Chemical Conversion of α-Keto Acids in Relation to Flavor Formation in Fermented Foods

Article abstract of DOI:10.1021/jf035147z

Formation of flavor compounds from branched-chain α-keto acids in fermented foods such as cheese is believed to be mainly an enzymatic process, while the conversion of phenyl pyruvic acid, which is derived from phenylalanine, also proceeds chemically. In this research, the chemical conversion of α-keto acids to aldehydes with strong flavor characteristics was studied, with the main focus on the conversion of α-ketoisocaproic acid to the aldehyde 2-methylpropanal, and a manganese-catalyzed reaction mechanism is proposed for this conversion. The mechanism involves keto-enol tautomerism, enabling molecular oxygen to react with the β-carbon atom of the α-keto acid, resulting in a peroxide. This peroxide can react in several ways, leading to unstable dioxylactone or noncyclic intermediates. These intermediates will break down into an aldehyde and oxalate or carbon oxides (CO and CO₂). All the α-keto acids tested were converted at pH 5.5 and in the presence of manganese, although their conversion rates were rather diverse. This chemical reaction might provide new ways for controlling cheese flavor formation with the aim of acceleration of the ripening process or diversification of the flavor characteristics.

Full text of DOI:10.1021/jf035147z

J. Agric. Food Chem. 2004, 52, 1263−1268 1263
Chemical Conversion of r-Keto Acids in Relation to Flavor
Formation in Fermented Foods
BART A. SMIT,† WIM J. M. ENGELS,† MARTIN ALEWIJN,‡
GIJS T. C. A. LOMMERSE,‡ ERWIN A. H. KIPPERSLUIJS,†
JAN T. M. WOUTERS,‡ AND GERRIT SMIT*,†
NIZO Food Research, Department of Flavor, Nutrition, and Ingredients, P.O. Box 20,
6710 BA Ede, The Netherlands, and Wageningen University, Department of Agrotechnology and
Food Sciences, Wageningen, The Netherlands
Formation of flavor compounds from branched-chain R-keto acids in fermented foods such as cheese
is believed to be mainly an enzymatic process, while the conversion of phenyl pyruvic acid, which is
derived from phenylalanine, also proceeds chemically. In this research, the chemical conversion of
R-keto acids to aldehydes with strong flavor characteristics was studied, with the main focus on the
conversion of R-ketoisocaproic acid to the aldehyde 2-methylpropanal, and a manganese-catalyzed
reaction mechanism is proposed for this conversion. The mechanism involves keto-enol tautomerism,
enabling molecular oxygen to react with the â-carbon atom of the R-keto acid, resulting in a peroxide.
This peroxide can react in several ways, leading to unstable dioxylactone or noncyclic intermediates.
These intermediates will break down into an aldehyde and oxalate or carbon oxides (CO and CO2).
All the R-keto acids tested were converted at pH 5.5 and in the presence of manganese, although
their conversion rates were rather diverse. This chemical reaction might provide new ways for
controlling cheese flavor formation with the aim of acceleration of the ripening process or diversification
of the flavor characteristics.
KEYWORDS: r-Keto acids; 2-oxo acids; flavor; leucine; aroma formation; 2-methylpropanal; dairy; cheese
INTRODUCTION
sion of PPA was demonstrated by Villablanca et al. (15) and
by Nierop Groot et al. (16), and the conversion of KMBA to
methylthioacetaldehyde has been identified by Yvon et al.
(personal communication). The conversion of phenylpyruvic
acid is due to oxidation by molecular oxygen and is catalyzed
by divalent cations such as manganese (15, 16).
Many flavors in fermented products are derived from amino
acids. Examples of important amino-acid-derived flavor com-
pounds in cheese are aldehydes such as benzaldehyde, 2-meth-
ylpropanal and 3- and 2-methylbutanal, but also methionine-
derived sulfur compounds, such as methanethiol and DMDS
(1-4). Most flavor forming reactions in fermented products such
as cheese are enzymatic (5-9). Protein degradation is initiated
by proteolysis and peptidolysis, leading to free amino acids (10).
