Determination of 4-Hydroxy-2,5-dimethyl-3(2H)-furanone and 2(or 5)-Ethyl-4-hydroxy-5(or 2)-methyl-3(2H)-furanone in Pentose Sugar-Based Maillard Model Systems by Isotope Dilution Assays

Article abstract of DOI:10.1021/jf960997i

The formation of 4-hydroxy-2,5-dimethyl-3(2H)-furanone (HDMF) and 2(or 5)-ethyl-4-hydroxy-5(or 2)-methyl-3(2H)-furanone (EHMF) from pentose sugars was studied in Maillard model systems. The amounts generated at 90°C for 1 h were determined by isotope dilution assay (IDA). The internal standards used for IDA, i.e., [¹³C₂]HDMF and [²H₃]EHMF, were prepared in good overall yields in three steps: addition of labeled acetaldehyde or propionaldehyde to tert-butyloxycarbonyl (Boc)-protected and lithiated 3-butyn-2-ol; oxidation of the Boc-protected diol with permanganate to 1,2-dione; and finally cyclization to the target molecules after removal of the protective groups under acidic conditions. Quantitative data confirmed previous findings that HDMF and EHMF are preferentially formed in the presence of glycine and L-alanine, respectively. The yields obtained were 2.6-5.1 μg of HDMF and 6.8-10 μg of EHMF per mmol pentose. Formation of both furanones was favored in phosphate-buffered solutions at pH 7 compared to pH 5, particularly in the presence of an excess of amino acid. These data are well in agreement with the previously proposed formation mechanism of HDMF and EHMF via Strecker-assisted chain elongation of the pentose moiety. However, both furanones were also produced to a lesser extent by sugar fragmentation-condensation reactions.

Full text of DOI:10.1021/jf960997i

2642
J. Agric. Food Chem. 1997, 45, 2642−2648
Deter m in a tion of 4-Hyd r oxy-2,5-d im eth yl-3(2H)-fu r a n on e a n d 2(or
5)-Eth yl-4-h yd r oxy-5(or 2)-m eth yl-3(2H)-fu r a n on e in P en tose
Su ga r -Ba sed Ma illa r d Mod el System s by Isotop e Dilu tion Assa ys
Imre Blank,† Laurent B. Fay,† Frederick J . Lakner,‡,§ and Manfred Schlosser*,‡
Nestle´ Research Center, Nestec Ltd., Vers-chez-les-Blanc, P.O. Box 44, CH-1000 Lausanne 26, Switzerland,
and Institut de Chimie Organique, BCh, Universite´ de Lausanne, CH-1015 Lausanne, Switzerland
The formation of 4-hydroxy-2,5-dimethyl-3(2H)-furanone (HDMF) and 2(or 5)-ethyl-4-hydroxy-5(or
2)-methyl-3(2H)-furanone (EHMF) from pentose sugars was studied in Maillard model systems.
The amounts generated at 90 °C for 1 h were determined by isotope dilution assay (IDA). The
internal standards used for IDA, i.e., [13C2]HDMF and [2H3]EHMF, were prepared in good overall
yields in three steps: addition of labeled acetaldehyde or propionaldehyde to tert-butyloxycarbonyl
(Boc)-protected and lithiated 3-butyn-2-ol; oxidation of the Boc-protected diol with permanganate
to 1,2-dione; and finally cyclization to the target molecules after removal of the protective groups
under acidic conditions. Quantitative data confirmed previous findings that HDMF and EHMF
are preferentially formed in the presence of glycine and L-alanine, respectively. The yields obtained
were 2.6-5.1 µg of HDMF and 6.8-10 µg of EHMF per mmol pentose. Formation of both furanones
was favored in phosphate-buffered solutions at pH 7 compared to pH 5, particularly in the presence
of an excess of amino acid. These data are well in agreement with the previously proposed formation
mechanism of HDMF and EHMF via Strecker-assisted chain elongation of the pentose moiety.
However, both furanones were also produced to a lesser extent by sugar fragmentation-condensation
reactions.
Keyw or d s: Maillard reaction; pentose model system; 4-hydroxy-2,5-dimethyl-3(2H)-furanone;
Furaneol; 2(or 5)-ethyl-4-hydroxy-5(or 2)-methyl-3(2H)-furanone; Homofuraneol; synthesis; alcohol
protection; alkyne oxidation; 1,2-dione; stable-isotope label; isotope dilution assay; GC-MS
INTRODUCTION
water miscibility, and delicate gas chromatographic
behavior (Pickenhagen et al., 1981; Blank et al., 1992).
The isotope dilution assay (IDA) has been shown to be
a sensitive, accurate, and reliable quantification tech-
nique in flavor research (Schieberle and Grosch, 1987;
Guth and Grosch, 1990) and has also been applied to
HDMF and EHMF in strawberries (Sen et al., 1991),
cheese (Preininger and Grosch, 1994), and coffee (Sem-
melroch et al., 1995). This method involves spiking food
materials with known amounts of a labeled substance
prior to sample preparation and analysis by GC-MS.
In this way, losses can be accounted for because
whatever changes occur in the natural substance also
occur in the labeled version.
4-Hydroxy-2,5-dimethyl-3(2H)-furanone (HDMF, Fur-
aneol, a registered trademark of Firmenich S.A., Gene-
va, Switzerland) and 2(or 5)-ethyl-4-hydroxy-5(or 2)-
methyl-3(2H)-furanone (EHMF, Homofuraneol) are im-
portant naturally occurring food flavor compounds with
a caramel-like note. HDMF is especially prevalent, and
it has been found in fruits, fermented products, and
thermally processed foods (Blank and Fay, 1996, and
references cited therein).
The thermally induced formation of 3(2H)-furanones
from sugars is explained to occur via 2,3-enolization in
the Maillard reaction leading to 1-deoxyosones as
intermediates (Hodge et al., 1972; Ledl and Schleicher,
1990). In general, the carbon skeleton of the sugar is
determinant for the furanone formed, i.e., HDMF is
produced from hexoses. As recently shown in model
experiments using 13C-labeled precursors, HDMF and
EHMF can also be formed from pentose sugars in the
presence of amino acids such as glycine and L-alanine,
i.e., by Strecker-assisted chain elongation of the sugar
moiety (Blank and Fay, 1996) and by sugar fragmenta-
tion-condensation reactions (Blank et al., 1996a). How-
ever, no quantitative data have yet been published on
the amounts of HDMF and EHMF formed.
