Mechanistic Studies of the Photoinduced Quinone Trimethyl Lock Decaging Process

Article abstract of DOI:10.1021/jacs.6b12007

Mechanistic studies of a general reaction that decages a wide range of substrates on exposure to visible light are described. The reaction involves a photochemically initiated reduction of a quinone mediated by an appended thioether. After reduction, a trimethyl lock system incorporated into the quinone leads to thermal decaging. The reaction could be viewed as an electron-transfer initiated reduction of the quinone or as a hydrogen abstraction-Norrish Type II-reaction. Product analysis, kinetic isotope effects, stereochemical labeling, radical clock, and transient absorption studies support the electron transfer mechanism. The differing reactivities of the singlet and triplet states are determined, and the ways in which this process deviates from typical quinone photochemistry are discussed. The mechanism suggests strategies for extending the reaction to longer wavelengths that would be of interest for applications in chemical biology and in a therapeutic setting.

Full text of DOI:10.1021/jacs.6b12007

Subscriber access provided by University of Newcastle, Australia
Article
Mechanistic Studies of the Photoinduced
Quinone Trimethyl Lock Decaging Process
Clinton J Regan, David P. Walton, Oliver S Shafaat, and Dennis A. Dougherty
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.6b12007 • Publication Date (Web): 15 Feb 2017
Downloaded from http://pubs.acs.org on February 15, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 10
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Mechanistic Studies of the Photoinduced Quinone Trimethyl Lock
Decaging Process
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Clinton J. Regan, David P. Walton, Oliver S. Shafaat, and Dennis A. Dougherty*
Division of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena, CA 91125 USA
ABSTRACT: Mechanistic studies of a general reaction that decages a wide range of substrates on exposure to visible light are
described. The reaction involves a photochemically initiated reduction of a quinone mediated by an appended thioether. After re-
duction, a trimethyl lock system incorporated into the quinone leads to thermal decaging. The reaction could be viewed as an elec-
tron-transfer initiated reduction of the quinone or as a hydrogen abstraction – Norrish Type II – reaction. Product analysis, kinetic
isotope effects, stereochemical labeling, radical clock, and transient absorption studies support the electron transfer mechanism. The
differing reactivities of the singlet and triplet states are determined, and the ways in which this process deviates from typical qui-
none photochemistry are discussed. The mechanism suggests strategies for extending the reaction to longer wavelengths that would
be of interest for applications in chemical biology and in a therapeutic setting.
enhancements associated with the trimethyl lock system, re-
leasing HX as a generic alcohol, amine, thiol, or phosphate. Of
course, deprotection of the phenol can be accomplished photo-
chemically using established caging groups,5,6 but this ap-
proach does not lead naturally to longer wavelength systems.
Introduction and Synthesis
We recently described a new class of compounds that undergo
photochemical decaging of a wide range of substrates at wave-
lengths as long as 600 nm.1 Such compounds could find use as
chemical biology tools, and in a therapeutic setting, where
longer wavelengths lead to deeper tissue penetration. In an
effort to maximize decaging efficiency and to provide insights
into possible strategies for extending the photoreactivity to
longer wavelengths, we have conducted extensive mechanistic
studies of the photoreaction. Herein we describe those mecha-
nistic studies and the design strategies they suggest.
Scheme 1. Products of photolysis of 1 at 420 nm.
Figure 1. Variants of the trimethyl lock decaging process.
The initial approach sought to take a known chemical decag-
ing process and design systems that could be phototriggered.
Figure 1 shows two variants of the well-established trimethyl
lock system.2–4 Either reducing a quinone or revealing a phe-
nol produces a nucleophile that can exploit the remarkable rate
1
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 10
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Photoinduced quinone trimethyl lock derivatives discussed in this work.
With a goal of creating a photochemically triggered trimethyl
lock system, we considered compound 1 (Figure 2). The bi-
molecular photoreduction of quinones by sulfides has been
reported, but, in general, the reaction has not been extensively
studied.7,8 The process is believed to be initiated by an electron
transfer (ET) followed by a crucial C-H oxidation, similar to
photoreduction by amines.9–12 Intramolecular variants are
known,13,14 but for these reactions a direct hydrogen abstrac-
tion (Norrish Type II-like) process cannot be ruled out. For the
present purposes, we sought to employ an ET mechanism, as
this seemed better suited for longer wavelengths. Although
most quinone photoreductions involve amines,15–22 we chose a
sulfide as the potential electron donor in our initial design. We
anticipated a more facile synthesis of the desired systems,
more favorable redox properties, and perhaps greater stability
in air and in a biological system. The synthesis of 1 is efficient
and permits a wide variety of sulfide substituents to be intro-
duced in the last step (Figure 2). A representative UV/Vis
spectrum is shown in Figure 3 for the methyl derivative (1a).
Notably, a broad absorption band is observed at approximately
413 nm; the relevant data for this band (λmax and ε) for key
substrates are reported in Table 1. We have been unable to
observe luminescence from 1a, either in fluid media at room
temperature or at 77K in a frozen matrix, as is typical of qui-
nones.23–30
Steady-State Photolysis
Photolysis of 1a with, for example, a 420 or 455 nm LED, in
air-equilibrated methanol leads to the clean formation of thio-
acetal 2a (Scheme 1), as observed by H NMR of the crude
1
product. When the reaction is carried out in deuterated sol-
vent, ethanol can also be detected in the reaction mixture as
the exclusive byproduct. For some variants of 1, other reaction
products are observed in the crude and can be isolated by sili-
ca gel chromatography. Compound 3 is presumably produced
by intramolecular capture of the species that is trapped by
methanol. Photolysis in water (pure or buffered to pH 7.5) also
releases the caged alcohol and produces the disulfide 4, pre-
sumably via a thiohemiacetal intermediate. Both reactions are
very clean; quantum yields will be discussed below. In other
solvents such as acetonitrile, benzene, or hexane, the reaction
is slower and produces a complex mixture of products.
