Hypericin, hypocrellin, and model compounds: Primary photoprocesses of light-induced antiviral agents

Article abstract of DOI:10.1021/jp963046g

The excited-state photophysics of the light-induced antiviral agents hypericin and hypocrellin are compared with those of the hexa- and tetramethoxy analogues of hypericin. The results are consistent with the interpretation of the primary photoprocess in

Full text of DOI:10.1021/jp963046g

J. Phys. Chem. A 1997, 101, 3235-3240
3235
Hypericin, Hypocrellin, and Model Compounds: Primary Photoprocesses of Light-Induced
Antiviral Agents
D. S. English, K. Das, J. M. Zenner, W. Zhang, G. A. Kraus, R. C. Larock, and J. W. Petrich*
Department of Chemistry, Iowa State UniVersity, Ames, Iowa 50011
ReceiVed: October 3, 1996; In Final Form: February 3, 1997X
The excited-state photophysics of the light-induced antiviral agents hypericin and hypocrellin are compared
with those of the hexa- and tetramethoxy analogues of hypericin. The results are consistent with the
interpretation of the primary photoprocess in hypericin and hypocrellin as that of excited-state intramolecular
proton or atom transfer.
Introduction
structure of the tautomeric species (parts a and b of Figure 1)
in such a way as to destroy the mirror-image symmetry. High-
level quantum chemical calculations will have much to offer in
understanding this problem. It is also possible, as we have
argued elsewhere,8,9,15 that the ground state of hypericin is
already partially tautomerized and that this ground-state het-
erogeneity yields the observed mirror-image symmetry between
absorption and emission spectra. Another strategy is to study
the excited-state photophysics of synthetic analogues that are
unable to execute excited-state proton or atom transfer. We
have already begun such an examination by comparing the
photophysics of hypericin with that of its hexamethoxy analogue
(Figure 1e).15 Here, we pursue this line of investigation more
thoroughly by examining the tetramethoxy analogue of hypericin
(Figure 1f), which can be considered a methoxy hybrid of
hypericin and hypocrellin.
Hypericin and hypocrellin are naturally occurring pigments1,2
that are remarkable because of their light-induced antiviral
activity3 against enveloped lentiviruses such as the human
immunodeficiency virus.4,5 Although the requirement of light
for this activity in hypericin and hypocrellin is absolute, these
two molecules possess very different modes of reactivity.6
Consequently, we have undertaken the task6-15 of unraveling
the excited-state primary photophysical processes of hypericin
and hypocrellin.
It has been our thesis from the very first that a significant
nonradiative process in hypericin and its analogues is intramo-
lecular proton (or atom) transfer.7-9 The argument for such a
process is the following. The hypericin analogue lacking labile
protons, mesonaphthobianthrone (Figure 1g), is significantly
fluorescent and has optical spectra that resemble those of
hypericin only when its carbonyl groups are protonated. In
hypericin, the fluorescent state grows in on a time scale of
seVeral picoseconds, as measured by the rise time of stimulated
emission. Therefore, the combined observations of the require-
ment of protonated carbonyls for strong hypericin-like fluores-
cence and the rise time of fluorescence in hypericin were taken
as evidence for intramolecular excited-state proton transfer in
hypericin. The possible role of labile protons in the light-
induced antiviral activity of hypericin and its analogues is
discussed in detail elsewhere.6,10,11,13 In this article, we shall
frequently refer to the process in question as a proton transfer,
but it should be kept in mind that not enough data are available
to determine whether a proton or a hydrogen atom is being
transferred.
Experimental Section
A. Synthesis of the Tetramethoxy Analogue.28 Di(3,5-
dimethoxy-2-methoxycarbonylphenyl)acetylene. The compound
was obtained as a white solid from the palladium-catalyzed
reaction of methyl 2-ethynyl-3,5-dimethoxybenzoate and methyl
2-(trifluoromethanesulfonyloxy)-3,5-dimethoxybenzoate17 (mp
149 °C). 1H NMR (CDCl3): δ 3.82 (s, 12H), 3.92 (s, 6H),
6.48 (d, 2H, J ) 0.9 Hz), 6.62 (d, 2H, J ) 0.9 Hz).
