Planta (2014) 239:1187–1200
DOI 10.1007/s00425-014-2041-2
Original Article
Selection of candidate reference genes for real‑time PCR studies
in lettuce under abiotic stresses
Joyce Moura Borowski · Vanessa Galli · Rafael da Silva Messias ·
Ellen Cristina Perin · Julieti Hugh Buss · Sérgio Delmar dos Anjos e Silva ·
Cesar Valmor Rombaldi 
Received: 25 August 2013 / Accepted: 31 January 2014 / Published online: 27 February 2014
© Springer-Verlag Berlin Heidelberg 2014
Abstract The process of selection and validation of ref-
erence genes is the first step in studies of gene expression
by real-time quantitative polymerase chain reaction (RT-
qPCR). The genome of lettuce, the most popular leaf veg-
etable cultivated worldwide, has recently been sequenced;
therefore, suitable reference genes for reliable results in
RT-qPCR analyses are required. In the present study, 17
candidate reference genes were selected, and their expres-
sion stability in lettuce leaves under drought, salt, heavy
metal, and UV-C irradiation conditions and under the appli-
cation of abscisic acid (ABA) was evaluated using geNorm
and NormFinder software. The candidate reference genes
included protein-coding traditional and novel reference
genes and microRNAs (miRNAs). The results indicate that
the expression stability is dependent on the experimental
conditions. The novel protein-coding reference genes were
more suitable than the traditional reference genes under
Electronic supplementary material The online version of this
article (doi:10.1007/s00425-014-2041-2) contains supplementary
material, which is available to authorized users.
J. M. Borowski · V. Galli · R. da Silva Messias · E. C. Perin ·
J. H. Buss · S. D. dos Anjos e Silva  18S 18S ribosomal RNA
Embrapa Clima Temperado, Rodovia BR 396, Km 78, Caixa
Postal 403, Pelotas, RS CEP 96001‑970, Brazil
J. M. Borowski · R. da Silva Messias · E. C. Perin · J. H. Buss ·
C. V. Rombaldi (*)  APT1 Adenosine phosphoribosyl transferase
Faculdade de Agronomia Eliseu Maciel, Campus Universitário
s/n, Universidade Federal de Pelotas, Caixa Postal 354, Pelotas,
RS CEP 96010‑900, Brazil
e-mail: cesarvrf@ufpel.edu.br
V. Galli 
Universidade Federal do Rio Grande do Sul, Centro de
Biotecnologia, Av. Bento Gonçalves 9500, Campus do Vale,
Caixa Postal 15005, Porto Alegre, RS CEP 91501‑970, Brazil
drought, UV-C irradiation, and heavy metal conditions and
under the application of ABA. Only under salinity condi-
tions were the traditional protein-coding reference genes
more stable than the novel genes. In addition, the miRNAs,
mainly MIR169, MIR171/170 and MIR172, were stably
expressed under the abiotic stresses evaluated, represent-
ing a suitable alternative approach for gene expression data
normalization. The expression of phenylalanine ammo-
nia lyase (PAL) and 4-hydroxyphenylpyruvate dioxyge-
nase (HPPD) was used to further confirm the validated pro-
tein-coding reference genes, and the expression of MIR172
and MIR398 was used to confirm the validated miRNA
genes, showing that the use of an inappropriate reference
gene induces erroneous results. This work is the first survey
of the stability of reference genes in lettuce and provides
guidelines to obtain more accurate RT-qPCR results in let-
tuce studies.
Keywords ABA · Abiotic stresses (drought, salt, heavy
metal and UV-C stress) · GeNorm · Lactuca · miRNAs ·
NormFinder · Reference genes
Abbreviations
40S 40S ribosomal RNA
ABA Absicic acid
ACT Actin
EIF2A Elongation initiation factor gamma subunit
EIF4A1 Eukaryotic translation initiation factor 4A1
EXP45 Expansin 5 precursor
microRNA miRNA
HPPD 4-Hydroxyphenylpyruvate dioxygenase
PAL Phenylalanine ammonia-lyase
PP2A Serine/threonine protein phosphatase 2A
13

