for the rational design of
stimuli-responsive molecules and materials.
Our mechanistic investigation of QM rearrangement was
prompted by the kinetic behavior of the compounds 1 and
2. These molecules were designed specifically to integrate
(i) a Si-O group to be cleaved by F- anion functioning as
the triggering agent,7 (ii) a carbonate- (for 1) or ether-
extended (for 2) 1,2-QM linker as a signal relay, (iii) a
“masked” umbelliferone (7-hydroxycoumarin) unit to be
released as the final product and to elicit turn-on fluorescence
response,8 and (iv) an ethynyl group as an anchoring point
for future immobilization studies.
380 nm) after addition of F- (100 equiv) at t ) 10 s (indicated by
a gray arrow). T ) 293 K.
response kinetics was strikingly different from that of 1
(Figure 1b). Apparently, the release of CO2 as part of the
chain fragmentation cascade provides additional thermody-
namic driving force that is responsible for the experimentally
observed faster kinetic response of 1 relative to 2.
To derive kinetic parameters associated with each step in the
chain fragmentation (Scheme 3), subsequent measurements were
conducted in CH2Cl2. Under pseudo-first-order conditions, an
exponential increase in the emission intensity was observed from
1 + F- at 293 K (Figures 2a and S1, Supporting Information).9
A linear dependence of the pseudo-first-order rate constant k1′
() k1[F-]0; eq 1) on [F-]0 (Figure S2, Supporting Information)
(3) (a) Amir, R. J.; Pessah, N.; Shamis, M.; Shabat, D. Angew. Chem.,
Int. Ed. 2003, 42, 4494–4499. (b) Shamis, M.; Lode, H. N.; Shabat, D.
J. Am. Chem. Soc. 2004, 126, 1726–1731. (c) Amir, R. J.; Shabat, D. Chem.
Commun. 2004, 1614–1615. (d) Amir, R. J.; Popkov, M.; Lerner, R. A.;
Barbas, C. F., III.; Shabat, D. Angew. Chem., Int. Ed. 2005, 44, 4378–
4381. (e) Amir, R. J.; Danieli, E.; Shabat, D. Chem.-Eur. J. 2007, 13,
812–821. (f) Shamis, M.; Shabat, D. Chem.-Eur. J. 2007, 13, 4523–4528.
(g) Sagi, A.; Weinstain, R.; Karton, N.; Shabat, D. J. Am. Chem. Soc. 2008,
130, 5434–5435. (h) Weinstain, R.; Sagi, A.; Karton, N.; Shabat, D.
Chem.-Eur. J. 2008, 14, 6857–6861.
(4) (a) Li, S.; Szalai, M. L.; Kevwitch, R. M.; McGrath, D. V. J. Am.
Chem. Soc. 2003, 125, 10516–10517. (b) Szalai, M. L.; Kevwitch, R. M.;
McGrath, D. V. J. Am. Chem. Soc. 2003, 125, 15688–15689. (c) McGrath,
D. V. Mol. Pharmaceutics 2005, 2, 253–263.
(5) de Groot, F. M. H.; Albrecht, C.; Koekkoek, R.; Beusker, P. H.;
Scheeren, H. W. Angew. Chem., Int. Ed. 2003, 42, 4490–4494.
(6) (a) de Groot, F. M. H.; Loos, W. J.; Koekkoek, R.; van Berkom,
L. W. A.; Busscher, G. F.; Seelen, A. E.; Albrecht, C.; de Bruijn, P.;
Scheeren, H. W. J. Org. Chem. 2001, 66, 8815–8830. (b) Warnecke, A.;
Kratz, F. J. Org. Chem. 2008, 73, 1546–1552. (c) Erez, R.; Shabat, D. Org.
Biomol. Chem. 2008, 6, 2669–2672.
(7) (a) Zhou, Q.; Rokita, S. E. Proc. Natl. Acad. Sci. U.S.A. 2003, 100,
15452–15457. (b) Weinert, E. E.; Dondi, R.; Colloredo-Melz, S.; Franken-
field, K. N.; Mitchell, C. H.; Freccero, M.; Rokita, S. E. J. Am. Chem. Soc.
2006, 128, 11940–11947. (c) Kim, S. Y.; Hong, J.-I. Org. Lett. 2007, 9,
3109–3112.
(8) (a) Gao, W.; Xing, B.; Tsien, R. Y.; Rao, J. J. Am. Chem. Soc. 2003,
125, 11146–11147. (b) Meyer, Y.; Richard, J.-A.; Massonneau, M.; Renard,
P.-Y.; Romieu, A. Org. Lett. 2008, 10, 1517–1520.
In THF at 293 K, 1 showed broad UV-vis absorptions at
λ < 360 nm but remained essentially nonemissive. Upon
reaction with F- (100 equiv, delivered as TBAF salt),
(9) See Supporting Information.
