Isolation, identification, and quantification of roasted coffee antibacterial compounds

Article abstract of DOI:10.1021/jf0722607

Coffee brew is a widely consumed beverage with multiple biological activities due both to naturally occurring components and to the hundreds of chemicals that are formed during the roasting process. Roasted coffee extract possesses antibacterial activity against a wide range of microorganisms, including Staphylococcus aureus and Streptococcus mutans, whereas green coffee extract exhibits no such activity. The naturally occurring coffee compounds, such as chlorogenic acids and caffeine, cannot therefore be responsible for the significant antibacterial activity exerted by coffee beverages against both bacteria. The very low minimum inhibitory concentration (MIC) found for standard glyoxal, methylglyoxal, and diacetyl compounds formed during the roasting process points to these α-dicarbonyl compounds as the main agents responsible for the antibacterial activity of brewed coffee against Sa. aureus and St. mutans. However, their low concentrations determined in the beverage account for only 50% of its antibacterial activity. The addition of caffeine, which has weak intrinsic antibacterial activity, to a mixture of α-dicarbonyl compounds at the concentrations found in coffee demonstrated that caffeine synergistically enhances the antibacterial activity of α-dicarbonyl compounds and that glyoxal, methylglyoxal, and diacetyl in the presence of caffeine account for the whole antibacterial activity of roasted coffee.

Full text of DOI:10.1021/jf0722607

10208 J. Agric. Food Chem. 2007, 55, 10208–10213
Isolation, Identification, and Quantification of
Roasted Coffee Antibacterial Compounds
†
†
‡
†
MARIA DAGLIA, ADELE PAPETTI, PIETRO GRISOLI, CAMILLA ACETI,
,†
†
‡
VALENTINA SPINI, CESARE DACARRO, AND GABRIELLA GAZZANI*
Department of Pharmaceutical Chemistry, Faculty of Pharmacy, University of Pavia, Via Taramelli 12,
2
7100 Pavia, Italy, and Department of Experimental and Applied Pharmacology, Faculty of Pharmacy,
University of Pavia, Via Taramelli 14, 27100 Pavia, Italy
Coffee brew is a widely consumed beverage with multiple biological activities due both to naturally
occurring components and to the hundreds of chemicals that are formed during the roasting process.
Roasted coffee extract possesses antibacterial activity against a wide range of microorganisms,
including Staphylococcus aureus and Streptococcus mutans, whereas green coffee extract exhibits
no such activity. The naturally occurring coffee compounds, such as chlorogenic acids and caffeine,
cannot therefore be responsible for the significant antibacterial activity exerted by coffee beverages
against both bacteria. The very low minimum inhibitory concentration (MIC) found for standard glyoxal,
methylglyoxal, and diacetyl compounds formed during the roasting process points to these R-dicarbonyl
compounds as the main agents responsible for the antibacterial activity of brewed coffee against Sa.
aureus and St. mutans. However, their low concentrations determined in the beverage account for
only 50% of its antibacterial activity. The addition of caffeine, which has weak intrinsic antibacterial
activity, to a mixture of R-dicarbonyl compounds at the concentrations found in coffee demonstrated
that caffeine synergistically enhances the antibacterial activity of R-dicarbonyl compounds and that
glyoxal, methylglyoxal, and diacetyl in the presence of caffeine account for the whole antibacterial
activity of roasted coffee.
KEYWORDS: Roasted coffee; antibacterial activity; Staphylococcus aureus; Streptococcus mutans;
r-dicarbonyl compounds; caffeine
INTRODUCTION
caramelization, and organic compound pyrolysis during the
roasting process. Roasted coffee has been seen to exert both in
vitro and ex vivo antioxidant activity (5–8). Given its signifi-
cantly decreased chlorogenic acid content (up to 50–90%), the
antioxidant properties of roasted coffee are mainly attributable
to Maillard reaction products formed during the roasting process,
which include both high-molecular mass components (melano-
idins) and heterocyclic compounds such as furan, pyrroles, and
maltol (9).
Coffee brew is among the most widely consumed beverages
all over the world, in part because of its putative health benefits.
Epidemiological studies have shown that regular coffee drinking
has favorable effects on health aspects ranging from psycho-
active responses to neurological disorders (Alzheimer’s and
Parkinson’s diseases) and metabolic disorders (type 2 diabetes
and liver cirrhosis) to liver function (1–4). A high level of
worldwide coffee consumption has stimulated research into the
chemical composition and biological activities of green and,
especially, roasted coffee. Most investigations have addressed
naturally occurring compounds such as caffeine (and its
pharmacological activity) and chlorogenic acids. These phenolic
acids are found in greater amounts in green beans and may be
the main factor responsible for the antioxidant activity of green
coffee (5, 6).
