Characterization by LC-MSn of four new classes of chlorogenic acids in green coffee beans: Dimethoxycinnamoylquinic acids, diferuloylquinic acids, caffeoyl-dimethoxycinnamoylquinic acids, and feruloyl- dimethoxycinnamoylquinic acids

Article abstract of DOI:10.1021/jf0601665

LC-MS⁴ has been used to detect and characterize in green coffee beans 12 chlorogenic acids not previously reported in nature. These comprise three isomeric dimethoxycinnamoylquinic acids (7-9) (M_r 382), three caffeoyl-dimethoxycinnam

Full text of DOI:10.1021/jf0601665

J. Agric. Food Chem. 2006, 54, 1957−1969 1957
Characterization by LC-MSn of Four New Classes of
Chlorogenic Acids in Green Coffee Beans:
Dimethoxycinnamoylquinic Acids, Diferuloylquinic Acids,
Caffeoyl-dimethoxycinnamoylquinic Acids, and
Feruloyl-dimethoxycinnamoylquinic Acids
MICHAEL N. CLIFFORD,*,† SUSAN KNIGHT,† BIRGUL SURUCU,‡ AND
‡
NIKOLAI KUHNERT
School of Biomedical and Molecular Sciences, University of Surrey, Guildford,
Surrey GU2 7XH, United Kingdom
LC-MS4 has been used to detect and characterize in green coffee beans 12 chlorogenic acids not
previously reported in nature. These comprise three isomeric dimethoxycinnamoylquinic acids (7-9)
(Mr 382), three caffeoyl-dimethoxycinnamoylquinic acids (22, 24, and 26) (Mr 544), three diferuloylquinic
acids (13-15) (Mr 544), and three feruloyl-dimethoxycinnamoylquinic acids (28, 30, and 32) (Mr 558).
Structures have been assigned on the basis of LC-MS4 patterns of fragmentation and relative
hydrophobicity and, in the case of the dimethoxycinnamoylquinic acids, by comparison with authentic
standards. Several new structure-diagnostic fragmentations have been identified for use with diacyl-
chlorogenic acids, for example, m/z 299 and 255 for C4 caffeoyl, m/z 313 and 269 for C4 feruloyl,
nearly equal elimination of both cinnamoyl residues for vic-3,4-diacyl, and an increasing ratio of
“dehydrated” ions to “non-dehydrated” ions at MS2 with increasing methylation of those cinnamoyl
residues. Possible mechanisms have been proposed to account for the fragmentations observed.
The mass spectrometric resolution of six isomeric chlorogenic acids (Mr 544) in a crude plant extract
by fragment-targeted LC-MS2 and LC-MS3 experiments illustrates the analytical power and advantage
of ion trap mass spectroscopy.
KEYWORDS: Caffeoyl-dimethoxycinnamoylquinic acids; chlorogenic acids; coffee; diferuloylquinic acids;
dimethoxycinnamoylquinic acids; feruloyl-dimethoxycinnamoylquinic acids; LC-MSn
INTRODUCTION
acid conjugates of amino acids (8), the possibility of additional
quinic acid conjugates, for example, diferuloylquinic acids, could
not be excluded. Accordingly, the structure-diagnostic LC-MSn
procedures previously developed (5, 8, 9) were applied to
methanolic extracts of commercial green Robusta coffee beans.
Classically, chlorogenic acids are a family of esters formed
between quinic acid and certain trans-cinnamic acids, most
commonly caffeic, p-coumaric, and ferulic (1-3). Structures
are shown in Figure 1. In the IUPAC system (-)-quinic acid
is defined as 1L-1(OH),3,4/5-tetrahydroxycyclohexanecarboxylic
acid, but Eliel and Ramirez (4) now propose 1R,3R,4R,5R-
tetrahydroxycyclohexanecarboxylic acid. Chlorogenic acids are
widely distributed in plants (2, 3), but the coffee bean is
remarkably rich, containing at least 18 chlorogenic acids that
are not acylated at C1. These have been subdivided into five
classes, that is, three caffeoylquinic acids, three p-coumaroyl-
quinic acids, three feruloylquinic acids, three dicaffeoylquinic
acids, and six caffeoyl-feruloylquinic acids (5). A considerable
number of chlorogenic acid-like compounds have also been
reported in coffee (6, 7), and while some of these are cinnamic
MATERIALS AND METHODS
Methanolic Extracts of Coffee Beans. Methanolic extracts of green
Robusta coffee beans were prepared as previously described (5). The
extracts were treated with Carrez reagents (1 mL of reagent A plus 1
mL of reagent B) (10) to precipitate colloidal material, diluted to 100
mL with 70% v/v aqueous methanol, and filtered through a Whatman
no. 1 filter paper. The methanol was removed by evaporation with
nitrogen and the aqueous extract stored at -12 °C until required, thawed
at room temperature, centrifuged (1360g, 10 min), and used directly
for LC-MS.
As required, the methanolic extract of Robusta coffee beans was
treated with tetramethylammonium hydroxide (TMAH) to interesterify
and transesterify the diacyl-chlorogenic acids. Extract (200 µL) was
treated with TMAH (20 µL) at room temperature for periods of 1, 3,
and 5 min. The reaction was terminated by adding 3.5 M acetic acid
* Author to whom correspondence should be addressed (telephone 44
14 83 68 97 03; fax 44 14 83 68 64 01; e-mail m.clifford@surrey.ac.uk).
† Centre for Nutrition and Food Safety.
‡ Department of Chemistry.
10.1021/jf0601665 CCC: $33.50
© 2006 American Chemical Society
Published on Web 03/01/2006

1958 J. Agric. Food Chem., Vol. 54, No. 6, 2006
Clifford et al.

LC-MSn of Diferuloylquinic Acids
J. Agric. Food Chem., Vol. 54, No. 6, 2006 1959

1960 J. Agric. Food Chem., Vol. 54, No. 6, 2006
Clifford et al.
Figure 1. Structures of selected coffee bean chlorogenic acids (IUPAC numbering) (1).
(40 µL) essentially as previously described (5, 11, 12). The reaction
products were stored at -12 °C until required, thawed at room
temperature, centrifuged (1360g, 10 min), and used directly for LC-
MS.
Synthesis of Standards. Methyl 3,4-Dimethoxycinnamate. Acetyl
chloride (0.1 mL) was added at 0 °C to 10 mL of methanol and stirred
for 10 min. 3,4-Dimethoxycinnamic acid (300 mg; 1.44 mmol) was
added to the solution, and the mixture was stirred for 20 h at room
temperature. The solvent was removed under vacuum, and the residue
was recrystallized from methanol and toluene to give the title compound
as a white powder (266 mg, 83%): 1H NMR (CDCl3, 500 MHz) δH
7.64 (1H, d, J ) 15.9 Hz, HCdC), 7.11 (1H, d, J ) 8.3 Hz, ArH),
7.09 (1H, s, ArH), 6.87 (1H, d, J ) 8.3 Hz, ArH), 6.31 (1H, d, J )
15.9 Hz, CdCH), 3.91 (6H, s, OCH3), 3.80 (s, 3H, OCH3); MS (ESI),
m/z 221 [M - H+]-.
1-(3,4-Dimethoxycinnamoyl)quinic Acid (34). To a suspension of 214
mg (1 mmol) of 3,4-isopropylidenequinide and 12 mg of N,N-
(dimethylamino)pyridine (0.1 mmol) in 5 mL of dichloromethane was
added 110 mg of triethylamine followed by 250 mg (1.1 mmol) of
3,4-dimethoxycinnamoyl acid chloride, and the mixture was stirred at
20 °C for 48 h. HCl (1 M, 10 mL) and 10 mL of dichloromethane
were added; the organic phase was separated, washed with 5 mL of
brine, dried over Na2SO4, and filtered; and the solvent was removed
under reduced pressure to give 302 mg of a crude ester. This crude

