2,6,7-Trithiabicyclo[2.2.2]octanes as promising photolabile tags for combinatorial encoding

Article abstract of DOI:10.1021/jo702091e

(Chemical Equation Presented) The adducts of trithiabicyclo[2.2.2]octane (TTBO) and carbonyl compounds undergo efficient photoinduced fragmentation with quantum yields comparable to that of dithiane adducts. The effect of the third sulfur on the stability of the respective radical cations and radicals is examined computationally and experimentally in a laser flash photolysis study. A straightforward synthetic approach to a variety of 4-substituted trithiabicyclo[2.2.2]octanes from 3-bromo-2,2-bis-(bromomethyl)propanol is developed, making a diverse set of mass-differentiated photolabile tags readily available for combinatorial encoding.

Full text of DOI:10.1021/jo702091e

SCHEME 1
2,6,7-Trithiabicyclo[2.2.2]octanes as Promising
Photolabile Tags for Combinatorial Encoding
Roman A. Valiulin and Andrei G. Kutateladze*
Department of Chemistry and Biochemistry, UniVersity of
DenVer, DenVer, Colorado 80208
akutatel@du.edu
ReceiVed September 24, 2007
SCHEME 2
accessible via a simple and straightforward synthetic approach.
We have demonstrated that 2-alkyl-1,3-dithianes satisfy these
requirements and can be used for encoding as binary digits.
Low oxidation potential of dithianes, the very property which
aids the photoinduced electron-transfer fragmentation, makes
them also amenable for high-sensitivity MS detection with
electron impact ionization. Individual dithianes can be detected
with a greater than 10:1 signal-to-noise ratio at a subpicomolar
level using a standard mass-produced HP GCMS.
As a large set of tags may be needed for the encoding of
high-dimensional combinatorial libraries, we systematically
sought other di- or trisulfides that can be utilized as photoactive
molecular encoding tags. In this note, we report that a diverse
auxiliary set of mass-differentiated 2,6,7-trithiabicyclo[2.2.2]-
octanes (TTBO) can be readily synthesized and used for
combinatorial encoding. The photophysical properties of the
synthesized TTBO adducts with carbonyl compounds were
investigated experimentally and computationally.
The adducts of trithiabicyclo[2.2.2]octane (TTBO) and
carbonyl compounds undergo efficient photoinduced frag-
mentation with quantum yields comparable to that of dithiane
adducts. The effect of the third sulfur on the stability of the
respective radical cations and radicals is examined compu-
tationally and experimentally in a laser flash photolysis study.
A straightforward synthetic approach to a variety of 4-sub-
stituted trithiabicyclo[2.2.2]octanes from 3-bromo-2,2-bis-
(bromomethyl)propanol is developed, making a diverse set
of mass-differentiated photolabile tags readily available for
combinatorial encoding.
We have been developing photolabile systems capable of
detecting molecular recognition events. The methodology is
based on sensitized photofragmentations in R-hydroxyalkyl-
dithianes,1 which can be made contingent on an external
molecular recognition event and therefore can be used to detect
and report such events at a molecular level. Most recently, we
have adopted this methodology to the encoding and screening
of solution-phase combinatorial libraries, either unsupported or
immobilized on nanosized carriers too small for mechanical
manipulations.2 The encoding dithiane-based tags are released
into solution only when a binding event occurs between the
ligand and the receptor, outfitted with an electron-transfer
sensitizer. The released tags are detected by ubiquitous analytical
techniques, such as GCMS, and used as binary digits revealing
the identity of the lead ligand or narrowing the range of potential
leads. The following considerations were taken into account in
designing the tags: the tags should (i) be amendable for
detection at very low concentrations by ubiquitous analytical
techniques; (ii) not possess any functional groups implicated
in biological interactions; (iii) not interfere with synthetic steps;
(iv) be easy to separate from the aqueous screening environment,
for example, due to significant hydrophobicity; and (v) be
4-Methyl-TTBO (1) was first described in 1955 by Doering.3
A decade later, as a part of the umpolung strategy development,
reaction of lithiated TTBO with electrophiles was described by
Seebach.4 4-Methyl-TTBO (pKDa MSO 30.5) is over 8 orders of
magnitude more acidic than 1,3-dithiane (pKDa MSO ∼39),5
which makes it easier to deprotonate for subsequent reactions
with electrophiles.
We found that a diverse set of 4-alkoxymethyl TTBOs can
be synthesized starting from inexpensive commercially available
pentaerythritol tribromide as shown in Scheme 1. TTBOs were
then lithiated and reacted with benzaldehyde or benzophenone
to give photoactive adducts 3-6 (Scheme 2).