1188
RT-qPCR Real-time quantitative polymerase chain
reaction
TIP41 TAP42-interacting protein of 41 kDa
TUB Tubulin
UBC21 Ubiquitin-protein ligase
UBQ Ubiquitin
Introduction
Lettuce (Lactuca sativa L.) belongs to the Asteraceae fam-
ily and is the most popular leaf vegetable cultivated and
consumed throughout the year (Zdravković et al. 2014).
This vegetable is increasingly accepted by consumers
because it is easy to prepare for consumption (Borghi 2003)
and contains health-promoting compounds such as vita-
mins C and E, fibers and polyphenols (Nicolle et al. 2004;
Llorach et al. 2008). These health-promoting compounds
help prevent the incidence of many chronic diseases (Lee
and Aedin 2006); therefore, efforts involving the bioforti-
fication of this crop are of great interest (Hawrylak-Nowak
2013). Most of these compounds are secondary metabolites
synthesized during normal plant growth or in response to
environmental stresses (Llorach et al. 2008). Drought, heat,
salinity, UV radiation and hormones induce numerous
physiological stress reactions in plants that alter the chemi-
cal composition of crops and the quality of the harvested
products (Wang and Frei 2011). Therefore, understanding
the mechanisms involved in the lettuce response to differ-
ent stresses and their effect on the content and composition
of health-promoting compounds is of great importance.
The study of the expression of genes involved in the
stress response has provided useful information that has
been used to define biofortification approaches in plants
(Qin et al. 2011; Messias et al. 2013). To date, studies based
on gene expression in lettuce are scarce in the literature;
however, the recent genome sequencing of the Composi-
tae family (Lai et al. 2012) and its publication will increase
research in this area. Real-time quantitative polymerase
chain reaction (RT-qPCR) is a sensitive, specific and repro-
ducible technique that has been widely used to analyze the
expression of genes in different organisms, tissues and dif-
ferent conditions (Bustin 2002; Derveaux et al. 2010). How-
ever, the reliable normalization of RT-qPCR data depends
on the expression of reference genes. The analyses are
affected by several experimental conditions, including vari-
ations in the quality, stability and input of RNA as well as
variations in the efficiency of cDNA synthesis and PCR
(Derveaux et al. 2010). Therefore, it is crucial to amplify
a reference gene alongside the target gene to calibrate for
experimental variability (Chen et al. 2011). A suitable ref-
erence gene should be expressed at a constant level among
samples, and its expression is assumed to be unaffected by
13
Planta (2014) 239:1187–1200
the experimental conditions (Bustin 2002). The use of inap-
propriate reference genes radically changes the interpre-
tation of the expression pattern of a given target gene and
introduces flaws in our understanding of the role of the
gene. Therefore, the lack of validation of reference genes
in plants compromises the accuracy of results obtained by
RT-qPCR using non-valid references (Gutierrez et al. 2008).
Unlike the study of gene expression in humans (Malt-
seva et al. 2013), the number of plant reference genes still
remains limited, and their validation is controversial. The
most traditional reference genes used in RT-qPCR studies
with plants include actin (ACT) (Maroufi et al. 2010), tubu-
lin (TUB) (Wan et al. 2010), ubiquitin (UBI) (Chen et al.
2011), 18S ribosomal RNA (18S) (Jain et al. 2006) and 40S
ribosomal RNA (40S) (Cruz et al. 2009). These genes were
adopted from traditional techniques such as Northern blot-
ting with the assumption of highly uniform expression in
different cell types under different experimental conditions
(Jain et al. 2006). Nevertheless, studies have demonstrated
that traditional reference genes vary under several condi-
tions (Gutierrez et al. 2008); therefore, a search for novel
and suitable reference genes based on large-scale transcrip-
tomic approaches has been successfully performed (Libault
et al. 2008). Furthermore, the discovery of new gene prod-
ucts such as small non-coding RNAs, especially microR-
NAs (miRNAs), is also contributing to the selection of suit-
able reference genes. The potential of miRNAs as reference
genes to normalize gene expression in human tissues was
previously demonstrated (Peltier and Latham 2008). In plant
tissues, the first investigation into the suitability of miRNAs
as reference genes was described by Kulcheski et al. (2010).
However, currently, no studies have reported the stability of
putative reference genes in lettuce; therefore, the selection
of reference genes in this crop will benefit gene expression
studies by RT-qPCR and will be useful in studies regarding
biofortification through metabolic engineering.
Specialized software programs such as geNorm (Van-
desompele et al. 2002) and NormFinder (Andersen et al.
2004) exist to determine the stability of candidate reference
genes. In addition to these software programs, the stabil-
ity of a suitable reference gene should be evaluated with
target genes associated with the experimental conditions to
obtain reliable results. Phenolic compounds, derivatives of
phenylpropanoids pathways, and vitamin E are important
antioxidants in lettuce, and they are activated in response to
abiotic stresses (Yuan et al. 2010; Ren et al. 2011). Pheny-
lalanine ammonia-lyase (PAL) and 4-hydroxyphenylpyru-
vate dioxygenase (HPPD) are the main enzymes involved
in phenylpropanoid derivative (Xu et al. 2012) and vitamin
E synthesis (Ren et al. 2011), respectively; therefore, the
genes coding for these enzymes are interesting target genes
that may be able to confirm the reliability of the reference
genes under abiotic stress conditions.
Planta (2014) 239:1187–1200 1189
In the present study, we evaluated the stability of com-
monly used reference genes and novel candidate refer-
ence genes, including protein-coding genes and miRNAs
(Table 1), to identify potential reference genes suitable for
transcript normalization in experiments with lettuce under
different abiotic stresses. In addition, the expression of
two target genes related to the treatments evaluated, PAL
and HPPD, was investigated to demonstrate the effi-
cacy of the selected protein-coding reference genes, and
MIR172 and MIR398 were used to validate the miRNA
genes.
Materials and methods
Experimental conditions and plant materials
Lettuce seedlings (cv. Vera; seeds obtained from Sakata,
Gragança Paulista, SP, Brazil) were transplanted to 9-L
pots containing a mixture of soil and vermiculite as the
substrate (3:1). The required amount of fertilizer was cal-
culated and applied according to the technical recommen-
dations for the crop (CQFS/RS-SC 2004). Irrigation was
performed by water blade to maintain a constant soil mois-
ture content (20 ± 5 %) as measured by a Hydrofarm elec-
tronic meter (Falker™). Plants were watered according to
the evapotranspiration value of the culture as proposed by
Marouelli et al. (2008). The experiment was conducted in
a greenhouse in a completely randomized design, which
consisted of seven treatments [non-stressed plants (con-
trol), five abiotic stress treatments and a treatment with
exogenous application of abscisic acid (ABA)] with four
biological replications. For the drought stress treatment,
the plants were maintained while withholding irriga-
tion for 10 days. For the drought-recovering treatment,
the plants were submitted to drought stress (for 10 days)
followed by rewatering for 3 days according to the crop
evapotranspiration value for lettuce (Marouelli et al. 2008)
to verify the recovery ability after drought stress. For salt
stress treatment, the plants were submitted to the applica-
tion of 200 mM NaCl in the soil according to Mahmoudi
et al. (2012). For UV-C stress treatment, the plants were
irradiated for 15 min with a UV-C lamp with an inten-
sity of 0.29 mW cm−2, which corresponds to a hermetic
dose of 4.2 kj m−2 (Tiecher et al. 2013). For heavy metal
stress treatment, the plants were sprayed with 0.7 g mL−1
arsenic in the sodium arsenate form. For ABA treatment,
the plants were sprayed with ABA (Sigma®) at 300 ppm
as described by Lai et al. (2012). The ABA treatment was
included because this compound has been identified as a
messenger in stress–perception–response pathways (Wu
et al. 1997) and was used as a stress indicator in related
literature (Rabbani et al. 2003).
All the treatment applications were performed three times
during the last three weeks before the harvest (one applica-
tion per week). The leaf samples were collected in 2 h after
the last treatment application and were immediately frozen
in liquid nitrogen and stored at −80 °C until RNA isolation.
Total RNA isolation and first strand cDNA synthesis
Total RNA of lettuce leaves was isolated using the CTAB
(hexadecyltrimethylammonium bromide) protocol (Chang
et al. 1993). Three replicates of RNA isolation were con-
ducted for each biological replicate. RNA concentration and
RNA quality were evaluated by spectrometry using A260/A280
and A260/A230 ratios in a spectrophotometer Nanovue 4282
(GE Healthcare). Total RNA (1 μg) was digested with 1 U
DNase I and DNase 1 × reaction buffer (Invitrogen) before
cDNA synthesis. For the amplification of protein-coding tra-
ditional and novel candidate reference genes, all the digested
RNA was reverse transcribed using the M-MLV enzyme and
oligo-dT primers according to the manufacturer’s instructions
(Invitrogen). For the amplification of miRNAs, cDNA syn-
thesis was conducted using the multiplex technique according
to Kulcheski et al. (2010). Briefly, each RNA digested with
DNAse was primed with a pool of 10 mM of each stem–loop
primer and 1 μL oligonucleotide dT24V (Invitrogen) and
then reverse transcribed using the M-MLV enzyme according
to manufacturer’s instructions (Invitrogen).
Primer design and RT‑qPCR conditions
Seventeen candidate reference genes, including five pro-
tein-coding traditional reference genes (ACT, TUB, UBQ,
18S and 40S) that have been reported to be good poten-
tial candidates in previously published studies (Jain et al.
2006), seven protein-coding novel reference genes (adeno-
sine phosphoribosyl transferase—APT1, elongation initia-
tion factor gamma subunit—EIF2A, eukaryotic translation
initiation factor 4A1—EIF4AI, expansin 5 precursor—
EXP45, serine/threonine protein phosphatase 2A—PP2A,
TAP42-interacting protein of 41 kDa—TIP41 and ubiqui-
tin-protein ligase—UBC21) and five miRNAs (MIR156,
MIR169, MIR171/170, MIR172 and MIR398) were evalu-
ated. The sequences of lettuce protein-coding traditional
candidate reference genes and genes coding for HPPD and
PAL were extracted from the GenBank database. The novel
protein-coding reference genes were identified by screen-
ing the lettuce sequences from the Compositae Genome
Project Database (CGPDB 2013; http://cgpdb.ucdavis.edu/)
using the Arabidopsis homologs as the query because these
genes have been shown to be stably expressed in Arabidop-
sis (Czechowski et al. 2005) and in lettuce (Argyris et al.
2008; Soós et al. 2009) under different experimental condi-
tions. Specific primers were designed using Vector NTI10
13