LC-MSn of Diferuloylquinic Acids
J. Agric. Food Chem., Vol. 54, No. 6, 2006 1961
product [200 mg (≈0.50 mmol)] was dissolved in 8 mL of tetra-
hydrofuran and treated with 0.013 g of LiOH (0.54 mmol in 4 mL of
H2O). The solution was stirred for 3 days at room temperature and
was then quenched by the addition of 2 M HCl. The reaction mixture
was extracted with dichloromethane (3 × 30 mL) and 2 M HCl (10
mL). The organic phase was dried with MgSO4 and filtered. The solvent
was removed under reduced pressure to provide after recrystallization
from ethanol the title compound (160 mg, 84% yield) as an off-white
powder: Rf (TLC) 0.60 (MeCN); Rf (LC-MS, gradient as below) 32.3
min (UV λmax 325 nm; MS, m/z 381.0 [M - H+]-); mp 140 °C; IRνmax
(Nujol)/cm-1 3368 (OH), 2954 (COOH), 2854 (CH), 1797, 1712
(CdO), 1628 (CArdCAr), 1075, 1156 (C-O); 1H NMR (500 MHz, D2O)
δ 1.97 (1H, dd, J ) 13.7, 10.9; H-6ax), 2.29 (1H, dd, J ) 15.5,
3.4, H-2ax), 2.50 (1H, ddd, J ) 13.7, 4.4, 3.4 Hz, H-6eq), 2.59 (1H,
dt, J ) 15.5, 3.4 Hz, 1H, H-2eq), 3.63 (1H, dd, J ) 9.2, 3.4 Hz, H-4),
4.16, 3.86 (3H, s, OMe), 3.88 (3H, s, OMe) 4.16 (1H, ddd, J ) 10.9,
9.2, 4.4 Hz, H-5), 4.27 (1H, q, J ) 3.4 Hz, H-3), 6.43 (1H, d, J ) 15.9
Hz; CdCH), 6.99 (1H, d, J ) 8.3 Hz; CH), 7.18 (1H, d, J ) 1.9 Hz,
CH), 7.20 (1H, dd, J ) 8.3, 1.9 Hz, CH), 7.63 (1H, d, J ) 15.9 Hz;
CdCH); 13C NMR(125 MHz, D2O) δ 34.3 (C-6), 38.51 (C-2), 55.64
(OMe), 55.73 (OMe), 66.13 (C-5), 68.62 (C-3), 74.46 (C-4), 80.93 (C-
1), 110.26 (CHAr), 111.57 (CHAr), 114.77 (CdCH), 123.65 (C-9),
127.19 (C-12), 146.93 (C-10), 148.26 (CdC), 150.69 (CAr), 168.16
(CdO), 175.53 (CdO); accurate mass of C18H11O9 requires 382.1264,
found 382.1266.
Synthesis of Mixture of Isomers of (3,4-Dimethoxycinnamoyl)quinic
Acid (7-9, 34). To a suspension of 23 mg (0.1 mmol) of quinic acid
in 1 mL of dichloromethane was added 10 mg of triethylamine followed
by 12 mg (0.05 mmol) of 3,4-dimethoxycinnamoyl acid chloride, and
the mixture was stirred at 20 °C for 48 h; the solvent was removed
under reduced pressure, and the mixture was dried in a vacuum. The
crude reaction mixture was dissolved in 1 mL of methanol, filtered
and subjected to LC-MS conditions as described below to show in the
SIM mode at m/z 381 four isomers of dimethoxycinnamoylquinic acids
[estimated by TIC: 80% 5-dimethoxycinnamoylquinic acid (9), 10%
3-dimethoxycinnamoylquinic acid (7), 7% 4-dimethoxycinnamoylquinic
acid (8), and 3% 1-dimethoxycinnamoylquinic acid (34)].
LC-MSn. The LC equipment (ThermoFinnigan, San Jose, CA)
comprised a Surveyor MS pump, an autosampler with a 50 µL loop,
and a PDA detector with a light-pipe flow cell (recording at 320, 280,
and 254 nm and scanning from 200 to 600 nm). This was interfaced
with an LCQ Deca XP Plus mass spectrometer fitted with an ESI source
(ThermoFinnigan) and operating in zoom scan mode for the accurate
determination of parent ion m/z, and in data-dependent, full-scan, MSn
mode to obtain fragment ion m/z. For better discrimination of isomers
having Mr 544, additional MS2 and MS3 experiments were performed
that focused only on compounds producing a parent ion at m/z 543.
MS operating conditions (negative ion) had been optimized using
5-caffeoylquinic acid (3) with a collision energy of 35%, an ionization
voltage of 3.5 kV, a capillary temperature of 350 °C, a sheath gas flow
rate of 65 arbitrary units, and an auxiliary gas flow rate of 10 arbitrary
units.
Separations were achieved on 150 × 3 mm i.d. columns containing
Luna 5µ phenylhexyl packing (Phenonemex, Macclesfield, U.K.).
Solvent A was water/acetonitrile/glacial acetic acid (980:20:5, v/v, pH
2.68): solvent B was acetonitrile/ glacial acetic acid (1000:5, v/v).
Solvents were delivered at a total flow rate of 300 µL/min. The gradient
profile was 4% B to 33% B linearly in 90 min, a linear increase to
100% B at 95 min, followed by 5 min isocratic, a return to 4% B at
105 min, and 5 min isocratic to re-equilibrate.
located (5). Selected ion monitoring at m/z 543 immediately
located five chromatographic peaks eluting between 74 and 82
min, each with a UV spectrum typical of chlorogenic acids [λmax
325 nm with a shoulder (85%) at 290 nm]. However, because
coffee beans do not acylate quinic acid at C1 and produce only
three dicaffeoylquinic acid isomers, that is, 3,4-dicaffeoylquinic
acid, 3,5-dicaffeoylquinic acid, and 4,5-dicaffeoylquinic acid
(10-12), only three diferuloylquinic acids (13-15) were
expected. The five peaks produced MS3 fragment ions at either
m/z 173 or 191 consistent with the presence of a quinic acid
residue (5), and the absence of an MS3 fragment ion at m/z 205
confirmed that none of these substances were derivatives of
methyl quinate. Two of the five peaks produced MS2 ions at
m/z 367 ([feruloylquinic acid - H+]-) and m/z 349 ([feruloyl-
quinic acid - H2O - H+]-) and MS3 ions at m/z 193 ([ferulic
acid - H+]-) analogous to or identical with those produced by
dicaffeoylquinic acids and caffeoyl-feruloylquinic acids (5) and,
thus, totally consistent with the predictable behavior of di-
feruloylquinic acids.
The other three peaks variously produced MS2 ions at m/z
335, 353, 363, and 381 and MS3 ions at m/z 179 and 207. The
MS2 ions at m/z 335 and 353 and the MS3 ion at m/z 179 are
characteristic (5) of caffeic acid-derived diacyl-chlorogenic acids
and can be assigned as [caffeoylquinic acid - H+]-, [caffeoyl-
quinic acid - H2O - H+]-, and [caffeic acid - H+]-,
respectively. The remaining ions were tentatively assigned to a
similar series of fragments 28 amu larger than the caffeic acid-
related fragments (or 14 amu larger than the ferulic acid-related
fragments), strongly suggesting that they might be derived from
compounds in which both methyl residues are associated with
the same cinnamic acid residue, that is, with quinic acid bearing
one residue of caffeic acid and one residue of 3,4-dimethoxy-
cinnamic acid, the dimethyl ether of caffeic acid. Although 3,4-
dimethoxycinnamic acid had previously been reported in green
Robusta coffee beans (13), its quinic acid esters had not. Those
compounds of Mr 544 that produced MS2 ions at m/z 363 or
381 were tentatively assigned as caffeoyl-dimethoxycinnamoyl-
quinic acids. Although it is conceivable that a caffeoyl-
dimethoxycinnamoylquinic acid might produce ions at m/z 349
or 367, diferuloylquinic acids would not produce ions at m/z
363 or 381, and there can be no doubt that both series of
chlorogenic acids were present.
A post hoc search for ions at m/z 381 located an additional
four peaks in the spectrum from the Robusta extractsone was
a parent ion, and three were MS2 fragments from m/z 557. These
were tentatively assigned to a dimethoxycinnamoylquinic acid
and three feruloyl-dimethoxycinnamoylquinic acids. These
substances were further investigated using more sensitive and
more specific LC-MS protocols: (i) MS1 experiments targeting
m/z 381, 543, and 557; (ii) MS2 experiments targeting combina-
tions of m/z 543 + 381, m/z 543 + 367, m/z 543 + 363, m/z
543 + 353, m/z 543 + 349, and m/z 543 + 335; and (iii) MS2
experiments targeting combinations of m/z 557 + 381, m/z 557
+ 367, m/z 557 + 363, and m/z 557 + 349.
Characterization of Putative Dimethoxycinnamoylquinic
Acids (7-9). The Robusta extract contained three minor
components with molecular ions at m/z 381 that eluted between
39 and 50 min. All three peaks had UV spectra typical of
chlorogenic acids. MS2 and MS3 data are presented in Table 1
and Figure 2.
RESULTS AND DISCUSSION
Preliminary Assessment of Data. All data for chlorogenic
acids presented in this manuscript use the recommended IUPAC
numbering system (1), and structures are presented in Figure
1. When necessary, previously published data have been
amended to ensure consistency and avoid ambiguity.
Because the monoacyl chlorogenic acids so far examined on
this phenylhexyl packing elute in the sequence 3-acyl, 5-acyl,
and 4-acyl (5, 14), the first dimethoxycinnamoylquinic acid was
tentatively assigned as the 3-isomer (7). The MS2 base peak at
The Robusta coffee extract gave a typical chromatogram in
which the 18 previously reported chlorogenic acids were easily