As multiply labeled standards simplify mass spectro-
metric analyses, labeled HDMF and EHMF having at
least two stable isotopes per molecule are required.
While a number of preparations exist for the unlabeled
furanones (Mazenod et al., 1992; Huber, 1992), few are
well suited for multiply labeled versions. HDMF in
which both methyl groups were 13C-labeled has previ-
ously been synthesized (Sen et al., 1991). Deuterium-
labeled EHMF was similarly prepared, but in consid-
erably reduced yield, and complicated purifications of
intermediates were required (Preininger and Grosch,
1994).
The quantitative analysis of dihydrofuranones is a
challenging task because of their rather fragile nature,
In the present study we report on quantification of
HDMF and EHMF generated from pentose sugars in
Maillard model reactions. In addition, we present an
improved preparation mode of the labeled internal
standards, [13C2]HDMF and [2H3]EHMF, required for
IDA.
* Author to whom correspondence should be ad-
dressed.
† Nestec Ltd.
‡ Universite´ de Lausanne.
§ Present address: Department of Biochemistry, Uni-
versity of Illinois, Urbana, IL 61801.
S0021-8561(96)00997-1 CCC: $14.00
© 1997 American Chemical Society

3(2H)-Furanones Formed from Pentoses
J. Agric. Food Chem., Vol. 45, No. 7, 1997 2643
ganate to give 1,2-diones. The protective groups were removed
under acidic conditions while cyclization occurred as the final
step.
3-Butyn-2-yl tert-Butyl Carbonate (1). This was prepared
from 3-butyn-2-ol and di-tert-butyl dicarbonate according to
the method of Eren and Keinan (1988). 1 was distilled (36-
20
38 °C/0.3 mmHg) in 92% yield. nD ) 1.4195. 1H-NMR δ
(CDCl3): 1.49 (s, 9H), 1.52 (d, J ) 6.9 Hz, 3H), 2.47 (d, J )
2.0 Hz, 1H), 5.23 (dq, J ) 6.9, 2.0 Hz, 1H). IR (cm-1): 3260,
2960, 1730, 1260, 1085. MS (EI) m/z (% relative abundance):
155 (8), 115 (25), 114 (100), 93 (15), 57 (8); the molecular ion
was confirmed by CI, ammonia m/z (% relative abundance):
188 (43, [M + NH4]+).
[1,2-13C2]-3-Hexyn-2,5-diyl Bis(tert-butyl carbonate) (2). A
lithium diisopropyl amide (LIDA) solution was prepared by
adding butyllithium (15.2 mL, 1.43 M in hexane, 21.7 mmol)
to 25 mL of THF containing diisopropylamine (3.08 mL, 21.7
mmol) at -30 °C. After 30 min, it was added over 20 min by
cannula to the Boc-protected butynol 1 (3.90 mL, 3.70 g, 21.7
mmol) in 30 mL of THF and kept at -78 °C for 30 min. [1,2-
13C2]Ethanal (1.00 g, 21.7 mmol) was then added, stirring
continued for 1 h, and di-tert-butyl dicarbonate (5.68 g, 26.0
mmol) in 20 mL of THF was added by an addition funnel over
20 min. The reaction was kept at -78 °C for 1 h before it was
warmed to 25 °C, diluted with 100 mL of Et2O, and worked
up sequentially with 2 N NaOH, H2O, and brine. Drying (Na2-
SO4) and evaporation of the solvent provided a product in high
purity and quantitative yield (7.18 g). An equimolar mixture
of diastereomers was obtained as an oil. The resonances of
the corresponding nuclei of both diastereomers were isochro-
nous, that is, had coincidentally the same chemical shifts. 1H-
NMR δ (CDCl3): 1.48 (s, 18H), 1.52 (ddd, 1J CH ) 129 Hz, 2J CH
) 4.5 Hz, 3J HH ) 6.8 Hz, 3H), 1.52 (d, ∼6.8 Hz, 3H), 5.28 (dm,
1J CH ) 153 Hz, 1H), 5.3 (m, 1H). 13C-NMR δ (CDCl3): 21.2
(d, J ) 39 Hz), 63.0 (d, J ) 39 Hz). IR (cm-1): 2960, 1730,
1260, 1240, 1150. MS (EI) m/z (% relative abundance): 261
(1), 205 (3), 143 (100), 116 (32), 98 (58), 81 (93), 57 (83); the
molecular ion was confirmed by CI, ammonia m/z (% relative
abundance): 334 (100, [M + NH4]+).
F igu r e 1. Schematic procedure to synthesize 4-hydroxy-2(or
5)-[13C]methyl-5(or 2)-methyl-3(2H)-[2(or 5)-13C]furanone([13C2]-
HDMF, 4a b) and 2(or 5)-([2,2,2-2H3]ethyl)-4-hydroxy-5(or 2)-
methyl-3(2H)-furanone ([2H3]EHMF, 7a and 7b). Conditions:
a, lithium diisopropyl amide (LIDA), 13CH313CHO, (t-BuOCO)2O,
THF; b, LIDA, CD3CH2CHO, (t-BuOCO)2O, THF; c, KMnO4,
acetone/H2O/CH3COOH; d , oxalic acid, H2O.
EXPERIMENTAL PROCEDURES
The corresponding unlabeled compound, 3-hexyn-2,5-diyl
bis(tert-butyl carbonate), was prepared in the same manner
as the labeled material, but by substituting unlabeled acet-
aldehyde. 1H-NMR δ (CDCl3): 1.48 (s, 18H), 1.50 (d, ∼6.5 Hz,
6H), 5.28 (q, J ) 6.5 Hz, 2H). 13C-NMR δ (CDCl3): 21.2, 27.9,
63.0, 82.5, 83.4, 152.4. IR (cm-1): 2980, 1745, 1265, 1165. MS
(EI) m/z (% relative abundance): 259 (2), 203 (3), 141 (97),
114 (30), 96 (77), 79 (78), 57 (100); the molecular ion was
confirmed by CI, ammonia m/z (% relative abundance): 332
(100, [M + NH4]+).
Gen er a l. D-Xylose, D-ribose, D-arabinose, glycine, L-alanine,
and 4-aminobutyric acid of highest purity (g99%) were from
Fluka (Buchs, Switzerland) as well as reagents and solvents
for synthesis which were used as received. Butyllithium was
stored in a Schlenk tube. [1,2-13C2]Ethanal and [2,2,2-2H3]-
ethyl bromide were from Dr. Glaser AG (Basel, Switzerland).