Roasted coffee has been found to interfere with streptococcal
sucrose-independent adsorption to hydroxyapatite beads through
the involvement of small molecules naturally occurring in coffee,
such as trigonelline, nicotinic and chlorogenic acids, and high-
molecular mass compounds that are formed during the roasting
process (melanoidins) (10). Roasted coffee has also been shown
to possess antibacterial activity against a wide range of Gram-
positive and Gram-negative bacteria, including Streptococcus
mutans. Daglia et al. (11–14) found no detectable antibacterial
activity of green coffee and concluded that naturally occurring
compounds are not responsible for the antibacterial activity of
roasted coffee. Almeida et al. (15) showed that roasted coffee
is active against several strains of Enterobacteria and that
trigonelline, caffeine, and protocatechuic acid are potential
In recent years, research interest has been focused on bioactive
compounds induced by the Maillard reaction, carbohydrate
*
To whom correspondence should be addressed. Telephone: +39
382 987373. Fax: +39 0382 422 975. E-mail: gabriella.gazzani@
unipv.it.
0
†
Department of Pharmaceutical Chemistry.
Department of Experimental and Applied Pharmacology.
‡
1
0.1021/jf0722607 CCC: $37.00  2007 American Chemical Society
Published on Web 11/15/2007

Roasted Coffee Antibacterial Compounds
J. Agric. Food Chem., Vol. 55, No. 25, 2007 10209
natural antimicrobial agents against Salmonella enterica. In
particular, they reported that caffeine may account for up to
R-dicarbonyl compound content, SPE-F1 was derivatized with 1,2-
diaminobenzene (1.0 mg) to obtain quinoxaline derivatives using the
method described by de Revel et al. (21), with some modifications.
The pH of the reaction mixture was adjusted to 8.0 with NaOH (0.5
M), and the mixture was kept at 60 °C for 3 h; after cooling, the
solutions were injected into the HPLC system. SPE-F2 (diluted 1:5
with methanol) was directly analyzed by HPLC for the assessment of
chlorogenic acids, 5-hydroxymethylfurfural, and caffeine. All experi-
ments were performed in triplicate.
5
0% of the antimicrobial effect of coffee brew against S.
enterica. Dogazaki et al. (16) and Furuhata et al. (17) also
reported that roasted coffee was active against a strain of
Legionella pneumophila, a bacterium involved in respiratory
diseases, pointing out a role for caffeic, chlorogenic, and
protocatechuic acids.
This study was performed to isolate and identify the
compounds responsible for the antibacterial activity of roasted
coffee against Staphylococcus aureus and Streptococcus mutans
observed in previous studies.
Reverse-Phase High-Performance Liquid Chromatography with
Diode Array Detection (RP-HPLC–DAD). All experiments were
performed using a 1100 HPLC system (Agilent, Waldbronn, Germany)
equipped with a gradient quaternary pump, a thermostated column
compartment, and a DAD apparatus. The Agilent Chemstation software
was used for HPLC system control and data processing. Determinations
were carried out using a 250 mm × 4.6 mm, 5 µm C18 Hypersil column
(CPS Analitica, Milan, Italy) with a matching Lichrospher 100 RP-18,
5 mm guard column (Merck, Darmstadt, Germany).
Chromatographic Conditions for r-Dicarbonyl Compound As-
sessment. Chromatographic conditions for gradient elution were as
follows: flow rate, 0.6 mL/min; volume injected, 20 µL; column
temperature, 20 °C. UV spectra were recorded in the 190–600 nm range,
and chromatograms were acquired at 314 nm. Separations were
performed using a gradient of increasing methanol concentrations in
water acidified (pH 3.00 ( 0.01) with 0.5% acetic acid (v/v) as follows:
40 min linear gradient from 10 to 50% methanol, 10 min linear gradient
from 50 to 75% methanol, and 4 min increasing gradient segment to
100% methanol followed by a 4 min isocratic period with 100%
methanol. The composition of the mobile phase was taken to the initial
condition (10% methanol) in 4 min, and the column was equilibrated
for 10 min before the next injection.
Chromatographic Conditions for 5-Hydroxymethylfurfural, Chlo-
rogenic Acid, and Caffeine Assessment. Chromatographic conditions
for gradient elution were as follows: flow rate, 1 mL/min; volume
injected, 20 µL; column temperature, 20 °C. UV spectra were recorded
in the 190–600 nm range, and chromatograms were acquired at 280
nm (5-hydroxymethylfurfural and caffeine) and 324 nm (chlorogenic
acids). Separations were performed using a gradient of increasing
methanol concentrations in water acidified (pH 3.00 ( 0.01) with 0.1%
formic acid (v/v) as follows: 10 min at 5% methanol, 10 min linear
gradient from 10 to 30% methanol, 30 min linear gradient from 30 to
45% methanol, and 10 min linear gradient from 45 to 80% methanol.
