Food Chemistry 84 (2004) 383–388

www.elsevier.com/locate/foodchem

Does caffeine bind to metal ions?

Sevgi Kolaylı*, Miraç Ocak, Murat Küçük, Rıza Abbasoǧlu
Karadeniz Technical University, Faculty of Arts and Sciences, Department of Chemistry Trabzon, Trabzon, Turkey

Received 7 March 2003; accepted 6 May 2003

Abstract
The complex formation capacity of caffeine, a highly-consumed tea and coffee component, was determined for Ca, Mg, Fe, Zn,
Pb, Mn, Co and Cr metal ions. The binding constants of metal ion–caffeine complexes for the metals chosen were determined
spectrophotometrically. The results were compared with the known stability constants of metal ion–EDTA complexes, EDTA
being known for its high metal binding capacity. Furthermore, iron chelating activity of caffeine, using the ferrozine reference
method, was studied and compared with that of EDTA. The results showed very little complex formation capacity of caffeine with
binding constants of 29.6, 22.4, 59, 396, 55, 9.3, 83 and 592 M
1 for Ca, Mg, Fe, Zn, Pb, Mn, Co and Cr metal ions, respectively, in
contrast to that of EDTA. The iron chelating activity of caffeine was also found to be 6%, which was considered to be quite low
compared with EDTA.
# 2003 Elsevier Ltd. All rights reserved.
Keywords: Caffeine; EDTA; Ferrozine; Iron chelating activity; Metal ions; Complex formation; Binding constant

1. Introduction containing beverages and intestinal mineral absorption

(Heaney, 2002; Huang, Yang, Hsieh, & Liu, 2002;
Caffeine, 1,3,-trimethylxanthine, a purine alkaloid, is Rapuri, Gallagher, Kinyamu, & Ryschon, 2001; Yaz-
a key component of many popular drinks, mainly tea dani, Gottschalk, Ide, & Nakamoto, 2002), and urinary
and coffee, but most phytochemists know little about its and intestinal mineral secretion (Heaney & Recker,
biochemistry and molecular biology (Ashihara & Cro- 1982), many others suggest no correlation between caf-
zier, 2001). Green tea leaves contain 10–30% (w/w) tea feine consumption and metal absorption in intestines
polyphenols and 2–4% (w/w) caffeine and all are soluble and kidneys (Conlisk, Deborah, & Gakuska, 2000;
in water. A cup of tea or coffee typically contains Grainge, Coupland, Cliffe, Chilvers, & Hosking, 1998;
approximately 30–175 mg caffeine. While many reports Nafisi, Slamboo, Mohajerani, & Omidi, 2002) and bone
are available about antioxidant and anticancer proper- gain in adolescents (Lloyd, Rollings, Kieselhorst, Eggli,
ties and health benefits of tea, there is also a strong & Mauger, 1998). Hegarty, May, and Khaw (2000)
belief that the people consuming high amounts of caf- found a positive relationship between tea drinking and
feine tend to carry a higher risk of developing bone bone mineral density measurements in older women.
problems, including osteoporosis, as well as problems in The issue of whether or not caffeine consumption is
metal absorption, excretion and reabsorption processes harmful for humans still remains controversial.
in intestines and in kidney (Borse, Jagan Mohan Rao, Calcium, magnesium, iron, zinc, manganese, cobalt
Nagalakshmi, & Krishnamurthy, 2002; Chen & Whit- and chromium are essential elements for many organ-
ford, 1999; Massey, 2001; Pan, Guaguang, & Liu, 2003), isms. They are involved in a variety of biological reac-
and iron deficiency anemia (Hallberg & Rossander, tions (Turcot, Stintzi, Xu, & Raymond, 2000; Voet &
1982). Although some epidemiological studies show a Voet, 1995) and their deficiencies result in many medical
negative relationship between consumption of caffeine- problems (Prohaska & Brokate, 2001; Salgueiro, Zubil-
lage, Lysionek, Sarabia, Caro, & Paoli, 2000). Caffeine
* Corresponding author. Tel.: +90-462-377-24-87; fax: +90-462-
(Structure 1) has been suspected to bind many metal
325-31-96. ions and alter their balance in the human body (Chen &
E-mail address: [email protected] (S. Kolaylı). Whitford, 1999; Riesselmann, Rosenbaum, Roscher, &
0308-8146/03/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0308-8146(03)00244-9
384 S. Kolaylı et al. / Food Chemistry 84 (2004) 383–388

