ORIGINAL PAPER
The yeast Dekkera bruxellensis genome contains two orthologsof the ARO10 gene encoding for phenylpyruvate decarboxylase
Anna Theresa de Souza Liberal • Marcelo Falsarella Carazzolle •
Goncalo Amarante Pereira • Diogo Ardaillon Simoes •
Marcos Antonio de Morais Jr.
Received: 21 December 2011 / Accepted: 9 April 2012 / Published online: 22 April 2012
� Springer Science+Business Media B.V. 2012
Abstract The yeast Dekkera bruxellensis possesses
important physiological traits that enable it to grow in indus-
trial environments as either spoiling yeast of wine production
or a fermenting strain used for lambic beer, or fermenting
yeast in the bioethanol production process. In this work, in
silico analysis of the Dekkera genome database allowed the
identification of two paralogous genes encoding for phenyl-
pyruvate decarboxylase (DbARO10) that represents a unique
trait among the hemiascomycetes. The molecular analysis of
the theoretical protein confirmed its protein identity. Upon
cultivation of the cell in medium containing phenylpyruvate,
both increases in gene expression and in phenylpyruvate
decarboxylase activity were observed. Both genes were dif-
ferentially expressed depending on the culture condition and
the type of metabolism, which indicated the difference in
the biological function of their corresponding proteins.
The importance of the duplicated DbARO10 genes in the
D. bruxellensis genome was discussed and represents the first
effort to understand the production of flavor by this yeast.
Keywords Ehrlich pathway � Gene expression �RT-qPCR
Introduction
The yeast D. bruxellensis, through its anamorph Bretta-
nomyces bruxellensis, is considered to be the worst wine
contaminant. It is responsible for the formation of volatile
phenols and tetrahydropyridines that cause off-flavors
during the aging period in barrels. In contrast, its impor-
tance in the production of the traditional Belgian lambic
beer (Loureiro and Malfeito-Ferreira 2003) and fuel-
ethanol have been reported (Liberal et al. 2007; Blomqvist
et al. 2010). On the other hand, its cells can produce
4-ethyl-phenol that is regarded as an important component
of the wine flavor under certain concentrations (Chatonnet
et al. 1995). The accumulation of 4-ethyl-phenol and
4-vinylphenol at certain levels seems to be dependent on
the cell growth and on the activity of hydroxycinnamate
decarboxylase (Barata et al. 2008; Benito et al. 2009). In
addition, the assimilation by the yeast cells of the amino
acids present in wort or must promotes the production of
higher alcohols and esters with flavoring characteristics
(Dickinson et al. 2003). Aromatic amino acids (tryptophan,
tyrosine and phenylalanine), branched-chain amino acids
(leucine, isoleucine and valine) and methionine are
assimilated by the Ehrlich pathway. They are first transa-
minated by transaminases to produce glutamate from
2-oxoglutarate and the corresponding 2-oxo acid from the
assimilated amino acids. In the next step, the 2-oxo acids
Electronic supplementary material The online version of thisarticle (doi:10.1007/s11274-012-1054-x) contains supplementarymaterial, which is available to authorized users.
A. T. de Souza Liberal � D. A. Simoes � M. A. de Morais Jr.
Interdepartmental Research Group on Metabolic Engineering,
Federal University of Pernambuco, Recife, Brazil
M. F. Carazzolle � G. A. Pereira
Department of Genetics and Evolution, Institute of Biology,
University of Campinas, Campinas, Brazil
D. A. Simoes
Department of Biochemistry, Federal University
of Pernambuco, Recife, Brazil
M. A. de Morais Jr. (&)
Department of Genetics, Federal University of Pernambuco, Av.