The mobile phase composition was taken to the initial condition (5%
methanol) in 5 min, and the column was equilibrated for 5 min before
the next injection.
MATERIALS AND METHODS
Chemicals. HPLC-grade solvents (methanol, acetic acid, and formic
acid), glyoxal, methylglyoxal, diacetyl, quinoxaline, 2-methylquinoxa-
line, 2,3-dimethylquinoxaline, 5-methylquinoxaline, 1,2-diaminoben-
zene, 5-O-caffeoylquinic acid, caffeine, and 5-hydroxymethylfurfural
were purchased from Sigma-Aldrich (St. Louis, MO).
Coffee Beans and Coffee Extract (Coffee Brew) Preparation.
Green Coffea robusta beans from Java were roasted in a pilot roaster
apparatus (STA Impianti S.r.l., Bologna, Italy). The degree of roasting
was measured by the lost due to vapor formation and cell fragment
loss. Loss was 12%, corresponding to a medium degree of roasting.
Roasted coffee beans were ground in a laboratory scale mill and sieved
through a no. 30 sieve. Coffee extract (CR) was then obtained by the
coffee brewing procedure commonly used in Italy. Briefly, 6 g of
roasted coffee powder was boiled for 10 min in 100 mL of Millipore-
grade water (Millipore Corp., Billerica, MA). The extract (100 mL)
was filtered through a 0.45 µm Millipore membrane of cellulose acetate/
cellulose nitrate mixed esters and then subdivided into three aliquots.
One was freeze–dried, and the dry matter was dissolved in an
appropriate volume of Millipore-grade water for antibacterial activity
testing; the second was subjected to dialysis and the third to chemical
analysis.
Dialysis. Sequential dialysis was performed using a Spectra/Por
Biotech cellulose ester membrane (Spectrum Europe B.V., Breda, The
Netherlands) with molecular mass cutoffs of 3500, 1000, and 500 Da.
Aliquots (10 mL) of coffee extract were fractionated by dialysis in
1000 mL Millipore-grade water for 6 h at 4 °C. Dialysates and retentates
were freeze-dried; dry residues were assessed and then dissolved in 1
mL of Millipore-grade water. Recovered 5-O-caffeoylquinic acid
(
>95%) was used as a standard molecular mass marker. Dialysates
and retentates were tested against Sa. aureus and St. mutans.
Bacterial Strains, Media, and Buffers. The following strains were
used: Sa. aureus ATCC 25923 and St. mutans 9102 (18). St. mutans
was cultured in Todd Hewitt Broth (THB) (Oxoid, Basingstoke, U.K.)
at 37 °C and Sa. aureus in Tryptone Soya Broth (TBS) (Oxoid) at
Identification and Quantification of r-Dicarbonyl Compounds,
5
-Hydroxymethylfurfural, Chlorogenic Acids, and Caffeine. Reten-
tion times and UV spectra were used to identify the quinoxaline
derivatives, 5-hydroxymethylfurfural, 5-O-caffeoylquinic acid, and
caffeine. Stock standard solutions of quinoxaline, 2-methylquinoxaline,
3
7 °C.
Evaluation of the Coffee Brew Minimal Inhibitory Concentration
MIC) and Minimal Bactericidal Concentration (MBC). Concen-
2
,3-dimethylquinoxaline, 5-O-caffeoylquinic acid, 5-hydroxymethyl-
furfural, and caffeine were prepared by dissolving carefully weighed
amounts of each standard compound in a 50% (v/v) methanol/Millipore-
grade water mixture. Solutions were analyzed by RP-HPLC–DAD. Each
standard solution was diluted with mobile phase to five final concentra-
tions ranging from 5 to 100 µM for quinoxaline, 2-methylquinoxaline,
and 2,3-dimethylquinoxaline, from 10 to 200 µg/mL for 5-hydroxy-
methylfurfural, and from 50 to 250 mg/mL for 5-O-caffeoylquinic acid
and caffeine. Each concentration was analyzed in triplicate. Quantifica-
tion of individual compounds was performed by the external standard
method using a five-point regression curve.