the iron that remains unbound in the solution contain-

ing EDTA or caffeine. A 0.2 ml sample solution of
varying concentration, 1.2 ml deionized distilled water
and 0.2 ml 0.2 mM FeCl2.4H2O solution were mixed
and vortexed immediately. 0.4 ml 1 mM ferrozine was
added to the reaction mixture and change in colour was
monitored at 593 nm with a Unicam UV2 UV/vis spec-
trometer after a 10 min incubation period at room tem-
Structure 1. Caffeine.
perature. EDTA was used for comparison.
The values presented in this report are the means of
triplicate measurements.
Scheineder, 1999). The interactions between caffeine
and metal ions can be through its oxygen and nitrogen
atoms. Because of the blockage on N1, N3 and N7 3. Results
atoms by methyl groups, caffeine probably binds to
metal ions through its O2 and O6 atoms. Indeed our The binding constants were calculated using the data
previous theoretical study showed that caffeine would obtained from the UV spectrophotometric method as
bind to metal ions through its O2 and O6 atoms in the reported (Klotz & Hunston, 1971; Purcell, Neault, &
gaseous phase (Abbasoǧlu, Küçük, Kolaylı, & Ocak, Tajmir-Riahi, 2000). The equilibrium between caffeine
2002). Therefore, we aimed at finding the extent of in and a metal ion can be represented by the following
vitro metal ion interactions with caffeine in aqueous reaction:
solutions using EDTA for comparison. This study was
partly presented at an international meeting (Ocak, Caffeine þ Metal ion $ Caffeine : Metal ion
Kolaylı, Küçük, & Abbasoǧlu, 2002).
The binding constant of caffeine with a metal ion is
defined as in Eq. (1):
2. Materials and methods
½caffeine : metal ion
k¼ ð1Þ
Caffeine was purchased from Fluka and ferrozine, 3- ½caffeine½metal ion
(2-pyridyl)-5,6-bis(2-5-furylsulfonic acid-1,2,4-triazine,
was purchased from Sigma. Ethylenediaminetetraacetic In order to determine the binding constants of caf-
acid disodium salt (EDTA), and Ca(II), Mg(II), Fe(II), feine for the metal ions studied, Klotz plots were uti-
Pb(II) and Zn(II) standard solutions were obtained lized. In order to obtain these plots, Eq. (2) was used.
from Merck. All of the solutions were prepared with 1 1 1 1
deionized distilled water. ¼ þ ð2Þ
c n:a:k L n:a
The stability constants of caffeine–metal ion com-
plexes were determined according to the method used where c is the concentration of metal ion–caffeine com-
by Nafisi et al. (2002) with a slight modification. The plex, n is the number of metal binding sites on caffeine,
solutions of Ca (II), Mg (II), Fe(II), Pb(II), Zn(II), a is caffeine concentration, L is total metal ion concen-
Co(II) and Mn(II) of 0.5 mM concentration were pre- tration and k is binding constant for metal ion–caffeine
pared by diluting the standard metal ion solutions in complex. The absorbance of metal ion–caffeine complex
deionized distilled water. 5 ml, from each appropriately (Ac) can be used instead of c by multiplying it with the
diluted metal ion solution, were added separately into 5 inverse of the extinction coefficient of caffeine, 1/", at
ml caffeine solutions to attain the desired metal ion that wavelength (Bobrovnik, 2003), by which Eq. (3)
concentrations of 0.001, 0.01, 0.1, and 1.0 mM with a was obtained.
final caffeine concentration of 0.0005% w/v, remaining 1 1 1 1
the same throughout all of the experiments. Following ¼ þ ð3Þ
1 n:a:k L n:a
vortexing and incubation for 10 min at room tempera- Ac
"
ture, absorbances were measured at 273 nm with a
Unicam UV2 UV/vis spectrometer. The absorbance of a By placing 1/Ac on one side of the equation, Eq. (4) was
caffeine solution of the same concentration containing obtained and used to calculate binding constants by
no metal ion was also measured. plotting 1/Ac against 1/L.
Iron chelating activity of caffeine was determined by
the method of Dinis, Madeira, and Almeida (1994).
This is based on the measurement of absorbance at 593 1 1 1 1
¼ þ ð4Þ
nm resulting from a colourful complex of ferrozine with Ac ":n:a:k L ":n:a
 S. Kolaylı et al. / Food Chemistry 84 (2004) 383–388 385