1190 Planta (2014) 239:1187–1200

Table 1  Candidate reference gene description and parameters derived from RT-qPCR analysis
Gene Acession Primer sequence (5′–3′) (foward/reverse) Product Amplicon RT-qPCR R2c
symbol numbera Tm (°C)b lenght efficiencyc

ACT AY260165.1 AGGGCAGTGTTTCCTAGTATTGTTG/CTCTTTTGGATTGTG 82.71 106 2.04 0.999

CCTCATCT
APT1 AT1G27450 CTGTACAAGAAGGAGAACGAGC/ACGAGCACATACAGTG 86.35 184 1.96 0.998
GCTT
EIF2A EU028334 TAGGCGAGTGGAGAAGCATT/GTAGAAACAGCAACAGGC 83.80 71 1.95 0.991
AAA
EIF4A1 AT3G13920 CTTCATAGGATTGGGCGAAG/TATGAGATCCGCAACATTCG 84.32 150 1.87 0.998
EXP45 EU103628 TGGAACCAACAAAGCCATTC/ACTCTCCCAAGAGCAAG 83.88 215 1.98 0.999
ACG
PP2A AT1G10430 TTGACGGAATCGGAGGTAAAA/CCGGCTGCACATTCCATT 82.51 76 1.93 0.995
TIP41 LACT_SATI. GAGAGATTTGCTGGAGGGAAACTA/CCTTTGACTGATGAT 80.90 101 1.98 0.999
CST1.6123 GTTTGGA
TUB AT5G62690 TAGGCGTGTGAGTGAGCAGT/AACCCTCGTACTCTGCCT 84.80 209 1.96 0.995
CTT
UBC21 AT5G25760 TCTTAGATCACCGTCCCATCGT/TCTGAGATTGTCCGAGG 85.09 89 2.19 0.997
ATATGAG
UBQ GW397659 AAGACCTACACCAAGCCCAA/AAGTGAGCCCACACTTA 86.70 197 1.99 0.993
CGA
18S AT3G41768 GCATGAGTGGTGTTTGGTTTGT/GAGGTAAGAACGCATGG 82.99 74 2.05 0.999
AAGTTG
40S HS586765 CAAGATTCGGTGACAGGGATG/CACCACCTCCAAATCCA 89.60 137 1.97 0.998
CCA
MIR156 MIRNA TGACAGAAGAGAGTGAGCAC/GTCGTATCCAGTGCAGGGT 82.35 72 2.05 0.999
CCGAGGTATTCGCACTGGATACGACGTGCTC
MIR169 MIRNA CAGCCAAGGATGACTTGCCGG/GTCGTATCCAGTGCAGGG 84.48 71 1.95 0.997
TCCGAGGTATTCGCACTGGATACGACCCGGCT
MIR171/170 MIRNA TGATTGAGCCGAGCCAATATC/GTCGTATCCAGTGCAGGGT 82.59 71 1.95 0.999
CCGAGGTATTCGCACTGGATACGACGATATT
MIR172 MIRNA AGAATCTTGATGATGCTGCAT/GTCGTATCCAGTGCAGGGT 81.87 71 1.99 0.995
CCGAGGTATTCGCACTGGATACGACATGCAG
MIR398 MIRNA TGTGTTCTCAGGTCACCCCTT/GTCGTATCCAGTGCAGGGT 82.71 71 1.99 0.998
CCGAGGTATTCGCACTGGATACGACAAGGGG
PAL AF299330.1 GAGAACTAAGCAAGGCGGT/CGCTTACAGTTTCTCAGG 87.13 200 2.03 0.998
TGG
HPPD FJ194493.1 CCGGCGCCTTCGTTGTGTTCCAGATAC/GCCCGGGTTTGA 85.6 309 1.90 0.999
ACCAGTTGAAAAG
a
 GenBank database or compositae genome project database (http://cgpdb.ucdavis.edu/)
b
 The melting temperature was calculated by SDS version 1.1 software in a 7500 Real time Fast thermocycler (applied biosystems)
c
 The RT-qPCR efficiency and correlation coefficients (R2) were determined by LinReg PCR program

