Effect of caffeine complexation on the photolysis of riboflavin in aqueous solution: A kinetic study

Article abstract of DOI:10.1248/cpb.57.1363

The effect of caffeine complexation with riboflavin on the kinetics of riboflavin photolysis in the pH range 2.0-10.5 has been studied. The photolysis of riboflavin solutions (5×10^-5 M) was carried out in the presence of caffeine (0.5-2.5x10

Full text of DOI:10.1248/cpb.57.1363

December 2009
Chem. Pharm. Bull. 57(12) 1363—1370 (2009)
1363
Effect of Caffeine Complexation on the Photolysis of Riboflavin in
Aqueous Solution: A Kinetic Study
a
a
,a
a
Iqbal AHMAD, Sofia AHMED, Muhammad Ali SHERAZ,* Muhammad AMINUDDIN, and
b
Faiyaz Hussain Madni VAID
a
Institute of Pharmaceutical Sciences, Baqai Medical University; Toll Plaza, Super Highway, Gadap Road,
b
Karachi–74600, Pakistan: and Department of Pharmaceutical Chemistry, Faculty of Pharmacy, University of Karachi;
Karachi–75270, Pakistan. Received June 6, 2009; accepted July 21, 2009; published online September 14, 2009
The effect of caffeine complexation with riboflavin on the kinetics of riboflavin photolysis in the pH range
ꢁ
5
2
.0—10.5 has been studied. The photolysis of riboflavin solutions (5ꢀ10 M) was carried out in the presence of
ꢁ
4
caffeine (0.5—2.5ꢀ10 M) using a visible radiation source. A specific multicomponent spectrophotometric
method was used for the determination of riboflavin and photoproducts in photolysed solutions. The apparent
ꢁ4
ꢁ2
ꢁ1
first-order rate constants (k) for the photolysis reactions range from 2.71ꢀ10 to 4.26ꢀ10 min . The values
of the rate constants decrease with increasing concentrations of caffeine indicating its inhibitory effect on the re-
actions. The second-order rate constants (kꢂ) for the caffeine inhibited reactions lie in the range of 0.13 to
ꢁ
3
ꢁ1
ꢁ1
5
.10ꢀ10
M
min . The log k–pH profiles for the photolysis reactions at various caffeine concentrations involve
multiple steps indicating a gradual increase in the rate up to pH 10. The lower rates at pH 2.0 and 10.5 are due to
the ionization of riboflavin as evident from fluorescence measurements. The kꢂ–pH profile for the interaction of
riboflavin with caffeine represents a bell-shaped curve in the pH range 3—6 followed by a sigmoid curve in the
pH range 7—10. The inhibition of photolysis of riboflavin in the presence of caffeine is a result of the monomeric
interaction and complex formation of caffeine with riboflavin. The photochemical interaction of riboflavin with
caffeine suggests that a pH around 6 is most appropriate for the stabilization of the vitamin. At this pH the com-
plex shows the highest stability constant.
Key words photolysis; kinetics; complexation; riboflavin; caffeine
5)
ꢁ5
Riboflavin is a highly photosensitive compound and is de- a similar study. The photolysis of 5ꢀ10 M riboflavin (RF) solutions was
ꢁ
4
1
—3)
carried out in the presence of 0.5—2.5ꢀ10 M caffeine using a Philips
graded in aqueous solution by several mechanisms
in-
4,5)
HPLN 125 W high pressure mercury vapor fluorescent lamp (emission at
cluding intramolecular photoreduction and intramolecular
9)
4
05 and 435 nm) as the irradiation source. The spectral measurements on
RF and degraded solutions were performed on a Shimadzu UV-1601 record-
ionic strength, buffer ing spectrophotometer using silica cells of 10 mm path length. The intensity
and light intensity and wave- of the fluorescent lamp was determined by potassium ferrioxalate actinome-
6—9)
4,5,10—12)
photoaddition.
These reactions are affected by pH,
1
3,14)
15)
solvent polarity and viscosity,
6
—9,16)
kind and concentration
6
—9,17)
lengths.
Various methods including the use of photosta-
bilizers, incorporation into liposomes and cyclodextrin com-
plexation have been used to stabilize riboflavin solutions in
1
8—21)
the presence of light.
Another approach to achieve sta-
bilization of drugs is through complexation with caffeine and
2
2—26)
other agents.
complexes with caffeine
ence the rate of its chemical and photochemical degrada-
Riboflavin is known to form molecular
2
7—32)
which has been found to influ-
3
3)
17)
tion. Riboflavin–caffeine complexation may improve the
bioavailability and therapeutic activity of the vitamin.
NMR studies have been conducted to understand the nature
34)
35)
36)
of association between caffeine and flavin mononucleotide
that may help in the understanding of the mode of riboflavin
stabilization. The present work is based on a detailed study
of the kinetics of riboflavin photolysis in the presence of caf-
feine over a wide range of pH using a specific multicompo-
12)
nent spectrophotometric method for the determination of
riboflavin and photoproducts. The kinetic data have been
used to determine the stability constants for riboflavin–caf-
feine complex. The quantitative and kinetic evaluation of
such interactions may suggest the possibility of achieving the
stabilization of photolabile compounds and thus enhance
their biological action. The chemical structures of riboflavin,
its photoproducts and caffeine are shown in Fig. 1.
