Photo-degradation behaviour of roseoflavin in some aqueous solutions

Article abstract of DOI:10.1016/j.chemphys.2010.02.004

An absorption and emission spectroscopic characterization of roseoflavin (8-dimethylamino-8-demethyl-riboflavin, RoF) in aqueous solutions was carried out. The studies were concentrated on roseoflavin in pH 8 phosphate buffer. Absorption cross-section spectra, fluorescence excitation spectra, fluorescence quantum distributions, fluorescence quantum yields and fluorescence lifetimes were determined. The fluorescence of RoF is quenched by photo-induced intra-molecular charge-transfer at room temperature. The photo-degradation of RoF in un-buffered water, in Tris-HCl buffer, and in phosphate buffer was studied. Phosphate buffer and to a smaller extent Tris buffer catalyse the RoF photo-degradation. Photo-excitation of the primary photoproduct, 8-methylamino-riboflavin (8-MNH-RF), enhanced the RoF degradation by triplet 8-MNH-RF - singlet RoF excitation transfer with subsequent triplet-state RoF degradation.

Full text of DOI:10.1016/j.chemphys.2010.02.004

Chemical Physics 369 (2010) 27–36
Contents lists available at ScienceDirect
Chemical Physics
journal homepage: www.elsevier.com/locate/chemphys
Photo-degradation behaviour of roseoflavin in some aqueous solutions
b
b
A. Tyagi a, A. Penzkofer a, , T. Mathes , P. Hegemann
*
a Institut II – Experimentelle und Angewandte Physik, Universität Regensburg, Universitätsstraße 31, D-93053 Regensburg, Germany
b Institut für Biologie/Experimentelle Biophysik, Humboldt Universität zu Berlin, Invalidenstraße 42, D-10115 Berlin, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
An absorption and emission spectroscopic characterization of roseoflavin (8-dimethylamino-8-demethyl-
riboflavin, RoF) in aqueous solutions was carried out. The studies were concentrated on roseoflavin in pH
8 phosphate buffer. Absorption cross-section spectra, fluorescence excitation spectra, fluorescence quan-
tum distributions, fluorescence quantum yields and fluorescence lifetimes were determined. The fluores-
cence of RoF is quenched by photo-induced intra-molecular charge-transfer at room temperature. The
photo-degradation of RoF in un-buffered water, in Tris–HCl buffer, and in phosphate buffer was studied.
Phosphate buffer and to a smaller extent Tris buffer catalyse the RoF photo-degradation. Photo-excitation
of the primary photoproduct, 8-methylamino-riboflavin (8-MNH-RF), enhanced the RoF degradation by
triplet 8-MNH-RF – singlet RoF excitation transfer with subsequent triplet-state RoF degradation.
Ó 2010 Elsevier B.V. All rights reserved.
Received 23 December 2009
In final form 3 February 2010
Available online 10 February 2010
Keywords:
Roseoflavin
8-Methylamino-riboflavin
Absorption spectroscopy
Fluorescence spectroscopy
Photo-induced charge-transfer
Photo-degradation
Triplet–singlet excitation transfer
1. Introduction
tral un-buffered water. The photo-stability of RoF was found to
be highest in pure water, reduced by Tris–HCl buffer, and more
Roseoflavin (8-dimethylamino-8-demethyl-
D
-riboflavin, RoF) is
strongly reduced by phosphate buffer. The photo-degradation
was found to be enhanced by the intermediate photoproduct 8-
methylamino-riboflavin (8-MNH-RF).
a riboflavin derivative with modified absorption and emission
spectroscopic behaviour [1] (first absorption band red-shifted with
larger absorption strength than riboflavin, fluorescence emission
very weak). The biological function of roseoflavin is different from
that of riboflavin (vitamin B2) [2]. It acts as an antibiotic [3] and
vitamin B2 antagonist [4]. Substitution of riboflavin, flavin mono-
nucleotide (FMN) and flavin adenine dinucleotide (FAD) by RoF
and its counterparts RoFMN and RoFAD in flavoproteins during
protein expression has been achieved ([5] and references therein).
In most cases the biological cofactor action was lost or strongly re-
duced by the cofactor exchange [6–9].
Bi-anionic phosphate ðHPO24ꢀÞ and sulphate ðSO42ꢀÞ catalysed
photo-degradation of riboflavin and FMN is well established in
the literature [19–22]. The bi-anionic catalytic conversion of ribo-
flavin to 20,9-cyclo-dehydro-riboflavin was identified [20].
The structural formulae of roseoflavin (RoF) and of derivative
classes are displayed in Fig. 1.
2. Experimental
The absorption behaviour of roseoflavin in several solvents was
reported in [1,3,7,10–16]. The fluorescence behaviour of roseofla-
vin was studied in [11,13,16,17]. The photo-stability of roseoflavin
in aqueous solution under different pH conditions was studied in
[1,11,17,18].
The dye RoF was bought from Toronto Research Chemicals Inc.,
North York, Canada. The specified purity was >97%. It was dis-
solved in aqueous solutions. If not stated different experiments
were carried out at room temperature under aerobic conditions.
Transmissions were measured with a commercial spectropho-
tometer (Cary 50 from Varian), fluorescence emission and fluores-
In this paper, the absorption and emission spectroscopic behav-
iour of roseoflavin was mainly studied in aqueous pH 8 10 mM so-
dium phosphate buffer (abbreviated by NaPi, composition:
cence excitation spectra were recorded with
spectrofluorimeter (Cary Eclipse from Varian).
a commercial
0.68 mM NaþH2POꢀ, 9.32 mM Na22þHPO24ꢀ, and 10 mM NaCl).
4
Fluorescence lifetime measurements were performed with sec-
ond harmonic pulses of a mode-locked titanium–sapphire laser
system (Hurricane from Spectra Physics) and an ultrafast streak-
camera (type C1587 temporal disperser with M1952 high speed
streak unit from Hamamatsu) [16]. Picosecond excitation pulses
at 456 nm were generated by stimulated Raman scattering of
second harmonic pulses (wavelength 402 nm, duration 4 ps) of
Photo-degradation studies were carried out on roseoflavin in pH
8 phosphate and Tris–HCl buffers (Tris = (CH2CH2OH)3C–NH2) of
different molarity, in pH 10 10 mM phosphate buffer, and in neu-
* Corresponding author. Tel.: +49 941 943 2107; fax: +49 941 943 2754.
E-mail address: alfons.penzkofer@physik.uni-regensburg.de (A. Penzkofer).