software (Invitrogen), according to the following param-
eters: melting temperature (Tm) of 58–62 °C and GC con-
tent of 45–55 %. The sequences of lettuce miRNAs were
obtained from Han et al. (2010). The stem–loop primers
used for miRNA cDNA synthesis were designed according
to Chen et al. (2005). The stem–loop sequence consists of
44 conserved and 6 variable nucleotides that are specific to
the 3′ end of the miRNA sequence. All primer sequences
and relevant information regarding the genes are presented
in Table 1. The specificity of the amplicons was verified for
the presence of primer dimers or non-specific amplicons
by the presence of a single peak in RT-qPCR melting curve
products and a single band with the expected size in the
3 % agarose gel after electrophoresis.
The cDNAs at a concentration of 100 ng μL−1 were
amplified by RT-qPCR in a final volume of 20 μL con-
taining 1 μL cDNA, 10 μL of Platinum Sybr green UDG
(Invitrogen), and 3–5 ρmol of each primer. Amplification
was standardized in a 7500 Real time Fast thermocycler
(applied biosystems) using the following conditions: 50 °C
for 20 s, 95 °C for 10 min followed by 40 cycles of 15 s at
95 °C and 60 s at 60 °C. The PCR products for each primer

13
Planta (2014) 239:1187–1200 1191
set were subjected to melting curve analysis to verify the
presence of primer dimers or non-specific amplicons. The
melting curve analysis ranged from 60 to 95 °C with 1 %
stepwise temperature increases. No-template controls were
included to ensure that no reagent or genomic DNA con-
tamination occurred.
Reference gene expression stability determination
and statistical analysis
To estimate the expression stability of the seventeen can-
didates reference genes, all amplification plots were ana-
lyzed with a threshold fluorescence value of 0.1 to obtain
the amplification cycle (Cq) values using the SDS version
1.1 software (applied biosystems). The PCR efficiency was
estimated using the LinReg PCR program (Ramakers et al.
2003). The raw Cq data were processed via a linear scale
using the ΔCq method, and the expression stability was
evaluated using the geNorm (Vandesompele et al. 2002)
and NormFinder (Andersen et al. 2004) software packages
for Microsoft Excel. RT-qPCR data were exported into an
Excel datasheet (Microsoft Excel 2010), and the Ct values
were converted according to the requirements of the soft-
ware. Each of these approaches generates a measure of ref-
erence gene stability, which can be used to rank the order
of stability for reference genes.
The geNorm software calculates the average of the pair-
wise variation for a candidate reference gene with all other
genes tested; the M value is used to express the result (Van-
desompele et al. 2002). Genes with high variability result
in a high M value, which indicates low expression stabil-
ity, and vice versa, using a cut-off of 1.5. The geNorm soft-
ware also calculates a normalization factor for each sample
and suggests the optimal number of reference genes nec-
essary to normalize the experiment. NormFinder software
uses an ANOVA-based model to consider intra- and inter-
group variation of the candidate reference genes to evaluate
expression stability and provide a direct measure of varia-
tion (Andersen et al. 2004).
Validation of reference gene analysis
To confirm the reliability of the potential reference genes,
the relative expression profiles of PAL and HPPD were
measured and normalized with the most stable and least
stable protein-coding reference genes according to Nor-
mFinder in samples of lettuce leaves under the application
of ABA (for PAL) and drought stress (for HPPD). Simi-
larly, the relative expression of MIR172 and MIR398 was
measured and normalized with the most stable and least
stable miRNA reference genes, as determined by Nor-
mFinder in samples of lettuce leaves under salt stress. The
RT-qPCR amplification conditions were the same as those
described above. The relative expression data were calcu-
lated according to the 2−∆∆Ct method and were presented
as fold changes (Livak and Schmittgen 2001). Samples
from non-stressed plants were used as reference samples.
Statistical analyses were performed using the computer
program SAS System for Windows version 9.1.3 (SAS,
2000). Data were subjected to variance analysis (P ≤ 0.05).
In case of statistical significance, we compared the relative
quantitation results from the more stable and more unstable
reference genes by t test (P ≤ 0.05).
Results and discussion
Description and expression profiling of candidate reference
genes
RT-qPCR has become the most popular method for the
quantification of transcript accumulation and must be pre-
ceded by validated normalization to prevent non-biological
event variations. Doubtful results of gene expression arise
when target genes are normalized against single reference
genes without adequate justification and when non-val-
idated reference genes are used (Bustin et al. 2009; Der-
veaux et al. 2010). Normalization with validated reference
genes controls variations in RNA extraction yield, reverse
transcription yield and amplification efficiency, allow-
ing comparisons of mRNA concentrations across different
samples (Bustin et al. 2009). This variation requires that
the most suitable reference gene be chosen to normalize the
results for each species and under each experimental con-
dition. Therefore, a group of traditional and novel protein-
coding reference genes as well as miRNAs were tested to
select the most reliable genes to use in RT-qPCR analysis
in experiments with lettuce under abiotic stresses.
Robust and precise results of gene expression are usu-
ally correlated with high RT-qPCR efficiency (Bustin et al.
2009). In the present study, the efficiency of RT-qPCR
varied from 1.87 (EIF4A1) to 2.19 (UBC21) and was cal-
culated as the mean value obtained from the technical and
biological replicates. The Tm for all PCR products ranged
from 80.90 °C (TIP41) to 89.60 °C (40S) (Table 1). The
specificity of the candidate reference genes and target
genes was confirmed by the existence of a single peak in
the melting curves (Online Resource 1) and by a single
band with the expected size in 3 % agarose gel electropho-
resis (Online Resource 2).
The expression levels of the candidate reference genes
were determined as quantification cycle (Cq) values.
According to this approach, the seventeen tested genes
showed variation in the expression levels caused by differ-
ent stresses applied and by biological variations (Fig. 1).
MIR156 was the most expressed gene with the lowest mean
13