Experimental
The materials and methods used are the same as previously described for Fig. 1. Chemical Structures of Riboflavin, Its Photoproducts and Caffeine
∗
To whom correspondence should be addressed. e-mail: ali_sheraz80@hotmail.com
© 2009 Pharmaceutical Society of Japan

1
364
Vol. 57, No. 12
37)
17
ꢁ1
try as 1.15ꢂ0.10ꢀ10 quanta s
.
The fluorescence measurements were performed on a Versa Fluor fluo-
rimeter (Bio-Rad Laboratories, U.S.A.) equipped with a 20 W quartz halo-
gen lamp (350—390 nm), a photodiode detector and excitation and emission
filters. The measurements were carried out at 25—27 °C using 10 mm quartz
cells. The 0.05 mM RF solutions without caffeine and maintained at a partic-
ular pH were used to calibrate the fluorescence intensity scale of the instru-
ment. The excitation wavelength (374 nm) and the emission wavelength
10)
(
525 nm) were set by using appropriate filters and the relative fluorescence
intensity of RF solutions in the presence of caffeine in the pH range 2.0—
0.5 was measured.
1
Results and Discussion
Riboflavin Photoproducts In order to confirm the pres-
ence of known photoproducts of RF in the degraded so-
lutions (30—50% photolysis), thin-layer chromatography
(TLC) was performed on cellulose plates using the solvent
systems: (A) 1-butanol–acetic acid–water (40 : 10 : 50, v/v,
organic phase) and (B) 1-butanol–1-propanol–acetic acid–
38)
water (50 : 30 : 2 : 18, v/v). The photoproducts were identi-
fied by their characteristic fluorescence under UV (365 nm)
excitation and comparison of the Rf values with those of
the reference standards as follows (the fluorescence is given
in parenthesis):
Fig. 2. UV and Visible Absorption Spectra of (a) the Aqueous Phase (pH
2
4
.0) at 0, 15, 35, 60, 90 and 120 min and (b) the Chloroform Extract (pH
ꢁ5
.5) at 15, 35, 60, 90 and 120 min during the Photolysis of 5ꢀ10 M Ri-
ꢁ4
Acid Photolysis (pH 2.0—6.0). Major: RF (undegraded)
and formylmethylflavin (FMF, yellow green), lumichrome
boflavin Solution at pH 8.0 in the Presence of 2.5ꢀ10 M Caffeine
(LC, sky blue). Minor: Carboxymethylflavin (CMF, yellow mophoric system of RF causing a decrease in absorbance at
green).
the maxima. A typical set of the absorption spectra of RF so-
Neutral and Alkaline Photolysis (pH 7.0—10.5). Major: lutions photolysed at pH 8.0 in the presence of caffeine and
RF (undegraded), FMF, LC and lumiflavin (LF, yellow measured after chloroform extraction at pH 2.0 (aqueous
green). Minor: CMF and trace amounts of unknown com- phase) (Fig. 2a) showed gradual disappearance of the 445 nm
pounds with yellow green fluorescence.
band with the progress of photolysis indicating the formation
The products detected in this study are known and have of the major final product, LC. This is evident from the grad-
previously been identified in the photolysis of RF and FMF ual increase in absorbance at the maximum of this compound
3
9)
12)
(
an intermediate in this reaction) by Smith and Metzler,
(356 nm) in the chloroform extract determined at pH 4.5.
40)
5,12,14,38)
Treadwell et al. and Ahmad et al.
The confirmation An increase in the concentration of caffeine results in a
of the products formed on the photolysis of RF in the pres- slower decrease in absorbance at 445 nm in the aqueous
ence of caffeine is necessary to carry out the assay of RF and phase indicating the inhibition of RF photolysis as a result of
the photoproducts in degraded solutions. The intensity of the complexation. A similar change in the formation of LC is ob-
spots of the photoproducts in degraded solutions appeared to served due to a relative increase in absorbance at 356 nm in
decrease with an increase in the concentration of caffeine in- the chloroform extract (Fig. 2b). The overall spectral changes
dicating its inhibitory effect on the rate of photolysis of RF of the photolysed solutions are similar to those observed by
44)
in the pH range studied.
Spectral Characteristics of Fresh and Degraded Solu- aqueous solution.
tions of RF The absorption spectra of aqueous solutions of Assay of RF and Photoproducts The assay of RF and
RF (5ꢀ10 M) containing various concentrations of caffeine photoproducts (FMF, LC, LF) was carried out by a specific
Holzer et al. for the photo-induced degradation of RF in
ꢁ5
ꢁ
4
(
5
0.5—2.5ꢀ10 M) at pH 2.0—10.5 show variations in the multicomponent spectrophotometric method developed by
12)
00 to 300 nm region before irradiation. A loss of ab- Ahmad and Rapson and previously applied to the study of
sorbance of RF solutions in the presence of caffeine at pH RF and FMF photolysis reactions.