Bioactive Roasted Coffee Compound Mixture Preparations. Six
aqueous solution mixtures containing glyoxal, methylglyoxal, and
diacetyl (mixture 1), 5-O-caffeoylquinic acid, 5-hydroxymethylfurfural,
and caffeine (mixture 2), glyoxal, methylglyoxal, diacetyl, and 5-O-
caffeoylquinic acid (mixture 3), glyoxal, methylglyoxal, diacetyl, and
5-hydroxymethylfurfural (mixture 4), glyoxal, methylglyoxal, diacetyl,
and caffeine (mixture 5), and glyoxal, methylglyoxal, diacetyl, 5-O-
caffeoylquinic acid, 5-hydroxymethylfurfural, and caffeine (mixture 6)
were prepared at concentrations corresponding to a 10× coffee extract,
like those used to evaluate coffee brew MIC and MBC values (Table
(
trated coffee extract (10×) was filtered through a 0.22 µm Millex GP
membrane. MIC and MBC values were determined in Iso-Sensitest
broth (ISB) (Oxoid) according to Clinical and Laboratory Standards
Institute (CLSI, formerly NCCLS) procedures (19). The MIC was the
lowest coffee solution concentration (% by volume or milligrams per
milliliter) inhibiting observable growth; the MBC (% by volume or
milligrams per milliliter) was the lowest concentration resulting in a
>
99.9% reduction of the initial inoculum (20). All experiments were
performed in triplicate.
Sample Preparation for RP-HPLC–DAD Analysis. A C18 Sep-
Pak cartridge (Waters, Milford, MA) was conditioned with methanol
(
10 mL) and distilled water (2 × 10 mL). Aliquots (2 mL) of coffee
extract were passed through the cartridge at a flow rate of e2 mL/
min. Polar substances, among which were coffee R-dicarbonyl com-
pounds, i.e., glyoxal, methylglyoxal, and diacetyl, were eluted first,
with 2 mL of Millipore-grade water (SPE-F1). Less polar substances,
including chlorogenic acids, 5-hydroxymethylfurfural, and caffeine,
were then eluted with 4 mL of methanol (SPE-F2). To determine the

10210 J. Agric. Food Chem., Vol. 55, No. 25, 2007
Daglia et al.
Table 1. R-Dicarbonyl Compounds, 5-Hydroxymethylfurfural, Caffeine,
-O-Caffeoylquinic Acid, and Total Chlorogenic Acids in Roasted C.
be active. These data agree with a previous report in which we
attributed the antibacterial activity of coffee to unidentified
compounds with a molecular mass of <200 Da (14).
5
robusta Extract
a
Antibacterial Compound Identification. Since green coffee
did not exhibit detectable antibacterial activity, the active
compound(s) must arise during heat treatment, thus ruling out
caffeine and other natural components. This conclusion was
further supported by the fact that the concentrations of most
natural compounds found in green coffee decreased in roasted
coffee. We thus turned to R-dicarbonyl compounds and 5-hy-
droxymethylfurfural, the presence of which in roasted coffee
has been reported in a number of papers (22–24). R-Dicarbonyl
compounds are produced by the multiple-fragmentation reaction
of sugar moiety; 5-hydroxymethylfurfural is formed by different
pathways, mainly via dehydration of hexoses either in the
presence of amines (Maillard reaction) or in their absence
compound in SPE-F1
HPLC tR (min)
content in CR
glyoxal
methylgyoxal
diacetyl
36.35
43.52
51.22
4.73 ( 0.14 µg/mL
11.64 ( 0.29 µg/mL
2.72 ( 0.06 µg/mL
compound in SPE-F2
HPLC tR (min)
content in CR
5
-hydroxymethylfurfural
caffeine
-O-caffeoylquinic acid
17.78
25.94
25.94
–
180.92 ( 3.61 µg/mL
2.18 ( 0.09 mg/mL
0.62 ( 0.02 mg/mL
1.68 ( 0.04 mg/mL
5
chlorogenic acids
a
All experiments were performed in triplicate.
(
charbohydrate caramelization), giving 3-deoxyhexuloses that
1
). With regard to 5-O-caffeoylquinic acid, it was added to each mixture
can further react to give 5-hydroxymethylfurfural (25). The tests
carried out on standard glyoxal, methylglyoxal, diacetyl, and
5
at the total concentration of chlorogenic acids determined in coffee
brew. Solutions were analyzed for antibacterial (MIC) and bactericidal
-hydroxymethylfurfural aqueous solutions showed that all
(
MBC) activity against Sa. aureus and St. mutans.
Statistical Analysis. The reported data represent the means of the
values obtained from the analysis of three coffee extracts or fractions,
analyzed in triplicate.
R-dicarbonyl compounds possessed strong antibacterial activity,
particularly against St. mutans, whereas 5-hydroxymethylfurfural
exhibited weak antibacterial activity against both bacteria; Sa.
aureus MBC values could not be defined (Table 3). Standard
5-O-caffeoylquinic acid and caffeine aqueous solutions were
also tested. 5-O-Caffeoylquinic acid exhibited weak antibacterial
activity against both bacteria and no bactericidal activity against
Sa. aureus (MBC > 50%), whereas caffeine exhibited low
activity against St. mutans (high MIC and MBC) and no activity
against Sa. aureus (MIC and MBC > 50%) (Table 3).