1192
Fig. 1  Expression levels of
different candidate reference
genes. Expression data dis-
played as RT-qPCR quantifica-
tion cycle (Cq) values for each
reference gene in lettuce leaves
submitted to abiotic stresses.
The line across the box depicts
the median. The box indicates
the 25th and 75th percentiles,
and whisker caps represent
the maximum and minimum
values. The higher the boxes
and whisker, the greater the
variations
Cq value (14.86), and EXP45 was the least expressed gene
with the highest mean Cq value (30.32) under the evalu-
ated abiotic stress conditions (Fig. 1). Among the seventeen
candidate reference genes tested in this study, TUB showed
the most variation in expression levels when both biologi-
cal variation and the variation among treatments were con-
sidered. This variation is shown in Fig. 1, where the boxes
and whisker taps are larger for TUB compared to those for
the other genes; the lowest Cq value is for the heavy metal
treatment (21.64), and the highest Cq value is for drought
stress treatment (27.56), representing almost six cycles of
difference in Cq values. In contrast, MIR172 and MIR398
showed narrow ranges in Cq values below two cycles
(Fig. 1).
Expression stability of all candidate reference genes
under abiotic stresses
The expression stability of the candidate reference genes
was determined by quantifying the mRNA levels by qRT-
PCR. For each gene, we calculated the Cq value, which
represents the cycle at which a significant increase of the
PCR product occurs. In general, the stability is marked by
the middle of the exponential phase of amplification (Bus-
tin 2000). GeNorm (Vandesompele et al. 2002) and Nor-
mFinder (Andersen et al. 2004) were used to evaluate the
data.
The geNorm program recommends using reference
genes with an M value below the threshold of 1.5 (Vande-
sompele et al. 2002). According to this criterion, the sev-
enteen candidate reference genes tested in this work were
considered to be rather stable, presenting M values lower
than 1.5 (Fig. 2). Among these candidates, the miRNA
genes were more stable than the protein-coding traditional
and novel reference genes. The M values of MIR171/170
and MIR169 were lower than those of the other miRNAs,
indicating that they are the most stable candidate reference
genes according to geNorm. The traditional protein-coding
13
Planta (2014) 239:1187–1200
reference gene TUB was the most unstable gene under abi-
otic stresses, with an M value over 1.4 (Fig. 2).
To confirm the geNorm results, we used NormFinder
software, which performs an ANOVA-based analysis to
consider intra- and inter-group variation of the candidate
reference genes to evaluate expression stability and pro-
vide a direct measure of variation (Andersen et al. 2004).
According to this approach, the candidate reference genes
were ranked based on their stability value, with a low
standard error (Table 2). The novel protein-coding TIP41,
UBC21 and EIF2A reference genes showed the best stabil-
ity, followed by MIR172, MIR169 and MIR398 (Table 2).
TUB was the least stable gene (Table 2), as was previously
observed in the geNorm analysis (Fig. 2). The observed dif-
ferences between the geNorm and NormFinder results were
expected because the two software programs are based on
distinct statistical algorithms. Despite the divergent results
from the geNorm and NormFinder rankings, the results
of the present study indicate that lettuce miRNAs are sta-
bly expressed and are suitable for normalizing RT-qPCR
expression analysis, which is in agreement with the results
observed by Kulcheski et al. (2010). However, it is impor-
tant to consider that not all miRNAs are constitutively
expressed under stress conditions, and the universality of
these results needs to be demonstrated in other plants and
under other experimental conditions.
Expression stability of protein‑coding genes under abiotic
stresses
Because some studies may only evaluate the expression of
protein-coding sequences, we decided to perform stabil-
ity analysis using the geNorm and NormFinder software
programs using only the traditional and novel candidate
reference genes as input, without any miRNA reference
genes. As a result, the M values of all the protein-coding
genes were still lower than 1.5; however, the novel candi-
date genes APT1 (whose function involves the conversion
Planta (2014) 239:1187–1200 1193

Fig. 2  Average expression sta-

bility (M value) of the seventeen
candidate reference genes using
geNorm software. Expression
stability was evaluated in sam-
ples from lettuce submitted to
abiotic stresses. The most stable
reference genes were measured
during the stepwise exclusion of
the least stable reference genes.
A lower average expression
stability M value indicates more
stable expression

of adenine to adenosine monophosphate in the purine sal- Table 2  Ranking of candidate reference genes in order of their
vage pathway), EIF2A (which acts as an initiation factor expression stability calculated by NormFinder
in eukaryotes), and TIP41 (a TAP42-interacting protein Rank Candidate reference gene
of 41 kDa) were the three most stable genes (Fig. 3a). In Gene name Stability value Standard error
contrast, EXP45 was the most unstable reference gene,
followed by the traditional reference genes 18S, TUB and 1 TIP41 0.287 0.067
ACT. 2 UBC21 0.380 0.082
Considering the alternative of evaluating a single abiotic 3 EIF2A 0.403 0.085
stress, we also separately analyzed the expression stabil- 4 MIR172 0.407 0.086
ity of the twelve candidate protein-coding reference genes 5 MIR169 0.418 0.088
during drought, salinity, UV-C irradiation, heavy metal 6 MIR398 0.425 0.089
and ABA application conditions in separately. For this 7 APT1 0.431 0.090
approach, we considered drought and rewatering together 8 EIF4A1 0.461 0.095
because the related literature usually addresses drought 9 MIR171/170 0.476 0.098
stress and the ability of plants to recover after this condi- 10 MIR156 0.499 0.101
tion (Zou et al. 2010). According to geNorm, the most 11 UBQ 0.507 0.103
stable reference genes are not the same for all treatments 12 ACT 0.565 0.113
(Fig. 3), confirming that no single reference gene had con- 13 PP2A 0.570 0.114
stant expression in all the experimental conditions, thus 14 40S 0.583 0.116
justifying the selection of candidate reference genes for RT- 15 18S 0.809 0.157
qPCR studies of lettuce under abiotic stresses. 16 EXP45 0.811 0.157
Overall, the novel reference genes were better candi- 17 TUB 0.898 0.173
dates than the traditional genes regarding expression sta-
bility under drought, salt and UV-C stresses (Fig. 3b–d).
In contrast, traditional reference genes outperformed the genes 18S and ACT showed more stability than the novel
novel reference genes under heavy metal stress and the genes in the heavy metal treatment; however, 40S and UBQ
ABA application (Fig. 3e, f). When samples from drought showed more stability than the novel genes under ABA
stress were analyzed using geNorm, the results were application (Fig. 3e, f).
similar to the results found with samples from all abiotic In addition to the geNorm analysis, we evaluated the con-
stress conditions (Fig. 3b), and APT1, TIP41 and EIF4A1 sistency of candidate reference gene expression using Nor-
were the most stable genes. Under salinity conditions, the mFinder under individual stress conditions. The results of this
most stable genes used as reference genes in lettuce were analysis are summarized in Table 3, where the best candidate
EIF2A and TIP41 (Fig. 3c). When UV-C irradiation was is the one with the average expression value closest to zero.
applied, APT1 and TIP41 showed more stability than the Again, the geNorm and NormFinder results presented sev-
other genes (Fig. 3d). Interestingly, the traditional reference eral differences because these programs are based on distinct