Compound 1j contains a tethered alcohol, and upon photolysis
in methanol it cleanly produces both 2j and the expected cy-
clic product 5 in a 1:4 ratio (Scheme 1). Photolysis of 1j in
acetonitrile or benzene does produce the cyclic product, but
there are also other intractable products. Compounds 1h and
1i, both possessing electron-withdrawing groups, are peculiar
in that they produce decomposition products upon photolysis
in methanol, and display visible absorption bands that are
weak and blue-shifted (Table 1). It is clear that some of the
decomposition products have not undergone trimethyl lock
lactonization, suggesting that unmasking of the phenol has not
occurred. Compound 1d lacks the necessary γ-hydrogen on
sulfur and is found to be nonreactive to photolysis at 420 nm.
By comparison, compound 6 (Figure 2) lacks a sulfur substitu-
ent altogether, and it is found to undergo γ-hydrogen abstrac-
tion from the trimethyl lock side chain. Similar processes
have been reported previously.7,31–37
10
8
6
4
Mechanistic Studies
2
Most of our mechanistic studies have been conducted in meth-
anol, where the reaction is clean and solubility is not an issue.
We will examine the mechanism by working backwards from
the final product. It is clear that the actual ring closure of the
trimethyl lock and the release of the caged compound is the
final and slowest step of the process. One could have imagined
that an initial photochemical ET from the sulfide to the qui-
none would build enough negative charge on the quinone ox-
ygens such that a trimethyl lock closure could occur before
250
300
350
400
450
500
550
wavelength (nm)
Figure 3. Absorbance spectrum of 1a in methanol.
2
ACS Paragon Plus Environment

Page 3 of 10
Journal of the American Chemical Society
further reduction of the quinone,18 but that is not the case. In
compounds (1f,1g) do not follow a simple trend, and we note
again that the p-nitrobenzyl substrate 1h produces a complex
mixture of products.
1
2
3
4
5
6
7
8
deuterated methanol, the hydroquinone (7, Scheme 2) can be
directly observed by 1H NMR prior to trimethyl lock ring clo-
sure. In aqueous systems the ring closure is rapid for an alco-
hol leaving group, but not for an amine leaving group, again
allowing the hydroquinone to be observed prior to trimethyl
lock closure.1 These results could be anticipated based on
known trimethyl lock rates.2
We next considered the role of the hydrogen shift on the over-
all process. There is an isotope effect (ΦH/ΦD) on the quantum
yield for quinone disappearance. A value of 4.0 is obtained for
the methyl compound (1a vs. 1a-d3), and 2.5 for the benzyl
compound (1e vs. 1e-d2) (Table 1). These observations high-
light the critical role of the hydrogen-shift in the overall pro-
cess.
The mechanistic issue then becomes the conversion of quinone
1 to the methanol adduct hydroquinone, 7. The requirement
for a solvent capture step implicates zwitterion 8 as the likely
precursor to 7. Conceptually, the conversion of 1 to 8 then
requires reduction of the quinone by two electrons and the
shift of a proton from the carbon attached to the sulfur to the
quinone oxygen. The matter in question is whether the reac-
tion proceeds through ionic intermediates, similar to bimolecu-
lar photoreductions by sulfides and amines, or through a tradi-
tional radical mechanism similar to the Norrish II (Scheme 2).
To keep the semantics straight, we will use the term hydrogen-
shift as noncommittal regarding all steps in the process 1 ꢀ 8.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 1. Spectroscopic and photolysis data for 1 in MeOH.a
Scheme 2. Potential intermediates and pathways in the
photolysis of 1.
aAll quantum yields are reported at 420 nm relative to ferrioxalate
actinometer with a standard deviation of < 10%. bLongest wave-
length absorption band in air-equilibrated methanol. cQuantum
yield for disappearance of 1. dLifetime of the transient observed
at 480 nm upon pulsed laser irradiation at 355 nm in degassed
methanol. eQuantum yield for disappearance of 1 through the
f
singlet (S) and triplet (T) pathways; ref to Eq. 1. Quantum effi-
ciency of intersystem crossing for 1. gMinimum value of the
quantum efficiency for disappearance of 1 from the sensitized
state. hRate constant for reaction from the sensitized state.
Either a radical or ionic mechanism could potentially generate
biradical 10 (Scheme 2), with the former being a conventional
Norrish II reaction. To probe for the intermediacy of 10, we
incorporated radical clocks into the system, preparing the 5-
hexenyl
(1k),
cyclopropylmethyl
(1l),
and
2-
phenylcyclopropylmethyl (1m) derivatives. These are stand-
ard probes that have been used successfully in conventional
Norrish II reactions.39–43 For both 1k and 1l, no radical rear-
rangement is seen; the products 2 and 3 are cleanly produced.
The phenylcyclopropyl clock shows a very fast intrinsic ring
opening rate of 1011 s-1.39,44 Photolysis of 1m produces a 20%
yield of the expected methanol-trapping product with the phe-
nylcyclopropyl ring still intact. The remaining material is a
mixture of products that has not been fully characterized.