0301-0104/$ - see front matter Ó 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.chemphys.2010.02.004

28
A. Tyagi et al. / Chemical Physics 369 (2010) 27–36
R82
R10
N
N
H
O
N
O
O
R81
5'
NH
H
H
H
H
H
H
H3C
N
4'
OH
OH
OH
H
3'
2'
8-amino-flavin derivatives (AF)
1'
CH3
N
R82
R1
N
N
O
H3C
9
6
1
10
N
2
8
7
N
N
N
O
R81
3NH
5
N
4
H3C
NH
H3C
O
O
Roseoflavin (RoF)
8-amino-lumichrome derivatives
(ALC)
R8
R82
R10 R1
R10 R1
N
N
N
N
O
NH
N
N
O
R81
NH
H3C
N
H3C
N
O
R5
O
8-amino-quinone-flavin derivatives (AQF)
Fully reduced 8-amino-flavin
derivatives (AFred2
)
R82
R82
R10
R10 R1
N
N
N
N
O
N
N
O
R81
R81
H3C
NH
NH
H3C
N
N
R5
O
O
Semi-reduced 8-amino- C4a-flavin
Semi-reduced 8-amino- C10a-
flavin derivatives (C10a-AFred1)
derivatives (C4a-AFred1
)
R82
R10a
R10
R82
N
R10 R1
N
N
O
R81
N
N
N
O
R81
NH
H3C
N
NH
H3C
N
R5
O
R4a
O
8-amino- C4a-flavin adducts
8-amino- C10a-flavin adducts
Fig. 1. Structural formulae of roseoflavin (RoF), 8-amino-flavin derivatives (AF), 8-amino-lumichrome derivatives (ALC), 8-amino-quinone-flavin derivatives (AQF), fully
reduced 8-amino-flavin derivatives (AFred2), semi-reduced 8-amino-C4a-flavin derivatives, semi-reduced 8-amino-C10a-flavin derivatives, 8-amino-C4a-flavin adducts, 8-
amino-C10a-flavin adducts. Abbreviations: R81, R82 = CH3, H. R10 = ribityl, C20-keto-ribityl, C40-keto-ribityl, CH2CHO. R1 = CH3, H, OH. R5 = H, OH. R4a, R10a = H, OH.
the Ti:sapphire laser system in ethanol (Stokes shift 2928 cmꢀ1) [8]
using a Raman generator–amplifier arrangement [9] with two 5 cm
cells in series (distance between cells approximately 10 cm for
divergence reduction). Part of the fluorescence lifetime measure-
ments were also carried out with a mode-locked picosecond
Nd–glass laser system [23] (second harmonic pulses at 527 nm of
6 ps duration).
The photo-degradation was studied by sample excitation with a
high-pressure mercury lamp in combination with interference fil-
ters. Part of the photo-degradation studies were carried out by
excitation with second harmonic light of a small cw Nd:YAG laser
(wavelength 532 nm, power 4.5 mW).
Tris–HCl buffer [16]. The S0–S1 absorption strength is
R
¼
½
1 raðkÞ=kꢁdk ¼ ð2:3 ꢂ 0:2Þ ꢃ 10ꢀ17 cm2 (upper wavelength
ꢀ
ra
ku
position set to 420 nm). Included in Fig. 2 are the absorption
cross-section spectra of 8-MNH-RF in neutral aqueous solution
(from [24]) and of 8-amino-riboflavin (8-NH2-RF) in aqueous pH
8 buffer (from [17]).
Fluorescence quantum distributions, EF(k), of RoF in pH 8 phos-
phate buffer for some excitation wavelengths are shown in Fig. 3.
For kF,exc = 450 nm the fluorescence quantum yield is largest due
to fluorescence contribution from riboflavin contamination
(fluorescence quantum yield of riboflavin at pH 8 is 0.26 [25]).
The presence of a mole fraction of xRF = 0.0025 is calculated. At
kF,exc = 540 nm the fluorescence quantum distribution is domi-
nantly determined by RoF emission. At kF,exc = 505 nm EF(k) is dom-
inated by the degradation product 8-MNH-RF which is already
present in a small amount.
3. Results
3.1. Absorption and fluorescence characterization
The radiative lifetime of RoF is determined from the S0–S1
ꢀ
absorption strength ra and the mean fluorescence wavelength,
The absorption cross-section spectrum of roseoflavin (RoF) in
aqueous pH 8 10 mM NaPi buffer is shown in Fig. 2. It fully agrees
with the absorption cross-section spectrum of RoF in aqueous pH 8
R
R
1=3
kF ¼ ½ EFðkÞk3dk= EFðkÞdkꢁ
,
by the Strickler–Berg formula
ꢀ
[26–28],

A. Tyagi et al. / Chemical Physics 369 (2010) 27–36
29
Fig. 2. Absorption cross-section spectra, ra(k), of RoF in aqueous pH 8 buffer (this
Fig. 4. Temporal fluorescence traces of RoF in aqueous pH 8 10 mM NaPi buffer.
Fluorescence excitation occurred at 456 nm with light pulses of 4 ps duration.
work), 8-MNH-RF in neutral aqueous solution (from [24]), and of 8-NH2-RF in
Millipore water (from [37]). The fluorescence excitation spectrum of RoF is included
(detection wavelength: 600 nm, normalized to absorption spectrum at 488 nm).
Curves were measured with
a streak-camera. (a) Unexposed sample, streak
speed of 10 ps/pixel. Dashed curve: calculated riboflavin contribution (SF/SF,max
=
=
0.108exp(ꢀt/5 ns)). Dash-dotted curve: nonlinear regression fit (SF/SF,max
0.108exp(ꢀt/5 ns) + 0.48exp(ꢀt/0.1 ns) + 0.412exp(ꢀt/1.98 ns)). (b) Unexposed sam-
ple, streak speed of 0.33 ps/pixel. (c) Sample exposed in wavelength range from
425 nm to 503 nm with intensity of 0.085 W cmꢀ2 for 65 min. Trace is caused by
fluorescence decay of 8-MNH-RF. Dash-doted curve: single-exponential fit with
sF = 2.18 ns.
fluorescence signal detection with our streak-camera. The trace
in the main frame (a) was measured with a streak speed of
10 ps/pixel, and the trace in the small frame (b) was recorded with
a streak speed of 0.33 ps/pixel. In the main frame the riboflavin
contribution to the fluorescence signal is indicated by the dashed
line. The short peaks at time t = 0 in Fig. 4a and b are not time-re-
solved because of streak speed limitations. The dash-dotted curve
in Fig. 4a shows a three-exponential regression fit,
ꢀ
ꢁ
SFðtÞ=SFð0Þ ¼ x1 exp ꢀt=s0F;1 þ x2 exp ðꢀt=sF;2Þ þ x3
ꢃ exp ðꢀt=sF;3Þ;
ð2aÞ
with x1 = 0.48, s0F;1 = 100 ps, x2 = 0.412,
sF,2 = 1.97 ns, x3 = 0.108,
sF,3 = 5 ns. The true fluorescence lifetime of the short component
is extracted by involving the fluorescence quantum yield [16]. The
total fluorescence quantum yield is
/F ¼ /F;1 þ /F;2 þ /F;3
;
ð2bÞ
Fig. 3. Fluorescence quantum distributions, EF(k), of RoF sample in pH 8 10 mM
NaPi buffer at different fluorescence excitation wavelengths, kF,exc. The obtained
fluorescence quantum yields, /F, are listed in the figure. Differences are due to the
presence of trace amounts of riboflavin and 8-MNH-RF.
with a value of /F(456 nm) = 0.0012 (see Fig. 3). The fluorescence
quantum yield of the short component is
x1s0
x1s0 þ x2sFF;;12 þ x3sF;3
/
¼
/F;
ð2cÞ
F;1
F;1
3
ꢀ
nAkF
giving /F,1 = 6.7 ꢃ 10ꢀ5. The fluorescence lifetime is
srad
¼
;
ð1Þ
3 ꢀ
8pc0nF ra
sF;1 ¼ /F;1srad
;
ð2dÞ
where nA and nF are the average refractive indices in the S0–S1
absorption and emission region (nA ꢄ nF ꢄ 1.33), respectively, and
which gives
s
F,1 = 0.44 0.05 ps. The component with
s
F,1 = 0.44 ps
c0 is the speed of light in vacuum. The result is
s
rad = 6.5 0.4 ns.