The amino acids are mostly enzymatically transaminated in the
bacterial cell to the corresponding R-keto acid (11, 12). These
R-keto acids can be converted to various metabolites, such as
the aldehydes and sulfur compounds mentioned above. The
R-keto acid of leucine (KICA) for example, can be enzymati-
cally decarboxylated by a number of Lactococcus lactis strains
to 3-methylbutanal (13, 14). Although most of these reactions
are enzymatic, the R-keto acids of phenylalanine (phenylpyruvic
acid ) PPA) and methionine (KMBA) are also chemically
converted to flavor compounds such as benzaldehyde and
methylthioacetaldehyde. The occurrence of the chemical conver-
The reactivity of ketones such as R-keto acids is mainly due
to the existence of their enol tautomers (17-19). Molecular
oxygen is able to react with the enol, leading to the formation
of a peroxide on the â-carbon atom (20). The nature of this
peroxidation is not totally clear and might as well proceed
directly as via a radical mechanism (20, 21). The very reactive
peroxide intermediary of p-methoxyphenylpyruvate can react
with the acid or keto carbon atom of the molecule, resulting in
a dioxylactone or dioxyethanol, although the latter reaction only
occurs under basic, non nucleophilic conditions (22). The
unstable dioxylactone formed from p-methoxyphenylpyruvate
will decompose into an aldehyde and the carbon oxides (CO2
and CO) (22). A hydrophilic solvent, like water, is also able to
react with the keto function of the peroxide intermediate,
resulting in a noncyclic intermediate (22). Oxalic acid and water
can be split off, leaving an aldehyde. To summarize, the
conversion of several keto acids depends strongly on the
circumstances (e.g., solvent and pH), but the products of these
reactions are an aldehyde and oxalate or carbon oxides.
* To whom correspondence should be addressed. Tel.: +31 318 659538.
Fax: +31 318 650400. E-mail: gerrit.smit@nizo.nl.
† NIZO Food Research.
‡ Wageningen University.
10.1021/jf035147z CCC: $27.50 © 2004 American Chemical Society
Published on Web 02/11/2004

1264 J. Agric. Food Chem., Vol. 52, No. 5, 2004
Smit et al.
This research focuses on revealing the characteristics and
reaction mechanism of the chemical conversion of R-keto acids,
leading to compounds relevant for fermented foods. It is
important to note that the conversion of substrates present in
the food takes place under mild conditions. The branched chain
R-keto acid of leucine, R-ketoisocaproic acid, was chosen for
this study because its conversion product, 2-methylpropanal,
has an important flavor impact in foods.
MATERIALS AND METHODS
The Chemical Conversion was assayed in a 20-mL headspace vial
with a reaction volume of 5 mL or in 1 mL HPLC vials, with a reaction
volume of 0.5 mL. Final concentrations of reactants in the reaction
mixture were 50 mM buffer, 10 mM metal chloride salt, 10 mM
substrate. Standard conditions were succinate buffer (pH 5.5), manga-
nese chloride, and R-ketoisocaproic acid as substrate. Buffers were
succinate (pH 3.5-6.5), glycerolphosphate (pH 4.5-7.2), bis-Tris-
propane (pH 6.4-7.4), HEPES (pH 6.4-7.8), and Tris-HCl (pH 7-8).
Metal salts: MnCl2, MnSO4, MgCl2, CaCl2, CuCl2, FeCl2, FeCl3, CoCl2,
ZnCl2, and NaCl. Substrates used in this study were R-ketoisocaproic
acid (KICA), 2-oxo-3-methyl pentanoic acid, R-ketoisovaleric acid,
R-ketohexanoic acid, R-ketopentanoic acid, R-ketobutanoic acid, pyruvic
acid, 4-methylthio-2-oxobutanoic acid (KMBA), phenylpyruvic acid
(PPA), 3-(4-hydroxyphenyl)-2-oxopropanoic acid, 3-indol-3-yl-2-oxo-
propanoic acid (indole pyruvate), 2-oxopentanedioic acid, 4-methyl-
pentanoic acid, 3-methylbutanoic acid, and 2-methylpropanoic acid.
All substrates were analytical grade and obtained from Sigma (Zwijn-
drecht, The Netherlands) or Fisher (Landsmeer, The Netherlands). At
t ) 0, substrate was added to the rest of the reaction mixture, and the
vial was closed (including 15 mL air in the headspace), and placed in
a incubator which was shaking at 500 rpm for 10 s on/5 s off intervals
at 40 °C (Combi Pal, CTC Analytics, Zwingen, Switzerland).