Tetrahydrofuran (THF) was distilled from benzophenone/
sodium prior to use.
1H-NMR spectra were recorded with a Bruker AC-250
spectrometer at 250 MHz with CDCl3 as solvent and the CHCl3
resonance set at 7.27 ppm. Proton-decoupled 13C-NMR spectra
were recorded with a Bruker AMX-400 spectrometer at 100
MHz with CDCl3 solvent set at 77.1 ppm. Infrared spectros-
copy was performed with a Perkin-Elmer 1420. The sample
was applied to NaCl disks as a thin film.
Qualitative mass spectrometry (MS) was performed on a
Finnigan MAT 8430 mass spectrometer (Bremen, Germany).
Electron impact (EI) mass spectra were generated at 70 eV,
and positive chemical ionization (CI) was at 150 eV with
ammonia as the reagent gas. Nonvolatile samples were
introduced directly into the ion source held at 200 °C. Volatile
components were sampled via a Hewlett-Packard HP-5890 gas
chromatograph (Geneva, Switzerland) using the following
conditions: cold on-column injector; fused silica capillary
column [DB-FFAP (J &W Scientific), 30 m × 0.25 mm; film
thickness, 0.25 µm; carrier gas, helium (90 kPa)]; temperature
program, 50 °C (2 min), 4 °C/min to 180 °C, 10 °C/min to 240
°C (10 min).
Syn th eses. Labeled Furaneol ([13C2]HDMF, 4a b) and Ho-
mofuraneol ([2H3]EHMF, 7a b) were prepared in three steps
from labeled acetaldehyde or propionaldehyde (Figure 1).
Aldehyde addition to tert-butyloxycarbonyl (Boc) protected and
lithiated 3-butyn-2-ol followed by trapping with di-tert-butyl
dicarbonate led to di-tert-butyloxycarbonyl-protected diol in-
termediates. These were oxidized by acidic aqueous perman-
[1,2-13C2]-3,4-Hexanedione-2,5-diyl Bis(tert-butyl carbonate)
(3). Potassium permanganate (4.87 g, 30.8 mmol) was dis-
solved in a solvent mixture consisting of 140 mL of acetone,
28 mL of H2O, and 4.2 mL of acetic acid chilled in an ice bath.
Alkyne 2 (7.15 g) was dissolved in 20 mL of acetone and poured
into the purple solution, resulting in an exothermic reaction
and precipitation of brown MnO4. After 3.5 h at ice temper-
ature, the mixture was filtered through Celite, and excess
permanganate was destroyed with a dilute solution of NaH-
SO3. After a second filtration and evaporation of acetone, the
aqueous remainder was extracted with CH2Cl2 (3 × 70 mL).
The combined organic layers were extracted with saturated
NaHCO3 and dried (Na2SO4), and the solvent was stripped to
give 6.38 g (84% from 13CH313CHO) of a yellow solid. 1H-NMR
δ (CDCl3): (both diastereomers) 1.48 (s, 18H), 1.49 (s, 18H),
1
2
1.50 (d, ∼7.0 Hz, 6H), 1.50 (ddd, J CH ) 130 Hz, J CH ) 4.0
3
Hz, J HH ) 7.0 Hz, 6H), 5.50 (q, J ) 7.0 Hz, 1H), 5.57 (q, J )
7.0 Hz, 1H), 5.53 (dm, J CH ) 154 Hz, 2H). 13C-NMR δ
1
(CDCl3): (both diastereomers) 15.5 (d, J ) 36 Hz), 15.6 (d, J
) 36 Hz), 72.5 (d, J ) 36 Hz), 73.6 (d, J ) 36 Hz). IR (cm-1):
2960, 1730, 1720, 1265, 1242. MS (EI) m/z (% relative
abundance): 277 (8), 175 (16), 141 (13), 130 (3), 113 (4), 57
(100); the molecular ion was confirmed by CI, ammonia m/z
(% relative abundance): 366 (35, [M + NH4]+).
The corresponding unlabeled compound, 3,4-hexanedione-
2,5-diyl bis(tert-butyl carbonate), was prepared by substituting

2644 J. Agric. Food Chem., Vol. 45, No. 7, 1997
Blank et al.
unlabeled alkyne. The yield of the yellow solid formed was
88% from acetaldehyde. 1H-NMR δ (CDCl3): (both diastereo-
mers) 1.46 (s, 18H), 1.48 (d, ∼7.0 Hz, 12H), 5.48 (q, J ) 7.0
Hz, 2H), 5.54 (q, J ) 7.0 Hz, 2H); 13C-NMR δ (CDCl3): (both
diastereomers) 15.8, 15.9, 27.8, 72.5, 73.5, 83.3, 152.6, 152.8,
193.9, 194.2. IR (cm-1): 2960, 1733, 1722, 1265, 1245. MS
(EI) m/z (% relative abundance): 275 (48), 173 (94), 141 (60),
128 (19), 111 (17), 57 (100); the molecular ion was confirmed
by CI, ammonia m/z (% relative abundance): 364 (65, [M +
NH4]+).
oxidation took place for 1.5 h at ice temperature and an
additional 1.0 h at 25 °C. The product was a bright yellow oil
(7.43 g, 87% from propanal-d3). 1H-NMR δ (CDCl3): (both
diastereomers) 1.47 (s, 18H), 1.48 (s, 18H), 1.5 (m, 6H), 1.8
(m, 4H), 5.36 (dd, J ) 8.0, 4.5 Hz, 1H), 5.44 (dd, J ) 8.0, 4.5
Hz, 1H), 5.54 (q, J ) 7.0, 1.0 Hz, 1H). 13C-NMR δ (CDCl3):
(both diastereomers) 15.8, 23.2, 27.6, 72.4, 73.5, 77.1, 78.4,
83.3, 152.4, 152.7, 152.9, 153.1, 193.7, 193.8, 194.0, 194.1. IR
(cm-1): 2960, 1730, 1720, 1270, 1240, 1150. MS (EI) m/z (%
relative abundance): 292 (43), 252 (20), 234 (15), 190 (95), 158
(64), 146 (45), 128 (30), 113 (30), 108 (35), 91 (50), 57 (100);
the molecular ion was confirmed by CI, ammonia m/z (%
relative abundance): 381 (25, [M + NH4]+).