where Ac=A
Ao, A and Ao representing the absor- shown in Figs. 1–8 for metal ion–caffeine complexation.
bances of metal ion–caffeine complex and caffeine The graphs show a linear relationship for all of the
alone, respectively. metal ions. The binding constants (k: M
1) were calcu-
The plots of 1/(absorbance of complexed ion) vs. 1/ lated as the ratio of the intercept on the vertical axis to
(total metal ion concentration)—1/A-Ao vs. 1/L—are the slope [Eq. (5)].

Fig. 1. The plot of 1/(A
A0) vs. (1/L) for caffeine–Ca+2 interaction. Fig. 4. The plot of 1/(A
A0) vs. (1/L) for caffeine–Zn+2 interaction.
A0 is the initial absorbance of caffeine at 273 nm and A is the recorded A0 is the initial absorbance of caffeine at 273 nm and A is the recorded
absorbance at varying ion concentration. absorbance at varying ion concentration.

Fig. 2. The plot of 1/(A
A0) vs. (1/L) for caffeine–Mg+2 interaction. Fig. 5. The plot of 1/(A
A0) vs. (1/L) for caffeine–Pb+2 interaction.
A0 is the initial absorbance of caffeine at 273 nm and A is the recorded A0 is the initial absorbance of caffeine at 273 nm and A is the recorded
absorbance at varying ion concentration. absorbance at varying ion concentration.

Fig. 3. The plot of 1/(A
A0) vs. (1/L) for caffeine–Fe+2 interaction. Fig. 6. The plot of 1/(A
A0) vs. (1/L) for caffeine–Mn+2 interaction.
A0 is the initial absorbance of caffeine at 273 nm and A is the recorded A0 is the initial absorbance of caffeine at 273 nm and A is the recorded
absorbance at varying ion concentration. absorbance at varying ion concentration.
386 S. Kolaylı et al. / Food Chemistry 84 (2004) 383–388

Table 1
Binding constants for the metal ion–caffeine and EDTA complexes

Element Caffeine EDTAa

Calcium 29.6 5.01 1010

Magnesium 22.4 4.90 108
Iron (Fe2+) 59 2.14 1014
Zinc 396 3.16 1016
Chromium 592 –
Lead 55 1.10 1018
Cobalt 83 2.04 1016
Manganese 9.3 1.10 1014
a
The data were obtained from literature (Christian, 1994).

Fig. 7. The plot of 1/(A
A0) vs. (1/L) for caffeine–Cr+2 interaction.

A0 is the initial absorbance of caffeine at 273 nm and A is the recorded
absorbance at varying ion concentration.