5
,7,11,14,38)
In the presence of
3
.0—10.0 has been observed at the absorption maxima of RF caffeine the assay results could be affected due to the loss of
41)
at 444 and 373 nm to the extent of 0.2—1.0%. The loss of absorbance at the analytical wavelengths as discussed above
absorbance at these wavelengths at pH 2.0 and 10.5 due to on the spectral characteristics of RF solutions. The assay
the ionized forms of RF (pK 1.9, pK 10.2) is greater and method is based on preadjustment of the photolysed solu-
42)
a1
a2
is in the range of 0.4—2.0%. A similar change in absorbance tions to pH 2.0 (HCl–KCl buffer) and extraction with chloro-
has been observed at 385 nm, the wavelength, along with form to remove LC and LF. The chloroform is evaporated
445 nm, used for the assay in this study. It has been observed and the residue dissolved at pH 4.5 (acetate buffer) to deter-
that the addition of caffeine in aqueous solutions of RF mine LC and LF by a two-component assay at 356 and
causes a decrease in the intensity of the absorption at 375 445 nm. Similarly the aqueous phase is used to determine RF
4
3)
33)
35)
and 450 nm and at 445 nm and 446 nm. These changes and FMF by a two-component assay at 385 and 445 nm. The
12)
were considered to occur due to the formation of an inter- reproducibility of the method is of the order of ꢂ5%.
molecular complex between the isoalloxazine ring of RF and Caffeine is highly soluble in chloroform (1 g/5.5 ml) and
the aromatic ring of caffeine. This may affect the chro- exists from 3.8 to 6.1% in the protonated form at pH 2.0

December 2009
1365
4
2,45,46)
(
amine pK 0.6).
On the extraction of photolysed solu- various reactions under the same experimental conditions
a
tions of RF at pH 2.0, 94—96% of the added caffeine would (i.e. pH, analytical wavelengths, irradiation conditions). This
be transferred to chloroform as indicated by an increase is necessary to avoid any variations in the use of kinetic data
in absorption around 273 nm (absorption maximum of caf- for comparative purposes. A kinetic plot for the photolysis of
42)
feine) on measurement at pH 4.5 (Fig. 2b). Since the aque- RF in the presence of caffeine at pH 10.0 is shown in Fig. 3.
ous phase (pH 2.0) (Fig. 2a) contains only about 4—6% of During photolysis, the RF values gradually decreased with a
the added caffeine, there would be negligible effect on the concomitant increase in FMF (up to 40 min) and LC and LF
absorption of the mixture of RF and FMF at the analytical values. The highest values are obtained for LC which is
wavelength (385, 445 nm) as observed in this work.
formed by the degradation of FMF (an intermediate in this
To examine the effect of caffeine on the absorption of the reaction), along with LF. The photolysis of RF in aqueous so-
5)
mixture of LC and LF, a RF solution (pH 9.0) photolysed in lution follows first-order kinetics and the values of the ap-
the absence of caffeine was extracted with chloroform and on parent first-order rate constants, k , for the photolysis reac-
obs
evaporation the residue dissolved in pH 4.5 acetate buffer. tions carried out at various pH values are reported in Table 2.
The absorbance of the solution was measured in the presence It is evident from these values that the rate of photolysis is
and absence of the highest concentration of caffeine inhibited in the presence of caffeine. Further treatment of the
ꢁ4
(
2.5ꢀ10 M) used in the photolysis reactions. A loss of ab- kinetic data by plotting the values of kobs versus caffeine con-
sorbance to the extent of 3% and 1% was observed at the an- centration at all the pH values resulted in straight lines indi-
alytical wavelengths of 356 and 445 nm, respectively, sug- cating that the rate of photolysis is a linear function of caf-
gesting some interaction between caffeine and LC and caf- feine concentration in the range studied (Fig. 4). The slopes
feine and LF. Since LC is the major final product in this reac- of these plots yielded the second-order rate constants (kꢃ) for
tion, there is a greater possibility of interaction between caf- the caffeine inhibited reaction (Table 3). The values of k0
feine and LC. In view of the absorbance loss in the chloro- were determined by back-extrapolation of the second-order
form extract in the presence of caffeine, the analytical data plots to a zero concentration on the vertical axis (Fig. 4) and
have been corrected for LC and LF concentrations to account are reported in Table 3. These values are about one and a half
for the respective absorbance loss at the two wavelengths. A to four times greater than those obtained at the highest caf-
ꢁ4
typical assay of RF and photoproducts in the solutions pho- feine concentration (2.5ꢀ10 M) except at pH 6.0 (about
tolysed at pH 9.0 in the presence of caffeine is reported in one and a quarter times greater) and indicate the gradual in-
Table 1, indicating a gradual decrease in the concentration of hibitory effect of caffeine on the rate of reaction.
RF and an increase in the concentrations of FMF (up to
0 min and then its gradual loss to form LC and LF) and LC
and LF, with time, giving an almost constant molar balance.