Figure 1. Effect of manganese concentrations on the conversion rate of
KICA to 2-methylpropanal (succinate buffer, pH 6.0).
220 nm to detect the compounds in both system configurations.
Concentrations were calculated using separate calibration samples
containing 4, 6, 8, and 10 mM of the acids in 50 mM succinate buffer
(pH ) 5.5).
Gas Chromatography. GC was used to determine the concentrations
of O2 and CO2 in the headspace. The reaction mixtures were prepared
as described above and incubated at 40 °C. Before each measurement,
samples were inverted four times and cooled to 25 °C. For 5 s, a
headspace sample was taken and analyzed on a CP2001 gas chromato-
graph (Chrompack, Middelburg, The Netherlands), equipped with a
HayeSep A column (25 cm) and a MS 5A column (4 m). Because the
sampled volume was about 0.5 mL, and samples were analyzed in
quadruplicate, a separate reaction vial was used each time interval.
Concentrations were calculated by using air (21% O2, 0.033% CO2)
and a calibration gas (1.00% CO2, 1.00% O2) for calibration.
Direct Inlet Mass Spectrometry (DI-MS) was used for the on-line
monitoring of product formation. The method used was a slightly
adapted version of the method published previously (14). At regular
intervals, a 150-µL headspace sample was taken from the incubator.
The sample was injected with a 1.0-mL syringe at 60 °C and 20 µL/s
by the Combi Pal auto sampler (CTC Analytics, Zwingen, Switzerland)
into the gas chromatograph (CE-Instruments, Milan, Italy). The GC
was equipped with an 8-m × 0.1-mm deactivated silica column
(Interscience, Breda, The Netherlands). Helium was used as carrier gas,
with a column flow of 2.5 mL/min and split flow of 10 mL/min. The
oven temperature was 150 °C. Single ions at m/z ) 58, 72, and 105
were recorded by a quadrupole mass spectrometer (Trace MS, CE-
Instruments, Milan, Italy). These specific m/z values represent fragments
of 3-methylbutanal, 2-methylpropanal, and benzaldehyde, respectively.
Acquisition time was 30 s. The height of the response at a given m/z
value is directly correlated to the concentration of the compound.
Quantification was done by using a calibration curve ranging from 10
to 1000 µM in the same buffer as that used in the experiments, which
was analyzed before and after each experiment in triplicate. The
conversion rate is defined as the change of concentration in time,
expressed in µM/h.
High-Pressure Liquid Chromatography. HPLC was used for the
determination of R-keto acids and organic acids. The reaction mixture
was prepared as described above and incubated in a model 717 auto
sampler (Waters, Milford, MA) at 40 °C. At regular intervals, samples
(25 µL) were taken and injected to the system. For the analysis of
polar molecules (pyruvic acid, 2-oxobutyric acid, 2-oxopentanedioic
acid, and oxalate) the system consisted of a LC-10AT pump (Shimadzu,
Tokyo, Japan), pumping 0.01 M methane sulfonic acid with an isocratic
flow of 0.6 mL/min over a Rezex Fast Fruit precolumn (100- × 7.8-
mm) and a double Rezex Organic Acid column (300- × 7.8-mm,
Phenomenex, Torrance, Ca) at 30 °C. The other acids were analyzed
using the same system, but equipped with a wide pore C18-RP column
(250- × 4.6-mm, Bio-Rad, Veenendaal, The Netherlands), and com-
ponents were eluted at 40 °C with a gradient from 5 to 30% acetonitrile
in demineralized water containing 0.1% trifluoracetic acid (TFA). A
model 1481 Lamda-max detector (Waters, Milford, MA) was used at
RESULTS AND DISCUSSION
The production of branched chain aldehydes from amino acids
in cheese is believed to be a two step process consisting of a
transaminating step and decarboxylating step. Leucine, the most
abundant branched-chain amino acid in cheese, is transaminated
to R-ketoisocaproic acid (KICA), which can subsequently be
decarboxylated to 3-methylbutanal (13). Under certain condi-
tions, fermented dairy products had unexpectedly high 2-meth-
ylpropanal concentrations (unpublished data). This flavor com-
pound is believed to be produced from valine via an enzymatic
pathway similar to the one elucidated for leucine, with R-keto-
isovaleric acid as the intermediate (5-7). To control cheese
flavor formation, it is important to understand the characteristics
of these enzymatic conversions. Studying the enzymatic decar-
boxylation of KICA by static headspace measurements (13) of
incubations of cell free extract (CFE) of Lactococcus lactis
B1157 with KICA revealed that the concentration of 2-meth-
ylpropanal also increased (data not shown) although no R-keto-
isovaleric acid was added to these incubations. Repetition of
the experiment with cooked CFE, thereby inactivating the
decarboxylating enzyme, still led to the same increase of
2-methylpropanal concentration. This suggested that KICA could
also be converted chemically with 2-methylpropanal as one of
the products. This reaction was studied in more detail using
direct inlet mass spectrometry.