9,10-Bis[(2′-methoxycarbonyl-3′,5′-dimethoxy)phenyl]phenan-
threne. To a 2 dram vial equipped with a stir bar is added Pd-
(OAc)2 (0.125 mmol), NaOAc (0.50 mmol), LiCl (0.25 mmol),
2-iodobiphenyl (0.25 mmol), di(3,5-dimethoxy-2-methoxycar-
bonylphenyl)acetylene (0.275 mmol), and 5 mL of DMF. The
vial is then flushed with N2, capped with a screw cap containing
a Teflon liner, and placed in an oil bath at 100 °C for 48 h. The
vial is then removed from the oil bath, diluted with Et2O and
EtOAc, washed with saturated NH4Cl and water, dried over
MgSO4, and concentrated. This residue was then purified by
column chromatography using 4:1 hexanes/EtOAc (49% yield,
mp 210-212 °C (toluene)). 1H NMR (CDCl3): δ 3.18 (s, 6H),
3.63 (s, 6H), 3.83 (s, 6H), 6.39 (d, 2H, J ) 3.0 Hz), 6.81 (d,
2H, J ) 3.0 Hz), 7.50 (dt, 2H, J ) 7.8, 0.9 Hz), 7.62 (m, 4H),
8.70 (d, 2H, J ) 7.8 Hz). 13C NMR (CDCl3): δ 51.5, 55.6,
55.9, 98.5, 106.9, 117.2, 122.3, 126.3, 126.5, 127.8, 129.8,
130.5, 134.8, 141.0, 157.7, 161.5, 167.9. IR (CDCl3): 3010,
1717, 1103, 1047 cm-1. HRMS for C34H30O8: calcd 566.1941;
found 566.1952.
Two potential arguments against intramolecular excited-state
proton transfer in hypericin are the lack of an isotope effect for
the process in question8,9 and the observation of near mirror-
image symmetry between its absorption and emission spectra
(Figure 2a). The first of these points is discussed in detail in
the companion article. As for the mirror-image symmetry, this
is typically taken as a signature of negligible structural changes
between the absorbing and the emitting species. Intramolecular
excited-state proton transfer usually generates a broad, shifted,
and structureless emission spectrum. A classic example of such
behavior is given by 3-hydroxyflavone.16
The issue of the shape of the spectra may be addressed in
several ways. It is possible that the structural changes induced
by proton transfer do not significantly affect the electronic
2,4,2′,4′-Tetramethoxynaphthodianthrone (“Tetramethoxy-
hypericin”, Figure 1f). This compound was obtained as an
orange-brown solid in 91% yield by treatment of 9,10-bis(2′-
* To whom correspondence should be addressed.
X Abstract published in AdVance ACS Abstracts, April 1, 1997.
S1089-5639(96)03046-0 CCC: $14.00 © 1997 American Chemical Society

3236 J. Phys. Chem. A, Vol. 101, No. 18, 1997
English et al.
TABLE 1: Fluorescence Parameters in DMSO
compound
Φf
τf (ps)
hypericin
hypocrellin
hexamethoxyhypericin
tetramethoxyhypericin
0.35 ( 0.02
0.014 ( 0.008a
0.07 ( 0.01c
0.11 ( 0.01c
5500 ( 500
1300
490b
820b
a In all other solvents investigated, the quantum yield was 0.07 (
0.02. b This is the major component of the fluorescence decay and
comprised at least 95% of the decay intensity. See Figure 7. c ) The
steady-state fluorescence quantum yields provide upper limits on the
true value because they are contaminated by the few percent of long-
lived fluorescent impurity in our samples. A more accurate comparison
of integrated fluorescence intensity could in principle be given by the
ratio of the fluorescence lifetimes in the table, assuming the radiative
rate is the same for all the compounds being compared. Such an
assumption is probably not warranted given the anomalously low value
for the quantum yield of hypocrellin in DMSO as opposed to other
solvents.