The unlabeled compound, 3,4-heptanedione-2,5-diyl bis(tert-
butyl carbonate), was prepared by substituting unlabeled
alkyne. Yield of bright yellow oil was 85%. 1H-NMR δ
(CDCl3): (both diastereomers) 0.99 (t, J ) 7.0 Hz, 6H), 1.43
(s, 18H), 1.45 (s, 18H), 1.5 (m, 6H), 1.8 (m, 4H), 5.32 (dd, J )
7.0, 4.2 Hz, 1H), 5.40 (dd, J ) 7.0, 4.5 Hz, 1H), 5.50 (q, J )
7.0 Hz, 1H), 5.54 (q, J ) 7.0 Hz, 1H). 13C-NMR δ (CDCl3):
(both diastereomers) 9.4, 9.7, 15.6, 23.4, 23.6, 27.5, 31.4, 72.5,
73.5, 77.0, 78.4, 83.1, 152.4, 152.7, 152.9, 153.1, 193.7, 193.8,
194.0. IR (cm-1): 2960, 1730, 1720, 1270, 1240, 1150. MS
(EI) m/z (% relative abundance): 289 (11), 249 (5), 231 (5),
187 (33), 155 (10), 143 (21), 125 (15), 110 (5), 105 (10), 91 (10),
86 (35), 84 (50), 57 (100); the molecular ion was confirmed by
CI, ammonia m/z (% relative abundance): 378 (60, [M +
NH4]+).
4-Hydroxy-2(or 5)-[13C]methyl-5(or 2)-methyl-3(2H)-[2(or 5)-
13C]furanone ([13C2]HDMF) (4a b). Labeled dione 3 (6.35 g,
18.2 mmol) was refluxed in 38 mL of H2O with 3.5 g anhydrous
oxalic acid for 5 h under nitrogen. After cooling, the reaction
mixture was diluted with 60 mL of saturated NaCl and
extracted with CH2Cl2 (5 × 50 mL). The combined organic
layers were extracted with saturated NaHCO3, dried (Na2SO4),
and the solvent was stripped to give 1.30 g of pale yellow
crystals. Further purification was performed by sublimation
(50 °C/0.3 mmHg) providing 1.02 g (43%) of colorless crystals,
mp 77-79.5 °C (Re et al., 1973: 77-79 °C). 1H-NMR δ
(CDCl3): (both tautomers) 1.45 (d, J ) 7.0 Hz, 3H), 1.45 (ddd,
2
3
1J CH ) 130 Hz, J CH ) 4.5 Hz, J HH ) 7.0 Hz, 3H), 2.26 (d, J
1
2
6
) 1.0 Hz, 3H), 2.26 (ddd, J CH ) 130 Hz, J CH ) 7.0 Hz, J HH
) 1.0 Hz, 3H), 4.5 (m, 1H), 4.48 (dm, 1J CH ) 150 Hz, 1H). 13C-
NMR δ (CDCl3): (both tautomers) 13.6 (d, J ) 48 Hz), 16.5
(d, J ) 39 Hz), 80.5 (d, J ) 39 Hz), 174.5 (d, J ) 48 Hz). MS
(EI) m/z (% relative abundance): 130 (100), 87 (12), 85 (13),
59 (32), 57 (41), 45 (47), 43 (45), 29 (18); the molecular ion
was confirmed by MS (CI), ammonia m/z (% relative abun-
dance): 148 (100, [M + NH4]+).
2(or 5)-([2,2,2-2H3]Ethyl)-4-hydroxy-5(or 2)-methyl-3(2H)-
furanone ([2H3]EHMF) (7a b). [2H3]Dione (6) (7.40 g) was
treated similarly to 13C-labeled dione (3). Reflux was main-
tained for 6 h. The crude product was purified by flash
chromatography (silica gel; 1:1 hexane/ethyl acetate) providing
2.02 g (74%) of a pale yellow oil. 1H-NMR δ (CDCl3): (major
tautomer) 1.70 (dd, J ) 15.0, 7.0 Hz, 1H), 1.95 (dd, J ) 15.0,
4.5 Hz, 1H), 2.26 (d, J ) 1.0 H, 3H), 4.38 (ddd, J ) 7.0, 4.5,
1.0, Hz, 1H); (minor tautomer) 1.44 (d, J ) 7.0 Hz, 3H), 2.62
(s, 2H), 4.48 (q, J ) 7.0 Hz, 1H). 13C-NMR δ (CDCl3): (both
tautomers) 13.5, 16.5, 21.0, 24.4, 80.0, 84.9, 133.1, 135.0, 176.0,
179.1, 198.1, 199.2. IR (cm-1): 3250 (br), 2205, 1680, 1605,
1185. MS (EI) m/z (% relative abundance): (minor tautomer)
145 (71), 89 (10), 85 (26), 60 (100), 57 (48), 29 (15); (major
tautomer) 145 (96), 127 (41), 102 (37), 73 (77), 55 (28), 43 (100);
both molecular ions were confirmed by MS (CI), ammonia m/z
(% relative abundance): 163 (100, [M + NH4]+).
The unlabeled compound, 2(or 5)-ethyl-4-hydroxy-5(or 2)-
methyl-3(2H)-furanone (EHMF), was prepared by using un-
labeled dione. Yield of crude isolated product was 88%. 1H-
NMR δ (CDCl3): (major tautomer) 0.98 (t, J ) 7.2 Hz, 3H),
1.73 (ddq, J ) 14.3, 7.2, 7.1 Hz, 1H), 1.98 (ddq, J ) 14.3, 7.2,
4.5 Hz, 1H), 2.27 (d, J ) 1.0 Hz, 3H), 4.38 (ddd, J ) 7.1, 4.5,
1.0 Hz, 1H); (minor tautomer) 1.25 (t, J ) 7.5 Hz, 3H), 1.45
(d, J ) 7.2 Hz, 3H), 2.65 (q, J ) 7.5 Hz, 2H), 4.49 (q, J ) 7.2
Hz, 1H). 13C-NMR δ (CDCl3): (both tautomers) 8.5, 10.0, 13.4,
16.5, 21.0, 24.5, 80.0, 84.9, 133.1, 134.9, 175.5, 179.0, 198.1,
199.4. IR (cm-1): 3250 (br), 1680, 1605, 1185. MS (EI) m/z
(% relative abundance): (minor tautomer) 142 (54), 85 (12),
57 (100); (major tautomer) 142 (87), 127 (34), 102 (32), 71 (64),
55 (25), 43 (100); both molecular ions were confirmed by MS
(CI), ammonia m/z (% relative abundance): 160 (100, [M +
NH4]+).