None of the tested metal-salts, except the mangenese-salts,
were able to catalyze the reaction at a concentration of 10 mM
either at pH 5.5 or pH 7.5. Addition of 10 and 20 mM EDTA
to reaction mixtures containing 10 mM MnCl2 resulted in a
decrease in the conversion rate to 12 and 4% respectively,
indicating the need for free manganese ions. The correlation

Conversion of R-Keto Acids in Foods
J. Agric. Food Chem., Vol. 52, No. 5, 2004 1265
Figure 2. Effect of pH on the conversion rate of KICA to 2-methylpropanal.
(10 mM KICA, 10 mM MnCl2, and the following buffers: succinate (O),
glycerophosphate (2), bis-TRIS-propane (0), HEPES (+) and TRIS (b)).
between the manganese concentration and the conversion rate
is shown in Figure 1. The optimal manganese concentration
(>10 mM) was much higher than expected for a catalyst. This
might suggest that Mn2+ acts as reactant instead of catalyst, or
only a fraction of the manganese dissolved is present in the
catalytic active form.
The correlation between pH and conversion rate was deter-
mined using several buffers for pH values that are relevant for
fermented dairy products (pH 4-8) (Figure 2). The pH of
Gouda cheese is about 5.5 (23), and the intracellular pH of the
microorganisms in cheese is close to neutral (24, 25). A local
maximum conversion was found at pH 5.5, but at pH 8 the
conversion rate was even higher. At pH > 8, even higher
conversion rates were expected, but the formation of a brown
precipitate (MnO2) at this pH resulted in noninterpretable results.
Varying the substrate concentration between 0.5 and 50 mM
resulted in an almost linear increase of the conversion rate under
the conditions tested (excess of oxygen is present in the air
present in the headspace, 10 mM MnCl2, pH 6.5) (data not
shown). Oxygen is essential for the reaction because very low
oxygen concentrations (0.1%) in the reaction vial led to a major
decrease in the conversion rate (>90%). Moreover, oxygen was
found to be consumed during the reaction (Figure 3).
Figure 3. Oxygen (0) consumption and production of 2-methylpropanal
(O), carbon dioxide (CO2) (2) and oxalate (×) during the chemical
conversion of KICA at pH 5.5 (A) and 7.5 (B) in the presence of 10 mM
manganese. The dotted line represents the sum of oxalate and CO2.
and the increase of the sum of oxalate and CO2 clearly correlate
with ratios close to 1. However, the oxygen consumption in
relation to the 2-methylpropanal production is slightly larger at
high pH. This could again be caused by the oxidation of
manganese to MnO2, as described earlier. No further quantitative
measurements were done to confirm this.
Our finding of chemical conversion of KICA shows that the
chemical conversion of R-keto acids in fermented foods will
not be limited to PPA and KMBA. More information on the
conversion of several other substrates is therefore desired to
estimate the impact of this reaction on the degradation of other
R-keto acids in fermented products. Therefore, we incubated a
variety of different R-keto acids and organic acids and deter-
mined their conversion by HPLC. The substrates were chosen
based on their involvement in the amino acid catabolism or just
as model substrates (linear keto acids and branched organic
acids). In Table 1, the relative conversion rates and structures
of the substrates tested are shown. The results indicate that the
reaction is limited to R-keto acids, and in general, substrates
with electron withdrawing side chains are converted faster. Also,
two hydrogen atoms on the â-carbon atom seem to be preferable
to one.