Rate–pH Profiles The present work involves the study
4
ꢁ5
Table 1. Photolysis of 5ꢀ10
M RF Solution in the Presence of
ꢁ4
CMF is a minor photoproduct formed on the oxidation of 2.5ꢀ10 M Caffeine at pH 9.0 Concentrations of RF and Photoproducts
40)
FMF and its presence does not appear to affect the assay
results as the molar balance of RF and photoproducts is al-
most constant during the reactions. Control solutions of RF
stored in the dark at all pH values during the photolysis reac-
tions (120—240 min at pH 2.0—8.0, 40—60 min at pH
Time
RF
(Mꢀ10 )
FMF
(Mꢀ10 )
LF
(Mꢀ10 )
LC
(Mꢀ10 )
Total
(Mꢀ10 )
5
5
5
5
5
(min)
0
10
0
0
0
0
5.00
3.75
2.84
2.17
1.64
1.31
0.00
0.29
0.42
0.49
0.50
0.48
0.00
0.09
0.17
0.22
0.26
0.34
0.00
0.89
1.60
2.12
2.58
2.84
5.00
5.02
5.03
5.00
4.98
4.97
2
3
4
5
9.0—10.5) did not show any change in absorbance at the an-
alytical wavelengths.
Kinetics of RF Photolysis The kinetics of RF photolysis
has been evaluated by using the analytical data obtained for
ꢁ5
ꢁ4
Fig. 3. Photolysis of 5ꢀ10 M Riboflavin Solution at pH 10.0 in the Presence of 2.5ꢀ10 M Caffeine

1
366
Vol. 57, No. 12
Table 2. First-Order Rate Constants (kobs) for the Photolysis of Riboflavin Table 3. First-Order Rate Constants (k ) for the Photolysis of Riboflavin in
0
ꢁ4
at pH 2.0—10.5 in the Presence of Caffeine (0.5—2.5ꢀ10 M)
the Absence of Caffeine and Second-Order Rate Constants for the Photoly-
sis of Riboflavin (kꢃ) in the Presence of Caffeine
2
ꢁ1
kobsꢀ10 min
2
ꢁ2
pH
k0ꢀ10
(min
kꢃꢀ10
(M min )
Correlation
coefficient
pH
a)
a)
a)
a)
a)
ꢁ1
ꢁ1
ꢁ1
0
.5
1.0
1.5
2.0
2.5
)
2
3
4
5
6
7
8
9
.0
0.052
0.997)
0.067
0.998)
0.263
0.996)
0.357
0.999)
0.304
0.996)
0.351
0.998)
1.270
0.997)
3.650
0.999)
4.260
0.998)
3.439
0.998)
0.046
(0.999)
0.059
(0.998)
0.226
(0.997)
0.295
(0.997)
0.294
(0.999)
0.329
(0.998)
1.179
(0.999)
3.402
(0.998)
4.032
0.040
(0.998)
0.050
(0.999)
0.181
(0.996)
0.243
(0.999)
0.284
(0.999)
0.301
(0.996)
1.091
(0.999)
3.160
(0.998)
3.840
0.032
(0.999)
0.039
(0.996)
0.131
(0.997)
0.192
(0.997)
0.274
(0.997)
0.276
(0.996)
0.995
(0.998)
2.894
(0.999)
3.503
0.027
(0.998)
0.030
(0.999)
0.088
(0.996)
0.141
(0.996)
0.266
(0.997)
0.245
(0.996)
0.920
(0.997)
2.671
(0.999)
3.250
2.0
3.0
4.0
5.0
6.0
7.0
8.0
9.0
10.0
10.5
0.056
0.078
0.310
0.406
0.314
0.380
1.360
3.890
4.540
3.650
0.013
0.019
0.089
0.105
0.019
0.053
0.177
0.468
0.510
0.458
0.998
0.997
0.998
0.999
0.998
0.997
0.999
0.999
0.999
0.996
(
(
(
(
(
(
(
(
(
(
.0
.0
.0
.0
.0
.0
.0
5)
RF in the absence of caffeine. The profiles indicate gradu-
ally rising steps and a pH-independent plateau extending
over the pH range 5—7. The rates are lowest at pH 2.0 and
1
1
0.0
0.5
decrease at pH 10.5 due to ionization of RF (pK 1.9, pKa2
a1
(0.997)
3.220
(0.998)
(0.997)
2.941
(0.997)
(0.998)
2.772
(0.996)
(0.997)
2.520
(0.996)
42)
1
0.2). The ionized forms of the molecule are less suscepti-
5)
ble to photolysis than the non-ionized form. The increase in
the rate of photolysis in the pH range of 2—5 is due to the
4
a) Caffeine concentration (Mꢀ10 ). The values in parenthesis are correlation coeffi- gradual deprotonation of the molecule (pK 1.9, Nꢁ10) and
a1
cients.
in the alkaline range as a result of the increased reactivity of
4)
the triplet state. The presence of a plateau in the pH range
—5 may be explained on the basis of the lowest redox po-
tentials of RF in this range (E° pH 5.0ꢄꢁ0.117, pH
.0ꢄꢁ0.207 V, pH 8.6ꢄꢁ0.274 V)
action involved in the photolysis of RF is photoreduction.