While it is likely true that the sulfur in our system perturbs the
radical rearrangements studied here, it seems safe to conclude
that if a biradical such as 10 is directly formed in this system,
it has a very short lifetime – much shorter than a typical Nor-
rish II biradical.42An especially telling probe of the role of the
hydrogen shift was provided by a stereochemical test. Enanti-
omerically pure phenethyl derivative 1n was prepared with
>95% ee as determined by chiral HPLC (Supporting Infor-
mation). Upon photolysis to 75% conversion, recovered start-
ing material showed no racemization within the detection lim-
its of the method. As such, we conclude that the hydrogen
shift is effectively irreversible, and that nonproductive decay
We have applied a number of mechanistic tools to this reac-
tion. First, the influence of sulfide substituents on the quantum
yield for product formation was probed. Using a ferrioxalate
actinometer38 we have determined the quantum yield (Φ) for
the conversion of 1 in degassed methanol solutions, and the
results are summarized in Table 1. The effect of added oxygen
on the quantum yield is generally small and will be discussed
further below. There is a trend of i-propyl (1c) > ethyl (1b) >
methyl (1a) in relative quantum yield, although the effect is
not large. A benzyl substituent (1e) shows the largest effect,
with a > 5-fold increase in quantum yield. These trends would
be consistent with either radical or cationic character building
up on the substituted carbon. However, substituted benzyl
3
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 10
of the excited state occurs prior to the hydrogen shift, making yield. We have undertaken several studies to probe the role
1
the hydrogen shift a key mechanistic event.
and nature of the triplet state in the photoreaction.
2
3
Nanosecond Transient Absorption Studies
4
5
6
7
8
9
2.2
2.0
1.8
1.6
1.4
1.2
1
To provide further insight into possible mechanisms for this
reaction, we have studied this system using nanosecond laser
flash photolysis with transient absorption. Briefly, samples
were excited at 355 nm with an 8 ns pulse at 10 Hz. On excita-
tion of 1a, a transient spectrum with an absorption λmax of 480
nm is observed (Figure 4). The transient absorption signal
decayed in a single exponential with a lifetime (ꢀ) of 930 ns in
degassed methanol. Similar transients are seen from a number
of structures (Table 1). In all cases we have observed that the
products formed in the laser experiments are the same as in the
steady-state photolysis. In air-equilibrated solutions, the same
transient is observed, but in all cases the lifetime is in the 100
– 200 ns range, consistent with diffusional quenching by oxy-
gen. The observed transient is also quenched by amine-based
quenchers. Considering the parent, 1a, in the presence of 10
mM triethylamine (TEA), an initial decay with a lifetime of
310 ns is seen, compared to 930 ns in the absence of TEA. As
shown in Table 1, there is a considerable variation in lifetime
(ꢀ) for the 480 nm transient. For the simple hydrocarbon sys-
tems (methyl, ethyl, isopropyl, benzyl), the transient lifetime
tracks the product quantum yield measured in bulk, with the
benzyl transient being significantly shorter lived than the me-
thyl. As in the bulk photolysis, a significant KIE is seen for the
transient lifetime for the benzyl compound (1e vs. 1e-d2).
However, a minimal KIE is seen for the methyl compound (1a
vs. 1a-d3).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
0
0.08
0.09
0.01
0.02
0.03
0.04
0.05
0.06
0.07
quencher concentration, [Q] / M
Figure 5. Stern-Volmer plot for the quenching of the quantum yield
for disappearance of 1a (white diamonds) and 1e (black squares) by
diethylaniline. Dotted curves are multivariable regression fits to Eq.
2. Inset is the low concentration region of the plot.
We initially considered the impact of triplet quenchers on the
overall process, and obtained clean quenching with diethylani-
line. Quenching by diethylaniline likely occurs through re-
versible electron transfer, as has been previously shown in
similar systems.45 Shown in Figure 5 are Stern-Volmer (SV)
plots for photoreaction of the methyl (1a) and benzyl (1e)
compounds, where Φꢁ is defined as the quantum yield in the
presence of quencher. At low concentrations of diethylaniline
(up to ~ 1 mM), roughly linear SV behavior is observed (Fig-
ure 5, inset). However, at higher concentrations (up to ~ 100
mM), the SV plot deviates from linearity and plateaus. This
indicates that the photoreaction proceeds through two different
pathways, one being much more efficiently quenched than the
other. We have assigned the less and more quenchable por-
tions of the photoreaction to those that occur through the sin-
glet (Φꢂ) and triplet (Φꢃ) state, respectively. The sum of these
pathways describes the overall quantum yield (Eq. 1).
Φ
=
=
Φꢄ + Φꢅ
(1)
(2)
ꢌ
ꢐ
ꢈꢉ ꢊ
ꢋ
[ꢎ]ꢏꢈꢉ ꢊ
ꢋ
[ꢎ]
ꢏ
ꢆ
ꢌꢍ
ꢌꢍ
ꢐ
[
ꢐ
]
ꢋ
ꢑ
ꢆ
ꢌꢍ
ꢇ
ꢉ ꢊ
ꢌ
ꢒꢓꢈΦ
/Φ
ꢏ
Eq. 2 has been previously derived to describe the variation in
quantum yield when there are two quenchable pathways.38 The
expression includes two SV quenching constants, ꢋꢂꢂꢔ and ꢋꢂꢃꢔ,
Figure 4. Transient absorption spectrum of 1a observed upon laser
flash photolysis at 355 nm in aerated acetonitrile. Inset: single expo-
nential fit (red) of transient decay at 480 nm.