The reactivity of R-keto acids is mainly due to the existence
of their enol tautomers (17, 18). This was tested by measuring
the conversion of the organic acid similar to KICA but lacking
the keto-group (isohexanoic acid). This organic acid and other
organic acids were not converted under the conditions tested.
However, “linearized” KICA (2-keto hexanoic acid) was
converted with approximately the same conversion rate as
KICA. The organic acids tested are probably not able to form
the reactive enol tautomers. Therefore, we conclude that the
keto group (and the formation of the enol tautomer) is most
likely essential for this oxidation reaction. The amount of the
enol present in the equilibrium can be stabilized by intramo-
For a better understanding of the reaction mechanism, it is
important to know which products are formed. For the conver-
sion of KICA to 2-methylpropanal, two carbon atoms have to
be cleaved off, and the expected products are oxalic acid and/
or carbon oxides (CO and CO2), based on the homology with
other reactions (20-22). The formation of these products was
measured for reactions proceeding at pH 5.5 and 7.5 by use of
direct inlet mass spectrometry, GC, and HPLC simultaneously.
The HPLC method was also optimized for detection of other
di- and mono-carbonic acids, in case the two carbon atoms
would split off in a different manner. Oxalate and carbon dioxide
were the only products detected, but the GC columns were not
suitable for measurement of CO. These two products are formed
via different mechanisms. Per molecule of 2-methylpropanal
formed, either one molecule of oxalate or one molecule of CO2
were formed. In the case where CO2 was formed, this must
have been accompanied by CO or formiate to complete the
balance of elements. Formiate was not detected by HPLC
analysis, which is in agreement with the findings of the
conversion of p-hydroxyphenylpyruvate and phenylpyruvate (21,
22, 26). The quantitative headspace data were processed, taking
gas-liquid equilibria (Henry’s law (KHCO ) 0.034 M/atm, KHO
2
2
) 0.0013 M/atm (27)) and acid-base equilibria (to calculate
[HCO3-] (KaHCO ) 2.3 ‚ 10-8 M)) into account. The results
-
3
are combined in Figure 3. The increase of 2-methylpropanal

1266 J. Agric. Food Chem., Vol. 52, No. 5, 2004
Smit et al.
Table 1. Relative Conversion Rates of Several (R-Keto) Acids under Standard Conditions (Succinate, pH 5.5, 10 mM Mn)a
a The standard deviation of the duplicate measurements is relative to the conversion of PAA. The amino acid substrates for the R-keto acids are mentioned in the last
column. *Relative standard deviations of the values presented are in all cases e 3.5%.
lecular hydrogen bonds and by conjugation of the carbon-
carbon double bond with the carbonyl group (17, 28). Keto-
enol tautomerization is generally catalyzed by acid or base, but
bivalent metal ions can also accelerate this reaction (17, 26,
29). A minimum in the conversion rates of KICA to 2-meth-
ylpropanal near neutral pH was confirmed in our experiments.
We found a major catalytic effect of manganese ions; however,
we did not find effects of other divalent metal ions, as was
described for PPA conversion (30). If the main catalytic effect
of manganese under the conditions tested is enhancement of
the tautomerism by, for example, increasing the conversion rate
or stabilizing the enol tautomer, keto-enol tautomery most
probably determines the rate of the overall conversion. This is
also indicated by the slow conversion of 2-oxo-3-methyl
pentanoic acid and R-ketoisovaleric acid compared to their linear
equivalent. The â-carbon atom of 2-oxo-3-methyl pentanoic acid
and R-ketoisovaleric acid is methylated, leaving only one
â-hydrogen atom, which is essential for the tautomerism, and
thereby might slow the conversion.
On the basis of homology with several other ketones, the
next step in the oxidation is most probably the formation of a
peroxide on the â-carbon atom (20). This oxidation might
proceed directly or via a radical mechanism (20, 21). In the
latter case, electron transfer of the nucleophilic enolate ion to
molecular oxygen results in an R-keto radical (28, 29, 31). The
propagation step involves the addition of molecular oxygen,
resulting in a hydroperoxy radical (20). This radical can transfer
the electron to another enol, leaving a new R-keto radical and
a R-keto hydroperoxide (29, 31).