RESULTS AND DISCUSSION
Antibacterial and Bactericidal Activities of Roasted Cof-
fee. Ten times concentrated roasted C. robusta extract (CR),
i.e., coffee brew (containing 124.6 mg/mL dry matter), was
tested for its antibacterial activity toward Sa. aureus and St.
mutans. MICs and MBC values, reported in Table 2, showed
that CR is active against both bacteria, with lower values for
St. mutans. The extract obtained from green coffee exhibited
no activity. These data agree with our previous results for green
and roasted coffee extracts (14).
The tested compounds and the total concentration of the
compounds of the broad chlorogenic acid family occurring in
coffee brew were then determined in CR. This analysis required
a preliminary step. The more polar compounds (glyoxal,
methylglyoxal, and diacetyl) were separated from the less polar
compounds (5-hydroxymethylfurfural, chlorogenic acids, and
caffeine) by solid-phase extraction (SPE) using a C -SPE
Antibacterial Compound Isolation. The compound(s) re-
sponsible for the antibacterial activity detected in roasted coffee
was investigated. To isolate the active coffee components and
obtain preliminary indications of their molecular mass, CR was
dialyzed using membranes with different cutoffs starting from
1
8
cartridge. Glyoxal, methylglyoxal, and diacetyl were recovered
with water (SPE-F1), whereas elution of the less water soluble
5-hydroxymethylfurfural, chlorogenic acids, and caffeine (SPE-
F2) required methanol.
3
500 Da. This procedure enabled the separation of naturally
occurring low-molecular mass coffee compounds (e.g., caffeine,
chlorogenic acids, and trigonelline) or Maillard reaction-induced
low-molecular mass products (e.g., 5-hydroxymethylfurfural and
R-dicarbonyl compounds) from the higher-molecular mass
components, such as polysaccharides, proteins, and medium-
and high-molecular mass melanoidins. Dialysates and retentates
were freeze-dried; dry residues were assessed, then dissolved
in 1 mL of Millipore-grade water (10× concentration), and
tested for antibacterial activity.
Assessment of R-dicarbonyl compounds by RP-HPLC–DAD
required the formation of quinoxaline derivatives (21). SPE-F1
(Figures 1 and 2B) was derivatized using 1,2-diaminobenzene
(pH 8.0) for 3 h at 60 °C after the addition of 5-methylqui-
noxaline, used as the internal standard (Figure 2C). Derivatives
were identified by comparison of their retention times (Figure
2A) and UV spectra with those of the standard compounds.
Quantitative analysis of R-dicarbonyl compounds by the external
standard method yielded the CR glyoxal, methylglyoxal, and
diacetyl concentrations reported in Table 1.
Results (Table 2) showed that only the dialysate (containing
9
4.8 mg/mL dry matter), corresponding to 76% of coffee brew
5-Hydroxymethylfurfural, chlorogenic acids, and caffeine
were analyzed by RP-HPLC–DAD by direct injection of diluted
SPE-F2 (Figure 2C,D) and were identified by comparing their
retention times (Figure 2A,B) and UV spectra with those of
the standard compounds. The other chlorogenic acids (26)
occurring in roasted coffee (such as 5-O-caffeoylquinic acid
isomers, feruoylquinic acids, and di- and tricaffeoyl/feruoylqui-
nic acids) were identified by their characteristic UV spectra (λmax
) 324 nm) and quantified by the external standard method. Their
content was expressed as 5-O-caffeoylquinic acid (Table 1).
dry matter, possesses antibacterial activity and that it has the
same MIC values (expressed as % by volume) as CR against
the two bacteria. The dialysate was then subjected to sequential
dialysis with 1000 and 500 Da membrane cutoffs; dialysates
and retentates were tested for antibacterial activity. Both <1000
and <500 Da dialysates, containing 58.1 and 54.1 mg/mL dry
matter, respectively, were found to be active (Table 2) and to
have the same MIC values (% by volume), indicating that the
molecular mass of roasted coffee antibacterial compounds is
<
500 Da. The decreased level of dialysate dry matter explains
To establish their actual role in roasted coffee antibacterial
activity, all the coffee components considered were tested in
different mixtures at the concentrations found in concentrated
the decrease in the dialysate MIC when expressed as milligrams
per milliliter. Conversely, none of the retentates were seen to

Roasted Coffee Antibacterial Compounds
J. Agric. Food Chem., Vol. 55, No. 25, 2007 10211
a
Table 2. Antibacterial (MIC) and Bactericidal (MBC) Activities of Roasted C. robusta Extract (CR)
Sa. aureus ATCC 25923
St. mutans 9102
MIC
MBC
MIC
MBC
coffee sample
coffee extract
% v/v
mg/mL
% v/v
mg/mL
% v/v
mg/mL
% v/v
mg/mL
20 ( 2
20 ( 1
>50
20 ( 2
>50
20 ( 1
>50
24.9 ( 2.4
30b ( 2
nd
37.2 ( 2.5
–
10 ( 1
10 ( 1
>50
10 ( 1
>50
10 ( 1
>50
12.4 ( 1.2
20b ( 2
nd
24.9 ( 2.5
–
>3500 Da dialysate
<3500 Da retentate
>1000 Da dialysate
<1000 Da retentate
>500 Da dialysate
<500 Da retentate
18.8 ( 2.3
9.4 ( 0.9
–
–
b
b
b
b
11.6 ( 1.2
–
10.8 ( 0.5
–
nd
nd
–
–
5.8 ( 0.6
–
5.4 ( 0.5
–
nd
nd
–
–
a
MIC values of CR dialysis fractions (dialysate and retentate) were obtained with different cutoff membranes against Sa. aureus and St. mutans. All experiments were
b
performed in triplicate. Not determined.