lifetime is attributed to locally excited direct fluorescence emission
of RoF. This short locally excited-state lifetime is thought to be
determined by efficient twisted intra-molecular charge-transfer
Two temporal fluorescence traces are shown in Fig. 4a and b for
RoF in pH 8 10 mM NaPi buffer. The traces were obtained by sam-
ple excitation at 456 nm with light pulses of 4 ps duration and
(TICT) [16]. The slow component with sF,2 = 1.97 ns is attributed

30
A. Tyagi et al. / Chemical Physics 369 (2010) 27–36
dominantly to the presence of a small fraction of 8-MNH-RF which
is a photo-degradation product of RoF (some time-delayed locally
excited-state fluorescence emission of RoF may contribute [16],
but it cannot be resolved from the dominant 8-MNH-RF contribu-
sure. The formed absorption band around 488 nm decreases with
continued light exposure. A spectral shoulder around 470 nm
develops (likely formation and decay of 8-amino-riboflavin, 8-
NH2-RF). A spectral structure forms out in the 340 nm to 420 nm
region (indication of formation of 8-dimethylamino-lumichrome,
8-methylamino-lumichrome, and 8-amino-lumichrome, see Fig. 1
for structural formulae). An absorption tail remains in the 540–
590 nm range (roseoflavin quinonoid structure formation by
photo-induced methyl-transfer from the 8-dimethylamino group
to the N1 position, see Fig. 1 for structural formulae). The photo-
degradation strongly reduces the absorption strength of the first
absorption band in the 400 to 550 nm region indicating dominant
roseoflavin reduction (fully reduced 8-amino-flavin structure for-
mation by hydration, hydroxylation or methylation at N5 and N1
positions, for structural formulae see Fig. 1). In Fig. 5b, the absorp-
tion spectra of Fig. 5a are shown after approximate subtraction of
the remaining RoF contribution. The initial increase and later de-
crease of the 8-MNH-RF contribution is clearly seen. Some indica-
tion of build-up and decrease of 8-NH2-RF is found (has peak
absorption at about 478 nm, see Fig. 2). The formation of a quino-
noid structure is revealed by the build-up of an absorption band
around 560 nm. Some 8-amino-C4a-flavin-semiquinone or 8-ami-
no-C10a-flavin-semiquinone formation with absorption around
560 nm is thought to be unlikely because of expected low photo-
stability. The absorption structure in the 340–420 nm region
shows the presence of an 8-amino-lumichrome component. The
presence of small amounts of 8-amino-riboflavin-C4a-adducts or
8-amino-riboflavin-C10a-adducts with H, OH, CH3 as adducts
cannot be excluded.
tion). The component with
contamination.
sF,3 = 5 ns belongs to the riboflavin
A normalized fluorescence excitation spectrum of our RoF in pH
8 NaPi buffer is included in Fig. 2. The fluorescence was detected
at kdet = 600 nm and it is normalized to the absorption
cross-section spectrum of RoF at 488 nm according to
Eex;nðkÞ ¼ ½Eex;600 nmðkÞ=Eex;600 nmð488 nmÞꢁrað488 nmÞ. The shape of
the presented excitation spectrum does not fit to the shape of
the RoF absorption cross-section spectrum. It better fits to the
absorption cross-section spectrum of 8-MNH-RF. This finding
clearly shows that the small fraction of present 8-MNH-RF domi-
nates the fluorescence emission since RoF is extremely weak emit-
ting. In the UV spectral range below 320 nm, the excitation
spectrum becomes smaller than the absorption cross-section spec-
trum indicating that the higher excited states only partly relax via
the first exited state (other decay channels are present, violation of
Vavilov–Kasha rule of excitation wavelength independent fluores-
cence emission [29,30]).
3.2. Photo-degradation studies
The photo-degradation behaviour of RoF was studied by long-
time sample exposure and measurement of absorption spectra,
fluorescence emission spectra, and fluorescence excitation spectra
at certain time points.
In Fig. 5a, the development of the absorption coefficient spectra
of 0.047 mM RoF in pH 8 10 mM NaPi buffer is shown. The sample
The development of the absorption coefficients of RoF in pH 8
10 mM NaPi at the selected wavelengths kpr = 540 nm (only RoF
absorption) and 488 nm (peak absorption of 8-MNH-RF) versus
the exposed incident energy density is shown by the line-con-
nected circles in Fig. 6a and b, respectively (data taken from
was excited at kexc = 425–503 nm with an intensity of Iexc
=
0.085 W cm2. The original absorption band of RoF peaking at
504 nm transforms to an absorption band peaking at 488 nm (for-
mation of 8-MNH-RF). The change starts slowly within the first
hour, then speeds up, and is roughly completed after 2 h of expo-
4
3
6
Fig. 6. Development of absorption coefficients of RoF in phosphate buffer versus
exposed excitation energy density wexp. (a) Excitation in the wavelength range
kexc = 425–503 nm and absorption probing at kpr = 540 nm (only absorption of RoF).
(b) Excitation in the wavelength range kexc = 425–503 nm and absorption probing at
kpr = 488 nm (absorption peak of 8-MNH-RF). (c) Excitation at different wave-
Fig. 5. Absorption spectra development due to blue-light exposure. Excitation
wavelength kexc = 425–503 nm. Excitation intensity Iexc = 0.085 W cmꢀ2. Exposure
times, texp, are indicated in the sub-figures. (a) Measured absorption coefficient
spectra at different times of exposure. (b) Obtained absorption coefficient spectra
after subtraction of remaining RoF contribution.
lengths (kexc = 425–503 nm, Iexc = 0.25 W cmꢀ2; kexc = 532 nm, Iexc = 0.22 W cmꢀ2
kexc = 546 nm, Iexc = 0.16 W cmꢀ2) and absorption probing at kpr = 540 nm.
;

A. Tyagi et al. / Chemical Physics 369 (2010) 27–36
31
Fig. 5a, wexp ¼ Iexctexp). The trace for kpr = 540 nm shows the photo-
degradation of RoF. It starts with low efficiency and then becomes
rather effective. The absorption decrease levels off at high exposed
energy density because of absorption of the formed 8-methyl-
amino quinone flavin and fully reduced roseoflavin (RoFred2). The
trace for kpr = 488 nm shows the formation and decrease of 8-
MNH-RF, likely with some contribution of 8-NH2-RF formation
and decay.
rescence band peaking in the 520–540 nm region belongs mainly
to 8-MNH-RF and 8-NH2-RF emission and resembles the 8-MNH-
RF and 8-NH2-RF build-up and decrease.
In Fig. 8, fluorescence excitation spectra are shown for the same
experimental situation as in Figs. 5 and 7. The emission was de-
tected at kdet = 600 nm (a) and at kdet = 450 nm (b). In part (a) the
build-up and decay of the fluorescent photoproducts 8-MNH-RF
and 8-NH2-RF is manifested in the excitation spectrum develop-
ment in the 420–520 nm range. Some formation of 8-amino-lumi-
chromes is revealed by the excitation spectra shape changes in the
330–420 nm range. The 8-amino-lumichrome formation is more
clearly seen in the excitation spectra development in part (b).
The inset (c) shows the fluorescence excitation signal Eex,450nm
(k = 360 nm) as a function of exposure time. The signal (amino-
lumichrome formation) rises during RoF degradation (texp < 3 h)
and during 8-MNH-RF and 8-NH2-RF degradation (texp > 1 h).