The very reactive peroxide intermediary can react with the
acid or keto carbon atom of the molecule resulting in a
dioxylactone or dioxetanol. The cyclization to the dioxethanol
is not a spontaneous reaction and only occurs under basic, non
nucleophilic conditions (22). The unstable dioxylactone will
decompose into an aldehyde (RCdO), and the carbon oxides
(CO2 and CO) (22). A hydrophilic solvent like water is also
able to react with the keto function of the peroxide intermediary,
resulting in a noncyclic intermediary (22). Oxalic acid and water

Conversion of R-Keto Acids in Foods
J. Agric. Food Chem., Vol. 52, No. 5, 2004 1267
LITERATURE CITED
can be split off, leaving an aldehyde (RCdO). We identified
both CO2 and oxalate during conversion of KICA. This indicates
that the conversion of KICA most probably proceeds via both
pathways described for a nucleophilic environment. We also
found that the sum of oxalate and carbondioxide correlated with
the 2-methylpropanal production with ratios close to 1, and that
the pathway via the noncyclic intermediary is enhanced at higher
pH.
(1) Engels, W. J. M.; Dekker, R.; de Jong, C.; Neeter, R.; Visser,
S. A comparative study of volatile compounds in the water-
soluble fraction of various types of ripened cheese. Int. Dairy J.
1997, 7, 255-263.
(2) Rychlik, M.; Bosset, J. O. Flavour and off-flavour compounds
of Swiss Gruyere cheese. Identification of key odorants by
quantitative instrumental and sensory studies. Int. Dairy J. 2001,
11, 903-910.
All R-keto acids, including the R-keto acids corresponding
to amino acids, were converted under the mild conditions used
(Table 1). This implies that the products of the other amino
acid derived R-keto acids are also expected to be found in
fermented products. In cheese and/or in the starter bacteria, not
all parameters are as optimal as those under the conditions tested.
The pH, for instance, varies during the fermentation and differs
between the inside and outside of the bacterial cell. During
growth, the pH of lactic acid bacteria is neutral, but after growth
they are sometimes not able to maintain this pH, or will even
lyse, which results in a decrease in the pH (24, 25). The pH of
Gouda cheese, for example, decreases from 7 to below 5.4 in
the first 6 h of cheese production, after which it stays more or
less stable (23). This pH is optimal for the chemical reaction to
proceed. In smear-ripened and mould cheeses, the pH might
increase to 7.5, due to the metabolic activities such as deami-
nation (32). Some lactic acid bacteria, like Lactobacillus casei
subsp. casei and Lactobacillus plantarum are able to accumulate
manganese, leading to high local intracellular manganese
concentrations (up to 50 mM) (16, 33), which is also favorable
for the chemical reaction. Particularly when transamination is
stimulated, R-keto acids accumulate in the cheese matrix (up
to 17.9% of the initial leucine concentration (34)), resulting in
reasonable amounts of substrates present for the chemical
reaction to proceed. However, the low oxygen concentration in
cheese is very unfavorable, but some 2-methylpropanal was still
formed at low oxygen concentrations (0.1% in headspace, which
equals 1.3 µM in the reaction mixture), and depending on the
type of cheese, particularly at the outside, some oxygen is
available up to several weeks (35). The reaction will most
probably proceed much more slowly under cheese ripening
conditions than under the conditions used in this study, although
under simulated Cheddar cheese conditions, spontaneous deg-
radation of hydroxyphenylpyruvate to hydroxybenzaldehyde
occurs (36), and both benzaldehyde and hydroxybenzaldehyde
were found in semihard cheeses (34). Cheese ripening is a long
process, which may take up to a year, and therefore we expect
that the conversion rate will still be sufficient to change cheese
flavor characteristics. In future work, this will be tested in cheese
model systems, pilot cheese productions using selected starter
cultures containing manganese accumulating strains, and tran-
saminase overexpression mutants in combination with the
addition of R-ketoglutarate to increase R-keto acid formation.
A large difference in reaction rates between the substrates exist,
the availability of the substrates in a product is very different,
and the flavor characteristics of the aldehydes differ largely.