TABLE 2: Global Fitting Parameters for
Tetramethoxyhypericin
probe
wavelengtha
(nm)
a1
τ1 (ps)
a2
τ2 (ps)
a3
τ3 (ps)
500c
520d
530d
540b
550b
560b
570b
620b
630b
0.23
0
0
0
0
0.10
0.25
0.25
0.16
9.9
9.9
9.9
9.9
9.9
9.9
9.9
9.9
9.9
0
0.35
0.31
0
0
0
0
0
0
820
820
820
820
820
820
820
820
820
0.77
0.65
0.69
1.0
∞
∞
∞
∞
∞
∞
∞
∞
∞
1.0
0.90
0.75
0.75
0.84
a The functional form used to fit the data is indicated by a superscript
in the leftmost column, which gives the probe wavelength. The
preexponential factors are all normalized and should not be interpreted
in terms of absolute optical density changes owing to changes in pump
intensities and gain settings on data acquisition hardware. b ∆A(t) )
a1 exp(-t/τ1) + a2 exp(-t/τ2) + a3 exp(-t/τ3). c ∆A(t) ) a1[1 - exp(-
t/τ1) + a2 exp(-t/τ2) + a3 exp(-t/τ3). d ∆A(t) ) a2[1- exp(-t/τ2) +
a3 exp(-t/τ3).
at >98% purity as determined from the supplied TLC and NMR
measurements. The hexamethoxy analogue was prepared as
described elsewhere.15 DMSO was dried over 4 Å molecular
sieves. Care was taken to prevent sieve dust contamination.
All experiments were performed at ambient temperature (22 °C).
The time-correlated single-photon counting experiments were
performed with an apparatus described elsewhere.20 The
instrument response function is typically about 70 ps. Samples
were prepared for pump-probe experiments with an optical
density of 0.4-0.7 at the pump wavelength; for time-correlated
single-photon counting, samples were diluted by 10-fold or
more. The time-resolved fluorescence data provide a powerful
assay of the purity of the tetramethoxy analogue. In addition
to a 820 ps decay component, there is a small amount (∼5%)
of a ∼2 ns component.
Figure 1. Two-dimensional structures of (a) “normal” form of
hypericin, (b) hypericin bitautomer, (c) hypocrellin, (d) hypocrellin
tautomer, (e) hexamethoxyhypericin, (f) tetramethoxyhypericin, (g)
mesonaphthobianthrone. We have suggested, and our results are
consistent with the idea, that hypericin and hypocrellin exist in more
than one tautomeric form in the ground state. We have proposed,
however, that the hypericin bitautomer (b) is thermodynamically
unstable in the ground state. Thus, any excited-state hypericin bitau-
tomer that returns to the ground state will not accumulate, but will
rapidly form a lower energy tautomeric form.
methoxycarbonyl-3′,5′-dimethoxyphenyl)phenanthrene with PPA18
and then concentrated H2SO4 and sunlight.18 1H NMR
(HCCl3): δ 4.18 (s, 6H), 4.26 (s, 6H), 6.94 (s, 2H), 7.90 (t,
2H, J ) 7.8 Hz), 8.77 (d, 2H, J ) 7.5 Hz), 8.93 (d, 2H, J ) 8.1
Hz). The HRMS for C32H20O6 was not obtained because of
limitations in the HRMS ionization techniques. Electrospray
low-resolution mass spectroscopy found 500.8 (m + 1). UV-
vis absorption λmax (DMSO): 398, 419, 475, 503 nm. UV-
vis fluorescence excitation λmax (DMSO): 396, 418, 474, 502
nm. UV-vis emission λmax (DMSO): 531, 559 nm. The
fluorescence excitation spectrum is superimposable on the
absorption spectrum. Further characterization was not possible
owing to the low solubility of this product.
Steady-state fluorescent measurements were made using a
Spex Fluoromax with a 4 nm band-pass, and corrections were
made for detector response and excitation lamp spectrum.
Steady-state absorbance measurements were made using a
Perkin-Elmer Lambda 18 double beam UV-vis spectropho-
tometer with 1 nm band-pass. Fluorescence quantum yield
measurements were made relative to hypericin by exciting
solutions of each compound at several wavelengths where they
had equal optical densities and integrating the emission spectrum
over the entire band on a wavenumber scale.