A temporal fluorescence trace measured on 0.047 mM roseofla-
vin in 10 mM NaPi pH 8 buffer after 65 min of light exposure at
kexc = 425–503 nm with Iexc = 0.085 W cmꢀ2 is displayed in Fig. 4c
(fluorescence creation with 4 ps laser pulse at 456 nm). It is fitted
From the absorption decrease with light exposure quantum
yields of photo-degradation, /D, may be determined. The quantum
yield of photo-degradation, /D, is defined by
dNDðtexp
Þ
½aprðtexp þ dtÞ ꢀ aprðtexpÞꢁ=ra;pr
;
Iexcdt
/DðtexpÞ ¼
¼
ð3Þ
dnph;absðtexp
Þ
a
mexc
ðtexpÞ
exc
h
where dND is the number density of degraded molecules (unit cmꢀ3
)
in a time interval dt at texp, and dnph,abs is the number density of ab-
sorbed photons (unit cmꢀ3) in the same time interval dt at texp. We
estimate an initial quantum yield of photo-degradation of RoF from
the initial absorption decrease at 540 nm (see Figs. 5 and 6a) of
/
D,0(RoF) ꢄ 1.15 ꢃ 10ꢀ6. With time of exposure the quantum yield
with
F = 2.18 ns. It gives the fluorescence lifetime of 8-MNH-RF.
The absorption coefficient development at kpr = 540 nm (Fig. 6a)
a single-exponential decay function of time constant
of photo-degradation increased. For texp = 75 min
a value of
s
/D(75 min, RoF) ꢄ /D,max(RoF) ꢄ 2.3 ꢃ 10ꢀ5 is determined. The
quantum yield of photo-degradation of the formed 8-MNH-RF is
estimated from the absorption decrease at kpr = 488 nm for
and kpr = 488 nm (Fig. 6b) versus exposed energy density was stud-
ied for different NaPi buffer and RoF concentrations (seen by differ-
ent initial absorption coefficients). Light excitation occurred in the
wavelength range of 425–503 nm. The line-connected circle curves
(0.047 mM RoF in 10 mM NaPi pH 8 buffer) were alredy discussed
above. The dashed-line-connected triangles belong to 100 mM
NaPi pH 8 buffer (sample temperature 0 = 4 °C). The efficiency of
photo-degradation increased with the phosphate buffer concentra-
tion (steeper initial absorption decrease at kpr = 540 nm). The dot-
ted-line connected diamonds belong to higher concentrated RoF
(C = 0.092 mM) in 10 mM NaPi pH 8 buffer (0 = 4 °C). The initial
texp = 2.5 h (wexp = 765 J cmꢀ2).
RF) ꢄ 2 ꢃ 10ꢀ5 is estimated.
A value of /D(2.5 h, 8-MNH-
In Fig. 7 the fluorescence development of 0.047 mM roseoflavin
in 10 mM NaPi pH 8 buffer due to light exposure at kexc = 425–
503 nm with Iexc = 0.085 W cmꢀ2 is shown. In part (a) the fluores-
cence was excited at kF,exc = 470 nm. The fluorescence signal rises
and decreases with the formation and degradation of 8-MNH-RF.
It shifts a little bit to the blue side for texp P 2 h likely due to the
formation and degradation of 8-NH2-RF. In part (b) the fluores-
cence was excited at kF,exc = 360 nm. The fluorescence signal in
the 400–500 nm range belongs to 8-amino-lumichromes. The fluo-
300
360
210
Fig. 8. Fluorescence excitation spectra development of RoF im 10 mM NaPi pH 8
buffer due to blue-light exposure (kexc = 425–503 nm, Iexc = 0.085 W cmꢀ2). Para-
meters are the same as in Fig. 5. (a) Fluorescence detection at kdet = 600 nm. The
curves belong to indicated exposure times. Dominant emission occurs from 8-
MNH-RF and 8-NH2-RF. (b) Fluorescence detection at kdet = 450 nm (dominant
emission from 8-dimethylamino-lumichrome, 8-methylamino-lumichrome, and 8-
amino-lumichrome).
Fig. 7. Fluorescence-emission-spectra development of RoF im 10 mM NaPi pH 8
buffer due to blue-light exposure (kexc = 425–503 nm, Iexc = 0.085 W cmꢀ2). Para-
meters are the same as in Fig. 5. (a) Fluorescence excitation at 470 nm (dominant
emission from 8-MNH-RF and 8-NH2-RF). (b) Fluorescence excitation at 360 nm
(dominant short-wavelength emission from 8-dimethylamino-lumichrome,
8-methylamino-lumichrome, and 8-amino-lumichrome).

32
A. Tyagi et al. / Chemical Physics 369 (2010) 27–36
RoF degradation was about the same as in the case of lower con-
centration (line-connected circles), but with time of exposure it be-
came faster than at lower RoF concentration (shorter distance
between RoF and 8-MNH-RF). The dash-dotted-line connected
squares for RoF at 4 °C belong to 10 mM NaPi at pH 10 (composi-
tion: 0.14 mM Na3PO4, 9.84 mM Na2HPO4, and 10 mM NaCl). The
absorption changes were smaller than at pH 8 indicating a higher
photo-stability. The concentration of Na2þHPO42ꢀ in 10 mM pH 10
buffer (9.84 mM) is only slightly larger than at pH 8 (9.32 mM).
The increased photo-stability seems to be caused by the higher
pH. It should be noted that at pH 10 nearly half of the RoF mole-
cules are in the anionic state (pKa = 10.2 [17]).
In Fig. 6c, the photo-degradation of RoF in 10 mM NaPi pH 8
buffer was studied for different excitation wavelengths
(kexc = 425–503 nm, 532 nm, and 546 nm) at 4 °C. The absorption
coefficient at kpr = 540 nm is plotted versus incident exposed en-
ergy density. The initial efficiency of photo-degradation was
approximately the same in all three cases, but the speed-up of
photo-degradation with exposed energy density was different. At
kexc = 425–503 nm the light exposure excites both RoF and the pri-
mary photoproduct 8-MNH-RF, which enhances the photo-degra-
dation of RoF. At kexc = 532 nm the photoproduct 8-MNH-RF is
still slightly absorbing, and the photo-excited 8-MNH-RF enhances
the photo-degradation of RoF. In the case of excitation at
kexc = 546 nm, 8-MNH-RF is no longer absorbing, therefore no long-
er influenced the photo-degradation of RoF, and the quantum yield
of photo-degradation of RoF remained constant independent of the
exposed energy density.
both phosphate buffer and Tris buffer. The photo-stability of RoF
in the Tris buffer is roughly an order of magnitude higher than that
in the phosphate buffer.
In Fig. 9b, the maximum quantum yield of photo-degradation,
/
D,max, in the course of light exposure is plotted versus RoF concen-
tration, C0. Excitation occurred in the wavelength range 425–
503 nm. With the exception of 47 M RoF in pH 8 10 mM NaPi buf-
l
fer the samples were measured at 4 °C. The solid-line-connected
circles belong to RoF in 10 mM NaPi pH 8 buffer. At low dye con-
centration the photo-degradation is independent of concentration
and it is /D,max = /D,0. In a medium dye concentration range the
quantum yield of photo-degradation rises steeply with concentra-
tion. In the range of highest dye concentration, a linear rise of
photo-degradation with dye concentration is observed. This behav-
iour will be interpreted below as diffusion controlled degradation
of RoF by collision with photo-excited 8-MNH-RF or 8-NH2-RF.