Fig. 9. Chelating activity of various concentrations of caffeine and

EDTA. Each value is the average of tree determinations.

ation was found with respect to the type of metal ions,

Fig. 8. The plot of 1/(A
A0) vs. (1/L) for caffeine–Co+2 interaction. r2=0.06.
A0 is the initial absorbance of caffeine at 273 nm and A is the recorded Iron chelating activities of caffeine and EDTA were
absorbance at varying ion concentration. calculated according to Eq. (6).
A0
A
% Chelating activity ¼ ð6Þ
A0
1
intercept on ordinate axis "na where A is the absorbance in the presence of ligand
k¼ ¼ ð5Þ
slope 1 while A0 is the absorbance with no ligand present.
"nak The Fe(II) binding capacities of the two ligands were
compared in terms of iron chelating activity (Dinis et
In order to calculate the concentrations of complexed al., 1994). While EDTA showed a concentration-depen-
metal ions, the absorbance of complexed caffeine was dent activity, from 99% metal binding at 1–10 mg/ml
subtracted from that of uncomplexed caffeine at 273 EDTA concentrations to less then 5% at low con-
nm. The concentrations of free metal ions could be cal- centrations, caffeine showed a low but steady chelating
culated by subtracting that of complexed metal ion from activity with around 6% binding at all caffeine con-
the total metal ion concentration used in the experi- centrations (Fig. 9).
ments. The calculated binding constants of metal ion–
caffeine complexes and the binding constants for metal
ion–EDTA complexes are presented in Table 1. 4. Discussion
The binding constants of metals with EDTA were
orginally reported by Christian (1994) and presented in Caffeine, a well known alkaloid, is taken daily with
Table 1 for comparison. The binding constants of caf- beverages and known to have effects on energetic arou-
feine with all the metal ions studied appear to be very sal, to stimulate the autonomic nervous system and to
low when compared to the binding constants of EDTA. increase alertness (Quinlan, Lane, Moore, Aspen,
When the binding capacities of EDTA and caffeine in Rycroft, & O’Brien, 2000). There is a belief among the
terms of k values were compared, no significant correl- public that those who drink too much tea or coffee are
 S. Kolaylı et al. / Food Chemistry 84 (2004) 383–388 387