B. Optical Measurements. Hypericin was used as received
from Carl Roth GmbH (distributed by Atomergic Chemetals
Corp.). Hypocrellin A (Molecular Probes) was used as received
Pump-probe transient absorption measurements were per-
formed with laser systems producing pulses of 150-200 fs

Hypericin, Hypocrellin, and Model Compounds
J. Phys. Chem. A, Vol. 101, No. 18, 1997 3237
Figure 3. Absorption transients of hypericin in DMSO, λpump ) 415
nm; probe wavelengths are indicated in the panels. The “spike” at zero
time that is most apparent in the data in this figure is a result of cross-
phase modulation of the pump pulse with the probe pulse.21 The time
constant obtained by globally fitting these and other data is 11.6 ps.15
duration at 2 kHz15 or of 1-3 ps duration at 30 Hz.14 Sample
preparation for the tetra and hexamethoxy analogue samples15
required bubbling with argon and filling the spinning sample
holder in a glovebag of argon to reduce oxygen content. This
was done after observing rapid degradation of the hexamethoxy
sample in the presence of oxygen. Deoxygenation sufficiently
preserved the samples to withstand pump incidence for several
hours. Hypericin is stable for hours without deoxygenation and
gives the same experimental results in either case.
Kinetic traces acquired at various wavelengths were fit using
a global fitting procedure described in detail elsewhere.15 The
time-resolved absorption/stimulated emission data are displayed
in Figures 3-6. The results of the global fits, and the functional
forms used to obtain them, are summarized in Table 2. The
“spike” that appears at zero time in some of the data arises from
the more intense pump pulse modulating the phase of the less
intense probe pulse.21
Results
A. Steady-State Spectra. The steady-state absorption and
emission spectra of hypericin, hexamethoxyhypericin, hypo-
crellin, and tetramethoxyhypericin are presented in Figure 2.
The spectra of hypericin differ significantly from those of the
other three analogues. The absorbance maximum of hypericin
in DMSO arising from the S0 f S1 transition occurs at 598
nm. Hypericin is unlike the other three compounds investigated
Figure 2. Normalized absorption (-) and emission (- - -) spectra in
DMSO of (a) hypericin (λex ) 550 nm), (b) hexamethoxyhypericin
(λex ) 410 nm), (c) hypocrellin (λex ) 400 nm), and (d) tetramethoxy-
hypericin (λex ) 412 nm). The extinction coefficients for hypericin8,25
and hypocrellin27 are taken or estimated from the literature. The arrows
indicate the wavelengths at which pump-probe transient absorption
experiments were performed.

3238 J. Phys. Chem. A, Vol. 101, No. 18, 1997
English et al.
Figure 4. Absorption transients of hexamethoxy hypericin in DMSO,
λpump ) 415 nm; probe wavelengths are indicated in the panels. The
full scale is 320 ps. The full scale of the inset is 40 ps. The time constant
obtained by globally fitting these and other data is 2.5 ps.15
Figure 5. Absorption transients of hypocrellin in DMSO, λpump ) 588
nm; probe wavelengths are indicated in the panels. The full scale is
200 ps. Comparable results are obtained using a pump wavelength at
415 nm. The time constant obtained by globally fitting these and other
data is 90 ps.26 For λprobe ) 550 nm, ∆A(t) ) 0.68[1 - exp(-t/90 ps)]
+ 0.32 exp(-t/∞). For λprobe ) 595 nm, ∆A(t) ) -0.99 exp(-t/90 ps)
- 0.01 exp(-t/∞).
here in that it lacks an intense transition in the 400-500 nm
region, which is most likely due to an S0 f Sn transition. This
band is almost as intense as that at 530 nm for the hexamethoxy
analogue. In hypocrellin and tetramethoxy hypericin, this band
contains the absorbance maximum. These data indicate that
methylation of the hydroxyl groups adjacent to the carbonyls
(or their removal) provides relative enhancement of the oscillator
strength at higher energies. Consequently, these data suggest
that interaction of the enol proton with the keto oxygen strongly
determines the characteristics of the absorption spectrum. The
methoxy analogues are also less fluorescent than their coun-
terparts. This is consistent with our earlier proposals concerning
the role of a proton interacting with the keto oxygen for
establishing hypericin-like fluorescence.8,9 Note that in Figure
2 we have normalized the emission spectra so that their maxima
are identical with those of the peaks (or the shoulder, in the
case of hypocrellin) of the lowest energy absorption features.