The initial quantum yield of photo-degradation, /D,0, was found
to be independent of dye concentration within our experimental
accuracy (no 8-MNH-RF or 8-NH2-RF is present for photo-degrada-
tion enhancement).
The quantum yields of photo-degradation, /D,max, of RoF in
10 mM NaPi at pH 10 and of RoF in 100 mM NaPi at pH 8 were mea-
sured only at one dye concentration. The efficiency of photo-degra-
dation was higher at higher buffer concentration. It was lower at
pH 10 than at pH 8.
In the case of RoF in pure Millipore water and in the case of
Tris–HCl buffer, the RoF photo-degradation was only followed over
the initial start phase of degradation (less than 10% of RoF de-
graded) because of the high photo-stability. The photo-degradation
products are expected to be 8-MNH-RF and fully reduced roseofla-
vin (RoFred2) as in the case of RoF in pH 8 NaPi buffer because of ob-
served remarkable fluorescence rise with exposure which agrees
with 8-MNH-RF formation.
In Fig. 9a, the initial quantum yield of photo-degradation, /D,0
,
of RoF is depicted versus buffer concentration, Cb, for phosphate
buffer at pH 8 and Tris–HCl buffer at pH 8 (excitation at 425–
503 nm). One data point for phosphate buffer at pH 10 is included.
The data point at Cb = 0 belongs to pure Millipore water. The mea-
surements were carried out at 4 °C. The quantum yield of photo-
degradation of RoF in Millipore water is /D,0 = 9 ꢃ 10ꢀ8. The
photo-stability of RoF decreases with the buffer concentration for
After long-time photo-excitation of RoF in phosphate buffer, a
component with absorption out to 590 nm remained (Fig. 5). Its
fluorescence quantum yield was measured by fluorescence excita-
Fig. 9. (a) Dependence of initial quantum yield of photo-degradation, /D,0, of RoF on buffer concentration, Cb. Excitation wavelength kexc = 425–503 nm, excitation intensity
Iexc = 0.25 W cmꢀ2. The pH values and buffer compositions are indicated in the figure legend. (b) Dependence of maximum quantum yield of photo-degradation, /D,max, of RoF
on its initial concentration, C0. Excitation wavelength kexc = 425–503 nm, excitation intensity Iexc = 0.25 W cmꢀ2. pH values and buffer compositions are indicated in the figure
legend.

A. Tyagi et al. / Chemical Physics 369 (2010) 27–36
33
tion at 550 nm. A value of /F ꢄ 0.016 was estimated. Its fluores-
cence decay was determined by picosecond laser pulse excitation
at 527 nm (duration ꢄ 6 ps) and streak-camera fluorescence detec-
tion (curves not shown). A bi-exponential fluorescence decay was
intra-molecular charge-transfer (TICT state formation) with non-
radiative charge recombination. The same fluorescence behaviour
was found here for roseoflavin in the investigated solvents (Milli-
pore water, phosphate buffer solution, Tris–HCl buffer solution).
The photo-stability of roseoflavin in organic solvents has not yet
been studied in detail [24]. A high photo-stability in neutral water
was reported in [24]. Our own photo-degradation studies on RoF in
pure Millipore water in this work give a quantum yield of photo-
degradation of /D,0 ꢄ 9.1 ꢃ 10ꢀ8 (see above). The photolysis prod-
ucts of roseoflavin in alkaline and acidic aqueous solutions were
identified in [24]: In 0.1 M NaOH RoF photo-degraded to 8-MNH-
RF, while in 6 M HCl RoF photo-degraded to 7-methyl-8-dimethyl-
amino-alloxazine (8-dimethylamino-lumicrome, 8-M2 N-LC).
Our studies on RoF in aqueouꢀs pH 8 phosphate buffer (phos-
found with a short component of lifetime
grated amount of /F,1//F = 0.25) and a slow component of lifetime
F,2 = 2.2 0.1 ns (integrated amount of /F,2//F = 0.75). The fast
component has an approximate S1–S0 radiative lifetime of
F,1//F,1 ꢄ 4 ns. It likely belongs to a quinonoid roseoflavin
photoproduct (likely 8-MNH-1-methyl-quinone-riboflavin). The
slow component has radiative lifetime of srad,2
F,2//F,2 ꢄ 200 ns. It may belong to fully reduced 8-amino-
riboflavins.
sF,1 = 15 5 ps (inte-
s
srad,1 = s
a
=
s
A quantitative analysis of the fluorescence spectra in Fig. 7a and
of the absorption coefficient spectra in Fig. 5b gives a fluorescence
quantum yield of /F = 0.40 0.04 for 8-MNH-RF in aqueous pH 8
10 mM NaPi. Combined with the fluorescence lifetime of
phate composition 6.8% NaþH2PO and 93.2% Na22þHPO42ꢀ) clearly
4
show phosphate buffer catalysed photo-degradation (see Fig. 9a).
In photolysis studies on riboflavin it was found a di-anionic catal-
ysis of conversion of riboflavin to 20,9-cyclo-dehydro-riboflavin by
PO24ꢀ and SO24ꢀ [20] (peak absorption at 407 nm). We think that
PO24ꢀ also catalyses the photo-degradation of RoF, but resolvable
formation of 20,9-cyclo-dehydro-roseoflavin formation was not ob-
served. Here, different 8-amino-flavin derivatives were formed as
intermediates or final products during light exposure. Structural
formulae of involved classes of 8-amino-flavin derivatives are
shown in Fig. 1 [8-amino-isoalloxazines (flavins), 8-amino-alloxa-
zines (lumichromes), 8-amino-quinone-flavins, fully reduced 8-
amino-flavins)]. Possible constituents of the groups Ri are listed
in the caption of Fig. 1.
s
F = 2.18 0.1 ns for 8-MNH-RF (Fig. 4c) a radiative lifetime of
srad
=
s
F//F = 5.5 0.6 ns (Eq. (2d)) and an S0–S1 absorption
strength of ra ¼ ð2:6 ꢂ 0:3Þ ꢃ 10ꢀ17 cm2 (Eq. (1)) are calculated
ꢀ
ꢂ
ꢃ
ꢀ
kF ¼ 577 nm .
For RoF in pH 8 10 mM NaPi buffer the photo-stability was also
investigated under anaerobe sample conditions (de-aerating by Ar
bubbling for 1 h). The de-aerated samples were photo-excited at
kexc = 425–503 nm with an intensity of Iexc = 0.25 W cmꢀ2. The
temperature was kept at 4 °C. The initial quantum yield of photo-
degradation, /D,0, increased roughly a factor of three compared to
the aerobe sample. A data point is included in Fig. 9a. The maxi-
mum quantum yield of photo-degradation, /D,max, during exposure
also increased for the anaerobe samples compared to the aerobe
samples. The result is included in Fig. 9b (two dashed-line con-
nected dots). The increased photo-sensitizing of RoF degradation
by 8-MNH-RF under anaerobe conditions will be explained below
by an enlarging of the RoF and 8-MNH-RF triplet lifetime.
Some enhancement of photo-degradation of RoF by the Tris–
HCl buffer was also observed. Only the efficiency of enhancement
was at least an order of magnitude reduced. The atomistic catalytic
action of the phosphate buffer and Tris buffer on the photo-conver-
sion of RoF is not yet known.