13

1194
Fig. 3  Average expression stability (M value) of the twelve protein-
coding candidate reference genes using geNorm software. Expression
stability was evaluated in samples from lettuce submitted to abiotic
stress (a), drought (b), salinity (c), UV-C irradiation (d), heavy metal
statistical algorithms, as indicated above. Only under UV-C
stress did geNorm (Fig. 3d) and NormFinder (Table 3) indi-
cate the same reference genes, APT1 and TIP41, in their top
rankings. According to the NormFinder analysis (Table 3),
TIP41 and APT1 appeared among the three most suitable pro-
tein-coding genes for the validation of gene expression data
in lettuce submitted to drought conditions, UV-C stress, and
when all abiotic stress were considered in the analysis. APT1
was also the most stable gene under heavy metal conditions,
13
Planta (2014) 239:1187–1200
(e) and ABA (f). The most stable reference genes were measured dur-
ing stepwise exclusion of the least stable reference genes. A lower
average expression stability M value indicates more stable expression
and TIP41 was one of the most stable genes under ABA
applications, showing that under these conditions, the novel
reference genes proposed were better than the traditional
genes. The results of this work corroborate the results of Gut-
ierrez et al. (2008) and Wei et al. (2013), who determined that
APT1 is the optimal reference gene during Arabidopsis and
sesame (Sesamum indicum L.) development, respectively.
TIP41 has already been described as the most stable reference
gene in tomato development (Expósito-Rodríguez et al. 2008)
Planta (2014) 239:1187–1200 1195

Table 3  Ranking of candidate reference genes under different stress conditions in order of their expression stability calculated by NormFinder
Rank Abiotic stressesa Drought Salinity UV-C irradiation Heavy metal ABA

Protein-coding reference gene

1 TIP41 (0.259) EIF2A (0.117) TUB (0.068) TIP41 (0.137) APT1 (0.100) EIF4A1 (0.105)
2 EIF2A (0.318) TIP41 (0.188) UBQ (0.098) APT1 (0.159) PP2A (0.1430) TUB (0.115)
3 APT1 (0.343) APT1 (0.190) 40S (0.161) EIF4A1 (0.163) ACT (0.166) TIP41 (0.155)
4 UBQ (0.369) UBQ (0.300) UBC21 (0.190) ACT (0.172) EIF2A (0.196) UBC21 (0.289)
5 UBC21 (0.393) UBC21 (0.336) TIP41 (0.208) EIF2A (0.182) 18S (0.221) PP2A (0.355)
6 EIF4A1 (0.403) EIF4A1 (0.357) EIF2A (0.287) UBC21 (0.193) TIP41 (0.228) APT1 (0.399)
7 40S (0.419) 40S (0.374) APT1 (0.340) PP2A (0.214) UBQ (0.319) EIF2A (0.413)
8 PP2A (0.555) PP2A (0.527) EIF4A1 (0.342) UBQ (0.216) EIF4A1 (0.338) UBQ (0.438)
9 ACT (0.613) ACT (0.570) ACT (0.545) 40S (0.299) TUB (0.356) 40S (0.486)
10 TUB (0.837) 18S (0.757) PP2A (0.564) TUB (0.337) UBC21 (0.456) ACT (0.614)
11 18S (0.865) EXP45 (1.069) 18S (0.646) 18S (0.427) 40S (0.598) EXP45 (0.742)
12 EXP45 (0.873) TUB (1.245) EXP45 (1.043) EXP45 (0.882) EXP45 (1.181) 18S (0.783)
miRNA reference gene
1 MIR169 (0.046) MIR169 (0.010) MIR169 (0.070) MIR171/170 (0.045) MIR169 (0.049) MIR169 (0.078)
2 MIR172 (0.065) MIR172 (0.030) MIR172 (0.070) MIR169 (0.047) MIR171/170 (0.052) MIR156 (0.093)
3 MIR171/170 (0.088) MIR171/170 (0.054) MIR171/170 (0.075) MIR172 (0.049) MIR172 (0.106) MIR172 (0.094)
4 MIR398 (0.128) MIR398 (0.132) MIR398 (0.106) MIR398 (0.082) MIR398 (0.166) MIR171/170 (0.104)
5 MIR156 (0.186) MIR156 (0.163) MIR156 (0.151) MIR156 (0.101) MIR156 (0.250) MIR398 (0.107)
a
 Abiotic stresses analyses include all stress abiotic together