1962 J. Agric. Food Chem., Vol. 54, No. 6, 2006
Clifford et al.
Table 1. MS3 Fragmentation Data for the Putative Dimethoxycinnamoylquinic Acids
MS2
parent base
MS3
base
peak
m/z
MS2 secondary ions
MS3 secondary ions
ion
m/z
peak
m/z
compd Na
m/z
intensity
m/z
intensity
m/z
intensity
12
m/z
intensity
m/z
intensity
25
m/z
intensity
m/z
intensity
40
7
6
6
6
381.0 207.0 173.1
381.0 173.1
381.1 193.0
5
100
bpb
207.1
100
10
149.1 162.8
93.4 162.7
134.1 149.0
93.4
65
85
25
133.1
8
bp
155.0
15
111.0
9
134.1
34
381.1 173.1 143.1
15
298.9
10
a Number of LC-MS analyses used to collect MS data. b Occurs as base peak.
Figure 2. MS2 spectra of the putative dimethoxycinnamoylquinic acids.
m/z 207 [dimethoxycinnamoylquinic acid - quinic acid - H+]-,
which subsequently decarboxylates and demethylates at MS3,
is analogous in behavior to 3-feruloylquinic acid (4) (5). The
most strongly retained of the dimethoxycinnamoylquinic acid
isomers has the fragmentation characteristic of a 4-acylchloro-
genic acid (MS2 and MS3 base peaks at m/z 173 and 93,
respectively) and can be reliably assigned as 4-dimethoxy-
cinnamoylquinic acid (8). Peak 13, by a process of elimination
and on the basis of its retention time, should be 5-dimethoxy-
cinnamoylquinic acid (9). However, its fragmentation is atypical,
giving an MS2 base peak of m/z 193 (rather than the expected
m/z 207), indicative of demethylation.
We have proposed previously (5) that the elimination of a
cinnamoyl residue from either C3 or C5 involves quinic acid
chairs with a 1,3-syn diaxial conformation such that the C1
carboxyl protonates the C5 cinnamoyl residue or the C1
hydroxyl protonates the C3 cinnamoyl residue. For C3 the ease
with which the cinnamoyl residue is lost and the nature of the
subsequent base peak are functions of the identity of the
cinnamoyl residue.
To explain the behavior of the 3,4-dimethoxycinnamoyl
residue we must consider two alternative fragmentation mech-
anisms, but we are currently unable to distinguish between them.
First, deprotonation at the most acidic site will produce an [M
- H+]- anion A (Figure 3). If the negative charge is localized
at the COO- group, this functionality can act as a nucleophile
in an intermolecular acylation reaction to give bicyclic inter-
mediate B via pathway A. Loss of the methyl cation can then
lead to quinoid structure C. This assumption accounts for the
ease of demethylation of 5-dimethoxycinnamoylquinic acid and
its derivatives. In an alternative scenario the resulting anion is
after proton transfer localized at an alcoholate oxygen, for
example, in D-1 or its resonant structure D-2. The [M - H+]-
anion D-1 is in this case stabilized by resonance and an
additional hydrogen bond from the carboxyl at C1 of the quinic
acid moiety. This hydrogen-bonded species appears to be
particularly stable and does not fragment easily in the case of
the caffeoyl derivatives (see D-2 for R ) H in Figure 3). In
the case of a 3,4-dimethoxycinnamoyl residue the hydrogen
To gain further evidence to support this assignment, we
synthesized pure 1-dimethoxycinnamoylquinic acid from a
literature-known 3,4-acetal-protected quinide (15). The synthetic
derivative displayed the expected NMR spectroscopic features
and eluted under the same gradient conditions at 32 min. It
showed the characteristic UV absorption and MS fragmentation
pattern expected for 1-dimethoxycinnamoylquinic acid with a
base peak in MS2 at m/z 173.0. Furthermore, we obtained a
mixture of the four possible isomers of dimethoxycinnamoyl-
quinic acid by acylating quinic acid with the acid chloride of
3,4-dimethoxycinnamic acid. All four isomers could be readily
detected by LC-MS using SIM at m/z 381.0 and, with the
exception of 1-dimethoxycinnamoylquinic acid, were identical
with respect to retention time, UV absorbance, and MS2
fragmentation to the compounds detected in coffee.