a
Table 3. Antibacterial (MIC) and Bactericidal (MBC) Activities of Bioactive Coffee Components against Sa. aureusand St. mutans
Sa. aureus ATCC 25923
St. mutans 9102
MIC
MBC
MIC
MBC
coffee compound
% v/v
µg/mL
% v/v
µg/mL
% v/v
µg/mL
% v/v
µg/mL
glyoxal
methylglyoxal
diacetyl
15 ( 2
15 ( 1
10 ( 1
88.3 ( 11.8
30 ( 2
40 ( 3
10 ( 2
176.7 ( 11.8
294.0 ( 22.1
114.2 ( 11.4
2 ( 0.2
2 ( 0.2
3 ( 0.2
11.8 ( 1.2
14.7 ( 1.5
34.3 ( 2.3
2 ( 0.2
2 ( 0.2
6 ( 0.3
11.8 ( 1.2
14.7 ( 1.5
68.5 ( 3.4
110.2 ( 5.9
114.2 ( 11.4
Sa. aureus ATCC 25923
St. mutans 9102
MIC
MBC
MIC
MBC
coffee compound
% v/v
mg/mL
% v/v
mg/mL
% v/v
mg/mL
% v/v
mg/mL
5
-hydroxymethylfurfural
30 ( 2
35 ( 2
8.4 ( 0.6
6.3 ( 0.4
–
>50
>50
>50
–
–
–
10 ( 1
15 ( 3
20 ( 1
2.8 ( 0.3
2.7 ( 0.5
5.0 ( 0.2
20 ( 1
20 ( 3
50 ( 3
5.6 ( 0.3
3.6 ( 0.5
12.5 ( 0.7
5-O-caffeoylquinic acid
caffeine
>50
a
All experiments were performed in triplicate.
Figure 1. RP-HPLC–DAD chromatograms. (A) Standard quinoxaline derivative solution detected at 314 nm and SPE-F1 results obtained from roasted coffee
CR (B) before and (C) after derivatization with 1,2-diaminobenzene: (1) quinoxaline, (2) 2-methylquinoxaline, (3) 5-methylquinoxaline, and (4) 2,3-dimethylquinoxaline.
1
0× CR (Table 4). Analysis of the R-dicarbonyl compound
affectMICvalues,indicatingthatchlorogenicacidsand5-hydroxy-
methylfurfural at the concentrations found in roasted coffee do
not influence its antibacterial activity. Conversely, the presence
of caffeine (mixture 4) in the previously tested R-dicarbonyl
compound mixture considerably reduced the R-dicarbonyl
compound mixture MIC values, which were close to the CR
MIC values. This indicates a synergistic effect of caffeine with
respect to R-dicarbonyl compound activity. Therefore, in the
mixture showed that together they account for about 50% of
the antibacterial activity of roasted coffee, the MIC values of
the R-dicarbonyl compound mixture being about twice the CR
MIC values. In contrast, analysis of the mixture of naturally
occurring compounds revealed no activity. The addition of 5-O-
caffeoylquinic acid (mixture 2) or 5-hydroxymethylfurfural
(
mixture 3) to the R-dicarbonyl compound mixture failed to

10212 J. Agric. Food Chem., Vol. 55, No. 25, 2007
Daglia et al.
Figure 2. RP-HPLC–DAD chromatograms. Standard 5-hydroxymethylfurfural, 5-O-caffeoylquinic acid, and caffeine solution detected at (A) 280 and (B)
24 nm. SPE-F2 results obtained from roasted coffee CR detected at (C) 280 and (D) 324 nm: (1) 5-hydroxymethylfurfural, (2) caffeine, and (3) 5-O-
caffeoylquinic acid.