4
47,48)
7
since the primary re-
3)
Thus, the ionization state, redox potentials and triplet state
reactivity appear to influence the rate of photolysis of RF
over the pH range studied. These characteristics may also
play a significant role in the photolysis of RF in the presence
of caffeine and thus influence the rate of reaction. The
rate–pH profile for the photochemical interaction of RF and
caffeine represents a bell-shaped curve in the pH range 3—6
followed by a sigmoid curve in the pH range 7—10 (Fig. 6).
The initial step of the bell-shaped curve (around pH 2—5)
may result from the gradual increase in the non-ionized form
of RF and a subsequent greater interaction between RF and
caffeine. This is followed by a decrease in the rate (around
pH 5—6) which may be due to the involvement of an inter-
mediate species and consequently a change in the rate-deter-
mining step of the reaction as observed in the hydrolysis of
49)
hydrochlorothiazide. The involvement of an intermediate
species in this range may be substantiated by the fact that RF
shows lower reactivity (lowest redox potentials around pH
5—6) in this region and caffeine has greater influence (high-
est stability constant around pH 6) in suppressing the rate of
reaction. The sigmoid curve extending over the pH range 7—
10 suggests a lower interaction of caffeine with RF vis-à-vis
greater reactivity of RF triplet state in enhancing the rate of
reaction. Above pH 10 the ionization of RF (Nꢁ3) appears
to be the major factor in causing the decrease in the rate of
reaction. The photolysis of RF in the presence of caffeine
Fig. 4. Plots of kobs versus Caffeine Concentration for the Photolysis of
Riboflavin at pH 2.0—10.5
of the kinetic behavior of RF in the presence of caffeine over suggests that the molecule is most stable at a pH value
a wide range of pH. The rate–pH profiles (Fig. 5) depict the around 6. This is in agreement with a previous study on the
5)
highest rate at a pH value around 10, with a gradual decrease photolysis of pure RF solutions.
on increasing the caffeine concentration at all pH values and Fluorescence Studies Aqueous solutions of RF exhibit
exhibit a pattern similar to that observed for the photolysis of intense yellow green fluorescence which is destroyed on the

December 2009
1367
4
Fig. 5. log kobs–pH Profiles for the Photolysis of Riboflavin in the Presence of Caffeine (Mꢀ10 )
Fig. 6. kꢃ–pH Profile for the Photolysis of Riboflavin in the Presence of Caffeine
1
,41,43,50)
ꢁ5
addition of mineral acids or alkalis.
In the present Table 4. Fluorescence Intensity of 5ꢀ10 M Riboflavin Solutions at pH
ꢁ4 a)
2
.0—10.5 in the Presence of 0.5—2.5ꢀ10 M Caffeine
work a loss of RF fluorescence has been observed in the
presence of caffeine at pH 2.0 (5%) and 10.5 (18%) (Table 4)
compared to that of the pure solutions. This may result from
the formation of a complex between RF and caffeine as re-
Caffeine Relative fluorescence intensity at pH
concentration
4
2
7,30,33,43,51)
(Mꢀ10 )
ported by previous workers.
The relatively greater
2.0
4.0
7.0
9.0
10.5
loss of fluorescence at pH 2.0 and 10.5, compared to the neg-
43,50)
0
0
1
1.5
2.0
2
100.0
98.5
97.2
96.4
95.6
95.0
100.0
100.0
100.0
99.7
99.4
99.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
99.8
100.0
95.2
90.6
87.5
84.8
82.2
ligible loss at pH 4—9, is due to the ionization of RF,
in
.5
.0
addition to complex formation. The loss of RF fluorescence
due to complex formation probably involves the interaction
of the chromophoric systems of RF and caffeine and thus
may affect the RF excitation process resulting in the inhibi-
tion of the rate of photolysis.
.5
99.5
a) Measurements were made on 1 : 10 dilute solutions using appropriate buffers at all
Stability Constants of RF–Caffeine Complex The pH values.
complexation of RF with caffeine may be represented by the
following reaction:

1
368
RFꢅcaffeine
Vol. 57, No. 12
→
RF: caffeine
(1) of RF in this region. The redox potentials of RF increase
←
53)
with pH and appear to influence the rate of reaction and
The stability constants of 1 : 1 RF–caffeine complex can the interaction with caffeine. In the pH range 7—10 the K
be determined by using the following Eq. 2 developed by values are decreasing while the kꢃ values are increasing with
22)
Carstensen :
pH. In this region the predominant factor appears to be the
4)
RF triplet state reactivity leading to a greater increase in kꢃ
values although the K values are decreasing in this region.
Thus several factors such as RF ionization, redox potentials
and triplet state reactivity contribute to the values of the sta-
bility constants in the pH range studied.
k ꢅk*K[B]
(k*ꢁk)K[B]
1ꢅ K[B]
kobs
ꢄ
ꢄk ꢅ
(2)
(
1ꢅ K[B])
which may be expressed reciprocally by Eq. 3 as:
Structure of 1 : 1 RF–Caffeine Complex Caffeine is a
hydrotropic agent and it forms a hetero-association complex
1
1
1
1
ꢄ
ꢅ
⋅
(3)
kobsꢁk
k*ꢁk
K(k*ꢁk) [B]
where k is the rate constant for the degradation of RF (k0),
kobs is the overall rate constant for the degradation of RF in
the presence of caffeine, k* is the rate constant for the degra-
dation of RF–caffeine complex (k ꢁk), [B] is the molar
obs
concentration of caffeine in solution, and K is the stability
constant of RF–caffeine complex.