The most likely mechanism of photoinduced fragmentation
in adducts 3-6 in the presence of benzophenone as an electron-
(3) Doering, W. v. E.; Levy, L. K. J. Am. Chem. Soc. 1955, 77, 509-
513.
(4) Seebach, D.; Geiss, K. H.; Beck, A. K.; Graf, B.; Daum, H. Chem.
Ber. 1972, 105, 3280-3300.
(5) (a) Bordwell, F. G.; Bares, J. E.; Bartmess, J. E.; Drucker, G. E.;
Gerhold, J.; McCollum, G. J.; Van, der Puy, M.; Vanier, N. R.; Matthews,
W. S. J. Org. Chem. 1977, 42, 326-332. (b) Bordwell, F. G.; Drucker, G.
E.; Andersen, N. H.; Denniston, A. D. J. Am. Chem. Soc. 1986, 108, 7310-
7313.
(1) McHale, W. A.; Kutateladze, A. G. J. Org. Chem. 1998, 63, 9924-
9931.
(2) Kottani, R.; Valiulin, R. A.; Kutateladze, A. G. Proc. Natl. Acad.
Sci. U.S.A. 2006, 103, 13917-13921.
10.1021/jo702091e CCC: $40.75 © 2008 American Chemical Society
Published on Web 12/11/2007
J. Org. Chem. 2008, 73, 335-338
335

TABLE 2. Electron-Transfer Rate Constants for TTBO or
Methyldithiane Adducts of Benzophenone and Benzaldehyde in
MeCN
SCHEME 3
transfer sensitizer is shown in Scheme 3. It involves photoin-
duced electron transfer from the trithiabicyclo[2.2.2]octane
moiety to excited triplet benzophenone, followed by mesolytic
fragmentation in the generated radical cation. The fragmentation
is assisted by the benzophenone radical anion deprotonating the
hydroxy group in radical cation B. In the fragmentation
mechanism for the similar dithiane-based adducts, we found
evidence6 that such deprotonation is coupled with intramolecular
electron transfer, furnishing the oxygen centered radical of type
D, which fragments spontaneously, although the fragmentation
concerted with deprotonation or the involvement of the charge
separated species C was not completely ruled out.
accessible and also is expected to produce a more stable radical
cation, improving the driving force for the forward electron
transfer and rendering the BET less likely. Both the improved
quantum yields for the benzophenone adducts and the faster
rate of initial electron-transfer quenching seem to corroborate
this conclusion. In the benzaldehyde adducts, the fragmentation
channel appears to be more sensitive to the lower stability of
the TTBO radical.
TABLE 3.a
TABLE 1. Relative Quantum Yields of Fragmentation
CD3CN
CD2Cl2
Benzophenone adducts
5
6c
6f
1.69
1.01
1.53
(1.0)
1.83
-
-
benzophenone-methyldithiane
(1.0)
Benzaldehyde adducts
3
0.64
0.69
0.49
0.39
-
4a
4c
4e
0.61
-
-
(1.0)
benzaldehyde-methyldithiane
(1.0)
As the primary focus of this study was to develop an auxiliary
set of encoding tags other than dithiane adducts, we compared
the photophysical properties of adducts 3-6 with those of
dithiane adducts. As follows from Table 1, the overall quantum
yields of fragmentation for TTBO-benzophenone adducts are
improved over dithianes, whereas their benzaldehyde adducts
fragmented with somewhat lesser relative efficiency.
Our nanosecond laser flash photolysis (LFP) experiments
showed that the rate of initial electron-transfer quenching of
the sensitizer’s triplet state was also faster for the TTBO adducts
(Table 2). In the sterically hindered benzophenone adduct, the
partitioning of the ion radical pair B between the productive
fragmentation channel and the wasteful back electron transfer
(BET) primarily controls the quantum yield of fragmentation.
The TTBO moiety, which contains three sulfur atoms, is more
a Zpe (zero point energy)-corrected b3lyp/6-311+g(2d,p)//b3lyp/6-
311+g(2d,p) energies (except for the vertical IP, which is not corrected).
b Radical resonance stabilization energy. c Calcd IP is 0.36 eV lower than
the reported experimental value (ref12); the relative values are more
significant.