More research has to be done on this topic. The chemical
reaction described might be used as a new control point for
aroma formation and flavor diversification in several fermented
food products, not only by increasing flavor formation but also
for preventing off-flavors. Increasing the oxygen concentration
by using more permeable coatings, or selecting manganese
accumulating strains might be relevant control parameters in
this respect.
(3) 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-
244.
(4) Banks, J. M.; Brechany, E. Y.; Christie, W. W.; Hunter, E. A.;
Muir, D. D. Volatile components in steam distillates of cheddar
cheese as indicator indices of cheese maturity, flavour and odour.
Food Res. Int. 1992, 25, 365-373.
(5) Smit, G.; Verheul, A.; van Kranenburg, R.; Ayad, E.; Siezen,
R.; Engels, W. J. M. Cheese flavour development by enzymatic
conversions of peptides and amino acids. Food Res. Int. 2000,
33, 153-160.
(6) McSweeney, P. L. H.; Sousa, M. J. Biochemical pathways for
the production of flavour compounds in cheeses during ripen-
ing: A review. Lait 2000, 80, 293-324.
(7) Yvon, M.; Rijnen, L. Cheese flavour formation by amino acid
catabolism. Int. Dairy J. 2001, 11, 185-201.
(8) Weimer, B.; Seefeldt, K.; Dias, B. Sulfur metabolism in bacteria
associated with cheese. Ant. Leeuwenhoek 1999, 76, 247-261.
(9) Williams, A. G.; Noble, J.; Tammam, J.; Lloyd, D.; Banks, J.
M. Factors affecting the activity of enzymes involved in peptide
and amino acid catabolism in nonstarter lactic acid bacteria
isolated from Cheddar cheese. Int. Dairy J. 2002, 12, 841-852.
(10) Visser, F. M. W. Contribution of enzymes from rennet, starter
bacteria and milk to proteolysis and flavour development in
Gouda cheese. II. Development of bitterness and cheese flavour.
Neth. Milk Dairy J. 1977, 31, 188-209.
(11) Engels, W. J. M. Volatile and nonvolatile compounds in ripened
cheese: their formation and their contribution to flavour. Ph.D.
Thesis, Wageningen Agricultural University, 1997.
(12) Gao, S.; Oh, D. H.; Broadbent, J. R.; Johnson, M. E.; Weimer,
B. C.; Steele, J. L. Aromatic amino acid catabolism by lactococci.
Lait 1997, 77, 371-381.
(13) Smit, B. A.; Engels, W. J. M.; Wouters, J. T. M.; Smit, G.
Diversity of L-leucine catabolism in various microorganisms
involved in dairy fermentations and identification of the rate
controlling step in 3-methylbutanal formation. Appl. Microbiol.
Biotechnol. 2003, Online First, DOI: 10.1007/s00253-003-1447-
8.
(14) Smit, B. A.; Engels, W. J. M.; Bruinsma, J.; Van Hylckama
Vlieg, J. E. T.; Wouters, J. T. M.; Smit, G. Development of a
High Throughput-screening method to test flavour-forming
capability of food microorganisms, in Press.
(15) Villablanca, M.; Cilento, G. Oxidation of phenylpyruvic acid.
Biochim. Biophys. Acta 1987, 926, 224-230.
(16) Nierop Groot, M. N.; de Bont, J. A. M. Involvement of
manganese in conversion of phenylalanine to benzaldehyde by
lactic acid bacteria. Appl. EnViron. Microbiol. 1999, 65, 5590-
5593.
(17) Carey, F. A.; Sundberg, R. J. 7.3 Enols and enamines. In
AdVanced organic chemistry; F. A. Carey and R. J. Sundberg,
Eds.; Plenum Press: London, 1990; pp 416-421.
(18) Pitt, B. M. Oxidation of phenylpyruvates to aromatic aldehydes
and oxalate. Nature 1962, 196, 272-273.
(19) Doy, C. H. Alkaline conversion of 4-hydroxyphenylpyruvic acid
to 4-hydroxybenzaldehyde. Nature 1960, 14, 1529-1531.