Moraes Rego, 1235, Cidade Universitaria, Recife,
PE 50670-901, Brazil
e-mail: [email protected]
123
World J Microbiol Biotechnol (2012) 28:2473–2478
DOI 10.1007/s11274-012-1054-x
are decarboxylated into the corresponding aldehydes which
are further reduced to the corresponding fusel alcohol by
the alcohol dehydrogenases, at the expense of NAD(P)H
(Dickinson et al. 2003; Hazelwood et al. 2008). The partial
genome of D. bruxellensis was recently published and a
number of orthologous and non-orthologous S. cerevisiae
genes were identified (Woolfit et al. 2007). However, no
studies have been reported on the identification and char-
acterization of decarboxylase-encoding genes. Thus, the
search for these genes in the D. bruxellensis genome may
help to better understand the contribution that this yeast to
the final sensory quality of the end product. In this study,
we provide an identification of two paralogs of phenyl-
pyruvate decarboxylase-encoding genes, which differs not
only in their nucleotides and proteins, but also in their
gene expression profile in response to the type of carbon
and nitrogen available in the medium. Since no other
hemiascomycete yeasts seem to carry two copies of this
gene, we start to question the biological and evolutionary
role of this enzyme activity in the yeast metabolism.
Materials and methods
Strains
Dekkera bruxellensis GDB 248 (Liberal et al. 2007) and
Saccharomyces cerevisiae JP1 (Silva-Filho et al. 2005)
were collected from a bioethanol production plant. The
type strain D. Bruxellensis CBS 74/CLIB 316 was provided
by the CLIB collection (INRA-Grignon, France).
Dekkera genome project
The genome of the D. bruxellensis wine strain CBS 2499
was described by Woolfit et al. (2007). The nucleotide
sequences were kindly provided by Prof. Jure Piskur (Lund
University, Sweden) and assembled in a local database at
the University of Campinas (http://www.lge.ibi.unicamp.
br/dekkera/), Brazil.
Computational analysis
Nucleotide sequences of the 2-oxo acid decarboxylases
encoding genes PDC1, PDC5, PDC6, THI3 and ARO10
were collected from the Saccharomyces Genome Data-
base—SGD (http://www.yeastgenome.org/) and used as a
means of searching for homologous genes in genetically
close ascomycetes by BLASTx analysis, using Genole-
vures (http://www.genolevures.org/), GenBank (http://
www.ncbi.nlm.nih.gov/) and Dekkera project (http://www.
lge.ibi.unicamp.br/dekkera/) databases. Sequences with
e-value lower than -20 were collected (Table S1—
supplementary data) from the yeasts D. bruxellensis (Db
genes), Kluyveromyces lactis (Kl genes), Candida glabrata
(Cg genes), Ashbya gossypii (Ag genes), Candida albicans
(Ca genes), Debaryomyces hansenii (Dh genes) and Yarr-
owia lipolytica (Yl genes), and their amino acid sequences
were obtained by theoretical translation using BioEdit
Sequence Alignment Editor Program v.7.0 (http://www.
mbio.ncsu.edu/BioEdit/bioedit.html).
The similarities and divergences among the sequences
(nucleotides and amino acids) were calculated by multiple
alignment analysis using the ClustalW tool of EMBL-EBI
database (http://www.ebi.ac.uk/) and used to construct
similarity matrices of protein sequences in the Bioedit pro-
gram (Table S2—supplementary data). Phylogenetic anal-
yses were performed in the Molecular Evolutionary Genetic
analysis—MEGA program (http://www.megasoftware.
net/). Gene clustering analysis was performed by the
Neighbor -Joining Method using p-distance. Confirmation of
nucleotide sequences were provided by DNA sequencing of
the entire genes (see primers at Table S3—supplementary
data) in ABI Prism 3100 device, using the Technological
Service Platform at the Osvaldo Cruz Foundation (Recife,
Brazil).
Culture conditions, RNA extraction and cDNA
synthesis
Yeast cells were pre-cultivated overnight at 30 �C in
synthetic defined (SD) medium (YNB at 1.7 g/L; ammo-
nium sulfate at 5 g/L; glucose at 20 g/L) with agitation
(150 rpm) and used for gene expression analysis. For long
term induction during cell growth, yeast cells were
transferred to YNB media containing different C and N
sources (Table 1) to initial concentration of 0.1 (A600nm).