The corresponding unlabeled compound, 4-hydroxy-2,5-
dimethyl-3(2H)-furanone (HDMF), was prepared by substitut-
ing unlabeled dione. Yield of crude isolated product, 85%. 1H-
NMR δ (CDCl3): 1.45 (d, J ) 7.0 Hz, 3H), 2.27 (d, J ) 1Hz,
3H), 4.50 (dq, J ) 7.0, 1.0 Hz, 1H). 13C-NMR δ (CDCl3): 13.6,
16.5, 80.5, 134.0, 175.2, 199.2. IR (cm-1): 3300 (br), 1680,
1600, 1295, 1185. MS (EI) m/z (% relative abundance): 128
(100), 85 (23), 57 (67), 43 (82), 29 (22); the molecular ion was
confirmed by MS (CI), ammonia m/z (% relative abundance):
146 (100, [M + NH4]+).
[7,7,7-2H3]-3-Heptyn-2,5-diyl Bis(tert-butyl carbonate) (5).
Proceeded as for 13C-labeled alkyne 2, but using CD3CH2CHO
(1.43 g) prepared from CD3CH2Br, magnesium, and N-formyl-
piperidine in THF according to the method of Olah and
Arvanaghi (1981). The aldehyde was codistilled with solvent,
dried with molecular sieves (3 Å), and used as such [1H-NMR
δ (CDCl3): 2.41 (2H), 9.78 (1H)]. Concentration was deter-
mined by 1H-NMR using benzene as internal standard and
integrating against the aldehydic hydrogen (23.4 mmol in 50
mL THF). 5 was obtained as a pale yellow oil (8.26 g)
consisting of a mixture of two diastereomers which was taken
directly to the next step. The diastereomers showed identical
NMR spectra. 1H-NMR δ (CDCl3): 1.50 (s, 18H), 1.5 (m, 3H),
1.80 (br d, J ) 6.5 Hz, 2H), 5.16 (t, J ) 6.5, 1.0 Hz, 1H), 5.3
(m, 1H). 13C-NMR δ (CDCl3): 21.2, 27.6, 63.0, 68.0, 82.1, 82.6,
84.0, 152.3, 152.5. IR (cm-1): 2960, 1730, 1260, 1240, 1150.
MS (EI) m/z (% relative abundance): 175 (10), 158 (100), 131
(30), 127 (20), 116 (20), 113 (85), 96 (80), 70 (15), 57 (75); the
molecular ion was confirmed by CI, ammonia m/z (% relative
abundance): 349 (100, [M + NH4]+).
The corresponding unlabeled compound, 3-heptyn-2,5-diyl
bis(tert-butyl carbonate), was obtained by substituting un-
labeled propanal. It quantitatively yielded a pale yellow oil
consisting of a mixture of diastereomers showing identical
NMR spectra. 1H-NMR δ (CDCl3): 0.96 (t, J ) 7.2 Hz, 3H),
1.45 (s, 18H), 1.5 (m, 3H), 1.78 (pentet, J ) 7.0 Hz, 2H), 5.10
(t, J ) 7.0 Hz, 1H), 5.2 (m, 1H). 13C-NMR δ (CDCl3): 9.2, 21.2,
27.7, 28.0, 63.0, 68.0, 82.0, 82.6, 84.0, 152.3, 152.5. IR (cm-1):
2960, 1730, 1260, 1240, 1150. MS (EI) m/z (% relative
abundance): 172 (5), 155 (75), 128 (15), 126 (10), 113 (10), 110
(40), 93 (25), 67 (10), 57 (100); the molecular ion was confirmed
by CI, ammonia m/z (% relative abundance): 346 (100, [M +
NH4]+).
Qu a n tifica tion Exp er im en ts. Sample preparation was
performed as recently described (Blank and Fay, 1996) with
some modifications. The precursors were allowed to react in
water and phosphate- and malonate-buffered solutions (0.2 M,
Na2HPO4, Na2H2C3O4) at different pHs (5, 6, 7) at 90 °C for 1
h. After the reaction mixture was cooled rapidly, water (100
mL) and the labeled internal standards (9.6-48.2 µg of [13C2]-
HDMF and 10.0-50.2 µg of [2H3]EHMF) were added. The
solution was saturated with NaCl, and the pH was adjusted
to 4.0 (aqueous HCl, 2 M) which is the pH optimum for the
stability of HDMF in aqueous solutions (Hirvi et al., 1980).
Neutral compounds were continuously extracted with Et2O
overnight using a rotation perforator. The organic phase was
dried over Na2SO4 at +4 °C and concentrated to 0.5-1 mL.
All experiments were performed in duplicate.
[7,7,7-2H3]-3,4-Heptanedione-2,5-diyl Bis(tert-butyl carbon-
ate) (6). Proceeded as for 13C-labeled dione except that

3(2H)-Furanones Formed from Pentoses
J. Agric. Food Chem., Vol. 45, No. 7, 1997 2645
F igu r e 2. Quantification of 4-hydroxy-2,5-dimethyl-3(2H)-furanone (HDMF) and 2(or 5)-ethyl-4-hydroxy-5(or 2)-methyl-3(2H)-
furanone (EHMF) by isotope dilution assay using [13C2]HDMF (4a b) and [2H3]EHMF (7a and 7b) as internal standards. The
mass chromatograms were recorded by GC-MS operating in the selected ion monitoring mode to measure the molecular ions m/z
128 of HDMF, m/z 142 of EHMF (two tautomers), m/z 130 of [13C2]HDMF, and m/z 145 of [2H3]EHMF (two tautomers).
GC-MS analyses were performed on an HP-5971 mass
spectrometer connected to an HP-5890 gas chromatograph
equipped with an HP-7673 autosampler. The interface was
kept at 220 °C, and the ion source working in EI mode at 70
eV was held at about 180 °C. The samples were injected via
a splitless injector onto a Carbowax capillary column using
the chromatographic conditions as previously described (Blank
et al., 1996b). As shown in Figure 2, unlabeled HDMF and
EHMF and the corresponding labeled internal standards were
detected by selected ion monitoring (SIM) of their molecular
ions, i.e., m/z 128 (HDMF), m/z 142 (EHMF, two tautomers),
m/z 130 ([13C2]HDMF), and m/z 145 ([2H3]EHMF, two tau-
tomers).