3
Table 4. Antibacterial (MIC) and Bactericidal (MBC) Activities of Standard
Bioactive Coffee Component Mixtures at the Same Concentrations Found
in CR (10× concentrated) against Sa. aureus and St. mutans
In conclusion, naturally occurring coffee compounds do not
explain the antibacterial activity of coffee brew. Therefore,
R-dicarbonyl compounds, formed during the roasting process,
were investigated. Glyoxal, methylglyoxal, and diacetyl have
so far been studied largely for their cytotoxic, mutagenic, and
genotoxic activities, whereas to the best of our knowledge, their
antibacterial activity has never been investigated. The very low
MIC values of standard glyoxal, methylglyoxal, and diacetyl
point to these R-dicarbonyl compounds as the main agents
responsible for the beverage’s antibacterial activity. Neverthe-
less, their low concentrations in roasted coffee do not account
for all of the beverage’s antibacterial activity. However, their
MICs are close to coffee brew MIC values when the synthetic
mixture of R-dicarbonyl compounds added with caffeine is
tested at the same concentrations found in coffee. These findings
indicate that caffeine can synergistically enhance R-dicarbonyl
compound activity and that glyoxal, methylglyoxal, and diacetyl
in the presence of caffeine account for the whole antibacterial
activity of roasted coffee.
a
Sa. aureus ATCC 25923
St. mutans 9102
MIC
% v/v
MBC
% v/v
MIC
% v/v
MBC
% v/v
coffee compound mixture
mixture 1
mixture 2
mixture 3
mixture 4
mixture 5
mixture 6
50 ( 2
>50
20 ( 1
40 ( 1
>50
>50
>50
>50
50 ( 2
50 ( 2
>50
>50
50 ( 2
50 ( 3
30 ( 1
40 ( 2
20 ( 1
20 ( 2
9 ( 1
40 ( 3
40 ( 2
30 ( 2
20 ( 3
9 ( 1
a
All experiments were performed in triplicate.
presence of R-dicarbonyl compounds, caffeine, which has weak
intrinsic antibacterial activity, is capable of enhancing their total
activity and appears to be responsible for approximately half
of the antibacterial activity of roasted coffee. The synergistic
activity of caffeine is observed even when a lower concentration
of caffeine (5 mg/mL) is added to R-dicarbonyl compound
mixture, corresponding to approximately the residue mean
caffeine content occurring in decaffeinated coffee (10× con-
centrated) (27). This findings justify the results obtained by our
group in a previous investigation (data not shown) that indicated
that decaffeinated coffee brew exhibited antibacterial activity
not significantly different from that in coffee brew.
ABBREVIATIONS USED
MIC, minimal inhibitory concentration; MBC, minimal
bactericidal concentration; SPE, solid-phase extraction; THB,
Todd Hewitt broth; ISB, Iso-Sensitest broth; CR, C. robusta
coffee extract.

Roasted Coffee Antibacterial Compounds
J. Agric. Food Chem., Vol. 55, No. 25, 2007 10213
ACKNOWLEDGMENT
selected coffee chemical compounds against Enterobacteria. J.
Agric. Food Chem. 2006, 54, 8738–8743.
16) Dogazaki, C.; Shindo, T.; Furuhata, K.; Furuyama, M. Identifica-
tion of chemical antibacterial components against Legionella
pneumophila in a coffee beverage. J. Pharm. Soc. Jpn. 2002, 122,
We thank MOKA SIR’S srl (Pavia, Italy) for providing the
coffee samples.
(
(
(
LITERATURE CITED
4
87–494.
17) Furuhata, K.; Dogazaki, C.; Hara, M.; Furuyama, M. Inactivation
of chemical antibacterial components against Legionella pneu-
mophila by phenol compounds contained in coffee. Bokin Bobai
(
(
(
1) Higdon, J. V.; Frei, B. Coffee and health: A review of recent
human research. Crit. ReV. Food Sci. Nutr. 2006, 46, 101–123.
2) Dorea, J. G.; da Costa, T. H. M. Is coffee a functional food? Br. J.
Nutr. 2005, 93, 773–782.
3) Cadden, I. S. H.; Partovi, N.; Yoshida, E. M. Review article:
Possible beneficial effects of coffee on liver disease and function.
Aliment. Pharmacol. Ther. 2007, 26, 1–7.
2
002, 30, 291–297.
18) Tarsi, R.; Muzzarelli, R. A. A.; Guzmàn, C. A.; Pruzzo, C.
Inhibition of Streptococcus mutans adsorption to hydroxyapatite
by low molecular weight chitosans. J. Dent. Res. 1997, 76, 665–
6
72.
(
(
4) Bonita, J. S.; Mandarano, M.; Donna, S.; Vinson, J. Coffee and
cardiovascular disease: In vitro, cellular, animal, and human
studies. Pharmacol. Res. 2007, 55, 187–198.