The cell cultures were incubated at 30 �C in a rotator
shaker (200 rpm) and the cells were collected when they
reached a cell density of 1.0 (A600nm) at the exponential
growth phase in the same physiological state, for RNA
extraction. The growth time varied according to the car-
bon and nitrogen sources. For short-term gene induction,
cells from the pre-inoculum in synthetic defined medium
were recovered when reached concentration of 1.0
(A600nm). Five milliliters were recovered and centrifuged,
and the yeast cells were washed in water and re-sus-
pended to the same volume in YNB media with different
C and N sources (Table 1). After 1 h incubation at 30 �C
the cells were collected for RNA extraction. Each con-
dition was tested in duplicate and each sample was col-
lected in triplicate and immediately frozen in liquid
nitrogen for RNA extraction.
Yeast total RNA was isolated by using the NucleoSpin�
RNA II kit and following the manufacturer’s instructions
(Macherey–Nagel, Germany). RNA was quantified by
2474 World J Microbiol Biotechnol (2012) 28:2473–2478
123
means of the spectrophotometric method (Nanovue, GE
HealthCare) and its integrity evaluated by agarose gel 1 %
electrophoresis. For cDNA synthesis, 500 ng of total RNA
was used for each reverse transcription reaction tube
(40 ll) using ImProm-IITM Reverse Transcription System
Promega II kit and following the manufacturer’s instruc-
tions (Promega, USA).
Primer design for qPCR
The nucleotide sequence of the D. bruxellensis DbEFA1
(translation Elongation Factor, Alpha chain) and DbYNA1
(transcription factor of the nitrate assimilation regulon)
genes were recovered from the GenBank Nucleotide
database (accession numbers EF552481 and EF364427,
respectively) and the nucleotide sequences of the D. brux-
ellensis genes DbEFB1 and the ARO10 homologous were
recovered from D. bruxellensis database (http://www.lge.
ibi.unicamp.br/dekkera), as described above. Primer design
was performed by on-line Genscript primer design software
at advanced mode (http://www.genscript.com/cgi-bin/tools
) with the following parameters: primer length between 17
and 25 bases, Tm value around 59 �C and amplicon length
between 70 and 110 bp. Primer pairs were analyzed by
means of the on-line Netprimer tool (www.premierbiosoft.
com/netprimer/netprlaunch/netprlaunch.html) for self-hybrid,
duplex, hairpins and loops The primer pairs ranked over
90 % were selected for RT-qPCR experiments (Table S4—
supplementary data). Reference genes DbEFA1, DbEFB1
and DbYNA1 were used for data normalization.
Analysis of relative gene expression by RT-qPCR
Real-time PCR assays were optimized for maximum effi-
ciency via experimental design, using SYBR Green PCR
Master Mix (Applied Biosystems, USA). The chosen
parameters were SYBR Green (5 ll), primers 200 nM
(0.4 ll each), H2O (3.2 ll) and cDNA (1 ll). The tempera-
ture–time profile (95 �C for 10 min and 40 cycles at 95 �C
for 15 s and 60 �C for 1 min) was optimized for the ABI
Prism 7300 (Applied Biosystems). The amplification curves
were analyzed with the aid of SDS v.2.0 software (Applied
Biosystems). Negative PCR control (without template) and
negative RT control (RNA as template in Real Time PCR
assays, i.e. RNA not reverse-transcribed to cDNA) were run
for quality improvement and the detection of contamination
by genomic DNA. The relative gene expression values are
the average of three technical replicates, with biological
duplicates, totalizing 6 samples for each condition. The mean
Cq values for each growth medium were then plotted in
Microsoft Excel 2007 worksheets to create a suitable input
file for geNorm applet (Vandesompele et al. 2002). The first
stage in the geNorm analysis was to determine the stability of
each reference gene, which represents the gene expression
stability as the average pairwise variation for that gene with
all the other tested references (Vandesompele et al. 2002).