(20) Carey, F. A.; Sundberg, R. J. 12 Oxidations. In AdVanced organic
chemistry; F. A. Carey and R. J. Sundberg, Eds.; Plenum Press:
London, 1990; pp 651-657.
(21) Gersmann, H. R.; Bickel, A. F. Autoxidation of ketones and esters
in basic solution. J. Chem. Soc. B 1971, 2230-2237.
ACKNOWLEDGMENT
We thank Jildert Bruinsma for technical assistance.

1268 J. Agric. Food Chem., Vol. 52, No. 5, 2004
Smit et al.
(22) Jefford, C. W.; Kno¨pfel, W.; Cadby, P. A. Oxygenation of methyl
p-methoxyphenylpyruvate. alpha-keto-â-peroxylactone and di-
oxetanol intermediates. Tetrahedron Lett. 1978, 38, 3585-3588.
(23) Walstra, P.; Noomen, A.; Geurts, T. J. Dutch type varieties. In
Cheese: Chemistry, physics, and microbiology; P. FOX, Ed.;
Chapman & Hall: London, 1993; pp 39-82.
(24) Molenaar, D.; Abee, T.; Konings, W. N. Continuous measure-
ment of the cytoplasmic pH in Lactococcus lactis with a
fluorescent pH indicator. Biochim. Biophys. Acta 1991, 1115,
75-84.
(25) Hugenholtz, J.; Starrenburg, M. J. C.; Abee, T. Control of internal
pH in lactic acid bacteria. In Proceedings: Lactic acid bacteria;
Adria Normandie: Caen, France, 1997; pp 469-470.
(26) Bu¨chner, T.; Kirberger, E. U¨ ber die enol-keto-tautomerie der
p-oxy-phenylbrenztraubensaure. Biochim. Biophys. Acta 1952,
8, 401-406.
(27) Sander, R. Compilation of Henry’s law constants for inorganic
and organic species of potential importance in environmental
chemistry. www..mpch-mainz.mpg.de/∼sander/res/henry.html.
1999.
(30) Nierop Groot, M. N.; De Bont, J. A. M. Conversion of
phenylalanine to benzaldehyde initiated by an aminotransferase
in Lactobacillus plantarum. Appl. EnViron. Microbiol. 1998, 64,
3009-3013.
(31) House, H. O. Modern Synthetic Reactions, 2nd ed.; Benjamin-
Cummings: Menlo Park, CA, 1972.
(32) Bockelmann, W. The production of smear cheeses. In Dairy
Processing: ImproVing quality; G. Smit, Ed.; Woodhead Publish-
ing Limited: Cambridge, 2003; pp 470-491.
(33) Archibald, F. Manganese: Its aquisition by and function in lactic
acid bacteria. Crit. ReV. Microbiol. 1986, 13, 63-109.
(34) Yvon, M.; Berthelot, S.; Gripon, J. C. Adding alpha-ketoglutarate
to semi-hard cheese curd highly enhances the conversion of
amino acids to aroma compounds. Int. Dairy J. 1998, 8, 889-
898.
(35) Van den Tempel, T.; Gundersen, J. K.; Nielsen, M. S. The
microdistribution of oxygen in Danablu cheese measured by a
microsensor during ripening. Int. J. Food Microbiol. 2002, 75,
157-161.
(36) Gao, S.; Mooberry, E. S.; Steele, J. L. Use of 13C nuclear
magnetic resonance and gas chromatography to examine me-
thionine catabolism by lactococci. Appl. EnViron. Microbiol.
1998, 64, 4670-4675.
(28) Russel, G. A.; Bemix, A. G. The oxidation of carbanions. I.
Oxidation of triaryl carbanions and other tertiairy carbanions. J.
Am. Chem. Soc. 1966, 88, 5491-5497.
(29) Bailey, E. J.; Barton, D. H. R.; Elks, J.; Templeton, J. F.
Compounds related to the steroid hormones. part IX. Oxygen-
ation of steroid ketones in strongly basic medium: A new method
of preparation of 17a-hydroxypregnan-20-ones. J. Chem. Soc.
1962, 1578-1591.
Received for review October 7, 2003. Revised manuscript received
January 9, 2004. Accepted January 9, 2004. This work was financially
supported by Stichting J. Mesdag Fonds, The Netherlands.
JF035147Z