5) Daglia, M.; Racchi, M.; Papetti, A.; Lanni, C.; Govoni, S.;
Gazzani, G. In vitro and ex vivo antihydroxyl radical activity of
green and roasted coffee. J. Agric. Food Chem. 2004, 52, 1700–
(19) NCCLS. Methods for dilution antimicrobial susceptibility tests
for bacteria that grow aerobically. In ApproVed Standard M7-A6,
6th ed.; NCCLS: Wayne, PA, 2003.
(20) NCCLS. Methods for determining bactericidal activity on anti-
microbial agents. In ApproVed Guidelines; NCCLS: Wayne, PA,
1999; Vol. 19.
1
704.
(
6) Daglia, M.; Papetti, A.; Gregotti, C.; Bertè, F.; Gazzani, G. In
vitro antioxidant and ex vivo protective activities of green and
roasted coffee. J. Agric. Food Chem. 2000, 48, 1449–1454.
(21) de Revel, G.; Pripis-Nicolau, L.; Berbe, J. C.; Bertrand, A. The
detection of R-dicarbonyl compounds in wine by formation of
quinoxaline derivatives. J. Sci. Food Agric. 2000, 80, 102–108.
(
(
(
7) Morales, F. J.; Babbel, M. B. Melanoidins exert a weak antiradical
(
(
(
(
(
22) Shibamoto, T. Analytical methods for trace levels of reactive
carbonyl compounds formed in lipid peroxidation system.
J. Pharm. Biomed. Anal. 2006, 41, 12–25.
23) Murkovic, M.; Pichler, N. Analysis of 5-hydroxymethylfurfural
in coffee, dried fruits and urine. Mol. Nutr. Food Res. 2006, 50,
activity in water fluids. J. Agric. Food Chem. 2002, 50, 4657–
4
661.
8) Wen, X.; Takenaka, M.; Murata, M.; Homma, S. Antioxidative
activity of zinc-chelating substance in coffee. Biosci., Biotechnol.,
Biochem. 2004, 68, 2313–2318.
9) Fuster, M. D.; Mitchell, A. E.; Ochi, H.; Shibamoto, T. Antioxi-
dative activities of heterocyclic compounds formed in brewed
coffee. J. Agric. Food Chem. 2000, 48, 5600–5603.
8
42–846.
24) Murkovic, M.; Bornik, M. A. Formation of 5-hydroxymethyl-2-
furfural (HMF) and 5-hydroxymethyl-2-furoic acid during roasting
of coffee. Mol. Nutr. Food Res. 2007, 51, 390–394.
25) Antal, M. J.; Mok, W. S. L.; Richards, G. N. Mechanism of
formation of 5-hydroxymethyl-2-furaldehyde from D-fructose and
sucrose. Carbohydr. Res. 1990, 199, 91–109.
26) Clifford, M. N.; Johnston, K. L.; Knight, S.; Kuhnert, N.
Hierarchical scheme for LC-MSn identification of chlorogenic
acids. J. Agric. Food Chem. 2003, 51, 2900–2911.
(
10) Daglia, M.; Tarsi, R.; Papetti, A.; Grisoli, P.; Dacarro, C.; Pruzzo,
C.; Gazzani, G. Antiadhesive effect of green and roasted coffee
on Streptococcus mutans adhesive properties on saliva-coated
hydroxyapatite beads. J. Agric. Food Chem. 2002, 50, 1225–1229.
11) Daglia, M.; Cuzzoni, M. T.; Dacarro, C. Antibacterial activity of
coffee. J. Agric. Food Chem. 1994, 42, 2270–2272.
12) Daglia, M.; Cuzzoni, M. T.; Dacarro, C. Antibacterial activity of
coffee: Relationship between biological activity and chemical
markers. J. Agric. Food Chem. 1994, 42, 2273–2277.
(
(
(27) McCusker, R. R.; Fuehrlein, B.; Goldberger, B. A.; Gold, M. S.;
Cone, E. J. Caffeine content of decaffeinated coffee. J. Anal.
Toxicol. 2006, 30, 611–613.
(
(
(
13) Dacarro, C.; Daglia, M.; Cuzzoni, M. T.; Bonferoni, B. Antibacte-
rial activity of coffee against Streptococcus mutans. Ig. Mod. 1995,
1
04, 379–387.
Received for review July 27, 2007. Revised manuscript received
September 17, 2007. Accepted September 24, 2007. The work was
supported by a grant from MIUR 2005 to G.G. and took place within
the Interdepartmental Center of Food Quality and Safety (CISQUA).
14) Daglia, M.; Papetti, A.; Dacarro, C.; Gazzani, G. Isolation of an
antibacterial component from roasted coffee. J. Pharm. Biomed.
Anal. 1998, 18, 219–225.
15) Amelia, A.; Almeida, P.; Farah, A.; Silva, D. A. M.; Nunan, E. A.;
Gloria, M. B. A. Antibacterial activity of coffee extracts and
JF0722607