for the singlet and triplet pathways, and the ratio of their quan-
tum yields (Φꢃ Φ ). A regression analysis of Eq. 2 results in
ꢂ
⁄
the fits shown in Figure 5. This analysis has been performed
for key substrates, producing values of Φꢂ and Φꢃ after inser-
tion of the determined ratio Φꢃ Φ into Eq. 1 (Table 1). As
ꢂ
⁄
Triplet Sensitization/Quenching Studies
expected, the triplet pathways of 1a and 1e are efficiently
quenched, with similar ꢋꢂꢃꢔ values of 1800 and 1700 M-1, re-
spectively. In contrast, the singlet states are negligibly
quenched with ꢋꢂꢂꢔ values of 0.45 and 0.52 M-1. These two
Excitation of simple quinones typically produces a triplet state
with near unit efficiency.23,24 The present system, however, is
significantly perturbed, electronically by the sulfur substituent
and geometrically by the bulky trimethyl lock system. The
long lifetime of the transient from the flash photolysis studies,
and the fact that it is quenched efficiently by oxygen, suggest
that the transient is a triplet. However, in steady-state photoly-
sis studies, oxygen has only a small effect on the quantum
compounds are found to differ largely in their ratios of singlet
to triplet reactivity, Φꢃ Φ . The calculated ratios reveal that,
ꢂ
⁄
for 1a, the reaction proceeds 88% through the singlet and 12%
through the triplet. For 1e, the corresponding singlet and tri-
plet values are 47% and 53%. If we assume that the transient
observed in laser flash photolysis is the triplet, we can use the
transient lifetimes (ꢀ) and ꢋꢂꢃꢔ values to obtain kq, the second-
4
ACS Paragon Plus Environment

Page 5 of 10
Journal of the American Chemical Society
described by Eq. 5, where τꢗꢞꢜ is the intrinsic lifetime of the
order rate constant for quenching. We find kq to be 10 x 109 M-
1s-1 and 2 x 109 M-1s-1 for 1a and 1e, respectively, which are
similar to diffusion-controlled values in methanol, where kdiff
= 1.2 x 1010 M-1s-1.46
1
2
3
triplet sensitizer, and ꢠꢟꢞꢟ[ꢙ] and ꢠꢜꢡ[ꢙ] are pseudo-first order
rate constants for deactivation of the sensitizer through pro-
ductive and nonproductive collisions with quinone, respective-
ly.
4
5
6
7
8
ꢗꢞꢜ
Φꢗꢞꢜ
ϕꢟꢞꢟ
=
=
ꢕ
• ꢕꢟꢞꢟ • ꢕꢚꢃꢛꢜ
(4)
(5)
ꢖꢗꢘ
ꢢ
[ꢙ]
ꢣꢤꢣ
ꢣꢤꢣ
ꢉ
ꢨ
[
]
ꢙ
[ ]
ꢙ
ꢧꢩ
ꢊ ꢢ
ꢊ ꢢ
ꢦꢤꢧ
ꢥ
9
Insertion of Eq. 5 into Eq. 4, and taking the inverse reveals a
double-reciprocal linear relationship between the sensitized
quantum yield (Φꢗꢞꢜ) and quinone concentration ([1]) (Eq. 6),
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
where the product ꢠꢟꢞꢟ • τꢗꢞꢜ is recognized as a Stern-Volmer
ꢟꢞꢟ
constant ꢋ for the productive quenching of the triplet sensi-
ꢂꢔ
tizer by 1. The reciprocal of the y-intercept in Eq. 6 is desig-
ꢪꢖꢫ
nated as Φ
(Eq. 7) and describes the sensitized quantum
ꢗꢞꢜ
yield for the disappearance of 1 in the limit where deactivation
of the triplet sensitizer occurs exclusively through collisions
with the quinone. Similar analyses have been presented previ-
ously.47
Figure 6. Processes that contribute to the direct (solid) and sensitized
(dashed) quantum yields for the disappearance of 1 upon excitation at
420 nm. Nonproductive processes have been omitted for clarity.
ꢉ
ꢢ
ꢊ ꢢ
ꢣꢤꢣ
ꢉ
ꢉ
ꢉ
ꢙ
ꢣꢤꢣ
=
=
ꢲ ꢳ
•
+
]
ꢧꢩ ꢶ
(6)
(7)
ꢦꢤꢧ
ꢭꢦꢮ
ꢱ
ꢣꢤꢣ
[
ꢬ
• ꢬ
ꢢ
ꢆ
ꢯꢰꢧ
ꢴꢵ
ꢦꢤꢧ
The results from the quenching experiments reveal that the
overall quantum yield (Φ, Table 1) measured in the steady-
state photolysis has components from both the singlet (ΦS) and
triplet (ΦT) pathways (Eq. 1). A minimal model describing the
relevant steps that contribute to these pathways is shown in
Figure 6, where nonproductive processes have been omitted
for clarity. According to this model, the quantum yield from
the singlet (ΦS) due to direct photolysis of the quinone is de-
fined by the pathway 1 ꢀ *1 ꢀ 8 ꢀ 7 ꢀ 2. The conversion
of 8 ꢀ 7 ꢀ 2 occurs with unit efficiency, as evidenced by the
clean formation of product and the irreversibility of the hydro-
gen-shift. ΦS is therefore simply defined by the efficiency of
*1 ꢀ 8.
ꢢ
ꢣꢤꢣ
ꢗꢞꢜ
ꢖꢗꢘ
ꢪꢖꢫ
Φ
ꢕ
• ꢕꢷꢃꢛꢜ
•
ꢗꢞꢜ
ꢢ
ꢊ ꢢ
ꢣꢤꢣ
ꢧꢩ
Representative double-reciprocal plots are shown in Figure 7
for 1a, 1e, and the deuterated anologs 1a-d3 and 1e-d2. Deter-
ꢪꢖꢫ
mination of Φ
from the fit is accomplished by averaging
ꢗꢞꢜ
three independent samples. Although the standard deviation in
ꢪꢖꢫ
Φ
is consistently less than 10% (Figure 7, error bars), we
ꢗꢞꢜ
note that the slope in the double-reciprocal plots is unexpect-
edly sensitive to the concentration of thioxanthone. This fact is
demonstrated explicitly for 1a, where three samples containing
1, 2, and 3 mM thioxanthone resulted in incremental shifts in
the slope. The y-intercept, however, is clearly unaffected.