DFT computations at the B3LYP/6-311+G(2d,p) level show
that the gas-phase vertical ionization potential of the parent
4-methyl-TTBO is almost 0.5 eV lower than dithiane’s (Table
3). A simple MO picture helps understand this in semi-
qualitative terms. The radical cation in dithianes is stabilized
due to the formation of 3-electron-2-center bond between two
sulfurs.7 The C3 symmetric TTBO framework offers additional
stabilization in a form of cyclic array of atomic orbitals which
(7) For a review on the 2-center-3-electron sulfur-sulfur bonds, see:
Asmus, K.-D. Acc. Chem. Res. 1979, 12, 436-442.
(6) Li, Z.; Kutateladze, A. G. J. Org. Chem. 2003, 68, 8236-8239.
336 J. Org. Chem., Vol. 73, No. 1, 2008

FIGURE 1. Typical LFP waveform at 520 nm (left) and a typical
concentration series for determination of the second-order electron-
transfer rate constant, kET (right), for the benzophenone adduct 5.
FIGURE 4. DFT, B3LYP/6-311+G(2d,p), potential energy surface
of C3-symmetric (HS)3C• radical with a global minimum at R ) 98.1°
and γ ) 51.3°.
FIGURE 2. Mo¨bius cyclopropenyl topology.
theory based on Pasto’s9 isodesmic reaction R-H + CH3• f
R• + CH4 (Table 3). The RSE for E (7.9 kcal/mol) is a factor
of 2 smaller, which can be rationalized in terms of unfavorable
overlap between sulfurs’ p lone pairs and the singly occupied
orbital on the pyramidalized carbon radical center.
Zpe-corrected DFT activation energy for the fragmentation
of the oxy-radical D (benzophenone adduct) is 4.8 kcal/mol.
Such DFT barrier for fragmentation in benzophenone-eth-
yldithiane is much smaller, 1.7 kcal/mol for the most favorable
conformation (experimental ∆Hq ) 1.5 kcal/mol).10 Due to
decreased resonance stabilization of radical E, fragmentation
of the oxygen centered radical D has a late transition state, with
the stretched CC bond length of 1.927 Å, compared to only
1.745 Å for the corresponding methyl-dithiane adduct and 1.735
Å for the ethyldithiane adduct.
To better understand the effect of geometry on stabilization
of the radical center in tris-alkylthiomethyl radicals we have
scanned the potential energy surface for the parent tris-
mercaptomethyl radical, (HS)3C•, as a function of two vari-
ables: (i) the pyramidalization angle R, defined as the angle
between the singly occupied orbital on carbon and the C-S
bond, that is, S-C- • and (ii) the torsional dihedral γ, defined
as H-S-C- • . The rest of internal variables were optimized
without constraints at the B3LYP/6-311+G(2d,p) level of theory
(Figure 4).
The radical’s relaxed geometry is more planar with only a
slight pyramidalization (R ) 98.1°). The SH bonds, however,
deviate from the SSS plane, with torsional dihedral γ ) 51.3°.
In the TTBO radical R and γ are constrained by the bicyclic
framework to 103.9 and 166.5°, respectively. Although this
pyramidalization of 103.9° is less pronounced than in the parent
hydrocarbon for which R ) 107.1°, the geometry constraints
decrease the radical resonance stabilization energy for TTBO
radical by 10 kcal/mol compared with dithianyl radicals (Table
3).
FIGURE 3. TTBO: (a) C3-symmetric (twisted) ground state and its
HOMO (b) C3V-symmetric radical cation and its SOMO.
can be loosely referred to as a 3-center-5-electron bond. In the
neutral precursor, the three p lone pairs are considerably
destabilized by repulsion. A more helpful way of thinking about
this is to apply Zimmerman’s circular mnemonic8 to the
6-electron Mo¨bius antiaromatic arrangement of sulfurs’ p lone
pairs in TTBO (i.e., Mo¨bius cyclopropenyl topology), Figure
2. Removal of an electron from this system relieves the
unfavorable orbital interaction, leading to lower ionization
potential.
This rationale finds support in DFT computations with an
extended triple-ú basis set, 6-311+G(2d,p). The neutral ground
state TTBO has a twisted C3 symmetry, where the lone pairs
repulsion is somewhat minimized by a screw-type distortion
from the C3V structure. On the contrary, the radical cation of
TTBO has a perfect C3V geometry, which is a true minimum as
shown by vibrational analysis. Its antisymmetric singly occupied
HOMO (i.e., SOMO) is shown in Figure 3b.