The value of K can be obtained from a linear plot of
1
/(k ꢁk) versus 1/[caffeine] according to Eq. 3 as the slope-
obs
to-intercept ratio as shown for the photolysis reaction at pH
6.0 (Fig. 7).
The values of K for 1 : 1 RF–caffeine complex, determined
at pH 2.0—10.5 (Table 5), range from 0.121—2.490ꢀ
ꢁ
2
ꢁ1
1
0
M
and vary with pH (Fig. 8). The highest value is
ꢁ2
ꢁ1
close to the value (2.72ꢀ10
M
, 25 °C) determined by
3
3)
Fig. 7. A Plot of 1/(kobsꢁk) versus 1/[caffeine] for the Photolysis of Ri-
boflavin at pH 6.0
Guttman in alkaline solution (0.05 M NaOH). A compari-
son of the graphs of second-order rate constants (kꢃ) for the
interaction of RF and caffeine versus pH (Fig. 6) and the
graph of the stability constants (K) versus pH (Fig. 8) indi-
cates a nearly inverse relationship between the two values.
The K values increase at pH 2.0 and 10.5 (cationic and an-
ionic forms of RF, respectively) probably due to higher inter-
action (e.g. dipole–dipole interaction, van der Waal’s force,
hydrogen bonding) between the two molecules. These val-
ues are lower in the pH range 3—5 probably due to lower in-
teraction and reach the maximum value at pH 6 vis-à-vis the
lowest kꢃ value at that pH due to the lowest redox potentials
Table 5. Stability Constants of Riboflavin–Caffeine Complex (K) at pH
2
.0—10.5
pH
2.0
2
ꢁ1
2
ꢁ1
Kꢀ10 (M
)
pH
Kꢀ10 (M
)
1.533
0.121
0.140
0.614
2.488
7.0
8.0
9.0
10.0
10.5
0.140
0.554
0.415
0.175
0.812
52)
3
4
5
6
.0
.0
.0
.0
Fig. 8. A Plot of Stability Constants for Riboflavin–Caffeine Complex versus pH

December 2009
1369
with flavin mononucleotide (FMN) in aqueous solution as a in the presence of caffeine at a pH value around 6.
result of vertical stacking interaction between the aromatic
Acknowledgements The authors are grateful to Prof. Dr. Sheikh Arshad
Saeed of Dr. Panjwani Centre for Molecular Medicine and Drug Research,
University of Karachi, for providing facilities to carry out fluorescence
chromophores of the two molecules. The possible structure
of the 1 : 1 FMN–caffeine complex has been studied by mo-
lecular dynamics simulations and an analysis of the induced measurements. They are thankful to Dr. Masood Chowhan of Alcon Re-
36)
proton chemical shifts for these molecules. Similar hetero- search Ltd., Fort Worth, Texas, for helpful suggestions.
54)
association complexes of caffeine with nucleic acids, acti-
References
5
5)
56)
57,58)
nomycin D, theaflavin, novatrone (mitoxantrone)
and
1) Heelis P. F., Chem. Soc. Rev., 11, 15—39 (1982).
2) Heelis P. F., “Chemistry and Biochemistry of Flavoenzymes,” Vol. 1,
ed. by Muller F., CRC Press, Boca Raton, FL, 1991, pp. 171—193.
5
9—60)
aromatic drugs
have been studied by NMR spec-
troscopy. It has been suggested that the imidazole moiety of
caffeine is involved in the interaction with aromatic
3)
Ahmad I., Vaid F. H. M., “Flavins Photochemistry and Photobiology,”
ed. by Silva E., Edwards A. M., Royal Society of Chemistry, Cam-
bridge, 2006, pp. 13—40.
25)
species. The structure of RF–caffeine complex may be sta-
bilized by hydrogen bonds between caffeine oxygens (in po-
sitions 2 and 6) and NH group (in position 3) of riboflavin as
4) Cairns W. L., Metzler D. E., J. Am. Chem. Soc., 93, 2772—2777
(1971).
5) Ahmad I., Fasihullah Q., Noor A., Ansari I. A., Ali Q. N. M., Int. J.
Pharm., 280, 199—208 (2004).
57)
observed in the case of caffeine–novatrone complex. Inter-
molecular hydrogen bonding has also been suggested in the
6)
Schuman Jorns M., Schollnhammer G., Hemmerich P., Eur. J.
Biochem., 57, 35—48 (1975).
61)
interaction of caffeine with biological compounds. The
crystal structure of caffeine complexes with inorganic bro-
mides and iodides has indicated the presence of a network of
7) Ahmad I., Fasihullah Q., Vaid F. H. M., J. Photochem. Photobiol. B:
Biol., 75, 13—20 (2004).
8) Ahmad I., Fasihullah Q., Vaid F. H. M., J. Photochem. Photobiol. B:
Biol., 78, 229—234 (2005).
62)
hydrogen bonds between the two molecules which may sta-
bilize these complexes.