This kind of behavior has been seen before,47 and it was not
probed further.
Likewise, disappearance of quinone through the triplet state
(ΦT) upon direct photolysis is defined by the pathway 1 ꢀ *1
ꢀ 31 ꢀ 8 ꢀ 7 ꢀ 2. The quantum yield for this process, giv-
en by Eq. 3, is constructed as the product of the contributing
efficiencies, namely ꢕꢖꢙꢗꢘ, the efficiency of triplet formation
via intersystem crossing, and ꢕꢚꢃꢛꢜ, conversion of triplet to
zwitterion 8.
If we assume that each quenching event of the sensitizer leads
to triplet energy transfer, then ꢠ_ꢸꢹꢸ/(ꢠ_ꢸꢹꢸ + ꢠ_ꢺꢻ ) = 1.
This, combined with the known efficiency of intersystem
48,49
ꢗꢞꢜ
crossing for thioxanthone (ꢕ ) of 0.56,
allows the deter-
ꢖꢗꢘ
mination of ꢕꢚꢃꢛꢜ values, which are reported in Table 1. These
values represent lower limits, given the assumption above. In
general, ꢕꢚꢃꢛꢜ is much larger than Φ, suggesting that the triplet
state is not efficiently generated by direct photolysis. Substi-
tution of ꢕꢚꢃꢛꢜ into Eq. 3 permits calculation of the efficiency
for intersystem crossing by the quinone ( ꢕꢖꢙꢗꢘ, Table 1). Alt-
hough quinones typically form triplets with high yields,23 the
efficiencies observed in this system do not exceed 10%, likely
due to the electronic and steric effects of the sulfide and trime-
thyl-lock substituents, respectively.
Φꢃ
=
ꢕꢖꢙꢗꢘ • ꢕꢚꢃꢛꢜ
(3)
We sought to explore the nature of this triplet pathway in more
detail through the use of triplet sensitizers. Many efforts to
employ sensitizers were either ineffective (acetophenone, ben-
zophenone, methylene blue) or produced undesired side prod-
ucts (biacetyl, naphthalene, anthracene, rose bengal). Howev-
er, thioxanthone produced clean and consistent results. Also
shown in Figure 6 are processes that contribute to the sensi-
tized quantum yield for the disappearance of 1. The pathway
begins with excitation and intersystem crossing of the sensitiz-
The first-order rate constants for formation of 8 from the tri-
plet state ( ꢠꢚꢃꢛꢜ) can be calculated using Eq. 8 if we assume
that the triplet considered here is the transient observed in
laser flash photolysis. Although the results, collected in Table
1, reflect broad trends in bond dissociation energy (BDE) with
the simple alkyl substituents 1a – 1c reacting slower than ben-
zylic substituent 1e, the rates are clearly complicated by other
factors. For instance, the fact that the methyl (1a) and isopro-
pyl (1c) derivatives have essentially the same rate constant
cannot be explained using simple BDE arguments alone.
er, ꢌꢝꢗꢞꢜ ꢀ ꢐꢗꢞꢜ, followed by bimolecular triplet energy trans-
ꢉ
fer to the quinone, ꢐꢗꢞꢜ + 1 ꢀ ꢌꢝꢗꢞꢜ + 31 . Conversion of the
ꢉ
triplet quinone to product then proceeds normally (31 ꢀ 8 ꢀ 7
ꢀ 2). The quantum yield for this pathway (Φꢗꢞꢜ) is given by
ꢗꢞꢜ
the product shown in Eq. 4, where ꢕ
is the efficiency of
ꢖꢗꢘ
intersystem crossing for the sensitizer, and ꢕꢟꢞꢟ is the efficien-
cy of triplet energy transfer. The efficiency of triplet energy
transfer depends upon the concentration of quinone, [1], and is
5
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 10
ϕꢃꢚꢛꢜ
=
ꢠꢚꢃꢛꢜ • ꢀ
(8)
1
2
3
4
800
600
400
∼
∼
5
6
7
8
40
35
30
25
20
15
10
5
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 8. The overall photochemical transformation and efficiencies
for the conversion of 1a (R = H; blue) and 1e (R = Ph; red) to the
zwitterion 8. The ester side chain is abbreviated TML.
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
x 104
1 / [1] (M-1)
Figure 7. Double reciprocal plots for the sensitized photolysis of 1 by
thioxanthone in degassed methanol. 1e (black squares), 1e-d2 (white
squares),1a (black diamonds), 1a-d3 (white diamonds). Dotted lines
are linear fits; error bars are the standard deviation in the y-int for
three independent samples. Three samples of 1a using different con-
centrations of thioxanthone are shown to demonstrate that the slope,
but not the y-intercept, is affected.