A proposed scheme of photo-degradation of RoF is described in
the following:
The primary process of photo-degradation, where initially only
RoF is present, is illustrated in Scheme 1. Light absorption excites
RoF in the S0 ground-state to the S1-state in locally excited-state
conformation 1RoFLꢅE. There occurs fast intra-molecular charge-
transfer (ICT) to RoFꢅCT. The charge-transfer state returns to the
ground-state by charge recombination (CR). During the lifetime
of 1RoFꢅLE there occurs some intersystem-crossing (ISC) to the
triplet state 3RoFꢅ (rate of intersystem-crossing kISC ꢄ 1 ꢃ 108 sꢀ1
[31–33]). The quantum yield of triplet formation is given by
4. Discussion
Some determined parameters of RoF, and of the primary RoF
photo-degradation product, 8-MNH-RF, in aqueous pH 8 10 mM
NaPi buffer are collected in Table 1.
The fluorescence behaviour of roseoflavin in aqueous solution
and in organic solvents was investigated in [9,16]. The weak
fluorescence efficiency was attributed to photo-induced twisted
/T ¼ kISCsF
;
ð4Þ
Table 1
which gives /T ꢄ 4.4 ꢃ 10ꢀ5 for RoF in pH 8 10 mM NaPi buffer
Parameters of RoF, and 8-MNH-RF in aqueous pH 8 10 mM sodium phosphate buffer
at room temperature under aerobe conditions.
(
sF ꢄ 0.44 ps). The lifetime sN of 3RoFꢅ is estimated below to be
sN ꢄ 4.5
ls under our aerobic experimental conditions (under
Parameter
RoF
8-NMH-RF
Comments
anaerobe condition a triplet lifetime of 36 ls was measured for
ka,max (nm)
504
3200
488
2500
Fig. 2
Fig. 2
Fig. 2
Figs. 3 and 7a
8-NH2–RF in pH 7 phosphate buffer at 26 °C [34]). During the
triplet-state lifetime degradation takes place. The initial quantum
efficiency of degradation in the triplet state is given by
~
D
ma ðcmꢀ1
Þ
ra,max (cm2)
kF,max (nm)
~
1.33 ꢃ 10ꢀ16
1.68 ꢃ 10ꢀ16
539
560
mF ðcmꢀ1
Þ
Þ
Þ
2860
1250
2940
2890
Figs. 3 and 7a
D
ꢀ1
ꢀ1
mSt ðcmꢀ1
~
~
d
d
mSt ¼ ka;max ꢀ kF;max
/
D;0
/
¼
ð5Þ
m01 ðcmꢀ1
ꢀ1
ꢀ1
D;0;T
~
19 180
19 300
~
m01 ¼ ðka;max þ kF;maxÞ=2
/
T
/
6.7 ꢃ 10ꢀ5
ꢄ0.4
Eqs. (2a)–(2c), Fig. 7a
Eq. (1)
F
srad (ns)
sF,1 (ps)
6.5
5.5
In the case of RoF in pH 8 10 mM NaPi buffer under aerobe condi-
tions, we estimate an efficiency of
0.44
2180
Eq. (2d)
Fig. 6a, Eq. (3)
/
D,0,T ꢄ 0.025
/
1.15 ꢃ 10ꢀ6
ꢄ2 ꢃ 10ꢀ5
D,0
(/D,0 ꢄ 1.15 ꢃ 10ꢀ6, /T ꢄ 4.4 ꢃ 10ꢀ5) The catalytic effect of the
phosphate buffer in the photo-degradation of RoF is active in the
triplet state of RoF. Dominant primary photoproducts are 8-methyl-
amino-riboflavin (8-MNH-RF, absorption maximum at 488 nm,
fluorescence quantum yield of about 0.4) and fully reduced roseo-
flavin (RoFred2). The 8-MNH-RF formation may proceed via
methyl-nitrogen bond fissure at the dimethylamino group and
~
Dma:
Abbreviations. ka,max: Wavelength position of maximum S0–S1 absorption.
Spectral half-width (FWHM) of S0–S1 absorption band. a,max = absorption cross-
section at ka,max. kF,max: Wavelength position of maximum fluorescence emission.
r
~
Dm
~
m
F: Spectral half-width (FWHM) of S0–S1 fluorescence band. d
St: Fluorescence
~
m
Stokes shift.
01: Electronic S0–S1 transition wavenumber. /F: Intrinsic fluorescence
rad: S1-state radiative lifetime. F,1: fluorescence lifetime (see text).
D,0: Initial quantum yield of photo-degradation.
quantum yield.
/
s
s

34
A. Tyagi et al. / Chemical Physics 369 (2010) 27–36
Scheme 1.
subsequent interaction of the radicals with the aqueous solution
(catalytic action of phosphate and Tris buffer) giving a formal reac-
tion according to
3RoFꢅ þ HOH ! 8-MN-RFÅ þ ÅCH3 þ HOH
! 8-MNH-RF þ CH3OH
ðR1Þ
*
The buffer catalysed reduction of RoF to RoFred2 may be due to in-
tra-molecular photo-reduction (hydrogen abstraction at C20 or C40
position of ribityl part) [35] according to
3RoFꢅ ! RoFred2
ðR2aÞ
where RoFred2 = 8-M2N-10-C20-keto-ribityl-1-H-5-H-isoalloxazine
or 8-M2N-10-C40-keto-ribityl-1-H-5-H-isoalloxazine. Or it may be
due to intermolecular photo-reduction like
Scheme 2.
ꢅ
:
:
38-MNH-RF þ HOH ! 8-NH-RF þ CH3 þ HOH
3RoFꢅ þ HOH ! RoFred2
ðR2bÞ
! 8-NH2-RF þ CH3OH:
ðR5Þ
In this case, RoFred2 is 1-OH-5-H-roseoflavin or 1-H-5-OH-roseofla-
vin. The formation of RoFred2 is concluded from the reduced sample
absorption (absorption of fully reduced flavin is small in blue and
violet spectral part [36]). A minor conversion of excited RoF to 8-
dimethylamino-lumichrome (splitting-off of ribityl part) according
to [9]
They degrade to fully reduced 8-MNH-RF in an intra-molecular
reaction according to
38-MNH-RFꢅ ! 8-MNH-RFred2;
ðR6aÞ
(8-MNH-RFred2 = 8-MNH-10-C20-keto-ribityl-1-H-5-H-isoalloxazine
or 8-MNH-10-C40-keto-ribityl-1-H-5-H-isoalloxazine) or an inter-
molecular reaction according to
3RoFꢅ ! 8-M2N-LC þ rybityl-ene
ðR3Þ
38-MNH-RF þ HOH ! 8-MNH-RFred2;
ðR6bÞ
is indicated by the 450 nm fluorescence excitation spectra (Fig. 8b)
and the 360 nm fluorescence emission spectra (Fig. 7b). Some ex-
cited RoF molecules seem to tautomerize to 8-amino-quinone-fla-
vins according to
(8-MNH-RFred2 = 8-MNH-1-OH-5-H-RF or 8-MNH-1-H-5-OH-RF)
similar to the reactions R2a and R2b. Also some amino-lumichrome
formation occurs according to
RoFꢅ ! 8-methylamino-1-methyl-quinone-flavin
ðR4Þ
38-MNH-RFꢅ ! 8-MNH-LC þ ribityl-ene:
ðR7Þ
(AQF formation). This formation is indicated by the build-up of an
absorption band in the 520–600 nm range (Fig. 5a and b).