The calibration curves were established with standard
mixtures containing defined amounts of labeled and unlabeled
compound in different ratios following the procedure described
by Guth and Grosch (1990). A good linearity was found in
the concentration range of 3-50 µg/mL (r2 ) 0.999). Samples
for establishing the calibration curves and for quantifying
HDMF and EHMF in the Maillard model reactions were
injected twice.
F igu r e 3. Electron impact mass spectrum of a tautomeric
mixture (1:1) of 4-hydroxy-2-[13C]methyl-5-methyl-3(2H)-[2-
13C]furanone and 4-hydroxy-5-[13C]methyl-2-methyl-3(2H)-[5-
13C]furanone ([13C2]HDMF) obtained by synthesis. Symbol 9
indicates the labeling position.
dicarbonate to give doubly-protected alkyne 2 as a 1:1
mixture of diastereomers. The crude product was
sufficiently pure to be directly oxidized with potassium
permanganate in aqueous acetone. The success of this
reaction depended upon low temperature and the pres-
ence of acetic acid. Dione 3 was then cyclized according
to the procedure of Re at al. (1973) resulting in tau-
tomers 4a and 4b (1:1 ratio). These conditions proved
efficacious for the concomitant removal of both Boc
protective groups, the prerequisite for cyclization. The
MS (EI) of 4a b is shown in Figure 3.
Similarly, the lithium acetylide derived from alkyne
1 was treated with [3,3,3-2H3]propanal followed by di-
tert-butyl dicarbonate to give doubly-protected alkyne
5. Oxidation to dione 6 and cyclization afforded EHMF
(7a /7b ) 1:3). The MS (EI) of 7a and 7b reported in
Figure 4 are in good agreement with those reported by
Preininger and Grosch (1994).
RESULTS AND DISCUSSION
Syn th esis of th e In ter n a l Sta n d a r d s [13C2]HDMF
a n d [2H3]EHMF . A simple three-step preparation has
been devised allowing labeled HDMF and EHMF to be
produced from a common unlabeled starting material
merely by substitution of one aldehyde reactant for
another (Figure 1). Yields were generally excellent, and
purification steps were not required until the final
products. In the case of HDMF, 13C-labeling was
indispensable to avoid exchange of hydrogen and oxygen
atoms during sample preparation (Sen et al., 1991). On
the contrary, EHMF could be deuterated in the terminal
position of the ethyl group.
The tert-butyloxycarbonyl (Boc) moiety is a frequent
choice for amine protection (Greene and Wuts, 1991) but
served perfectly well for our alcohol protection. It was
easily administered, withstood lithiation/addition and
oxidation steps, and was smoothly removed during the
final cyclization step. Thus, beginning with Boc-
protected alkynol 1, deprotonation with lithium diiso-
propyl amide and addition of [1,2-13C2]ethanal produced
an alkoxide which was trapped in situ with di-tert-butyl
Qu a n tifica tion of HDMF a n d EHMF in Ma illa r d
Mod el Rea ction s. As recently shown in model experi-
ments, HDMF and EHMF can be generated from
pentose sugars in the presence of glycine and L-alanine
(Blank and Fay, 1996; Blank et al., 1996a). This,
however, is a side reaction leading to relatively low
amounts of HDMF and EHMF. Therefore, isotope

2646 J. Agric. Food Chem., Vol. 45, No. 7, 1997
Blank et al.
F igu r e 4. Electron impact mass spectra of (A) 5-([2,2,2-2H3]ethyl)-4-hydroxy-2-methyl-3(2H)-furanone ([2H3]EHMF, 7a ) (B) and
2-([2,2,2-2H3]eth-1-yl)-4-hydroxy-5-methyl-3(2H)-furanone ([2H3]EHMF, 7b) obtained by synthesis. Symbol b indicates the labeling
position.
Ta ble 1. F or m a tion of HDMF a n d EHMF fr om P en tose
Su ga r s in Ma illa r d Mod el System s Con ta in in g Glycin e or
L-Ala n in ea
Ta ble 2. F or m a tion of HDMF a n d EHMF in Ma illa r d
Mod el System s Con ta in in g Xylose a n d Differ en t Am in o
Acid sa
model system
HDMFb
EHMFb
model system
HDMFb
EHMFb
arabinose/glycine
xylose/glycinec
ribose/glycine
5.1
2.6
3.6
1.3
0.3
0.7
xylose/4-aminobutyric acid
xylose/4-aminobutyric acid/glycine
xylose/4-aminobutyric acid/L-alanine
0.4
1.5
0.7
0.1
0.2
3.2
a
arabinose/L-alanine
xylose/L-alaninec
ribose/L-alanine
1.2
0.9
1.6
6.8
7.5
10.0
Equimolar amounts of precursors were used. Reaction condi-
b
tions: phosphate buffer (0.2 M, pH 6.0), 90 °C, 1 h. Data are
means of at least two assays, each of them injected twice;
maximum SD e 10% (in µg/mmol of sugar).
a
Equimolar amounts of precursors were used. Reaction condi-
tions: phosphate buffer (0.2 M, pH 6), 90 °C, 1 h. b Data are means
of at least two assays, each of them injected twice; maximum SD
e 10% (in µg/mmol of sugar). c The control sample (without amino
acid) resulted in less than 0.01 µg of HDMF and EHMF per mmol
of sugar.
HDMF and 0.3-1.3 µg of EHMF detected in the model
reactions pentose/L-alanine and pentose/glycine, respec-
tively. This, however, is a minor reaction pathway
where the amino acid indirectly contributes to the
formation of these furanones by accelerating the Mail-
lard reaction, i.e., by favoring sugar fragmentation and
condensation of the corresponding sugar fragments.
Several reactive sugar fragments have been reported
in the literature (review by Ledl and Schleicher, 1990).
First trials have shown that some of the well-known C3
fragments do generate HDMF, e.g., methylglyoxal,
acetol, and dihydroxyacetone (data not shown). These
results will be reported elsewhere.
4-Aminobutyric acid was employed to study the
formation of 3(2H)-furanones by sugar fragmentation.