This is done to indicate the mirror symmetry between the
absorption and emission spectra. Table 1 provides a summary
of fluorescence data.
even more striking. The tetramethoxy analogue exhibits no
stimulated emission and possesses a time constant that is 9 times
smaller than that observed in hypocrellin.
In all the data presented there is a component that does not
decay on the time scale of the experiment or that is characterized
by a long-lived time constant that agrees with the fluorescence
decay time (Table 1). Consequently, this long-lived component
is attributed to absorption by an excited-state singlet or formation
of triplet,23,24 depending on whether the signal is rising or de-
caying. Figure 7 presents the fluorescence decay of the
tetramethoxy analogue in DMSO. Its lifetime can be fit to a
time constant of 820 ps, which is consistent with the long-time
component extracted from the transient absorption data.
Discussion
B. Excited-State Kinetics. The excited-state kinetics of
hypericin, hexamethoxyhypericin, hypocrellin, and tetramethoxy-
hypericin are presented in Figures 3-6. The probe wavelengths
at which these kinetics were interrogated are indicated in the
steady-state spectra. Stimulated emission is observed in hy-
pericin but not in the hexamethoxy analogue. The difference
in the kinetics of the tetramethoxy analogue and hypocrellin is
We have argued in the Introduction and elsewhere8,9 that the
signature of proton transfer in hypericin is a rise time in
stimulated emission. Such a kinetic feature is absent in the
hexamethoxy analogue (Figure 1e), which cannot execute such
a nonradiative process. This absence consequently confirms
the presence of proton transfer in hypericin. A possible
objection to this argument is that the description of the traces

Hypericin, Hypocrellin, and Model Compounds
J. Phys. Chem. A, Vol. 101, No. 18, 1997 3239
Figure 7. Fluorescence decays obtained from time-correlated single-
photon counting at 22 °C. The topmost set of residuals corresponds to
hypericin. Curve a is for hexamethoxyhypericin in DMSO: λex ) 300
nm, λem > 530 nm; F(t) ) 0.97 exp(-t/490 ps) + 0.03 exp(-t/ 2.1
ns), ø2 ) 1.15. The full scale is 3 ns. Curve b is for tetramethoxyhy-
pericin in DMSO: λex ) 305 nm, λem > 530 nm; F(t) ) 0.95 exp(-
t/820 ps) + 0.05 exp(-t/2.0 ns), ø2 ) 1.15. The full scale is 3 ns.
Curve c is for hypocrellin in DMSO: λex ) 288 nm, λem > 455 nm;
F(t) ) 1.00 exp(-t/1.3 ns), ø2 ) 1.05. The full scale is 3 ns. Curve d
is for hypericin in DMSO: λex ) 305 nm, λem > 555 nm; F(t) ) 1.00
exp(-t/5.3 ns), ø2 ) 1.18. The full scale is 15 ns. To our knowledge,
the fluorescence decay of hypericin and its analogues does not depend
on excitation wavelength, visible and ultraviolet excitation yielding
similar results.8,25 This result should not be confused with the excitation
wavelength dependence of the picosecond transients in hypericin.15
These latter transients report on primary processes that are not resolvable
in the measurement of the longer lived fluorescence decay.
Figure 6. Absorption transients of tetramethoxyhypericin in DMSO,
λpump ) 415 nm; probe wavelengths are indicated in the panels. The
full scale is 320 ps. The fast component has a time constant of 9.9 ps.
The slowly rising transient apparent in the top panel has a time constant
of 820 ps, which matches the fluorescence decay of the analogue in
DMSO. The rise in absorbance at 500 nm has the same time constant
as the fast decaying component at 570 nm. A summary of absorption
data for all the probe wavelengths studied is given in Table 2.