Enzyme assay
Dekkera bruxellensis GDB 248 was cultivated in YNB-
glucose containing either ammonia or phenylalanine to cell
concentration of 1 (A600nm) as above and harvested by cen-
trifugation. The cells were washed in 200 mM HEPES (pH
7.2) and suspended in the same buffer containing 1 mM DTT
and 1 mM PMSF. Glass beads (Ø460–600 lm) were added
to the equivalent of one volume of cell suspension and the
cells were broken with five cycles of vortexing for 1 min,
with 1 min of interval on ice to avoid protein denaturation.
Following centrifugation at 10,000g for 5 min at 4 �C, cell-
free extracts were collected for protein determination by
Lowry’s method. Enzyme specific activity for phenylpyru-
vate decarboxylase using phenylpyruvate (Sigma) as a sub-
strate was measured according to Vuralhan et al. (2003).
Results
In silico gene identification
Five 2-oxo acid decarboxylases gene of S. cerevisiae were
recovered from the SGD and its orthologous from
Table 1 Relative expression of DbARO10 genes of D. bruxellensisGDB 248 cells in different conditions
Test condition Relative gene expression (RQ)
C-sourcea N-sourceb dbARO10-1 dbARO10-2
During cell growth
Glucosec Ammoniumc 1 1
Glucose Phenylalanine 95.7 35.6
Sucrose Ammonium 4.68 2.69
Glycerol Ammonium 1.42 11.08
Ethanol Ammonium -4.96 23.59
Short-term incubation
Glucose Ammonium 1 1
Glucose Phenylalanine 923 -2.69
Glucose Leucine 5.24 20.32
Glucose Proline 2.17 2.13
Glucose Tryptophan 1 1
Glucose Valine 1 -16.66
Sucrose Ammonium 19 -4.78
Glycerol Ammonium 10.41 -5.18
Ethanol Ammonium 7 -2.02
a Carbon source: glucose, 20 g/L; sucrose, 19 g/L; ethanol, 19.5 mL/
L; glycerol, 16.3 mL/Lb Nitrogen source: (NH4)SO2, 5 g/L; phenylalanine, 3.1 g/L; trypto-
phan, 3.8 g/L; leucine, 6 g/L; valine, 5.4 g/L; proline, 5.2 g/L
World J Microbiol Biotechnol (2012) 28:2473–2478 2475
123
genolevures ascomycetes database and Genbank (Table
S1—supplementary material). Afterwards, the Dekkera
genome database was browsed using keywords that resul-
ted in the identification of two contigs (e-values\-20) that
displayed complete sequences from ATG to the stop
codons. Phylogenetic analysis placed the contigs 2774 and
3332 at the ARO10 gene cluster (Fig. 1), which encodes for
phenylpyruvate decarboxylase in S. cerevisiae (Vuralhan
et al. 2003). These were designated DbARO10-1 (Gene-
Bank accession number HQ693757) and DbARO10-2
(GeneBank accession number HQ693758), respectively.
Both ARO10 paralogous genes clustered with their ortho-
logs from Debaryomyces hansenii and Candida albicans
(Fig. 1), in accordance with the previous whole genome
clustering analysis (Woolfit et al. 2007). Genetic similarity
matrices were prepared from the corresponding amino acid
sequences for all the genes that were recovered and
revealed that the two isoforms of Aro10p were only 37.7 %
similar to each other (Table S2—supplementary material).
The DbAro10-1p contained 583 aa (c.a. 66 kDa) and
included two thiamine pyrophosphate pyruvate/indolpyru-
vate binding domains at N-terminal (aa 20–125) and
C-terminal (aa 415–505). The isoform Db-Aro10-2p con-
tained 593 aa (c.a. 67 kDa) and also included two thiamine
pyrophosphate (TPP) pyruvate(PDC)/indolpyruvate(IPDC)
binding domains at N-terminal (aa 40–140) and C-terminal
(aa 425–520). Both TPP/PDC domains were overlapping
two dimer interface domains that might be involved in
protein–protein interactions. These data showed that both
contigs represented two different loci in the yeast genome.