9)
Ahmad I., Fasihullah Q., Vaid F. H. M., J. Photochem. Photobiol. B:
Biol., 82, 21—27 (2006).
It is evident from the above discussion that caffeine has
been used to form complexes with a number of compounds 10) Song P. S., Metzler D. E., Photochem. Photobiol., 6, 691—709 (1967).
1
1) Heelis P. F., Phillips G. O., Ahmad I., Rapson H. D. C., Photobiochem.
Photobiophys., 1, 125—130 (1980).
including RF. However, the efficacy of these complexes to
stabilize a drug depends on their relative strength and chemi-
cal reactivity. The complexation of RF with caffeine may im-
1
2) Ahmad I., Rapson H. D. C., J. Pharm. Biomed. Anal., 8, 217—223
(1990).
34)
35)
prove the bioavailability and therapeutic activity of the 13) Ahmad I., Tollin G., Biochemistry, 20, 5925—5928 (1982).
vitamin. Most of the investigations conducted on caffeine 14) Ahmad I., Fasihullah Q., Vaid F. H. M., Photochem. Photobiol. Sci., 5,
6
80—685 (2006).
5) Sato Y., Chaki H., Suzuki Y., Chem. Pharm. Bull., 32, 1232—1235
1984).
6) Ahmad I., Ahmed S., Sheraz M. A., Vaid F. H. M., J. Photochem. Pho-
tobiol. B: Biol., 93, 82—87 (2008).
complexes by earlier workers are based on structural consid-
erations. The present study provides a detailed account of the
kinetic behavior of the RF–caffeine complex over a wide
range of pH and may be useful in similar studies.
1
1
1
1
1
(
7) Sato Y., Yokoo M., Takahashi S., Takahashi T., Chem. Pharm. Bull.,
3
0, 1803—1810 (1982).
8) Asker A. F., Habib M. J., Drug Dev. Ind. Pharm., 16, 149—156
1990).
Conclusion
The rate of photolysis of RF in the pH range 2.0—10.5 is
inhibited by caffeine as a result of 1 : 1 complex formation
between the two molecules as indicated by the kinetic data.
(
9) Habib M. J., Asker A. F., J. Parenter. Sci. Technol., 45, 124—127
(1991).
The log k –pH profiles for the photolysis of RF in the pres- 20) Loukas Y. L., Jayasekera P., Gregoriadis G., Int. J. Pharm., 117, 85—
obs
9
4 (1995).
1) Loukas Y. L., Vraka V., Gregoriadis G., Int. J. Pharm., 144, 225—231
1996).
2) Carstensen J. T., “Drug Stability Principles and Practices,” 3rd ed., ed.
by Carstensen J. T., Rhodes C. T., Marcel Dekker, New York, 2000, pp.
ence of varying concentrations of caffeine are similar to that
obtained in the absence of caffeine and involve multiple steps
with a gradual increase in the rate reaching a maximum value
around pH 10. The increase in caffeine concentration results
in a gradual decrease in the rate of reaction at all pH values.
The lower rate of photolysis at pH 2 and above pH 10 ap-
pears to be due to the cationic and anionic forms of RF and a
lower susceptibility of the ionized form to undergo photoly-
sis in the presence of caffeine.
2
2
(
1
34—139.
2
3) Yoshioka S., Stella V. J., “Stability of Drugs and Dosage Forms,”
Kluwer Academic/Plenum Publishers, New York, 2000, pp. 126—133.
4) Amiji M. M., Sandman B. J., “Applied Physical Pharmacy,” McGraw-
Hill, New York, 2003, p. 264.
2
25) Florence A. T., Attwood D., “Physicochemical Principles of Phar-
macy,” 4th ed., Pharmaceutical Press, London, 2006, pp. 405—413.
The kꢃ–pH profile for the photochemical interaction of RF
and caffeine represents a bell-shaped curve in the acid region
and a sigmoid curve in the alkaline region. The photochemi-
26) Sinko P. J., “Martin’s Physical Pharmacy and Pharmaceutical Sci-
ences,” 5th ed., Lippincott Williams & Wilkins, Baltimore, 2006, pp.
2
70—277.
cal interaction of the two compounds may involve an inter- 27) Guttman D. E., Athalye M. Y., J. Am. Pharm. Assoc. Sci. Ed., 49,
mediate species in the rate-determining step resulting in the
suppression of the rate of reaction in the pH range 5—6. The
increase in the rate in the pH range 7—10 is due to the
greater reactivity of the RF triplet state in this region. The ki-
687—691 (1960).
2
8) Tsibris J. C. M., McCormick D. B., Wright L. D., Biochemistry, 4,
5
04—510 (1965).
2
3
9) Wilson J. E., Biochemistry, 5, 1351—1359 (1966).
0) Penzer G. R., Radda G. K., Q. Rev., 21, 43—65 (1967).
netic behavior of RF in the pH range 3—9 appears to be due 31) DeRitter E., J. Pharm. Sci., 71, 1073—1096 (1982).
to the interaction of non-ionized form of the molecule with 32) Rivlin R. S., “Handbook of Vitamins,” 4th ed., ed. by Zempleni J.,
Rucker R. B., McCormick D. B., Suttie J. W., CRC Press, Boca Raton,
FL., 2007, p. 240.