The latter approach is taken in this article, the most striking
result of which is the difference in the excited-state kinetics of
hypocrellin and the tetramethoxy analogue. The tetramethoxy
analogue exhibits no stimulated emission and a time constant
that is 9 times smaller than that for hypocrellin (10 ps instead
of 90 ps). The absence of a ∼100 ps time constant in the
transient absorbance of the tetramethoxy analogue is consistent
with the assignment of this component to excited-state proton
or atom transfer in hypocrellin.
A possible origin of the 2.5 and 10 ps rapid transients in the
hexa- and tetramethoxy analogues is dynamic solvation of an
excited state (or internal conversion). As the solvent dynami-
cally readjusts itself to the change in the dipole of the solute
upon optical excitation, the solute is lowered in energy. This
is manifested in the transient absorbance. For example, in both
of the methoxy analogues, a decay of the induced absorption at
red wavelengths matches a rise in induced absorption at blue
wavelengths (Figures 3 and 6). These data are consistent with
a picture in which the lowest excited singlet is stabilized relative
to a higher-lying state to which it is optically coupled. The
dynamic solvation time of DMSO is known to be 3 ps.22 It
may, of course, be fortuitous that the observed transients are of
similar magnitude. Ideally, a correlation of decay time in
various solvents with the known dynamic solvation times would
enable us to make this assignment unambiguously. Unfortu-
nately, DMSO is the only solvent in which we are able to
dissolve the analogues in sufficient quantities to perform these
experiments and at the same time avoid sample degradation.
in Figure 3 is not unique. For example, instead of attributing
the form of the absorption transients to a rising component in
stimulated emission, they may equally well be described by a
component of stimulated emission that appears instantaneously
and that does not decay on the time scale of the experiment
and a component of transient absorption that decays in 11 ps.
This latter explanation seems unlikely, since the traces obtained
with an excitation wavelength of 588 nm yield much more
complicated kinetics that would be difficult to rationalize by
this alternative. For excitation at 588 nm, two components of
stimulated emission are observed: one rising with a ∼10 ps
time constant and another that appears instantaneously and that
decays in ∼1-2 ps.9
It is clear that a unique and self-consistent description of the
hypericin photophysics requires a range of experiments. We
have already indicated the utility of varying the excitation
wavelength. Fluorescence upconversion experiments22 will also
provide important complementary information, since only
emission is detected. Comparison of a range of analogues also
proves to be effective in understanding the origin of the
photophysics.

3240 J. Phys. Chem. A, Vol. 101, No. 18, 1997
English et al.
(7) Gai, F.; Fehr, M. J.; Petrich, J. W. J. Am. Chem. Soc. 1993, 115,
Another possibility for the origin of this component is prompt
intersystem crossing to form triplets.
3384-3385.
(8) Gai, F.; Fehr, M. J.; Petrich, J. W. J. Phys. Chem. 1994, 98, 5784-
5795.
(9) Gai, F.; Fehr, M. J.; Petrich, J. W. J. Phys. Chem. 1994, 98, 8352-
8358.
Conclusions
The comparison of the transient absorption kinetics of the
tetramethoxy analogue with those of hypocrellin (taken in the
context of the comparison of those of the hexamethoxy analogue
and hypericin) is cogent evidence for the identification of the
transients in hypericin and hypocrellin with excited-state proton
or atom transfer. In the companion article, further evidence
will be provided for the assignment of these processes to proton
or atom transfer events. In addition, the order of magnitude
difference in the time constant between hypericin and hypo-
crellin will be discussed in detail.
(10) Fehr, M. J.; Carpenter, S. L.; Petrich, J. W. Bioorg. Med. Chem.
Lett. 1994, 4, 1339-1344.
(11) Fehr, M. J.; McCloskey, M. A.; Petrich, J. W. J. Am. Chem. Soc.
1995, 117, 1833-1836.
(12) Carpenter, S.; Fehr, M. J.; Kraus, G. A.; Petrich, J. W. Proc. Natl.