prone to have mineral deficiency problems (Conlisk et Acknowledgements

al. 2000; Horie, Nesumi, Tomomi, & Kohata, 2000).
Previous studies on this subject appear to be more of the This research was supported by the Research Fund
in vivo and epidemiological kind. Chen and Whitford of Karadeniz Technical University, (Project No:
(1999) investigated mineral metabolism for calcium, 21.111.002. 9).
fluorine and phosphorus in rats given caffeine at various
doses, and reported that they found no significant
change in mineral balance. Similarly, Prystai, Prystai, References
Constance, Kies, Judy and Driskell (1999) reported that Abbasoǧlu, R., Küçük, M., Kolaylı, S., & Ocak M. (2002). Theoretical
no change was observed in the balance of Ca2+, Cu2+, investigation of Mg2+ binding to caffeine and theophylline. In Pro-
Fe2+, Mg2+ and Zn2+ ions in young adults who reg- ceedings of the 1st Int. Congress on the Chemistry of Natural Pro-
ularly took tea and decaffeinated tea. Conlisk et al. ducts, Trabzon, Türkiye.
(2000) studied the effect of drinking beverages with caf- Ashihara, H., & Crozier, A. (2001). Caffeine: a well known but little
mentioned compound in plant science. Trends in Plant Science, 6(9),
feine on bone mineral density (BMD) on 77 healthy 407–413.
women, using decaffeinated beverages as control and Barger-Lux, M. J., & Heaney, R. P. (1995). Caffeine and calcium
reported that no change had occurred in BMD of the economy revisited. Osteoporosis International, 5, 97–102.
subjects. Yeh, Aloia, Semla, and Chen (1986) gave caf- Bobrovnik, S. A. (2003). Ligand-receptor interactions: a new method
feine to rats by injection in daily doses of 25 and 100 for determining the binding parameters. Journal of Biochemical and
Biophysical Methods, 55, 71–86.
mg/kg and compared the results with the control group. Borse, B. B., Jagan Mohan Rao, L., Nagalakshmi, S., & Krishna-
The urinary excretion of phosphorus was increased in murthy, N. (2002). Fingerprint of black teas from India: identifica-
both caffeine groups and the group injected with 100 tion of the regio-specific characteristics. Food Chemistry, 79, 419–
mg/kg caffeine daily showed a significant reduction in 424.
phosphorus balance. Barger-Lux and Heaney (1995) Chen, X., & Whitford, G. M. (1999). Effects of caffeine on fluoride,
calcium and phosphorus metabolism and calcified tissues in the rat.
reported that caffeine intake reduced the calcium bal- Archives of Oral Biology, 44, 33–39.
ance slightly in 39–69-year-old women but the effect was Christian, G. D. (1994). Analytical chemistry (5th ed.). New York:
due to the reductions in intake and absorption. Nasifi et John Wiley & Sons.
al. (2002) investigated the binding constants of caffeine Conlisk, A. J., Deborah, A., & Gakuska, M. P. H. (2000). Is caffeine
associated with bone mineral density in young adult women? Pre-
and theophylline with Ca2+ and Mg2+ and reported
ventive Medicine, 31, 562–568.
that both alkaloids form a very weak complex with Dinis, T. C. P., Madeira, V. M. C., & Almeida, L. M. (1994). Action
Ca2+ and Mg2+ ions. of phenolic derivatives as inhibitors of membrane lipid peroxidation
In the current study, we found that the eight metal and as peroxyl radical scavengers. Archives of Biochemistry and
ions investigated, Ca2+, Mg2+, Fe2+, Zn2+, Pb2+, Biophysics, 315, 161–169.
Mn2+, Co2+ and Cr2+, formed complexes with caffeine Grainge, M. J., Coupland, C. A. C., Cliffe, S. J., Chilvers, C. E. D., &
Hosking, D. J. (1998). Cigarette smoking, alcohol and caffeine con-
in varying capacities but these were very weak in sumtion, and bone mineral density in postmenopausal women.
strength when compared to EDTA (Table 1). EDTA Osteoporosis International, 8(4), 355–363.
shows a 1010-fold higher metal binding activity com- Hallberg, L., & Rossander, L. (1982). Effect of different drinks on the
pared to caffeine. Nafisi et al. (2002) reported the bind- absorption of non-heme iron from composite meals. Human Nutri-
ing constants of Ca2+ and Mg2+ with caffeine to be tion. Applied Nutrition, 36(2), 116–123.
Heaney, R. P., & Recker, R. R. (1982). Effects of nitrogen, phospho-
29.8 and 22.4 M
1, respectively. The results of our rus and caffeine on calcium balance in women. The Journal of
experiments and theirs were quite similar. Laboratory and Clinical Medicine, 99(1), 46–55.
Heaney, R. P. (2002). Effects of caffeine on bone and the calcium
economy. Food and Chemical Toxicology, 40(9), 1263–1270.
Hegarty, V. M., May, H. M., & Khaw, K. T. (2000). Tea drinking and
5. Conclusions
bone mineral density in older women. American Journal of Clinical
Nutrition, 71(4), 1003–1007.
The current study clearly showed that caffeine does Horie, H., Nesumi, A., Tomomi, U., & Kohata, K. (2002). Rapid
not bind to Ca, Mg, Fe, Zn, Pb, Mn, Co and Cr metal determination of caffeine in tea leaves. Journal of Chromatography,
ions in significant extent to alter their functions in the 942, 271–273.
human body. The interactions between caffeine and the Huang, T. H., Yang, R. S., Hsieh, S. S., & Liu, S. H. (2002). Effects of
caffeine and exercise on the development of bone: a densitometric
metal ions studied were very low and this fact was and histomorphometric study in young Wistar rats. Bone, 30(1),
proved earlier with quantum mechanical investigations 293–299.
(Abbasoǧlu et al., 2002). The results of this study sug- Klotz, I. M., & Hunston, D. L. (1971). Properties of graphical repre-
gest that caffeine does not change the metal ion balance sentations of multiple classes of binding sites. Biochemistry, 10,
in the human body directly by binding metal ions. 3065–3069.
Lloyd, T., Rollings, N. J., Kieselhorst, K., Eggli, D. F., & Mauger, E.
Instead, there may be other mechanisms which are (1998). Dietary caffeine intake is not correlated with adolescent
induced by caffeine, resulting in undesired consequences bone gain. Journal of the American College of Nutrition, 17(5), 454–
of caffeine consumption, including osteoporosis. 457.
388 S. Kolaylı et al. / Food Chemistry 84 (2004) 383–388