Although the quantum yields of fragmentation generally
correlate with the stability of radical cations, there is no
correlation with the stability of the resulting radical (of type E,
Scheme 3). We calculated the radical resonance stabilization
energies (RSE) for unsubstituted and substituted dithian-2-yl
radicals and radical E at the B3LYP/6-311+G(2d,p) level of
(9) (a) Pasto, D. J.; Krasnansky, R.; Zercher, C. J. Org. Chem. 1987,
52, 3062-3072. (b) Pasto, D. J. J. Am. Chem. Soc. 1988, 110, 8164-8175.
(10) Gustafson, T. P.; Kurchan, A. N.; Kutateladze, A. G. Tetrahedron
2006, 62, 6574-6580.
(8) Zimmerman, H. E. J. Am. Chem. Soc. 1966, 88, 1564-1565. (b)
Zimmerman, H. E. Acc. Chem. Res. 1971, 4, 272-280.
J. Org. Chem, Vol. 73, No. 1, 2008 337

(C) 4-Alkoxymethyl-2,6,7-trithiabicyclo[2.2.2]octane. The solu-
tion of 2,2-bis(mercaptomethyl)-3-alkoxypropanethiol (from step
B) and methyl orthoformate in toluene with several crystals of
TsOH•H2O was refluxed for 24 h, washed with saturated NaHCO3
and water, and dried over Na2SO4. The solvent was removed under
reduced pressure, and the resulting yellow solid was purified by
column chromatography (silica gel, hexane/EtOAc, 20:1).
However, it is not surprising that the relatively low stability
of TTBO radical (E) does not have as large a detrimental effect
on the efficiency of fragmentation as does the back electron
transfer. Enhanced stability of the TTBO radical cations seems
to play a critical role in the fragmentation mechanism. Ad-
ditionally, the overall effect on the quantum efficiency of
fragmentation is mitigated by yet another factor: the acces-
sibility of sulfurs for the sensitizer approach, which enhances
the rates of the initial electron-transfer quenching, kET, of triplet
benzophenone. The LFP experiments consistently show that kET
for the TTBO adducts are greater than for dithiane adducts.
These rates also correlate (i.e., are attenuated) with the steric
crowding introduced by the carbonyl component: the adducts
with benzaldehyde quench the triplet sensitizer with a 40-50%
higher rate than the more crowded adducts of benzophenone
(Table 2).
1-Bromo-2,2-bis(bromomethyl)-3-methoxypropane: Prepared
from 5.13 g (91.5 mmol) of KOH, 7.00 g (20.8 mmol) of 3-bromo-
2,2-bis(bromomethyl)propanol and 13.00 g (91.5 mmol) of io-
domethane in 35 mL of DMSO: 5.49 g, 78%.
2,2-Bis(mercaptomethyl)-3-methoxypropanethiol: Prepared
from 3.23 g (57.7 mmol) of NaSH‚9H2O, 2.6 g (81.3 mmol) of S
powder, 4.88 g (14.4 mmol) of 1-bromo-2,2-bis(bromo-methyl)-
3-methoxypropane, 2.5 g (39.3 mmol) of Cu powder, 13.7 g of Zn
1
dust (216 mmol) and 70 mL of conc. HCl. H NMR (500 MHz,
CDCl3) δ 3.33 (s, 3H), 3.29 (s, 2H), 2.62 (d, 6H, J ) 8.9 Hz), 1.22
The features discussed above account for the superior
quantum yields of sensitized fragmentation in benzophenone
adducts of TTBOs compared with benzophenone adduct of
methyldithiane. Although the benzaldehyde adducts of TTBOs
showed somewhat lesser quantum efficiencies, when compared
with benzaldehyde adduct of methyldithiane (Table 1), adducts
4 with alkoxy substituents were found to be adequate for tagging
as they cleanly produce TTBO tags upon sensitized irradiation.
Only allylic compound 4e produced traces of side product,
which is conceivably the result of Paterno`-Bu¨chi cycloaddition
of the sensitizer to the double bond.
Similar to dithianes, TTBOs 1 and 2 can be detected with
excellent sensitivity using mass-produced analytical instruments,
for example, GCMS. The enhanced stability of their radical
cations accounts for their low detection limit.
In conclusion, 4-substituted trithiabicyclo[2.2.2]octanes are
versatile and readily accessible mass-differentiated tags for the
encoding of combinatorial libraries, which can be photoreleased
with high quantum efficiencies and detected in minute amounts
by ubiquitous analytical techniques.
(t, 3H, J ) 8.9 Hz).