2066
Org. Lett., Vol. 11, No. 10, 2009

′
∆I
I
) 1 - e-k t
(1)
(2)
1
Scheme 3
.
Stepwise Fragmentation of 1-4 Triggered by
Fluoride-Induced Desilylation
′
∆I
I
1
′
-k2t
k (1 - e-k t) - k1(1 - e
)
1
)
{
}
2
k2 - k′1
Under similar conditions, however, the fluorescence
enhancement from 2 showed a markedly different kinetic
behavior (Figure 2a). Specifically, the early induction
period3c-f consistently observed under pseudo-first-order
reaction conditions (Figures 2a and S3-S4, Supporting
Information) pointed toward the formation of a significant
amount of phenoxide intermediate prior the QM rearrange-
ment and subsequent elimination,4a the kinetics of which
could be modeled using eq 2.10 To a first approximation,
the substituent on the methylene group of the QM precursor,
i.e., carbonate oxygen atom for 1 versus ether oxygen atom
for 2, was anticipated to have negligible effects on the
reactivity of the scissile Si-O bond off the o-phenoxy group.
Using the k1 value obtained from 1 as the intrinsic bimo-
lecular rate constant for the Si-O bond cleavage step, the
time-dependent fluorescence intensity change from 2 was
fitted to derive the 1,4-rearrangement rate constant of k2 )
1.43 × 10-2 s-1. As anticipated, this unimolecular rate
constant was independent of the [F-] () 0.02-0.10 mM).9
With 1 and 2 serving as convenient rearrangement clocks,
we decided to investigate the kinetic consequences of varying
the number of QM relays and the pathways sampled from
the covalent trigger to the fluorescent reporter. Toward this
end, compounds 3 and 4 were prepared and subjected to
kinetic studies. Under the same conditions as used for 2 (gray
curve, Figure 2b), the para-branched 3 involving consecutive
1,4- and 1,6-rearrangements responded much more slowly
toward F- (blue curve, Figure 2b) with an effective k2 value
of 1.38 × 10-3 s-1. This k2 parameter in a simplified two-
step reaction scheme (eq 2) reflects the collective time delay
via two QM rearrangement steps (Scheme 3) and is therefore
significantly (>10-fold) smaller than that (1.43 × 10-2 s-1)
of 2 undergoing only single QM rearrangement. Quite
unexpectedly, however, the ortho-branched isomer 4 (red
curve, Figure 2b) showed a remarkably faster release of the
end group compared with 3 (blue curve, Figure 2b). Notably,
the exponential growth of the self-fragmentation product
from 4 now indicated shifting of the mechanism to k1′ , k2
with the rate-limiting step taken over by the Si-O cleavage
reaction (k1 ) 159 M-1 s-1) similar to the behavior of 1
(Figures 2a and S3, Supporting Information).
Figure 2. Time-dependent changes in the fluorescence intensity at
λ ) 460 nm (λexc ) 380 nm) observed from 1 (open circle), 2
(open square), 3 (filled circle, blue) and 4 (open triangle, red) upon
reaction with F- (20 equiv) in CH2Cl2 at T ) 293 K. The gray
curves overlaid on the experimental data in (a) are theoretical fits
generated using k1 ) 155 M-1 s-1 (for 1 and 2) and k2 ) 1.43 ×
10-2 s-1 (for 2). The fitted curve of 2 is reproduced in (b) for
comparison with 3 and 4.
indicated a rate-determining Si-O bond cleavage by F- with
the second-order rate constant of k1 ) 155 M-1 s-1. This initial
bimolecular step is followed by a rapid 1,4-rearrangement
(Scheme 3) and subsequent release of CO2 and the fluorescent
reporter. Here, the requisite phenoxide intermediate apparently
builds up only to a very low and steady-state concentration,
thus validating the approximation of k1′ , k2 in eq 2 and the
use of eq 1 for kinetic modeling (Figure S3, Supporting
Information).10
Considering that the p- (for 3) and o-QM (for 4) precursor
would have similar propensity toward bond rearrangements,
their distinctively different kinetic behavior was difficult to
rationalize. More perplexingly, the self-fragmentation of 4
(red curve, Figure 2b) was even faster than that of 2 (gray
curve, Figure 2b), which is essentially impossible to explain
if the chain fragmentation of 4 proceeds via sequential
cleavage of the ether followed by the carbonate junction as
deduced from the molecular structure (Scheme 3). This
counterintuitive observation of a faster signal relay oVer a
longer distance suggested the presence of an efficient
“shortcut” in self-immolation and immediately called for a
complete reconsideration of our initial mechanistic view.