and in cucumber plants subjected to abiotic stress and growth
regulators (Migocka and Papierniak 2011).
EXP45 was the least suitable reference gene in all the
evaluated stress conditions, according to the geNorm
(Fig.  3a) and NormFinder (Table 3) results. The geNorm
analyses of each stress condition also showed that EXP45 is
not a suitable candidate to normalize RT-qPCR results across
drought, salinity, UV-C irradiation, heavy metal and ABA
conditions (Fig. 3b–f). Expansins are cell wall-loosening
proteins that regulate cell wall expansion and cell enlarge-
ment in a pH-dependent manner (Cosgrove 2000). Studies
suggest that expansins are regulated during stress conditions
(Wu et al. 2001), indicating a potential role in the control of
adaptive cell wall modification under stress (Dal Santo et al.
2013). Therefore, expansin genes do not satisfy the criteria
for use as reference genes according to Bustin (2000), who
suggests that these genes should be expressed at a constant
level among samples and that their expression cannot be
affected by the experimental conditions. Curiously, TUB was
the most unstable reference gene in drought conditions and
the most stable gene under salt stress conditions, according
to NormFinder (Table 3). TUB was also considered appropri-
ate to normalize expression under growth hormone and H2O2
treatments and in different cucumber tissue samples, accord-
ing to Wan et al. (2010), but its expression stability was sig-
nificantly lower than that of most of the other genes evalu-
ated using roots, leaves and stems of maize under different
stress conditions (Manoli et al. 2012).
Expression stability of miRNA genes under abiotic stresses
Although the number of studies involving miRNA expres-
sion has increased in recent years due to advances in the
ability to characterize such expression, in many research
areas, the experiments involving miRNA expression are
still challenging due to the lack of proper control genes for
normalizing these transcripts (Kulcheski et al. 2010). Thus,
the stability of miRNA genes was assessed using geNorm
and NormFinder software. The M values of the five miRNA
genes were lower than 1.5 according to geNorm (Fig. 4);
however, MIR169 and MIR171/170 were the two most sta-
ble genes under abiotic stress conditions (Fig. 4a). In con-
trast, MIR156 was the most unstable reference gene under
the same conditions.
As previously shown with the protein-coding genes,
the expression stability of the five candidate miRNA ref-
erence genes was also analyzed during drought, salinity,
UV-C irradiation, heavy metal and ABA application con-
ditions. Drought and rewatering were considered together.
According to geNorm, when the samples from drought
stress were analyzed (Fig. 4b), MIR169 and MIR172 were
the most stable genes. Under salinity conditions and UV-C
irradiation, the most stable miRNAs used as reference
genes in lettuce were MIR169 and MIR171/170 (Fig. 4c,
d). When heavy metal and ABA were applied, MIR172 and
MIR398 showed more stability compared to the other miR-
NAs (Fig. 4e, f).

13

1196
Fig. 4  Average expression stability (M value) of the five miRNA
candidate reference genes using geNorm software. Expression stabil-
ity was evaluated in samples from lettuce submitted to abiotic stress
(a), drought (b), salinity (c), UV-C irradiation (d), heavy metal (e)
According to NormFinder (Table 3), MIR169 and
MIR172 were the best miRNA reference genes under all
stress conditions analyzed together and when drought
and salinity were analyzed separately. MIR171/170 and
MIR169 were the most stable reference genes in UV-C
irradiation and heavy metal stress conditions, and when
ABA was applied, MIR169 and MIR156 were the best ref-
erence genes. Interestingly, MIR156 was the most unstable
miRNA reference gene when all abiotic stress conditions
were analyzed together and separately, except in the ABA
application, in which MIR156 was among the most stable
genes, after MIR169.
13
Planta (2014) 239:1187–1200
and ABA (f). The most stable reference genes were measured during
stepwise exclusion of the least stable reference genes. A lower aver-
age expression stability M value indicates more stable expression
Minimum number of reference genes
Using the protein-coding and miRNA genes described
above, we determined the number of optimal genes to
validate gene expression in lettuce leaves across abiotic
stresses using the geNorm application (Fig. 5). This soft-
ware calculates the pairwise variation Vn/Vn + 1 between
two sequential normalization factors to determine the
necessity of adding the next most consistent reference gene.
A large variation suggests that the added gene has a signifi-
cant effect on normalization and should be included for cal-
culation (Vandesompele et al. 2002). Considering a cut-off
Planta (2014) 239:1187–1200 1197

Fig. 5  Pairwise variation (V) calculated by geNorm to determine the minimum number of protein-coding reference genes (a) and miRNA refer-
ence genes (b) for accurate normalization in lettuce submitted to abiotic stress

Fig. 6  Expression levels of PAL (a) under the ABA application;
HPPD (b) under drought stress; and MIR172 (c) and MIR398 (d)
under salt stress. The protein-coding reference genes used to normal-
ize PAL were: EIF4A1 and TUB—the most stable reference genes
in ABA treatments and 18S—the most unstable reference gene in
ABA treatments. The protein-coding reference genes used to normal-
ize HPPD were EIF2A and TIP41—the most stable reference genes
under drought stress and TUB—the most unstable reference gene
under drought stress. The miRNA genes used to normalize MIR172
and MIR398 were MIR169 and MIR171/170—the most stable genes
under salinity stress and MIR156—the most unstable gene under
salinity stress. Asterisk (*) indicates significant difference and no sig-
nificance (ns) indicates no significant difference by t test (P ≤ 0.05),
between the treated and non-treated plants

value of 0.15, the use of two protein-coding reference genes
and two miRNA reference genes was determined to be suf-
ficient to normalize the results of gene expression using let-
tuce leaves submitted to different abiotic stresses because
the V2/3 values were 0.11 (Fig. 5a) and 0.056 (Fig. 5b),
respectively. The MIQE guidelines (minimum information
for publication of quantitative real-time PCR experiments)
suggest the use of three reference genes to generate more
reliable results (Bustin et al. 2009); however, the results of
this work showed that the addition of a third reference gene
does not improve the cut-off value because V3/4 was 0.128
for protein-coding candidate reference genes (Fig. 5a) and
0.043 for miRNA candidate reference genes (Fig. 5b).
Validation of protein‑coding reference genes
The expression analysis of two genes related to lettuce
quality was used to validate the selected reference genes. To