LC-MSn of Diferuloylquinic Acids
J. Agric. Food Chem., Vol. 54, No. 6, 2006 1963
Figure 3. Fragmentation mechanism proposed for dimethoxycinnamoylquinic acids.
bond still occurs in D-1, the quinoid resonance structure of
On the basis of its intermediate hydrophobicity, 14 has been
which is set up to demethylate easily to give via pathway B
fragment ion E, a quinic acid fragment F at m/z 173, and a
ketene quinide fragment at m/z 193 (Figure 3). Because all ESI
spectra obtained in this investigation arise from highly acidic
solvents, proton transfer should occur rapidly within the ESI
droplet. Demethylation of a dimethoxycinnamoyl residue at C5,
but not C3, can be attributed to the low pK of the protonating
C1 carboxyl (pK ∼ 3.5) compared with the C1 hydroxyl (pK
> 15), which seems to reduce the stability of the hydrogen-
bonded intermediate. Hence, a dimethoxycinnamoyl fragment
at m/z 207 is observed instead of its demethylated counterpart
at m/z 193.
Characterization of Diacyl-chlorogenic Acids with Mr 544.
The LC-MS experiment targeted on compounds producing m/z
543 parent ions located six peaks. Four eluted at 75.2, 75.8,
76.6, and 77.8 min followed by an incompletely resolved pair
at 80.3 and 80.7 min (Figure 4). The LC-MS experiments
targeting subsequent fragmentation to either m/z 381, 367, 363,
353, 349, or 335 identified (Figure 5) three with MS2 base peaks
at m/z 381 (eluting at ca. 75.1, 76.6, and 79.6 min), and these
were assigned tentatively as caffeoyl-dimethoxycinnamoyl qui-
nic acids (22-26). The peaks eluting at ca. 74.4, 75.5, and 79.6
min (with MS2 base peaks at m/z 349 or 367) were tentatively
assigned as diferuloylquinic acids (13-15).
assigned tentatively as 3,5-diferuloylquinic acid despite a
cinnamate-derived MS3 base peak at m/z 193 in contrast to the
quinic acid-derived m/z 191 produced by 3,5-dicaffeoylquinic
acid (11). On the basis of the fragmentation (5, 9) of 3,5-
dicaffeoylquinic acid (11) and the two 3,5-caffeoyl-feruloyl-
quinic acid isomers (18, 19) it is anticipated that 3,5-diferuloyl-
quinic acid (14) would lose the feruloyl residue at C5 before
that at C3, producing [3-feruloylquinic acid - H+]- (rather than
[5-feruloylquinic acid - H+]-) as its MS2 base peak. We have
previously demonstrated (5) that whereas 3-caffeoylquinic acid
(1) yields a quinic acid-derived m/z 191 base peak, 3-feruloyl-
quinic acid (4) yields [ferulic acid - H+]-. Accordingly, the
fragmentation of 14 is completely consistent with its assignment
as 3,5-diferuloylquinic acid (14).
Peak 13, incompletely resolved from the first eluting caffeoyl-
dimethoxycinnamoyl isomer (22), must logically be 3,4-di-
feruloylquinic acid (13). However, in contrast to 3,4-dicaffeoyl-
quinic acid (10), 13 produces a “dehydrated” MS2 base peak at
m/z 349, and this yields m/z 175 as the MS3 base peak. We
have previously observed that 3-feruloyl-4-caffeoylquinic acid
(16) produces a comparatively strong (45% of base peak) m/z
349 accompanied by m/z 175 at MS3, and the fragmentation of
13 is thus not inconsistent with its assignment as 3,4-di-
feruloylquinic acid (10). The tendency of some chlorogenic acids
to dehydrate during fragmentation, and the origin of the fragment
ion m/z 175, are further discussed below.
Characterization of the Putative Diferuloylquinic acids
(13-15). Table 2 and Figure 6 compare the fragmentation
patterns of the putative diferuloylquinic acids with data previ-
ously obtained for coffee bean dicaffeoylquinic acids (10-12).
Peak 15 produces MS2 and MS3 base peaks that are either
identical with those produced by 4,5-dicaffeoylquinic acid (12),
or 14 amu larger, and it has been assigned as 4,5-diferuloylquinic
acid (15).
Characterization of Putative Caffeoyl-dimethoxycinnamoyl-
quinic Acids (22-27). Of the six peaks that yielded molecular
ions at m/z 543, three produced an MS2 base peak at m/z 381
with an MS3 base peak or secondary fragment ion at m/z 207,
suggestive of caffeoyl-dimethoxycinnamoylquinic acids. Theo-
retically, six caffeoyl-dimethoxycinnamoylquinic acid isomers