DbARO10 gene expression
The phylogenetic analysis revealed that both D. bruxell-
ensis decarboxylase-encoding genes were orthologous to
the S. cerevisiae ARO10 gene, which is involved in phen-
ylalanine catabolism and regulated by this amino acid
(Vuralhan et al. 2003). Thus, gene expression of DbARO10
genes was determined at the exponential growth phase in
synthetic medium containing phenylalanine as the nitrogen
source. During the cell growth, both DbARO10-1 and
DbARO10-2 were induced by phenylalanine both in type
strain CBS 74 and in the industrial strain GDB 248
(Fig. 2). Overall, DbARO10-1 was more induced than
DbARO10-2 and both genes was four times higher in the
industrial strain GDB 248 than in the type strain CBS 74
and that induction of DbARO10-1 was overall 2.6 times
higher than DbARO10-2 in the exponential growth phase of
both strains (Fig. 2). Afterwards, relative gene expression
was measured only in the industrial strain since the type
Fig. 1 Phylogenetic analysis of
the phenylpyruvate
decarboxylase encoding genes
of D. bruxellensis and their
positioning within the
hemiascomycetes group.
Nucleotide sequences were used
to prepare the p-distance matrix
and clustered by the Neighbor-
Joining method
2476 World J Microbiol Biotechnol (2012) 28:2473–2478
123
strain did not grow, or grew badly, in most of the different
conditions tested. Since that strain was isolated from the
ethanol fermentation using sugar cane, gene expression
was measured upon growth on sucrose. Interestingly, both
genes were induced in that carbon source, even in the
presence of ammonia, and again DbARO10-1 was more
induced than DbARO10-2 (Table 1). Gene expression was
also measured when yeast cells were grown under fully
respiratory metabolism, using glycerol or ethanol as carbon
source. In this condition only DbARO10-2 gene was
induced, while DbARO10-1 displayed a five-fold repres-
sion in ethanol relative to glucose (Table 1). Despite that
huge increase in gene transcript, the overall specific
activity of phenylpyruvate decarboxylase in the yeast cell-
free extract increased only 2.7 times in phenylalanine
compared to ammonia, which was comparable to the
increase in S. cerevisiae (Table 2).
For short-term response, the results showed that
DbARO10-1 was highly induced by immediate exposure to
phenylalanine, leucine and proline as nitrogen source as
well as in the presence of sucrose, glycerol and ethanol as
carbon source (Table 1). On the other hand, DbARO10-2
gene was induced only in the presence of leucine and
proline, and repressed in most of incubation condition
tested in those conditions (Table 1), including in the
presence of p-cumaric acid (-12.5 times). Interestingly,
phenylalanine first repressed DBARO10-2, while trypto-
phan had no effect on short-term expression of both genes
(Table 1).
Discussion
In the present study, we identified two different genes that
encode for phenylpyruvate decarboxylases (Phe-Pdc) in the
yeast D. bruxellensis, while only one gene exists in the
genomes of S. cerevisiae (Vuralhan et al. 2003) and other
phylogenetically closed hemiascomycetes (Fig. 1). In
general terms, the phylogeny of ARO10 gene products
reflected the whole phylogenetic analysis performed by
Woolfit et al. (2007) for the same yeast species, and
corroborated the topology #2 showed in that work. In
addition, the level of protein similarity of 37.7 % for both
isoforms was close to the calculated average similarity of
35.2 % for the hemiascomycetes Aro10p orthologous
analyzed. Thus, it indicates that paralogous DbARO10
genes in D. bruxellensis genome evolved from a duplica-
tion of an ancestral gene, which attests to its importance for
the yeast metabolism. Why D. bruxellensis presents two
copies of this gene?