As a γ-amino acid, it does not decompose by Strecker
deamination. The amounts of HDMF and EHMF
formed were relatively low, i.e., 0.4 and 0.1 µg, respec-
tively (Table 2). The data indicate that the generation
of HDMF by sugar fragmentation is favored compared
to that of EHMF, most likely due to higher amounts of
reactive C3 fragments formed, whereas EHMF requires
both reactive C3 and C4 fragments. As shown in Table
2, the amount of HDMF was significantly increased in
the presence glycine (1.5 µg). Similarly, addition of
L-alanine resulted in higher yields of EHMF (3.2 µg).
dilution assays were applied to obtain reliable data
about the absolute concentrations of HDMF and EHMF
formed and the parameters influencing the reaction.
F or m a tion fr om P en tose Su ga r s. The formation
of HDMF from pentose sugars was favored in the
presence of glycine (Table 1) resulting in about 2-5 µg
of HDMF. On the contrary, EHMF was preferentially
generated in the presence of L-alanine (6.8-10 µg). The
quantitative results support the hypothesis (Blank and
Fay, 1996) that HDMF and EHMF are mainly formed
by Strecker-assisted chain elongation of pentoses, i.e.,
through combination of the latter with aldehydes pro-
duced by Strecker deamination of glycine and alanine.
In other terms, formaldehyde (active C1) liberated from
glycine is required for the formation of HDMF as is
acetaldehyde (active C2 from L-alanine) for EHMF.
However, in all samples analyzed, 4-hydroxy-5-methyl-
3(2H)-furanone was the main reaction product (data not
shown) directly formed from pentose sugars (Feather,
1981).
As indicated in Table 1, both furanones were also
generated by sugar fragmentation, i.e., 0.9-1.6 µg of

3(2H)-Furanones Formed from Pentoses
J. Agric. Food Chem., Vol. 45, No. 7, 1997 2647
Ta ble 3. Effect of Rea ction Med iu m on F or m a tion of
HDMF a n d EHMF fr om Xylose in Ma illa r d Mod el
System sa
Ta ble 5. Effect of Am in o Acid Con cen tr a tion on th e
F or m a tion of HDMF a n d EHMF fr om Xylose in Ma illa r d
Mod el System sa
model system
reaction medium
HDMFb
EHMFb
molar ratio
xylose/amino acid
model system
HDMFb
EHMFb
xylose
xylose
waterc
phosphate
waterc
phosphate
malonate
waterc
phosphate
malonate
<0.01
<0.01
<0.01
<0.01
xylose/glycine
xylose/glycine
xylose/glycine
1:1
1:2
1:4
2.6
3.2
4.2
0.3
0.4
0.5
xylose/glycine
xylose/glycine
xylose/glycine
0.06
2.6
0.6
<0.01
0.3
0.1
xylose/L-alanine
xylose/L-alanine
xylose/L-alanine
1:1
1:2
1:4
0.9
1.2
1.6
7.5
12.5
20.0
xylose/L-alanine
xylose/L-alanine
xylose/L-alanine
0.02
0.9
0.3
0.05
7.5
0.8
a
Reaction conditions: phosphate buffer (0.2 M, pH 6.0), 90 °C,
1 h. Data are means of at least two assays, each of them injected
twice; maximum SD e 10% (in µg/mmol sugar).
b
a
Equimolar amounts of precursors were used. Reaction condi-
b
tions: pH 6, 90 °C, 1 h. Data are means of at least two assays,
each of them injected twice; maximum SD e 10% (in µg/mmol of
sugar). c The pH was not controlled during the reaction.
may accelerate the Maillard reaction, producing reactive
intermediates such as 1-deoxyosones which may decom-
pose to smaller fragments and so generate furanones
by condensation reactions (Blank et al., 1996a).
Ta ble 4. Effect of p H on F or m a tion of HDMF a n d EHMF
fr om Xylose in Ma illa r d Mod el System s Con ta in in g
Glycin e or L-Ala n in ea
model system
pH
HDMFb
EHMFb
CONCLUSION
xylose/glycine
xylose/glycine
xylose/glycine
5
6
7
2.6
2.6
3.1
0.3
0.3
0.7
The formation of HDMF and EHMF from pentose
sugars via Strecker-assisted chain elongation has been
substantiated. Quantitative data support the hypoth-
esis that HDMF is preferably formed in the presence of
glycine and EHMF in the presence of alanine. However,
sugar fragmentation/condensation reactions represent
an alternative formation pathway, particularly for
HDMF, which will be studied in more detail.
xylose/L-alanine
xylose/L-alanine
xylose/L-alanine
5
6
7
0.3
0.9
2.5
2.0
7.5
13.5
a
Equimolar amounts of precursors were used. Reaction condi-
tions: phosphate buffer (0.2 mol/L), 90 °C, 1 h. Data are means
of at least two assays, each of them injected twice; maximum SD
e 10% (in µg/mmol of sugar).
b
ACKNOWLEDGMENT
P a r a m eter s Affectin g th e F or m a tion of HDMF
a n d EHMF . Xylose was selected to study the formation
of HDMF and EHMF from pentose sugars in more
detail. As shown in Table 3 the furanones were
preferentially formed in the phosphate-buffered sys-
tems, i.e., 2.6 µg of HDMF in xylose/glycine and 7.5 µg
of EHMF in xylose/alanine, thus indicating a catalytic
effect of phosphate, particularly in the presence of the
amino acids. The effect of malonate was less pro-
nounced, yielding 0.6 µg of HDMF in xylose/glycine and
0.8 µg of EHMF in xylose/alanine. The model systems
without buffer gave rise to less than 0.1 µg of HDMF
and EHMF per mmol of pentose.
The formation of both furanones was favored at pH 7
resulting in 3.1 µg of HDMF and 13.5 µg of EHMF
(Table 4). This is in good agreement with the more
intense caramel-like overall flavor of these samples
compared to those prepared at pH 5. In general, 3(2H)-
furanones are formed from 1-deoxyosones which are
reactive intermediates in the Maillard reaction gener-
ated by degradation of Amadori compounds via 2,3-
enolization (Hodge et al., 1972; Ledl and Schleicher,
1990). This reaction is favored under neutral and
slightly alkaline conditions as shown by Beck et al.
(1988), i.e., the ratio 1-deoxyglucosone to 3-deoxy-
glucosone (formed via 1,2-enolization) was 20:1 at pH 7
and 8:5 at pH 4.5.