1964 J. Agric. Food Chem., Vol. 54, No. 6, 2006
Clifford et al.
Figure 4. LC-MS of chlorogenic acids giving a parent ion at m/z 543.
Figure 5. Fragment-targeted LC-MS2 spectra of chlorogenic acids giving a parent ion at m/z 543.
(22-27) would have been expected, as seen previously (5) for
the caffeoyl-feruloylquinic acids (16-21).
cinnamoyl residues with nearly equal facility, giving an MS2
base peak (m/z 353) accompanied by a very intense (90% of
base peak) secondary ion at m/z 367. Significant amounts of
the corresponding dehydrated ions (m/z 335 and 349) were also
produced (5).
By analogy with the relative hydrophobicity of the caffeoyl-
feruloylquinic acids and dicaffeoylquinic acids (5, 9, 14) one
would expect that the first caffeoyl-dimethoxycinnamoylquinic
acid to elute would be one of the 3,4-caffeoyl-dimethoxy-
cinnamoylquinic acid isomers. The first eluting caffeoyl-
dimethoxycinnamoylquinic acid (22) loses its two cinnamoyl
residues with nearly equal facility (slightly favoring the caffeoyl
residue) and produces strong dehydrated ions at m/z 335 and
363, thus closely resembling 3-feruloyl-4-caffeoylquinic acid
(16) (Table 3) (5). Its MS3 base peak (m/z 207) is consistent
with the MS2 base peak being [3-dimethoxycinnamoylquinic
acid - H+]- rather than [4-dimethoxycinnamoylquinic acid -
Table 3 and Figure 6 compare the fragmentation patterns of
the three putative caffeoyl-dimethoxycinnamoylquinic acids with
data previously obtained for the six caffeoyl-feruloylquinic acids.
The caffeoyl-feruloylquinic acids were assigned by comparing
their fragmentation behavior at MS(n+1) with the MSn fragmen-
tation of caffeoylquinic acids and feruloylquinic acids and the
MS(n+1) fragmentation of the dicaffeoylquinic acids (5). For
chlorogenic acids not substituted at C1, it was observed that a
caffeoyl residue at C5 was the most easily eliminated, whereas
that at C4 was the most stable and that at C3 being of
intermediate stability. For five of the six caffeoyl-feruloylquinic
acids this resulted at MS2 in one of the two cinnamoyl residues
being eliminated almost exclusively (maximally 10% loss of
the second cinnamoyl residue in 3-caffeoyl-4-feruloylquinic
acid). In contrast, 3-feruloyl-4-caffeoylquinic acid lost both

LC-MSn of Diferuloylquinic Acids
J. Agric. Food Chem., Vol. 54, No. 6, 2006 1965
Table 2. MS4 Fragmentation Data for the Putative Diferuloylquinic Acids and Equivalent Dicaffeoylquinic Acids
MS2
parent base
MS2 secondary Ions
ion
peak
m/z
compd Na m/z
m/z intensity m/z intensity m/z intensity m/z intensity m/z intensity m/z intensity m/z intensity m/z intensity m/z intensity
13
14
15
10
11
12
6
6
6
3
3
3
543.1 349.0 bpb
543.0 367.1 349.1
543.0 367.0 349.1
515.1 353.0 335.1
515.0 353.1
100
35
17
367.0
bp
bp
17
100
100
299.0
2
255.1
5
193.0
10
193.0
179.1
5
15
173.1
173.1
10
18
16
299.0
299.0
3
255.1
255.1
3
8
203.1
203.9
203.1
10
10
18
191.8
20
515.1 353.0
317.5
7
15
179.1
5
173.1
8
MS3
base
peak
MS4
base
peak
MS3 secondary ions
MS4 secondary ions
m/z intensity m/z intensity m/z intensity m/z intensity
compd Na m/z
m/z intensity m/z intensity m/z intensity m/z intensity m/z intensity m/z
13
14
15
10
11
12
6
6
6
3
3
3
173.1 193.0
193.0 bp
173.1 193.0
173.5 179.5
191.5 179.5
173.5 179.4
27
100
70
91
53
bpb
173.1
bp
bp
173.1
bp
100
7
100
100
5
161.1
12
155.0
155.0
10
5
134.1
134.1
134.1
135.6
135.6
135.7
5
5
7
14
12
12
134.1 149.1
80
93.2 111.1
93.4 128.1
85.5 127.0
93.3 127.5
65
3
95
3
191.7
bp
191.6
53
100
27
172.9
172.9
172.9
2
90
15
111.1
110.7
111.4
50
60
38
bp
93.4
bp
100
80
100
80
100
a Number of LC-MS analyses used to collect MS data. b Occurs as base peak.
Figure 6. LC-MS2 spectra of the putative diferuloylquinic acids and caffeoyl-dimethoxycinnamoylquinic acids.
H+]- that by analogy with the fragmentation of 4-dimethoxy-
cinnamoylquinic acid would be expected to produce the
dehydrated m/z 173 at MS3. A distinctive feature of the MS2
spectrum is ions (∼15% of base peak) at m/z 299 and 255. This
combination of ions has previously been seen only in the MS2
spectra of dicaffeoylquinic acids and caffeoyl-feruloylquinic
acids with a C4 caffeoyl residue (Table 3) (5, 9). They have
been attributed respectively to full aromatization (- 3H2O) of
the quinic acid moiety in a caffeoylquinic acid fragment and
its subsequent decarboxylation. Such ions would be expected
in the MS2 spectrum of 3-dimethoxycinnamoyl-4-caffeoylquinic
acid (22) but not 3-caffeoyl-4-dimethoxycinnamoylquinic acid
(23). Accordingly, it is assigned as 3-dimethoxycinnamoyl-4-
caffeoylquinic acid (22).
peak, a fragmentation consistent with a 3,5-caffeoyl-dimethoxy-
cinnamoylquinic acid rather than a Vic-caffeoyl-dimethoxy-
cinnamoylquinic acid. It has been shown that a caffeoyl residue
at C5 is easily eliminated (5). Such an elimination would
produce [3-dimethoxycinnamoylquinic acid - H+]- as the MS2
base peak, and as argued above for 22, the MS3 base peak at
m/z 207 is consistent with such an assignment, suggesting that
the second eluting caffeoyl-dimethoxycinnamoylquinic acid is
3-dimethoxycinnamoyl-5-caffeoylquinic acid (24). Its fragmen-
tation pattern more closely resembles 3-feruloyl-5-caffeoylquinic
acid (18) than 3-caffeoyl-5-feruloylquinic acid (19) (Table 3)
(5).
By analogy with the sequence of elution and relative
hydrophobicity observed for dicaffeoylquinic acids and caffeoyl-
feruloylquinic acids (5, 9, 14), the most hydrophobic caffeoyl-
dimethoxycinnamoylquinic acid should be one of the 4,5-
isomers. Peak 26 (Table 3) has the characteristic fragmentation
The next eluting caffeoyl-dimethoxycinnamoylquinic acid
(24) also loses its caffeoyl residue before its dimethoxy-
cinnamoyl residue, but does not produce a dehydrated MS3 base