(10) Connors, K. A. Chemical Kinetics: The Study of Reaction Rates in
Solution; Wiley-VCH: Weinheim, 1990.
Org. Lett., Vol. 11, No. 10, 2009
2067

To delineate the molecular basis of this unusually fast
signal relay through 4, an expanded set of compounds 5-8
were prepared and deployed for kinetic analysis. The results
summarized in Table 111 reconfirmed consistently faster
Scheme 4. Chain Fragmentation by Repetitive QM Rearrangement
and Elemination (A) or by Direct Nucleophilic Attack (B)
Table 1. Kinetic Parameters for Self-Fragmentation of 1-8
(CH2Cl2, T ) 293 K)11
compound
k1 (M-1 s-1
)
k2 × 103 (s-1
)
1
2
3
4
5
6
7
8
155a
155c
155c
159a
54a
b
14.3c
1.38c
b
b
54d
6.3d
b
for direct nucleophilic attack by the phenoxide group (path-
way B) to release the reporter group. If this alternative signal
relay proceeds faster than the conventional QM rearrangements,
this unified mechanistic model could fully account for the highly
unsual kinetic behavior of the 4.12 Fragmentation of the
geometerically constrained para-linked isomer 3, on the other
hand, can proceed only through stepwise rearrangements
(Scheme 3), thereby resulting in slower kinetics than 2 or 4
(Figure 2b). In support of this notion, high-resolution MS
analysis of the products obtained from the reaction between 4
and F- indeed identified the internal cyclization product from
a direct nucleophilic attack (Figure S5, Supporting Informa-
tion).9,13
In summary, a systematic investigation of well-defined
small molecules suggested the operation of a dual mechanism
in controlled chain fragmentation reactions. For long-lived
intermediates with appropriate conformational bias, direct
nucleophilic attack can kinetically outperform stepwise bond
rearrangements. Alternative design principles capitalizing on
such an efficient signal relay cascade can readily be
envisioned for stimuli-responsive molecules and materials,
which are currently being investigated in our laboratory.
53a
53e
5.4e
a Fitted using eq 1 under pseudo-first-order conditions with k1′ ) [F-]0k1.
b With k2 . k1, eq 2 converges to eq 1. c Fitted using eq 2 with k1 fixed to
that of 1. d Fitted using eq 2 with k1 fixed to that of 5. e Fitted using eq 2
with k1 fixed to that of 7.
fragmentation rates of the carbonate-linked systems compared
with their ether-linked analogues, regardless of the cross-
linking pattern, i.e., ortho (5 vs 6) or para (7 vs 8). In
addition, the rate-limiting Si-O bond cleavage proceeded
with comparable k1 values for both ortho and para isomers
(5 vs 7) of the carbonate-linked system. Most importantly,
the comparable k2 values obtained for 6 and 8 established
that there is no intrinsic difference in the 1,4- Vs 1,6-
rearrangement rates, which stands in stark contrast to the
behavior of 3 vs 4 shown in Figure 2b.
Acknowledgment. This work was supported by the U.S.
Army Research Office (W911NF-07-1-0533). D.L. is an
Alfred P. Sloan Research Fellow.
The simplest and most intuitive explanation for the unusual
“ortho effect” in 4 is the operation of an alternative frag-
mentation mechanism that apparently outperforms the sequential
QM rearrangements. As proposed in Scheme 4, the initial Si-O
bond cleavage of 4 would generate the phenoxy intermediate.
In addition to taking the anticipated repetitive 1,4-rearrangement
route (pathway A), this intermediate could adopt a conformation
that brings the scissle benzylic carbonate group into a position
Supporting Information Available: Synthesis and char-
acterization of 1-8 and kinetic measurements and data
analysis. This material is available free of charge via the
Internet at http://pubs.acs.org.
OL900433G
(12) A conceptually related intramolecular nucleophilic attack of
“unmasked” phenoxide and subsequent liberation of fluorophore has
previously been reported: Chandran, S. S.; Dickson, K. A.; Raines, R. T.
J. Am. Chem. Soc. 2005, 127, 1652–1653.
(11) For compounds 3, 6, and 8, kinetic data at initial stages of the reaction
suffer excessively from interference of multiple intermediates that build up
and decay simultaneously.10 As such, intensity changes at t > 40 s were used
to derive an effective k2 value from global fitting of five independent data sets
to approximate the time delay between the initial Si-O bond cleavage event
and the eventual release of the final product umbelliferone.
(13) HR-MS: calcd for [M + H]+ 281.0808, found 281.0803. Full
characterization of the reaction products from preparatory scale reaction
was hampered by the high dilution (1.0 µM) required to simulate the reaction
conditions in which the kinetic measurements were carried out.
2068
Org. Lett., Vol. 11, No. 10, 2009