Based on several lines of evidence, we consider the initial
1
excited state to be the charge transfer state, 9, as in Figure 8,
rather than a simple n,π* state. The broad visible absorption
band observed in these compounds (Figure 3) is common for
benzoquinones bearing sulfide or amino groups, and is indica-
1
tive of charge-transfer.29 The dominant fate of 9 is nonemis-
Significant isotope effects on the sensitized quantum yield
(ϕꢃꢚꢛꢜ) are also observed in Figure 7 and Table 1. For the
1e/1e-d2 system, a magnitude of 1.7, given by the ratio ꢕꢼ/ꢕꢽ
is similar to the magnitude of 2.5 observed in direct photoly-
sis. Application of Eq. 8 to these data reveal a large normal
KIE (ꢠꢼ/ꢠꢽ) of 7. A similar analysis shows that the methyl
analogue, 1a/1a-d3, experiences a very large sensitized product
isotope effect (ꢕꢼ/ꢕꢽ) of 70, compared to a direct photolysis
isotope effect of 4. In particular, we find that the deuterated
analog (1a-d3) is very inefficient in the sensitized photolysis
(Figure 7), with a quantum yield (ϕꢃꢚꢛꢜ) of 0.03, comparable to
the efficiency for direct photolysis (Φ). Calculation of the KIE
for 1a/1a-d3 from the transient lifetimes and Eq. 8 yields a
value (ꢠꢼ/ꢠꢽ) of 70.
sive return to the ground state (> 90% efficiency), again in
contrast to simpler benzoquinones, which typically undergo
intersystem crossing to the triplet with near unit efficiency.23
The triplet state does form with low efficiency, and again it is
not like the triplet of a typical benzoquinone. Shown in Figure
9 are spin-density plots for relevant species computed using
DFT M06/6-311++G**, where the trimethyl lock side chain is
not displayed for clarity but is included in the calculation. The
parent benzoquinone is a conventional n,π* triplet state (con-
sistent with experiment),50 demonstrated by significant spin
density on the oxygen n-orbitals orthogonal to the π−system
(Figure 9A). In sharp contrast, the triplet state of 1 is well-
3
represented by the charge-transfer structure, 9, with no spin
density on the oxygen n-orbitals, indicating that they are dou-
bly-occupied. There is also significant spin-density on sulfur
(Figure 9B).
Mechanistic Interpretation
Based on the accumulated evidence, we argue that the most
plausible mechanism is the one outlined in Figure 8 (a more
complete version of the ionic path in Scheme 2). The penul-
timate intermediate is the zwitterion 8; once it is formed,
product formation involves solvent capture and subsequent
trimethyl lock ring closure. That 8 proceeds with high effi-
ciency (>95%) to the final products is evidenced by the stereo-
chemical labeling studies. Formation of 8 can occur in both
the singlet and the triplet manifolds. For the simple alkyl sub-
strates, product formation is dominated by the singlet pathway,
as evidenced by a large contribution of Φꢄ to the overall quan-
tum yield, Φ (Table 1). The more efficient reaction for the
benzylic substrates includes a more substantial contribution
from the triplet pathway.
Typically, when given the opportunity, excited states of simple
benzoquinones will readily undergo γ-hydrogen abstrac-
tion.36,37,51 The trimethyl lock unit of the present system pos-
sesses eight γ-hydrogens, but we see no products suggesting
that hydrogen abstraction occurs from that unit. In contrast,
compound 6, lacking sulfur, undergoes efficient abstraction of
the trimethyl lock hydrogens, but 1d, possessing a t-butyl sub-
stituent on sulfur, is remarkably photostable. These observa-
tions suggest formation of a charge transfer excited state that,
in both the singlet and triplet states, has a reactivity pattern
that is significantly different from that of a simple benzoqui-
none. While the charge transfer excited state converts to prod-
ucts with low efficiency, apparently it also suppresses the typ-
ical reactivity of benzoquinones, and so product formation is
very clean.
6
ACS Paragon Plus Environment

Page 7 of 10
Journal of the American Chemical Society
complicated. The p-chlorobenzyl derivative (1g) is more effi-
A
B
C
1
cient than the benzyl derivative (1e) not because the hydrogen-
shift is more favorable, but because the rate of nonproductive
triplet decay is lessened. In Figure 8 we show the triplet path-
2
3
4
5
6
7
8
9
3
3
3
way as 9 ꢀ 8 ꢀ 8. However, a direct conversion 9 ꢀ 8
seems possible. As noted above, the reaction is slower than a
typical hydrogen transfer. Additionally, since the triplet ac-
counts for less than 10% of the excited states, the stereochem-
ical labeling studies do not establish whether nonproductive
triplet decay occurs directly from 39 or through reverse hydro-
gen shift from 38.
10
11
Another intriguing feature of this system is the collection of
KIEs seen. The macroscopic quantum yield shows a moderate
isotope effect. More interestingly, combining the transient
absorption spectroscopy with the sensitized photolysis studies
allows a determination of the true KIE, kH/kD, for the triplet.
For the benzyl system, 1e/1e-d2, a value of 7 is observed,
which is relatively large for a reaction of this sort. Remarka-
bly, the parent methyl system, 1a/1a-d3, gives a kH/kD of 70
(Table 1). This effect is not evident in transient lifetime data
because the dominant decay path for the 1a/1a-d3 triplet is
return to S0. Such a large KIE strongly implies tunneling in
the hydrogen shift for this system. While γ-hydrogen shifts by
triplet ketones are generally understood to be activated pro-
cesses, certain constrained systems have been designed to
probe for tunneling at low temperature.55,56 The tunneling
rates observed in those systems by deuterium are found to be
similar to ꢠꢚꢃꢛꢜ for 4a-d3 (~ 103 s-1), indicating that nonclassical
effects may be observable for this slow reaction even at room
temperature.
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 9. Spin-density plots for the triplet state of (A) benzoqui-
none,(B) 1a, and (C) 6 computed using M06/6-311++G**. The trime-
thyl lock side chain is not displayed for clarity.
We conclude that the transient observed in the laser flash pho-
tolysis experiments is the triplet state, 9. That the transient is
3
a triplet is supported by quenching data for both oxygen and
diethylaniline, and by its long lifetime.52 That this triplet is on
the reaction pathway is supported by several lines of evidence.