1966 J. Agric. Food Chem., Vol. 54, No. 6, 2006
Clifford et al.
Table 3. MS4 Fragmentation Data for the Putative Caffeoyl-dimethoxycinnamoylquinic Acids, Putative Feruloyl-dimethoxycinnamoylquinic Acids, and
Equivalent Caffeoyl-feruloylquinic Acids
MS2
parent
ion
m/z
base
peak
m/z
MS3 secondary ions
compd Na
m/z
intensity
m/z
intensity
92
m/z
intensity
32
m/z
intensity
8
m/z
intensity
12
m/z
intensity
16
m/z
intensity
22
22
23
24
25
26
27
28
29
30
31
32
33
16
17
18
19
20
21
6
6
6
6
6
6
543.0
543.0
543.0
557.0
557.0
557.0
381.1
381.1
381.0
349.1
381.2
381.0
335.1
363.1
349.1
298.7
254.9
203.1
335.0
335.1
2
2
363.1
75
bp
100
75
349.2
349.1
30
6
6
6
6
6
6
529.1
529.0
529.0
529.0
529.0
529.0
353.0 367.1
367.0
367.0
353.0 367.0
367.0 bp
353.0 367.0
90
100
100
55
100
25
335.1
335.1
335.1
335.1
335.1
50
10
5
7
2
bpb
353.0
100
5
349.1
349.3
45
5
298.9
299.0
6
8
255.1
255.1
6
5
203.1
203.1
8
8
bp
bp
bp
100
349.1
7
MS3
base
peak
MS4
base
peak
MS3 secondary ions
MS4 secondary ions
m/z intensity m/z intensity m/z intensity m/z intensity m/z intensity
compd Na m/z
m/z intensity m/z intensity m/z intensity m/z
22
23
24
25
26
27
28
29
30
31
32
33
16
17
18
19
20
21
6
6
6
6
6
6
207.1 173.3
207.1 173.0
173.1 207.0
175.0 193.1
207.1 173.1
173.1 207.3
10
25
55
80
30
47
149.1
149.0
93.2
bpb
100
100
bp
162.7
162.6
80
70
133.1
133.1
22
50
155.1
155.1
20
10
111.1
35
269.0
55
313.0
25
160.1
149.0
93.2
bp
100
32
111.1
111.3
40
73
6
6
6
6
6
6
173.1 179.1
173.0 193.0
193.3 173.4
191.5 179.5
173.0 193.0
173.1 179.1
85
30
45
56
70
84
191.5
191.1
15
28
135.8
135.2
11
17
134.1 148.9
93.5
134.0 149.0
127.2 172.4
93.4
48
68
111.2
111.2
57
40
93.1
a Number of LC-MS analyses. b Occurs as base peak.
of a Vic-diacyl-chlorogenic acid (MS3 and MS4 base peaks at
m/z 173 and 93, respectively), and it clearly loses its caffeoyl
residue before its dimethoxycinnamoyl residue [MS2 base peak
at m/z 381 and a strong MS3 secondary ion (55% of base peak)
at m/z 207]. The MS2 base peak must be either [4-dimethoxy-
cinnamoylquinic acid - H+]- or [5-dimethoxycinnamoylquinic
acid - H+]-, and the subsequent fragmentation to the dehy-
drated ion m/z 173 indicates [4-dimethoxycinnamoylquinic acid
- H+]- and thus favors assignment as 4-dimethoxycinnamoyl-
5-caffeoylquinic acid (26). The absence of the MS2 ions at m/z
299 and 255, suggesting that 26 does not have a caffeoyl residue
at C4, is consistent. Its fragmentation pattern more closely
resembles 4-feruloyl-5-caffeoylquinic acid (20) than 4-caffeoyl-
5-feruloylquinic acid (21) (Table 3) (5).
Characterization of Putative Feruloyl-dimethoxy-
cinnamoylquinic Acids (28-33). The LC-MS data for the
feruloyl-dimethoxycinnamoylquinic acids (28-33) are sum-
marized in Table 3 and Figure 7. The fragmentation pattern of
the slowest eluting isomer (32) is identical to that of caffeoyl-
dimethoxycinnamoylquinic acid (26) if allowance is made for
the weak caffeic acid-derived ion at m/z 335 being replaced by
the analogous, but more intense, m/z 349, suggesting that
logically it can be assigned as 4-dimethoxycinnamoyl-5-
feruloylquinic acid (32). An identical argument can be made
for assigning the preceding feruloyl-dimethoxycinnamoylquinic
acid as 3-dimethoxycinnamoyl-5-feruloylquinic acid (30).
The fragmentation of the most rapidly eluting feruloyl-
dimethoxycinnamoylquinic acid (28) resembles the analogous
caffeoyl-dimethoxycinnamoylquinic acid (22) in that both lose
their two cinnamoyl residues with almost equal facility. Whereas
3-dimethoxycinnamoyl-4-caffeoylquinic acid (22) slightly favors
elimination of the caffeoyl residue, 28 shows a slight preference
for eliminating the dimethoxycinnamoyl residue. Fragmentation
of the MS2 base peak (m/z 349) produces m/z 175, an ion
previously seen in the fragmentation of 3,4-diferuloylquinic acid
(13), 3-feruloyl-4-caffeoylquinic acid (16), and 3-dimethoxy-
cinnamoyl-4-caffeoylquinic acid (22). Because this ion is not
seen during the fragmentation of 3,4-dicaffeoylquinic acid (10)
or 3-caffeoyl-4-feruloylquinic acid (17), it is associated clearly
with a methylated cinnamoyl residue at C3, but for a feruloyl-
dimethoxycinnamoylquinic acid it is necessary to decide whether
the feruloyl or the dimethoxycinnamoyl residue is at C3.