The Phe-Pdc activity can be attributed to the decar-
boxylation of phenylpyruvate to phenyl-acetaldehyde dur-
ing the catabolism of phenylalanine (Vuralhan et al. 2003,
2005). Phenylalanine is found in grape and barley wort and
can be converted to 2-ethyl-phenol by the yeast cells during
fermentation that is responsible for the floral bouquet of the
beverages (Hazelwood et al. 2008). The induction of
DbARO10-1 gene by phenylalanine (96-fold) was higher
than that observed for S. cerevisiae (30-fold) (Vuralhan
et al. 2005), although the two works used different
approaches. Phe-PDC activity was observed in our
S. cerevisiae industrial strain growing in ammonium, despite
its absence in CEN.PK113-7D laboratory strain (Vuralhan
et al. 2003). However, the similar enzyme activities in the
industrial (69 nmol min-1 mgPRT-1) (Table 2) and labo-
ratory (53 nmol min-1 mgPRT-1) (Vuralhan et al. 2003)
strains of S. cerevisiae grown in phenylalanine ensure the
Fig. 2 Expression of D. bruxellensis genes in the type strain CBS 74
(gray columns) and the industrial strain GDB 248 (black columns)
cultivated in synthetic medium with phenylalanine as a nitrogen
source, relative to the same genes expressed in ammonia-based
medium (reference expression level of 1.0)
Table 2 Specific enzyme activity of phenylpyruvate decarboxylase
in cell free extract of yeast cells grown with different nitrogen sources
Yeast Specific activity
(nmol min-1 mgPRT-1)
Fold increase
(Phe/Amm)
Ammonia Phenylalanine
D. bruxellensis GDB248 15.5 42.1 2.72
S. cerevisiae JP1 27.4 69.3 2.53
World J Microbiol Biotechnol (2012) 28:2473–2478 2477
123
reliability of the assay. The induction of both DbARO10
genes by leucine (Table 1) indicates that their enzymes
could also decarboxylase the aldehydes from branched-
chain amino acids, as was shown for S. cerevisiae Aro10p
(Vuralhan et al. 2005; Boer et al. 2007). Nevertheless, the
products of two paralogous genes seem to perform different
biological functions. DbARO10-1 gene is immediately
induced upon exposure to phenylalanine, and its expression
decrease almost 10 times as the cultivation continues while
DbARO10-2 is lately induced almost 95 times in this
N-source. Another remarkable result was observed when
calculating the ratio of gene expression between long-term
and short-term incubation in respiratory carbon sources: an
over 10-times decrease in gene transcript of DbARO10-1
and an over 50-times increase in gene transcript of
DbARO10-2. This indicates that DbARO10-2 is induced
lately during the cultivation under respiratory metabolism
and maybe in response to putative oxidative damages
produced during cell growth.
Other phenolic compounds such as p-coumaric acid are
also present in the must from red grapes and their metab-
olisms contribute to the flavor and taste of the wine (Benito
et al. 2009). The decarboxylation of p-coumaric acid
produces 4-vinyl-phenol in a reaction that is dependent on
the hydroxycinnamate (p-coumarate) decarboxylase activ-
ity (Benito et al. 2009). Moreover, other decarboxylases
should are expected to act on 3-hydroxyphenylpyruvate
(from tyrosine), 3-indole pyruvate (from tryptophan) and
2-oxo-c-methylthio-butyrate (from methionine). The gene
expression results showed that any of the DbARO10 genes
responded to two of these substrates (tryptophan and
p-cumaric acid). Thus, alternative decarboxylases should
be present in the proteome of D. bruxellensis that were not
yet identified.
In conclusion, the results showed an unequivocal exis-
tence of two paralogous genes that encode for phenyl-
pyruvate decarboxylase enzymes in the genome of
D. bruxellensis and that they are involved in different
biological mechanisms in this yeast. A detailed study on
the function of each one is necessary to answer the question
on why this yeast needs to copies of this gene.
Acknowledgments This work was supported by grants from the
Brazilian funding agencies CNPq (Conselho Nacional de Desen-
volvimento Cientıfico e Tecnologico) though the UFPE-UNICAMP
cooperation program ‘‘casadinho’’, CAPES (Coordenacao de Apoio
ao Pessoal de Ensino Superior) and FACEPE (Fundacao de Apoio a
Ciencia e tecnologia do Estado de Pernambuco).