3) Guttman D. E., J. Pharm. Sci., 51, 1162—1166 (1962).
4) McCormick D. B., “Present Knowledge in Nutrition,” 6th ed., ed. by
Brown M. L., International Life Sciences Institute, Washington, D.C.,
caffeine. The photolysis of RF in the presence of caffeine is
influenced by complex formation, ionization state, redox po-
tentials, triplet state reactivity and fluorescence characteris-
tics in the pH range studied. RF is most stable to photolysis
3
3

1
370
Vol. 57, No. 12
1
990, p. 146.
5) Bolotin P. A., Baranovsky S. F., Chernyshev D. N., Evstigneev M. P., 49) Mollica J. A., Rehm C. R., Smith J. B., Goven H. K., J. Pharm. Sci.,
Int. J. Phys. Sci., 2, 68—72 (2007). 60, 1380—1384 (1971).
6) Evstigneev M. P., Rozvadovskaya A. O., Hernandez Santiago A. A., 50) Drossler P., Holzer W., Penzkofer A., Hegemann P., Chem. Phys., 282,
Williams & Wilkins, Baltimore, 1960, p. 444.
3
3
Mukhina Yu V., Veselkov K. A., Rogovo O. V., Davies D. B., Veselkov
429—439 (2002).
A. N., Russ. J. Phys. Chem., 79, 573—578 (2005).
51) Sakai K., Nagoya J. Med. Sci., 18, 245—251 (1956).
3
3
3
4
4
4
7) Hatchard C. G., Parker C. A., Proc. R. Soc. (Lond.), A235, 518—536 52) Connors K. A., “Remington: The Science and Practice of Pharmacy,”
(
1956).
8) Ahmad I., Rapson H. D. C., Heelis P. F., Phillips G. O., J. Org. Chem.,
5, 731—733 (1980).
20th ed., ed. by Gennaro A. R., Lippincott Williams & Wilkins,
Philadelphia, 2000, pp. 187—188.
53) Janik B., Elving P. J., Chem. Rev., 68, 295—319 (1968).
4
9) Smith E. C., Metzler D. E., J. Am. Chem. Soc., 85, 3285—3288 54) Kan L. S., Borer P. N., Cheng D. M., Ts’o P. O. P., Biopolymers, 19,
1963). 1641—1654 (1980).
0) Treadwell G. E., Cairns W. L., Metzler D. E., J. Chromatogr., 35, 55) Veselkov D. A., Kodintsev V. V., Pakhomov V. I., Djimant L. N.,
76—388 (1968). Davies D. B., Veselkov A. N., Biofizika, 45, 197—206 (2000).
1) “British Pharmacopoeia,” Her Majesty’s Stationary Office, London, 56) Charlton A. J., Davis A. L., Jones D. P., Lewis J. R., Davis A. P.,
(
3
2
009, electronic version.
Haslam E., Williamson M. P., J. Chem. Soc. Perkin Trans. 2, 2000,
317—322 (2000).
57) Davies D. B., Veselkov D. A., Evstigneev M. P., Veselkov A. N., J.
Chem. Soc. Perkin Trans. 2, 2001, 61—67 (2001).
2) Moffat A. C., Osselton M. D., Widdop B., “Clarke’s Analysis of Drugs
and Poisons,” 3rd ed., Vol. 2, Pharmaceutical Press, London, 2004, pp.
7
36—737.
4
3) Weber G., Biochem. J., 47, 114—121 (1950).
58) Evstigneev M. P., Veselkov D. A., Djimant L. N., Davies D. B.,
Veselkov A. N., J. Struct. Chem., 42, 777—783 (2001).
59) Aradi F., Foldesi A., Magn. Reson. Chem., 23, 375—378 (1985).
44) Holzer W., Shirdel J., Zirak P., Penzkofer A., Hegemann P., Deutzmann
R., Hochsmuth E., Chem. Phys., 308, 69—78 (2005).
45) Delgado J. N., Remers W. A., “Wilson and Gisvold’s Textbook of Or- 60) Davies D. B., Veselkov D. A., Djimant L. N., Veselkov A. N., Eur. Bio-
ganic Medicinal and Pharmaceutical Chemistry,” 10th ed., Lippincott-
Raven, Philadelphia, 1998, p. 915.
6) Spiro M., Grandoso D. M., Price W. E., J. Chem. Soc. Faraday Trans.
phys. J., 30, 354—366 (2001).
61) Polter V. I., Grokhlina T. I., Deriabina A., Gonzalez E., Theor. Chem.
Acc., 110, 466—472 (2003).
4
4
4
1
, 85, 4259—4267 (1989).
62) Zubair M. U., Hassan M. M. A., Al-Meshal I. A., “Analytical Profiles
of Drug Substances,” Vol. 13, ed. by Florey K., Academic Press, New
York, 1986, pp. 71—150.
7) Wells J. I., “Pharmaceutical Preformulation, The Physicochemical
Properties of Drug Substances,” Wiley, New York, 1988, p. 171.
8) Clark W. M., “Oxidation—Reduction Potentials of Organic Systems,”