LC-MSn of Diferuloylquinic Acids
J. Agric. Food Chem., Vol. 54, No. 6, 2006 1967
Figure 7. LC-MS2 spectra of the putative feruloyl-dimethoxycinnamoylquinic acids.
The MS2 secondary fragment ions at m/z 299 and 255
(associated with aromatization and decarboxylation of a 4-
caffeoylquinic acid moiety) are replaced at MS3 by the
analogous m/z 313 and 269, suggesting that 28 is 3-dimethoxy-
cinnamoyl-4-feruloylquinic acid. The strong MS2 secondary ion
(m/z 363) fragmenting to m/z 207, suggesting that it is derived
from [3-dimethoxycinnamoylquinic acid - H2O - H+]- (rather
than [4-dimethoxycinnamoylquinic acid - H2O - H+]-), is
consistent with this suggestion. Accordingly, 28 has been
assigned tentatively as 3-dimethoxycinnamoyl-4-feruloylquinic
acid rather than 3-feruloyl-4-dimethoxycinnamoylquinic acid
(29).
3-Dimethoxycinnamoyl-4-feruloylquinic acid (28) differs
from 3-dimethoxycinnamoyl-4-caffeoylquinic acid (22) in that
at MS2 the dehydrated ions are seen without the associated “non-
dehydrated” ions. In the caffeoyl residue-containing series 3,4-
dicaffeoylquinic acid (10), 3-feruloyl-4-caffeoylquinic acid (16),
3-dimethoxycinnamoyl-4-caffeoylquinic acid (22), the ratio of
dehydrated to non-dehydrated MS2 ions increases in the
sequence 0.16, 0.6, and 1.8 with progressively increasing
methylation of the parent molecule. The equivalent ratios for
3,4-diferuloylquinic acid (13) and 3-dimethoxycinnamoyl-4-
feruloylquinic acid (28) (increasing methylation without a
caffeoyl residue) are 6.3 and >100, respectively, reinforcing
the relationship between extent of cinnamoyl methylation and
tendency to produce dehydrated ions.
The observed preference of the methylated derivatives for
dehydration is surprising. We have postulated a 1,2-acyl
migration mechanism for the dehydration step. In this mecha-
nism an inverted chair conformation allows facile loss of H2O
by 1,2-acyl participation using the acyl substituent in the
4-position of the quinic acid, leading to an oxonium ion
intermediate (4). Further H-transfer, presumably via an exo-
elimination, will give the dehydrated molecular ion. Because
neither the basicity of the OH functionality nor the C-O bond
strength is expected to change if OH is replaced by OMe, the
only rationale available is that increased nucleophilicity of the
ester carbonyl increases the rate of dehydration or the stability
of the dehydrated ion. In solution chemistry Hammet parameters
are frequently used to describe the electronic effects of aromatic
substituents on a particular reaction mechanism. However, both
+
σp0 [σp0 (OH) ) -0.22 and σp0 (OMe) ) -0.12] and σp+ [σp
+
(OH) ) -0.92 and σp (OMe) ) -0.78] indicate that OH
should stabilize a positive charge in an oxonium ion intermediate
better than OMe (16), and therefore it should marginally increase
the nucleophilicity of the ester substituent in comparison with
OMe. In the gas-phase reaction observed, it appears that this
trend is reversed.
Relative Retention Time of Diacyl-CGA. It has previously
been noted (11, 12) that a feruloylquinic acid isomer elutes from
a C18 HPLC column packing some 32% later than the equivalent
caffeoylquinic acid isomer and that a caffeoyl-feruloylquinic
acid elutes some 16% later than the equivalent dicaffeoylquinic
acid. On the phenylhexyl packing used in this study, the
retention times of the newly characterized diacyl-chlorogenic
acids, relative to the dicaffeoylquinic acids, are 1.19-1.20
(caffeoyl-feruloylquinic acids), 1.38-1.41 (diferuloylquinic
acids), 1.38-1.43 (caffeoyl-dimethoxycinnamoylquinic acids),
and 1.60-1.66 (feruloyl-dimethoxycinnamoylquinic acids), an
increase of ∼20% for each additional methyl group.
Treatment with Tetramethylammonium Hydroxide. The
ability of TMAH to facilitate isomerization of chlorogenic acids,
along with the partial formation of the pertinent methyl
cinnamate(s), has been used to characterize various classes of
chlorogenic acids [including caffeoyl-feruloylquinic acids (11)]
and to synthesize 1,4-dicaffeoylquinic acid (9). This procedure
was applied to the crude methanolic extracts of Robusta coffee
beans for two reasons. One was to generate methyl dimethoxy-
cinnamate from dimethoxycinnamoylquinic acids, caffeoyl-
dimethoxycinnamoylquinic acids, and feruloyl-dimethoxy-
cinnamoylquinic acids, and another was to confirm by spiking
with an authentic sample of methyl 3,4-dimethoxycinnamate
(data not shown) that TMAH treatment did, indeed, produce
the 3,4-dimethoxy isomer.
Whereas TMAH facilitates the acyl migration of a cinnamoyl
residue from C5 to C4 to C3 (or vice versa), it would not be
expected to move a residue from C3 to C5 (or vice versa) if C4
were occupied. For example, 3-dimethoxycinnamoyl-4-caffeoyl-
quinic acid (22) could be converted stepwise to 3-dimethoxy-
cinnamoyl-5-caffeoylquinic acid (24) and 4-dimethoxy-

1968 J. Agric. Food Chem., Vol. 54, No. 6, 2006
Clifford et al.
Figure 8. Effect of interesterification with tetramethylammonium hydroxide on the relative proportions of caffeoyl-dimethoxycinnamoylquinic acids and
feruloyl-dimethoxycinnamoylquinic acids.
cinnamoyl-5-caffeoylquinic acid (26), but not to 4-caffeoyl-5-
dimethoxycinnamoylquinic acid (27). The structures assigned
tentatively in this study to these hetero-diacyl-chlorogenic acids
all have the dimethoxycinnamoyl residue at the numerically
“low” position and the caffeoyl or feruloyl residue at the “high”
position and should, therefore, be interconvertible. Treatment
with TMAH for up to 4 min demonstrated (Figure 8) that the
three caffeoyl-dimethoxycinnamoylquinic acids and the three
feruloyl-dimethoxycinnamoylquinic acids interconvert, provid-
ing further evidence that they do indeed belong to one
“numerical” group as assigned on the basis of their fragmenta-
tion.
3-dimethoxycinnanmoyl-5-caffeoylquinic acid (26). Accord-
ingly, fragment-targeted MS2 (m/z 543 + 381 and m/z 557 +
381) and MS3 (m/z 543 + 353 + 191, m/z 543 + 381 + 173,
m/z 543 + 381 + 207, m/z 557 + 349 + 193, and m/z 557 +
381 + 207) experiments were performed. A suggestion of
heterogeneity in 28 was observed as weak MS4 signals (not
exceeding 103) detected by variation in the ratios of m/z 155 to
111 and m/z 149 to 163, in the m/z 543 + 381 + 173 and m/z
543 + 381 + 207 spectra, respectively. It was not possible to
detect any heterogeneity in 24 and 26 or in 28, 30, and 32. One
must therefore conclude that if the missing isomers are present,
then they are present in only trace amounts and lack a distinctive
pattern of fragmentation
Missing Isomers. Because all six theoretical isomers of the
caffeoyl-feruloylquinic acids (16-21) are easily located in the
Robusta extract (5), it is a little surprising that only three
caffeoyl-dimethoxycinnamoylquinic acids (22, 24, and 26) and
three feruloyl-dimethoxycinnamoylquinic acids (28, 30, and 32)
have been observed. This could reflect the failure of the coffee
plant to synthesize all six isomers, but because all three
dimethoxycinnamoylquinic acids (7-9) are produced, there is
no obvious impediment to this. It is possible that these other
isomers are present, but below the limits of detection. If that is
the case, then the most concentrated of the “missing” caffeoyl-
dimethoxycinnamoylquinic acids and feruloyl-dimethoxy-
cinnamoylquinic acids must be present at a concentration
considerably below that of the weakest isomer found, because
analysis of deliberately (×10) concentrated extracts still did not
allow them to be located. A third possibility is that they are
present at low concentration and not resolved chromatographi-
cally. However, not only can the six caffeoyl-feruloylquinic
acids (16-20) be resolved chromatographically, but critical
pairs, for example, 3-feruloyl-4-caffeoylquinic acid (16) and
3-caffeoyl-4-feruloylquinic acid (17), can be discriminated also
by their fragmentation patterns at MS2, MS3, and MS4 (5). At
MS2 the feruloyl-dimethoxycinnamoylquinic acid peaks appear
to be homogeneous, as do the caffeoyl-dimethoxycinnamoyl-
quinic acid peaks once an allowance is made for the coeluting
diferuloylquinic acids.
Inspection of the fragmentation data in Table 3 suggested
that if any of the missing compounds with a dimethoxycin-
namoyl residue at C5 (25, 27, 31, or 33) form a [5-dimethoxy-
cinnamoylquinic acid - H+]- MS2 base peak and it fragments
to m/z 193 [as observed for 5-dimethoxycinnamoylquinic acid
(9) itself], then their MS3 spectra should be distinguishable from
those isomers with either a caffeoyl (24 or 26) or feruloyl (30
or 32) residue at C5. Similarly, if 3-caffeoyl-5-dimethoxy-
cinnamoylquinic acid (25) fragments in the same manner as
3-caffeoyl-5-feruloylquinic acid (19), then it should be locatable
by MS4 ions at m/z 127 and 172 that should be absent from
As for the other mono-acyl chlorogenic acids, 5-dimethoxy-
cinnamoylquinic acid (9) dominates. For the diferuloylquinic
acids (13-15), caffeoyl-dimethoxycinnamoylquinic acids (22,
24, and 26) and feruloyl-dimethoxycinnamoylquinic acids (28,
30, and 32) there is within each group a progressive increase in
concentration (ca. ×6) with increasing hydrophobicity, and this
pattern remained essentially constant over five Robusta samples
of different origin (Angola, Madagascar, Sierra Leone, Sri
Lanka, and Tanzania) previously observed to have different
chromatographic profiles (6). In contrast, the caffeoyl-feruloyl-
quinic acids (16-21) increase, by a factor of ca. ×50, in the
sequence 3-caffeoyl-5-feruloylquinic acid (19), 3-feruloyl-4-
caffeoylquinic acid (16), 4-feruloyl-5-caffeoylquinic acid (20),
3-feruloyl-5-caffeoylquinic acid (18), 4-caffeoyl-5-feruloylquinic
acid (21), and 3-caffeoyl-4-feruloylquinic acid (17).
Although the generation of quantitative data was not a primary
objective of this investigation, crude estimates of the content
of these newly reported chlorogenic acids were obtained from
the relative absorbance of individual peaks at 325 nm relative
to 5-caffeoylquinic acid (3), assigned unit absorbance. Dimethoxy-
cinnamoylquinic acids (7-9) ranged from 0.005 to 0.020 unit,
caffeoyl-dimethoxycinnamoylquinic acids (22, 24, and 26) from
0.008 to 0.028 unit, feruloyl-dimethoxycinnamoylquinic acids
(28, 30, and 32) from 0.004 to 0.020 unit, and diferuloylquinic
acids (13-15) from 0.004 to 0.010 unit, similar to or slightly
lower than the caffeoyl-feruloylquinic acids (16-21) (0.005-
0.030 unit). Assuming ∼5% dry mass basis (dmb) as a typical
5-caffeoylquinic acid (3) content (17-19), this suggests that
none of these new chlorogenic acids are likely individually to
exceed ∼0.15% dmb, or ∼0.7% dmb in total.
ACKNOWLEDGMENT
We acknowledge a grant to B.S. from the Tobb Economics and
Technology University, Turkey, and technical assistance from
H. Roozendaal.