The photo-excitation of 8-NH2-RF (absorption cross-section
spectrum included in Fig. 2) is thought to cause degradation
according to Scheme 3. The fluorescence lifetime of 8-NH2-RF is
sF ꢄ 2.3 ns [37]. The quantum yield of triplet formation is esti-
mated to be /T ꢄ kISC sF ꢄ 0.23 using kISC ꢄ 108 sꢀ1. The photo-deg-
radation of 8-NH2-RF to 8-NH2-RFred2 is thought to be dominated
by intra-molecular photo-reduction
Without the presence of RoF, the photo-excitation of the
strongly absorbing 8-MNH-RF (Fig. 2) causes degradation to a fully
reduced amino-flavin (8-MNH-RFred2), to 8-amino-riboflavin (8-
NH2-RF), and to 8-methylamino-lumichrome (8-MNH-LC) as is
shown in Scheme 2. The photo-excitation of 8-MNH-RF forms sin-
glet excited 18-MNH-RFꢅ with a fluorescenꢅce lifetime of sF ꢄ 2.2 ns.
Intersystem-crossing from 18-MNH-RF to 38-MNH-RFꢅ takes
place during this fluorescence lifetime. The quantum yield of trip-
let formation is estimated to be /T ¼ kISCsF ꢄ 0.22 (kisc ꢄ 108 sꢀ1
used). The 38-MNH-RFꢅ molecules dominantly return to the singlet
ground-state by T1–S0 intersystem-crossing (triplet lifetime
38-NH2-RFꢅ ! 8-NH2-RFred2;
ðR8aÞ
or intermolecular photo-reduction
38-NH2-RFꢅ þ HOH ! 8-NH2-RFred2;
ðR8bÞ
similar to (R2a) and (R2b) or (R6a) and (R6b). Also some lumi-
chrome formation according to
sT ꢄ 4.5
ls under our aerobe conditions, see below). The quantum
yield of photo-degradation of 38-MNH-RFꢅin pH 8 10 mM NaPi buf-
fer is estimated to be /D,T = /D//T ꢄ 10ꢀ4 (/D ꢄ 2 ꢃ 10ꢀ5, see
above). During the excited-state lifetime 8-MNH-RF molecules de-
grade to 8-amino-riboflavin by methyl-nitrogen bond fissure at the
dimethylamino group in a formal (buffer catalized) reaction
according to
38-NH2-RFꢅ ! 8-NH2-LC þ ribityl-ene;
ðR9Þ
is expected to occur.
The proposed 8-MNH-RF sensitized RoF photo-degradation is
illustrated in Scheme 4: The intermediate 8-MNH-RF strongly ab-

A. Tyagi et al. / Chemical Physics 369 (2010) 27–36
35
Scheme 3.
sorbs in the blue spectral range. Its photo-excitation causes
intersystem-crossing to 38-MNH-RFꢅ followed by excitation
transfer (ET) [38–42] from 38-MNH-RFꢅ to 1RoF forming
18-MNH-RF and 3RoFꢅ according to
ꢅ
ꢅ
1
1
3
38-MNH-RF þ RoF ! 8-MNH-RF þ RoF :
ðR10Þ
The created triplet roseoflavin then degrades as described above by
(R1), (R2a), (R2b), (R3), (R4). The excitatiꢅon transfer from
38-MNH-RFꢅ to 1RoF with 18-MNH-RF and 3RoF formation may
occur via iso-energetic (resonant) Dexter-type energy transfer
[40–42] with subsequent vibronic relaxation as illustrated in
Fig. 10a, or by combined exothermic reductive and oxidative (cross)
electron transfers [38,39,42] as illustrated in Fig. 10b.
Fig. 10. Illustration of singlet–triplet inter-change (a) by Dexter-type energy
(excitation) transfer, and (b) by combined reductive and oxidative (cross) electron
transfer. DET = Dexter energy transfer. VR = vibrational intra-band relaxation.
ETred = reductive electron transfer. ETox = oxidative electron transfer. Hatched area
indicates vibronic structure of first excited band. Lower levels represent HOMO
states, and upper levels represent LUMO states (HOMO = highest occupied molec-
ular orbital, LUMO = lowest unoccupied molecular orbital).
The 8-MNH-RF intermediate sensitizing of RoF degradation ex-
plains the speed-up of roseoflavin photo-degradation with expo-
sure time: the more 8-MNH-RF is present the more efficient is
the RoF photo-degradation. It explains the wavelength dependence
of the RoF photo-degradation: when the excitation occurred in the
long-wavelength transparency region of 8-MNH-RF, then the
formed 8-MNH-RF could not be excited and the sensitization pro-
cess (R10) could not occur. The 8-MNH-RF sensitized RoF photo-
degradation also explains the decreasing sensitizing efficiency of
RoF photo-degradation with sample dilution: Dexter-type excita-
tion transfer and combined reductiꢅve and oxidative electron trans-
fer occur only when a 38-MNH-RF molecule collides with a 1RoF
molecule. This means that within the triplet lifetime sT of 8-
MNH-RF the molecules RoF and 38-MNH-RFꢅ have to collide by dif-
fusion motion.
1=2
‘d ¼ ð2DsT
Þ
;
ð6Þ
where D is the diffusion constant. D may be obtained from the
Stokes–Einstein equation [43],
kB#
pgam
D ¼
;
ð7Þ
6
where kB is the Boltzmann constant, 0 is the temperature,
viscosity of the solvent, and am is the (solute) molecule radius.
Using experimental parameters of 0 = 277 K,
g
is the
g
= 1.003 ꢃ 10ꢀ3 Pa s
(solvent water), and am = 0.633 nm (average molecule radius of
RoF) a diffusion constant of D = 3.38 ꢃ 10ꢀ6 cm2sꢀ1 is calculated.
In Fig. 9b, a steep rise in RoF photo-degradation sets in at a critical
molecule concentration of Cm,cr ꢄ 10ꢀ5 mol dmꢀ3 corresponding to
The collision condition for sensitized photo-degradation of RoF
may be applied to estimate the triplet lifetime of 8-MNH-RF. The
average (centre–centre) distance acc between two molecules in
solution of number density Nm is acc ꢄ Nꢀm1=3. This distance has to
be overcome by the diffusion length ‘d within the triplet-state life-
a
critical dye number density of Nm,cr = Cm,crNA/1000 ꢄ 6 ꢃ
1015 cmꢀ3 (NA is ꢀA1v=3ogadro constant) and a critical molecule dis-
tance of acc;cr ¼ Nm;cr ꢄ 55 nm. Setting ‘d = acc,cr in Eq. (6) we esti-
time
s
T. ‘d is given by [43]
mate a triplet-state lifetime of sT ꢄ 4.5
ls.
Scheme 4.

36
A. Tyagi et al. / Chemical Physics 369 (2010) 27–36
[2] W. Friedrich, Vitamins, Walter de Gruyter, Berlin, 1988. p. 445.
Under anaerobe conditions the initial quantum yield of photo-
[3] S. Otani, in: T.P. Singer (Ed.), Flavins and Flavoproteins, Elsevier, Amsterdam,
1976, p. 323.
[4] S. Kasai, S. Yamanaka, H.-C. Wang, K. Matsui, J. Nutr. Sci. Vitaminol. 25 (1979)
289.
degradation, /D,0, increased likely because of longer triplet-state
lifetime (longer 3RoFꢅ lifetime enables longer catalytic buffer inter-
action time for degradation). Also the maximum quantum yield of
8-MNH-RF sensitized photo-degradation of RoF, /D,max, increased
since the triplet-state lifetimes of 8-MNH-RF and RoF increased
[5] T. Mathes, C. Vogl, J. Stolz, P. Hegemann, J. Mol. Biol. 385 (2009) 1511.