I.B. and L.B.F. are grateful to Ms. S. Devaud and Mr.
Y. Krebs for expert technical assistance and to Dr. E.
Prior for critically reading the manuscript.
LITERATURE CITED
Beck, J .; Ledl, F.; Severin, T. Formation of 1-deoxy-D-erythro-
2,3-hexodiulose from Amadori compounds. Carbohydr. Res.
1988, 177, 240-243.
Blank, I.; Fay, L. B. Formation of 4-hydroxy-2,5-dimethyl-
3(2H)-furanone and 4-hydroxy-2(or 5)-ethyl-5(or 2)-methyl-
3(2H)-furanone through Maillard reaction based on pentose
sugars. J . Agric. Food Chem. 1996, 44, 531-536.
Blank, I.; Devaud, S.; Fay, L. B. New aspects on the formation
of 3(2H)-furanones through the Maillard reaction. In Fla-
vour Science. Recent Developments; Taylor, A. J ., Mottram,
D. S., Eds.; The Royal Society of Chemistry: Cambridge,
1996a; pp 188-193.
Blank, I.; Lin, J .; Fumeaux, R.; Welti, H. D.; Fay, L. B.
Formation of 3-hydroxy-4,5-dimethyl-2(5H)-furanone (so-
tolone) from 4-hydroxy-L-isoleucine and 3-amino-4,5-di-
methyl-3,4-dihydro-2(5H)-furanone. J . Agric. Food Chem.
1996b, 44, 1851-1856.
Blank, I.; Sen, A.; Grosch, W. Potent odorants of the roasted
powder and brew of Arabica coffee. Z. Lebensm. Unters.
Forsch. 1992, 195, 239-245.
Eren, D.; Keinan, E. Total synthesis of linear polyprenoids. 3.
Syntheses of ubiquinones via palladium-catalyzed oligomer-
ization of monoterpene monomers. J . Am. Chem. Soc. 1988,
110, 4356-4362.
The amounts of HDMF and EHMF can also be
influenced by the ratio of the precursors (Table 5). In
general, increasing concentrations of the amino acids
favor the formation the furanones, e.g., a 4-fold increase
of alanine resulted in 2.7-fold more EHMF. One pos-
sible explanation is that elevated concentrations of
amino acids give higher levels of aldehydes, the Strecker
deamination products, required for the formation of
HDMF and EHMF from pentose sugars (Blank and Fay,
1996). Alternatively, higher amounts of amino acids
Feather, M. S. Amine-assisted sugar dehydration reactions.
Prog. Food Nutr. Sci. 1981, 5, 37-45.
Greene, T. W.; Wuts, P. G. M. Protective Groups in Organic
Synthesis, 2nd ed.; Wiley: New York, 1991; p 387.
Guth, H., Grosch, W. Deterioration of soya-bean oil: Quanti-
fication of primary flavour compounds using a stable isotope
dilution assay. Lebensm. Wiss. Technol. 1990, 23, 513-522.
Hirvi, T.; Honkanen, E.; Pyysalo, T. Stability of 2,5-dimethyl-
4-hydroxy-3(2H)-furanone and 2,5-dimethyl-4-methoxy-

2648 J. Agric. Food Chem., Vol. 45, No. 7, 1997
Blank et al.
3(2H)-furanone in aqueous buffer solutions. Lebensm. Wiss.
Technol. 1980, 13, 324-325.
Hodge, J . E.; Mills, F. D.; Fisher, B. E. Compounds of browned
flavor derived from sugar-amine reactions. Cereal Sci.
Today 1972, 17, 34-40.
Huber, U. A. Homofuronol: A powerful tool to prepare and
improve sophisticated high quality flavors. Perfum. Flavor.
1992, 17 (4), 15-19.
Re, L.; Maurer, G; Ohloff, G. Simple access to 4-hydroxy-2,5-
dimethyl-3(2H)-furanone (Furaneol), aroma component of
the pineapple and strawberry. Helv. Chim. Acta, 1973, 56,
1882-1894 (in German); Chem. Abstr. 1973, 79, 115381y.
Schieberle, P.; Grosch, W. Quantitative analysis of aroma
compounds in wheat and rye bread crusts using stable
isotope dilution assay. J . Agric. Food Chem. 1987, 35, 252-
257.
Ledl, F.; Schleicher, E. New aspects of the Maillard reaction
in foods and in the human body. Angew. Chem., Int. Ed.
Engl. 1990, 29, 565-594.
Semmelroch, P.; Laskawy, G.; Blank, I.; Grosch, W. Determi-
nation of potent odourants in roasted coffee by stable isotope
dilution assays. Flavour Fragrance J . 1995, 10, 1-7.
Sen, A.; Schieberle, P; Grosch, W. Quantitative determination
of 2,5-dimethyl-4-hydroxy-3(2H)-furanone and its methyl
ether using a stable isotope dilution assay. Lebensm. Wiss.
Technol. 1991, 24, 364-369.
Mazenod, F.; Delay, F.; Na¨f, F. 2,5-Dimethyl-4-hydroxy-3(2H)-
furanone (Furaneol) from methyl R-D-glucopyranoside. Chim-
ia 1992, 46, 403-405 (and references therein).
Olah, G. A.; Arvanaghi, M. Aldehydes by formylation of
Grignard or organolithium reagents with N-formylpiperi-
dine. Angew. Chem., Int. Ed. Engl. 1981, 20, 878-879.
Pickenhagen, W.; Velluz, A.; Passerat, J .-P.; Ohloff, G. Estima-
tion of 2,5-dimethyl-4-hydroxy-3(2H)-furanone (Furaneol) in
cultivate of wild strawberries, pineapples, and mangoes. J .
Sci. Food Agric. 1981, 32, 1132-1134.
Preininger, M.; Grosch, W. Evaluation of key odorants of the
neutral volatiles of Emmentaler cheese by the calculation
of odour activity values. Lebensm. Wiss. Technol. 1994, 27,
237-44.
Received for review December 27, 1996. Accepted March 24,
1997.X F.J .L. and M.S. are indebted to the Schweizerische
Nationalfonds zur Fo¨rderung der wissenschaftlichen Fors-
chung, Bern, for financial support (Grant 20-36385-92).
J F960997I
X Abstract published in Advance ACS Abstracts, J une
1, 1997.