Product formation is very clean in the transient absorption
experiments. Additionally, a significant KIE for the transient
lifetime indicates that the transient is a species that undergoes
a hydrogen-shift reaction. Also, to a considerable extent, the
lifetime of the transient tracks with the quantum yield of the
bulk reaction (Table 1). We would not expect a perfect corre-
lation, because both singlet and triplet paths are involved in
the bulk photolysis. Based on its reactivity pattern and the
computations noted above, we conclude that this triplet has
considerable charge transfer character, unlike a typical Norrish
II triplet. Also, for the alkyl substrates, reaction from the tri-
plet is unusually slow (ꢠꢚꢃꢛꢜ ~ 105 s-1, Table 1) compared to
typical γ-hydrogen shifts, which have rate constants greater
than 107 s-1 for typical ketones53 and simple benzoquinones.51
We interpret this to reflect the decreased n,π* character in the
triplet state of the structures studied here. Since γ-hydrogen
shifts are typically initiated by the electrophilic oxygen that is
characteristic of an n,π* state,53 39 is expected to react more
slowly through this type of mechanism. Electron-donating
substituents are known to preferentially stabilize the lowest
π,π* triplet state of quinones,23,54 and the substituents of 1,
Conclusions
We describe mechanistic studies of a new reaction that allows
rapid photochemical decaging of a wide range of structures
using the well-established trimethyl lock lactonization process.
Key to the development of this system was the discovery of an
effective phototrigger based on an intramolecular redox reac-
tion of benzoquinones bearing a sulfide substituent. Our re-
sults indicate that the process begins with photoinduced elec-
tron transfer, followed by a critical and irreversible hydrogen
shift that ultimately results in two electron reduction to the
hydroquinone. The nucleophilic hydroquinone oxygen is then
capable of undergoing rapid, thermal trimethyl lock lactoniza-
tion with release of the caged compound. Both singlet and
triplet pathways are viable, with the latter proceeding to prod-
uct much more efficiently.
3
especially the sulfide, are likely to exhibit this effect in 9.
The triplet state of 6, lacking sulfur, is also predicted to pos-
sess a high degree of π,π* character (Figure 9C), suggesting
that alkyl substituents alone may have a substantial effect on
the electronics of the quinone. While this could be due to
geometric distortion brought on by the trimethyl lock, some
unstrained quinones bearing alkyl substituents are also known
to have lowest π,π* triplet states.50
Given our mechanistic conclusions, many strategies for ex-
tending the excitation wavelength can be envisioned, as pho-
toinduced electron transfer is a heavily studied and well-
understood process. Also, the modular synthesis of these
compounds allows the substituent on sulfur to be readily var-
ied, allowing the introduction of groups that impact solubility,
cell permeability, and biodistribution in general. Further stud-
ies along these lines are underway.
In the present system, strong electron-withdrawing groups,
such as p-nitrobenzyl (1h) and methyl acetate (1i), result in
considerably lower quantum yields for quinone consumption
and yield a complex mixture of photolysis products. Again,
this is consistent with the presence of an ionic structure on the
pathway, as it would be destabilized by electron withdrawing
substituents due to the significant positive charge on the car-
bon attached to sulfur.
ASSOCIATED CONTENT
Supporting Information
Synthetic procedures and characterization data, procedures for
quantum yield determination and transient absorption experiments
(PDF) are available free of charge on the ACS Publications web-
site.
Although the triplet is formed in low yields, it proceeds to the
product much more efficiently than the excited singlet. For
benzylic derivatives, values of ϕꢃꢚꢛꢜ can approach unity. The
effect of substitution on the triplet photoreaction as a whole is
7
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 10
(23) Barbafina, A.; Latterini, L.; Carlotti, B.; Elisei, F. J. Phys.
Chem. A 2010, 114, 5980.
1
AUTHOR INFORMATION
2
3
4
5
6
7
8
(24) Hubig, S. M.; Bockman, T. M.; Kochi, J. K. J. Am. Chem.
Soc. 1997, 119, 2926.
Corresponding Author
(25) Trommsdorff, H. P. J. Chem. Phys. 1972, 56, 5358.
(26) Asundi, R. K.; Singh, R. S. Nature 1955, 176, 1223.
* dadougherty@caltech.edu
Notes
(27) Bridge, N. K.; Porter, G. Proc. R. Soc. Lond. Math. Phys.
Eng. Sci. 1958, 244, 259.
The authors declare no competing financial interest.
9
ACKNOWLEDGMENT
(28) Orgel, L. E. Trans. Faraday Soc. 1956, 52, 1172.
(29) Braude, E. A. J. Chem. Soc. Resumed 1945, 490.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
This work was supported by the W. M. Keck Foundation. Addi-
tional support was provided by the Arnold and Mabel Beckman
Foundation through the Caltech Beckman Institute Laser Re-
source Center. We thank Jay Winkler for technical assistance and
helpful discussions, and Scott Virgil for assistance with studies of
compound 1n.
(30) Bridge, N. K.; Porter, G. Proc. R. Soc. Lond. Math. Phys.
Eng. Sci. 1958, 244, 276.
(31) Orlando, C. M.; Bose, A. K. J. Am. Chem. Soc. 1965, 87,
3782.
(32) Orlando, C. M.; Mark, H.; Bose, A. K.; Manhas, M. S. J.
Am. Chem. Soc. 1967, 89, 6527.
REFERENCES
(33) Orlando, C. M.; Mark, H.; Bose, A. K.; Manhas, M. S. J.
Org. Chem. 1968, 33, 2512.
(1) Walton, D.; Dougherty, D.A., submitted.
(34) Baxter, I.; Mensah, I. A. J. Chem. Soc. C Org. 1970, 2604.