[6] S. Otani, S. Kasai, K. Matsui, Methods Enzymol. 66 (1980) 235.
[7] K. Yorita, H. Misaki, B.A. Palfey, V. Massey, PNAS 97 (2000) 2480.
[8] R. Flores, A. Dederichs, E. Cerdá-Olmedo, R. Hertel, Plant Biol. 1 (1999) 645.
[9] P. Zirak, A. Penzkofer, T. Mathes, P. Hegemann, J. Photochem. Photobiol. B 97
(2009) 61.
(longer
time
for
collision
excitation
transfer
from
38-MNH-RFꢅ to 1RoF, and longer lifetime of formed 3RoFꢅ). The
photo-degradation sensitizing enhancement under anaerobe con-
ditions excludes the involvement of singlet excited molecular oxy-
gen in the photo-degradation, since 1O2 may be only formed by
excited-state triplet flavin – ground-state triplet oxygen annihila-
[10] M.K. Otto, M. Jayaram, R.M. Hamilton, M. Delbrück, PNAS 78 (1981) 266.
[11] S. Kasai, R. Miura, K. Matsui, Bull. Chem. Soc. Jpn. 48 (1975) 2877.
[12] K. Shiga, K. Horuke, Y. Nishina, S. Otani, H. Watari, T. Yamano, Biochemistry 85
(1979) 931.
[13] V.M. Berezovskii, A.I. Stepanov, N.A. Polyakova, L.S. Tul’chinskaya, A.Ya.
Kukanova, Bioorg. Chem. 3 (1977) 393.
[14] S. Shinkai, K. Kameoka, N. Honda, K. Ueda, O. Manabe, J. Lindsey, Bioorg. Chem.
14 (1986) 119.
[15] K. Matsui, N. Juri, Y. Kubo, S. Kasai, J. Biochem. 86 (1979) 167.
[16] P. Zirak, A. Penzkofer, T. Mathes, P. Hegemann, Chem. Phys. 358 (2009) 111.
[17] P.-S. Song, E.B. Walker, R.D. Vierstra, K.L. Poff, Photochem. Photobiol. 32 (1980)
393.
ꢅ
ꢅ
tion according to 3flavin þ O2 ! flavin þ O2 under aerobe con-
ditions [38,39]. In the case of singlet oxygen involvement, the
opposite behaviour should have been observed (i.e. higher photo-
degradation under aerobe conditions should have occurred).
3
1
1
5. Conclusions
[18] B.J. Fritz, S. Kasai, K. Matsui, Photochem. Photobiol. 45 (1987) 113.
[19] B. Holmström, Arkiv. Kemi. 22 (1964) 281.
[20] M. Schuman Jorns, G. Schöllnhammer, P. Hemmerich, Eur. J. Biochem. 57
(1975) 35.
Roseoflavin in aqueous solution was found to have very low
fluorescence quantum yield and sub-picoseond fluorescence life-
time at room-temperature due to photo-induced intra-molecular
charge-transfer [9,16].
[21] I. Ahmad, Q. Fasihullah, F.H.M. Vaid, J. Photochem. Photobiol. B 75 (2004) 13.
[22] I. Ahmad, Q. Fasihullah, F.H.M. Vaid, J. Photochem. Photobiol. B 78 (2005) 229.
[23] W. Scheidler, A. Penzkofer, Opt. Commun. 80 (1990) 127.
[24] K. Matsui, S. Kasai, in: T.P. Singer (Ed.), Flavins and Flavoproteins, Elsevier,
Amsterdam, 1976, p. 328.
[25] P. Drössler, W. Holzer, A. Penzkofer, P. Hegemann, Chem. Phys. 282 (2002) 429.
[26] S.J. Strickler, R.A. Berg, J. Chem. Phys. 37 (1962) 814.
[27] J.B. Birks, D.J. Dyson, Proc. Roy. Soc. Lond. Ser. A 275 (1963) 135.
[28] A.V. Deshpande, A. Beidoun, A. Penzkofer, G. Wagenblast, Chem. Phys. 142
(1990) 123.
The photo-stability of roseoflavin was studied in un-buffered
Millipore water, in pH 8 and pH 10 phosphate buffer, and in pH
8 Tris–HCl buffer. The highest photo-stability was observed for
Millipore water and the lowest photo-stability was found in phos-
phate buffer. The photo-degradation of roseoflavin is thought to
occur dominantly in the triplet state. The primary photoproduct
of roseoflavin photo-degradation, 8-methylamino-riboflavin,
exhibits reasonably high fluorescence quantum yield and fluores-
cence lifetime. It strongly enhances the degradation of roseoflavin
by photo-induced excitation transfer (Dexter-type and combined
HOMO and LUMO electron transfer) from triplet 8-methylamino-
riboflavin to singlet ground-state roseoflavin with subsequent deg-
radation of the formed triplet roseoflavin.
[29] S.I. Vavilov, Z. Phys. 42 (1927) 311.
[30] G.N. Lewis, M. Kasha, J. Am. Chem. Soc. 66 (1944) 2100.
[31] N.J. Chacon, J. McLearie, R.S. Sinclair, Photochem. Photobiol. 47 (1988) 647.
[32] A. Losi, E. Polverini, B. Quest, W. Gärtner, Biophys. J. 82 (2002) 2627.
[33] S.D.M. Islam, A. Penzkofer, P. Hegemann, Chem. Phys. 291 (2003) 97.
[34] B.J. Fritz, K. Matsui, S. Kasai, A. Yoshimura, Photochem. Photobiol. 45 (1987)
539.
[35] W.L. Cairns, D.E. Metzler, J. Am. Chem. Soc. 93 (1971) 2772.
[36] F. Müller, S. Ghisla, A. Bacher, in: O. Isler, G. Brubacher, S. Ghisla, K. Kräutler
(Eds.), Vitamine II, Thieme Verlag, Stuttgart, 1988, p. 50.
[37] A. Tyagi, P. Zirak, A. Penzkofer, T. Mathes, P. Hegemann, M. Mack, S. Ghisla,
Chem. Phys. 364 (2009) 19.
Acknowledgements
[38] B. Valeur, Molecular Fluorescence: Principles and Applications, Wiley–VCH,
Weinheim, 2002.
[39] N.J. Turro, V. Ramamurthy, J.C. Scaiano, Principles of Molecular
Photochemistry: An Introduction, University Science Books, Sausalito, CA,
2009.
[40] D.L. Dexter, J. Chem. Phys. 21 (1953) 836.
[41] S. Speiser, Chem. Rev. 96 (1996) 1953.
The work was supported by the Deutsche Forschungsgemeins-
chaft (DFG): GK640 (A.P.), FOR526 (A.P., P.H.), SFB 498 (P.H.) and
the Cluster of Excellence ‘‘Unifying Concepts in Catalysis” (P.H.).
A.P. is grateful to Profs. F.J. Gießibl and J. Repp for their kind
hospitality.
[42] J. Shirdel, A. Penzkofer, R. Procházka, Z. Shen, J. Daub, Chem. Phys. 336 (2007)
1.
[43] P. Atkins, J. de Paula, Atkin’s Physical Chemistry, eighth ed., Oxford University
Press, Oxford, 2006.
References
[1] K. Matsui, S. Kasai, in: F. Müller (Ed.), Chemistry and Biochemistry of
Flavoenzymes, vol. 1, CRC Press, Boca Raton, FL, 1991, p. 105.