Massey, L. K. (2001). Is caffeine a risk factor for bone loss in the tea and coffee: the role of caffeine level. Pharmacology Biochemistry
elderly. American Journal of Clinical Nutrition, 74, 569–570. and Behavior, 66(1), 19–28.
Nafisi, S., Slamboo, D. S., Mohajerani, N., & Omidi, A. (2002). A Rapuri, P. B., Gallagher, J. C., Kinyamu, H. K., & Ryschon, K. L.
comparative study of caffeine and teophylline binding to Mg(II) and (2001). Caffeine intake increases the rate of bone loss in elderly
Ca(II) ions: studies by FTIR and UV spectroscopic methods. Jour- women and interacts with vitamin D receptor genotypes. American
nal of Molecular Structure, 608, 1–7. Journal of Clinical Nutrition, 74, 694–700.
Ocak, M., Kolaylı, S., Küçük, M., Abbasoǧlu, R. (2002). A compara- Riesselmann, B., Rosenbaum, F., Roscher, S., & Scheineder, V.
tive study of caffeine binding to some metal ions. Proceedings 1st (1999). Fatal caffeine intoxication. Forensic Science International,
Int. Congress on the Chemistry of Natural Products (p. 94) Trabzon, 103, 549–552.
Türkiye. Salgueiro, M. J., Zubillage, B. M., Lysionek, A., Sarabia, I., Caro, R.,
Pan, X., Guaguang, N., & Liu, H. (2003). Microwave-assisted extrac- & Paoli, T. (2000). Zinc as an essential micronutrient. Nutrition
tion of tea polyphenols and tea caffeine from green tea leaves. Che- Research, 20(5), 737–755.
mical Engineering and Processing, 42(2), 129–133. Turcot, I., Stintzi, A., Xu, J., & Raymond, K. N. (2000). Fast biolo-
Prohaska, J. R., & Brokate, B. (2001). Lower copper, zinc-superoxide gical iron chelators: kinetics of iron removal from human diferric
dismutase protein but not mRNA in organs of copper-deficient rats. transferrin by multidentate hydroxypridonates. Journal of Biological
Archives of Biochemistry and Biophysics, 393(1), 170–176. Inorganic Chemistry, 5, 634–641.
Prystai, E., Prystai, M. S., Constance, V., Kies, P. D., Judy, A., & Voet, D., & Voet, J. G. (1995). Biochemistry (2nd ed.). New-York:
Driskell, P. D. (1999). Calcium, copper, iron, magnesium and John Wiley & Sons.
zinc utilization of humans as affected by consumption of black, Yazdani, M., Gottschalk, S., Ide, K., & Nakamoto, T. (2002). Asso-
decaffeinated black and green teas. Nutrition Research, 19(2), 167– ciation of different zinc concentrations combined with a fixed caf-
177. feine dose on plasma and tissue caffeine and zinc levels in the rat.
Purcell, M., Neault, J. F., & Tajmir-Riahi, H. A. (2000). Interaction of Biological Trace Elements Research, 87, 183–190.
taxol with human serum albumin. Biochimica et Biophysica Acta, Yeh, J. K., Aloia, J. F., Semla, H. M., & Chen, S. Y. (1986). Influence
1478, 61–68. of injected caffeine on the metabolism of calcium and the retention
Quinlan, P. T., Lane, J., Moore, K. L., Aspen, J., Rycroft, J. A., & and excretion of sodium, potassium, phosphorus, magnesium, zinc
O’Brien, D. C. (2000). The acute physiological and mood effects of and copper in rats. Journal of Nutrition, 116, 273–280.