4-Methoxymethyl-2,6,7-trithiabicyclo[2.2.2]octane (2a): Pre-
pared from 2.86 g (14.5 mmol) of 2,2-bismercaptomethyl-3-
methoxypropane-1-thiol, 2.3 g (21.7 mmol) of methyl orthofor-
1
mate: 2.77 g (92%). H NMR (500 MHz, CDCl3) δ 4.60 (s, 1H),
3.37 (s, 3H), 3.19 (s, 2H), 3.05 (s, 6H). 13C NMR (500 MHz,
CDCl3) δ 80.55, 59.43, 39.04, 33.29,30.68.
General Procedure for Preparation of Adducts 3-6. n-BuLi
(1.6 M in hexane) was added to TTBO in 5-10 mL of THF under
nitrogen and stirred for 10-15 min at 25 °C, at which point a 2
mL solution of benzaldehyde or benzo-phenone in THF was added
and stirred for 3 h, quenched with aqueous ammonium chloride,
extracted with CH2Cl2 (2 × 20 mL), and dried over Na2SO4. The
solvent was removed under reduced pressure, and the crude material
was purified by column chromatography on silica gel (hexane/
EtOAc).
Adduct 4a: Prepared from 100 mg (0.48 mmol) of 4-methoxy-
methyl-2,6,7-trithiabicyclo[2.2.2]octane (2a), 0.66 mL (1.06 mmol)
of n-BuLi (1.6 M in hexane), and 76 mg (0.72 mmol) of
benzaldehyde: 47 mg (31%) after purification on silica gel, hexane/
1
EtOAc ) 10:1, then 3:1. H NMR (500 MHz, CDCl3) δ 7.57 (m,
Experimental Section
2H), 7.38 (m, 3H), 5.00 (d, 1H, J ) 3.1 Hz), 3.34 (s, 3H), 3.18
(dd, 2H, J ) 14.9. Hz, J ) 9.01 Hz), 3.05 (dd, 3H, J ) 11.7 Hz,
J ) 2.0 Hz), 2.95 (d, 1H, J ) 3.3 Hz), 2.93 (dd, 3H, J ) 11.7 Hz,
J ) 2.0 Hz). 13C NMR (400 MHz, CDCl3) δ 136.87, 129.00,
128.32, 127.81, 79.83, 78.27, 65.90, 59.39, 34.76, 32.52.
General Procedure for Preparation of TTBOs. (A) 1-Bromo-
2,2-bis(bromomethyl)-3-alkoxypropanes. Fine-powdered KOH
was added to a vigorously stirred solution of 3-bromo-2,2-bis-
(bromomethyl)propanol and an alkyl halide in 35 mL of DMSO.
The temperature was kept below 60 °C using an ice bath. After
the initial exothermic step, the stirred solution was heated at 60 °C
for 2 h and cooled to 20 °C; 100 mL of water was added, and the
aqueous phase was extracted with 3 × 50 mL of CH2Cl2. The
organic layer was dried over Na2SO4 followed by the removal of
the solvent.
Computational Methods. The structures and the energies were
computed with Gaussian 03, Rev. C02, computational package.11
The values in Table 3 and elsewhere in the text represent zpe-
corrected B3LYP/6-311+G(2d,p) energies, obtained for geometries
optimized at the same theory level.
(B) 2,2-Bis(mercaptomethyl)-3-alkoxypropanethiol. NaHS‚
9H2O and elementary sulfur powder were dissolved in DMF upon
warming, heated at 80 °C for 1 h, and cooled to ∼ 40 °C, and then
the 1-bromo-2,2-bis(bromomethyl)-3-alkoxypropane (from step A)
was added. The temperature was raised to 80-90 °C and kept there
for 7-8 h. DMF was removed under reduced pressure, and copper
powder and toluene (30 mL) were added to the resulting crude
product and refluxed overnight. Black copper sulfide was filtered
off, then zinc dust and 20-30 mL of ethanol were added to the
yellow filtrate. Concentrated HCl was added dropwise under stirring.
The resulting solution was stirred for 24 h under nitrogen; the
toluene layer was separated from the ethanol-water mixture,
washed with water (20 mL), and dried over Na2SO4. The solvent
was removed under reduced pressure to furnish oil with a
characteristic smell, which was used without further purification.
Acknowledgment. This work is supported by NSF, CHE-
640838.
Supporting Information Available: Additional experimental
information, spectra, computational details. This material is available
free of charge via the Internet at http://pubs.acs.org.
JO702091E
(11) Frisch, M. J.; et al. Gaussian 03, revision C.02; Gaussian, Inc.:
Pittsburgh, PA, 2003.
(12) Sweigart, D. A.; Turner, D. W. J. Am. Chem. Soc., 1972, 94, 5599-
5603.
338 J. Org. Chem., Vol. 73, No. 1, 2008