(2) Levine, M. N.; Raines, R. T. Chem. Sci. R. Soc. Chem. 2010
2012, 3 (8), 2412.
(35) King, T. J.; Forrester, A. R.; Ogilvy, M. M.; Thomson, R. H.
J. Chem. Soc. Chem. Commun. 1973, 844.
(3) Greenwald, R. B.; Choe, Y. H.; Conover, C. D.; Shum, K.;
Wu, D.; Royzen, M. J. Med. Chem. 2000, 43, 475.
(36) Kraus, G. A.; Wu, Y. J. Am. Chem. Soc. 1992, 114, 8705.
(37) Görner, H. J. Photochem. Photobiol. Chem. 2005, 175, 138.
(4) Gomes, P.; Vale, N.; Moreira, R. Molecules 2007, 12, 2484.
(5) Pelliccioli, A. P.; Wirz, J. Photochem. Photobiol. Sci. 2002,
1, 441.
(38) Klán, P.; Wirz, J. In Photochemistry of Organic Compounds;
John Wiley & Sons, Ltd, West Sussex, UK, 2009; pp 73–135.
(6) Klán, P.; Šolomek, T.; Bochet, C. G.; Blanc, A.; Givens, R.;
Rubina, M.; Popik, V.; Kostikov, A.; Wirz, J. Chem. Rev. 2013, 113
(1), 119.
(39) Hu, S.; Neckers, D. C. J. Org. Chem. 1997, 62, 755.
(40) Griller, D.; Ingold, K. U. Acc. Chem. Res. 1980, 13, 317.
(41) Wagner, P. J.; Liu, K. C. J. Am. Chem. Soc. 1974, 96, 5952.
(42) Small, R. D.; Scaiano, J. C. J. Phys. Chem. 1977, 81, 2126.
(43) Engel, P. S.; Keys, D. E. J. Am. Chem. Soc. 1982, 104, 6860.
(7) Görner, H. Photochem. Photobiol. 2006, 82, 71.
(8) Kavarnos, G. J.; Turro, N. J. Chem. Rev. 1986, 86, 401.
(9) Görner, H. Photochem. Photobiol. 2003, 78, 440.
(10) Scheerer, R.; Graetzel, M. J. Am. Chem. Soc. 1977, 99, 865.
(44) Castellino, A. J.; Bruice, T. C. J. Am. Chem. Soc. 1988, 110,
7512.
(11) Baxter, I.; Phillips, W. R. J. Chem. Soc. [Perkin 1] 1973,
268.
(45) Amouyal, E.; Bensasson, R. J. Chem. Soc., Faraday Trans. 1
1977, 73, 1561.
(12) Bruce, J. M. Q. Rev. Chem. Soc. 1967, 21, 405.
(46) Klán, P.; Wirz, J. In Photochemistry of Organic Compounds;
John Wiley & Sons, Ltd, West Sussex, UK, 2009; pp 467–470.
(13) Iwamoto, H.; Takuwa, A.; Hamada, K.; Fujiwara, R. J.
Chem. Soc. [Perkin 1] 1999, 575.
(47) Balzani, V.; Ballardini, R.; Gandolfi, M. T.; Moggi, L. J.
Am. Chem. Soc. 1971, 93, 339.
(14) Kallmayer, H.-J.; Fritzen, W. Pharmazie 1994, 49, 412.
(15) Shi, M.; Yang, W.-G.; Wu, S. J. Photochem. Photobiol.
Chem. 2007, 185, 140.
(48) Allonas, X.; Ley, C.; Bibaut, C.; Jacques, P.; Fouassier, J. P.
Chem. Phys. Lett. 2000, 322, 483.
(16) Chen, Y.; Steinmetz, M. G. J. Org. Chem. 2006, 71, 6053.
(17) Chen, Y.; Steinmetz, M. G. Org. Lett. 2005, 7, 3729.
(18) Jones; Qian, X. J. Phys. Chem. A 1998, 102, 2555.
(49) Burget, D.; Jacques, P. J. Lumin. 1992, 54, 177.
(50) Kemp, D. R.; Porter, G. Proc. R. Soc. Lond. Math. Phys.
Eng. Sci. 1971, 326, 117.
(51) Görner, H. J. Photochem. Photobiol. Chem. 2004, 165, 215.
(19) Falci, K. J.; Franck, R. W.; Smith, G. P. J. Org. Chem. 1977,
42, 3317.
(52) Barbafina, A.; Elisei, F.; Latterini, L.; Milano, F.; Agostiano,
A.; Trotta, M. Photochem. Photobiol. Sci. 2008, 7, 973.
(20) Giles, R. G. F. Tetrahedron Lett. 1972, 13, 2253.
(53) Wagner, P. J. Acc. Chem. Res. 1971, 4, 168.
(21) Cameron, D. W.; Giles, R. G. F. J. Chem. Soc. C Org. 1968,
1461.
(54) Amouyal, E.; Bensasson, R.; Land, E. J. Photochem. Photo-
biol. 1974, 20, 415.
(22) Cameron, D. W.; Giles, R. G. F. Chem. Commun. Lond.
1965, 573.
(55) Garcia-Garibay, M. A.; Gamarnik, A.; Bise, R.; Pang, L.;
Jenks, W. S. J. Am. Chem. Soc. 1995, 117, 10264.
8
ACS Paragon Plus Environment

Page 9 of 10
Journal of the American Chemical Society
(56) Johnson, B. A.; Kleinman, M. H.; Turro, N. J.; Garcia-
Garibay, M. A. J. Org. Chem. 2002, 67, 6944.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
9
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 10
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Insert Table of Contents artwork here
10
ACS Paragon Plus Environment