Acad. Sci. U.S.A. 1994, 91, 12273-12277.
(13) Kraus, G. A.; Zhang, W.; Fehr, M. J.; Petrich, J. W.; Wannemuehler,
Y.; Carpenter, S. Chem. ReV. 1996, 96, 523-535.
(14) Das, K.; English, D. S.; Fehr, M. J.; Smirnov, A. V.; Petrich, J. W.
J. Phys. Chem. 1996, 100, 18275-18281.
(15) English, D. S.; Zhang, W.; Kraus, G. A.; Petrich, J. W. J. Am.
Chem. Soc., in press.
(16) (a) Brucker, G. A.; Swinney, T. C.; Kelley, D. F. J. Phys. Chem.
1991, 95, 3190-3195. (b) Strandjord, A. J. G.; Barbara, P. F. J. Phys. Chem.
1985, 89, 2355-2361. (c) McMorrow, D.; Kasha, M. J. Phys. Chem. 1984,
88, 2235-2243. (d) Brucker, G. A.; Kelley, D. F. J. Phys. Chem. 1987,
91, 2856-2861. (e) Schwartz, B. J.; Peteanu, L. A.; Harris, C. B. J. Phys.
Chem. 1992, 96, 3591-3598.
(17) Okita, T.; Isobe, M. Tetrahedon 1995, 51, 3737.
(18) Hatch, R. P.; Shringarpure, J.; Weinreb, S. M. J. Org. Chem. 1978,
43, 4172.
Acknowledgment. D.S.E. was supported by fellowships
from Amoco and GAANN. J.M.Z. was supported by a GAANN
fellowship. Parts of this work were supported by NSF Grants
BIR9413969 and CHE-9613962 to J.W.P. and by a grant from
the Center for Advanced Technology and Development to
G.A.K.
(19) Barton, J. W.; Walker, R. B. Tetrahedon Lett. 1975, 569.
(20) Gai, F.; Rich, R. L.; Petrich, J. W. J. Am. Chem. Soc. 1994, 116,
735.
References and Notes
(21) Diels, J.-C.; Rudolph, W. Ultrashort Laser Pulse Phenomena;
Academic Press: San Diego, 1996; p 142.
(22) Barbara, P. F.; Jarzeba, W. AdV. Photochem. 1990, 15, 1.
(23) Racinet, H.; Jardon, P.; Gautron, R. J. Chim. Phys. Phys.-Chim.
Biol. 1988, 85, 971-977.
(24) Eloy, D.; Le Pellec, A.; Jardon, P. J. Chim. Phys. Phys.-Chim. Biol.
1996, 93, 442-457.
(25) Yamazaki, T.; Ohta, N.; Yamazaki, I.; Song, P.-S. J. Phys. Chem.
1993, 97, 7870.
(26) Das, K.; English, D. S.; Petrich, J. W. J. Am. Chem. Soc., in press.
(27) Diwu, Z.; Lown, J. W. J. Photochem. Photobiol. A 1992, 64, 273.
(28) Zenner, J. M. Ph.D. Thesis, Iowa State University, Ames, IA, 1995.
(1) Dura´n, N.; Song, P.-S. Photochem. Photobiol. 1986, 43, 677-680.
(2) Diwu, Z. Photochem. Photobiol. 1995, 61, 529-539.
(3) Carpenter, S.; Kraus, G. A. Photochem. Photobiol. 1991, 53, 169-
174.
(4) Meruelo, D.; Degar, S.; Nuria, A.; Mazur, Y.; Lavie, D.; Levin B.;
Lavie, G. In Natural Products as AntiViral Agents; Chu, C. K., Cutler, H.
G., Eds.; Plenum Press: New York, 1992; pp 91-119, and references
therein.
(5) Hudson, J. B.; Zhou, J.; Chen, J.; Harris, L.; Yip, L.; Towers, G.
H. N. Photochem. Photobiol. 1994, 60, 248-254.
(6) Fehr, M. J.; Carpenter, S. L.; Wannemuehler, Y.; Petrich, J. W.
Biochemistry 1995, 34, 15845-15848.