LC-MSn of Diferuloylquinic Acids
J. Agric. Food Chem., Vol. 54, No. 6, 2006 1969
LITERATURE CITED
(12) Clifford, M. N.; Kellard, B.; Birch, G. G. Characterisation of
chlorogenic acids by simultaneous isomerisation and trans-
esterification with tetramethylammonium hydroxide. Food Chem.
1989, 33, 115-123.
(13) Andrade, P. B.; Leitao, R.; Seabra, R. M.; Oliveira, M. B.;
Ferreira, A. 3,4-Dimethoxycinnamic acid levels as a tool for
differentiation of Coffea canephora and Coffea arabica. Food
Chem. 1998, 61, 511-514.
(14) Clifford, M. N. The analysis and characterization of chlorogenic
acids and other cinnamates. In Methods in Polyphenol Analysis;
Santos-Buelga, C., Williamson, G., Eds.; Royal Society of
Chemistry: Cambridge, U.K., 2003; [14], pp 314-337.
(15) Haslam, E.; Haworth, R. D.; Makinson, G. K. Synthesis of 3-O-
p-coumaroylquinic acid. J. Chem. Soc. 1961, 5153-5156.
(16) Isaacs, N. I. Physical Organic Chemistry; Longman Scientific:
New York, 1987.
(17) Clifford, M. N. Chlorogenic acids. In Coffee 1. Chemistry; Clarke,
R. J., Macrae, R., Eds.; Elsevier Applied Science: London, U.K.,
1985; [5], pp 153-202.
(18) Clifford, M. N. Chemical and physical aspects of green coffee
and coffee products. In Coffee: Botany, Biochemistry and
Production of Beans and BeVerage; Clifford, M. N., Willson,
K. C., Eds.; Croom Helm: London, U.K., 1985; pp 305-
374.
(19) Balyaya, K. J.; Clifford, M. N. Individual chlorogenic acids and
caffeine contents in commercial grades of wet and dry processed
Indian green robusta coffee beans. J. Food Sci. Technol. Mysore
1995, 32, 104-108.
(1) IUPAC. Nomenclature of cyclitols. Biochem. J. 1976, 153, 23-
31.
(2) Clifford, M. N. Chlorogenic acids and other cinnamatessnature,
occurrence, dietary burden, absorption and metabolism. J. Sci.
Food Agric. 2000, 80, 1033-1042.
(3) Clifford, M. N. Chlorogenic acids and other cinnamatessnature,
occurrence and dietary burden. J. Sci. Food Agric. 1999, 79,
362-372.
(4) Eliel, E. L.; Ramirez, M. B. (-)-Quinic acid: configurational
(stereochemical) descriptors. Tetrahedron: Asymm. 1997, 8,
3551-3554.
(5) Clifford, M. N.; Johnston, K. L.; Knight, S.; Kuhnert, N. A
hierarchical scheme for LC-MSn identification of chlorogenic
acids. J. Agric. Food Chem. 2003, 51, 2900-2911.
(6) Clifford, M. N.; Jarvis, T. The chlorogenic acids content of green
robusta coffee beans as a possible index of geographic origin.
Food Chem. 1988, 29, 291-298.
(7) Clifford, M. N.; Williams, T.; Bridson, D. Chlorogenic acids
and caffeine as possible taxonomic criteria in Coffea and
Psilanthus. Phytochemistry 1989, 28, 829-838.
(8) Clifford, M. N.; Knight, S. The cinnamoyl-amino acid conjugates
of green robusta coffee beans. Food Chem. 2004, 87, 457-463.
(9) Clifford, M. N.; Knight, S.; Kuhnert, N. Discriminating between
the six isomers of dicaffeoylquinic acid by LC-MSn. J. Agric.
Food Chem 2005, 53, 3821-3832.
(10) Egan, H.; Kirk, R. S.; Sawyer, R. Pearson’s Chemical Analysis
of Foods; Churchill Livingstone: London, U.K., 1981.
(11) Clifford, M. N.; Kellard, B.; Birch, G. G. Characterisation of
caffeoylferuloylquinic acids by simultaneous isomerisation and
transesterification with tetramethylammonmium hydroxide. Food
Chem. 1989, 34, 81-88.
Received for review January 19, 2006. Accepted January 25, 2006.
JF0601665