References
Barata A, Pagliara D, Piccininno T, Tarantino F, Ciardulli W,
Malfeito-Ferreira M, Loureiro V (2008) The effect of sugar
concentration and temperature on growth and volatile phenol
productions by Dekkera bruxellensis in wine. FEMS Yeast Res
8:1097–1102
Benito S, Palomero F, Morata A, Calderon F, Suarez-Lepe JA (2009)
Factors affection the hydroxycinnamate decarboxylase/vinylphe-
nol reductase activity of Dekkera/Brettanomyces: application for
Dekkera/Brettanomyces control in red wine making. J Food Sci
74:M15–M22
Blomqvist J, Eberhard T, Schnurer J, Passoth V (2010) Fermentation
characteristics of Dekkera bruxellensis strains. Appl Microbiol
Biotechnol 87:1487–1497
Boer VM, Tai SL, Vuralhan Z, Arifin Y, Walsh MC, Piper MD, de
Winde JH, Pronk JT, Daran JM (2007) Transcriptional responses
of Saccharomyces cerevisiae to preferred and nonpreferred
nitrogen sources in glucose-limited chemostat cultures. FEMS
Yeast Res 7:604–620
Chatonnet P, Dubourdieu D, Boidron JN (1995) The influence of
Brettanomyces/Dekkera spp. yeasts and lactic acid bacteria on the
ethylphenol content of red wines. Am J Enol Vitic 46:463–468
Dickinson JR, Salgado LE, Hewlins MJ (2003) The catabolism of
amino acids to long chain and complex alcohols in Saccharo-myces cerevisiae. J Biol Chem 278:8028–8034
Hazelwood LA, Daran J-M, van Maris AJA, Pronk JT, Dickinson JR
(2008) The Ehrlich pathway for fusel alcohol production: a
century of research on Saccharomyces cerevisiae metabolism.
App Environ Microbiol 74:2259–2266
Liberal ATS, Basılio ACM, Monte Resende A, Brasileiro BTV, Silva-
Filho EA, Morais JOF, Simoes DA, Morais MA Jr (2007)
Identification of Dekkera bruxellensis as a major contaminant
yeast in continuous fuel ethanol fermentation. J Appl Microbiol
102:538–547
Loureiro V, Malfeito-Ferreira M (2003) Spoilage yeasts in the wine
industry. Int J Food Microbiol 86:23–50
Silva-Filho EA, Melo HF, Antunes DF, Dos Santos SKB, Resende
MA, Simoes DA, Morais MA Jr (2005) Isolation by genetics and
physiological characteristics of a fuel ethanol fermentative
Saccharomyces cerevisiae strain with potential for genetic
manipulation. J Ind Microbiol Biotechnol 32:481–486
Vandesompele J, De Preter K, Pattyn F, Poppe B, Van Roy N, De
Paepe A, Speleman F (2002) Accurate normalization of real-time
quantitative RT-PCR data by geometric averaging of multiple
internal control genes. Genome Biol 3:00341
Vuralhan Z, Morias MA, Tai SL, Piper MDW, Pronk JT (2003)
Identification and characterization of phenylpyruvate decarbox-
ylase genes in Saccharomyces cerevisiae. Appl Environm
Microbiol 69:4534–4541
Vuralhan Z, Luttik MAH, Tai SL, Boer VM, Morais MA, Schipper D,
Almering JH, Kotter P, Dickinson JR, Daran JM, Pronk JT (2005)
Physiological characterization of the ARO10-dependent, broad-
substrate—specificity 2-oxo acid decarboxylase activity of Sac-charomyces cerevisiae. Appl Environ Microbiol 71:3276–3284
Woolfit M, Rozpedowska E, Piskur J, Wolfe KH (2007) Genome
survey sequencing of the wine spoilage yeast Dekkera (Bretta-
nomyces) bruxellensis. Eukaryot Cell 6:721–733
2478 World J Microbiol Biotechnol (2012) 28:2473–2478
123