13

1198
this end, the relative expression profiles of PAL, which was
reported to be upregulated by ABA application (Wheeler
et al. 2009; Xu et al. 2012) and HPPD, which was reported
to be upregulated during drought stress (Ren et al. 2011)
were measured and normalized with two sets of reference
genes according to the NormFinder rankings as follows
(Table  3): (1) the most stable protein-coding reference
genes (TUB and EIF4A1 for ABA treatment and EIF2A and
TIP41 for drought stress) and (2) the most unstable protein-
coding reference gene in (18S for ABA treatment and TUB
for drought stress).
As shown in Fig. 6a, b, when we used the least stable
reference gene under the ABA application, no significant
difference in the expression of PAL was observed between
the treated and non-treated plants; however, PAL expression
was significantly induced (P < 0.5) when the most stable
reference genes were used. A similar result was observed
when the expression of HPPD was evaluated in drought-
stressed plants using the least and the most stable reference
genes. Thus, the use of unsuitable references leads to dif-
ferences in the relative expression profile. These results
further confirmed the importance of validating reference
genes prior to experimental applications, and no single ref-
erence gene was observed with constant expression in all
the tested conditions.
Validation of miRNA reference genes
To validate the selected miRNA reference genes, the rela-
tive expression profiles of MIR172 and MIR398 were
measured and normalized with the most stable reference
genes under salinity stress (MIR171/170 and MIR169) and
with the most unstable reference gene under salinity stress
(MIR156) according to the NormFinder ranking (Table 3).
When the most stable reference gene was used, the expres-
sion of both miRNAs was significantly reduced, which is
in accordance with the literature for other species (Sunkar
et al. 2012; Khraiwesh et al. 2012; Zhu et al. 2013). How-
ever, when the most unstable reference gene was used,
the target expression was not reduced and the standard
error was increased substantially for both miRNAs when
MIR156 was used as the reference gene (Fig. 6c, d).
Conclusions
In conclusion, we evaluated seventeen candidate reference
genes for the normalization of gene expression in lettuce
leaves under different abiotic stress conditions. To the best
of our knowledge, this work represents the first attempt to
validate reference genes in lettuce for the normalization
of gene expression analysis using RT-qPCR. Our results
suggested that different appropriate reference genes or a
13
Planta (2014) 239:1187–1200
combination of reference genes for normalization should
be selected according to the abiotic stress conditions being
evaluated. The novel protein-coding reference genes appear
to be more stable than the traditional genes; they occupied
the top two positions according to NormFinder under all
the stress conditions evaluated except salt stress. EXP45
was clearly regulated under abiotic stresses and was con-
sidered the most unstable gene. In addition, miRNAs were
stably expressed under abiotic stresses, representing a suit-
able alternative approach to normalized gene expression
data. The expression analysis of PAL, HPPD, MIR172 and
MIR398 confirmed the importance of validating reference
genes to achieve accurate RT-qPCR results. These results
will facilitate further analyses aiming to elucidate the
mechanisms involved in the lettuce response to different
stresses and their effect on the content and composition of
health-promoting compounds in studies that target lettuce
biofortification.
Acknowledgments The authors gratefully acknowledge the
technical and financial support of the Embrapa Clima Temperado
(Embrapa), Fundação de Amparo à Pesquisa do Estado do Rio Grande
do Sul (Fapergs), and Conselho Nacional de Desenvolvimento Cientí-
fico e Tecnológico (CNPq).
References
Andersen CL, Jensen JL, Orntoft TF (2004) Normalization of real-
time quantitative reverse transcription-PCR data: a model-based
variance estimation approach to identify genes suited for normal-
ization, applied to bladder and colon cancer data sets. Cancer Res
64:5245–5250
Argyris J, Dahal P, Hayashi E, Still DW, Bradford KJ (2008) Genetic
variation for lettuce seed thermoinhibition is associated with
temperature-sensitive expression of abscisic acid, gibberellin,
and ethylene biosynthesis, metabolism, and response genes. Plant
Physiol 148:926–947
Borghi S (2003) Special: IV range (vegetables). Colt Protette
32:21–43
Bustin SA (2000) Absolute quantification of mRNA using real-time
reverse transcription polymerase chain reaction assays. J Mol
Endocrinol 25:169–193
Bustin SA (2002) Quantification of mRNA using real-time reverse
transcription PCR (RT-PCR): trends and problems. J Mol Endo-
crinol 29:23–39
Bustin SA, Benes V, Garson JA et al (2009) The MIQE guidelines:
minimum information for publication of quantitative realtime
PCR experiments. Clin Chem 55:611–622
CGPDB(2013) Compositae genome project database. http://cgpdb.uc
davis.edu/. Accessed 26 July 2013
Chang S, Puryear J, Cairney J (1993) A simple and efficient
method for isolating RNA from pine trees. Plant Mol Biol Rep
11:113–116
Chen C, Ridzon DA, Broomer AJ et al (2005) Real-time quantifica-
tion of microRNAs by stem–loop RT-PCR. Nucleic Acids Res
33:e179. doi:10.1093/nar/gni178
Chen L, Zhong H, Kuang J et al (2011) Validation of reference genes
for RT-qPCR studies of gene expression in banana fruit under dif-
ferent experimental conditions. Planta 234:377–390
Planta (2014) 239:1187–1200 1199
COMISSÃO DE QUÍMICA E FERTILIDADE DO SOLO—CQFS
RS/SC (2004) Manual de adubação e calagem para os Estados
do Rio Grande do Sul e Santa Catarina. 10. Ed. Porto Alegre,
SBCS/Núcleo regional Sul/UFRGS, p 400
Cosgrove DJ (2000) Loosening of plant cell walls by expansins.
Nature 407:321–326
Cruz F, Kalaoun S, Nobile P et al (2009) Evaluation of coffee ref-
erence genes for relative expression studies by quantitative real-
time RT-PCR. Mol Breed 23:607–616
Czechowski T, Stitt M, Altmann T, Udvardi MK, Scheible WR (2005)
Genome-wide identification and testing of superior reference
genes for transcript normalization in Arabidopsis. Plant Physiol
139:5–17
Dal Santo S, Vannozzi A, Tornielli JB et al (2013) Genome-wide anal-
ysis of the expansin gene superfamily reveals grapevine-specific
structural and functional characteristics. Plos One 8(4):e62206.
doi:10.1371/journal.pone.0062206
Derveaux S, Vandersompele J, Hellemans J (2010) How to do suc-
cessful gene expression analysis using real-time PCR. Methods
50:227–230
Expósito-Rodríguez M, Borges AA, Borges-Pérez A, Pérez JA (2008)
Selection of internal control genes for quantitative real-time RT-
PCR studies during tomato development process. BMC Plant
Biol 8:131–142
Gutierrez L, Mauriat M, Guenin S, Pelloux J, Lefebvre JF (2008) The
lack of a systematic validation of reference genes: a serious pit-
fall undervalued in reverse transcription-polymerase chain reac-
tion (RT-PCR) analysis in plants. Plant Biotechnol J 6:609–618
Han H, Zhu B, Luan F et al (2010) Conserved miRNAs and their
targets identified in lettuce (Lactuca) by EST analysis. Gene
463:1–7
Hawrylak-Nowak B (2013) Comparative effects of selenite and
selenate on growth and selenium accumulation in lettuce plants
under hydroponic conditions. Plant Growth Regul 70:149–157
Jain M, Nijhawan A, Tyagi AK, Khurana JP (2006) Validation of
housekeeping genes as internal control for studying gene expres-
sion in rice by quantitative real-time PCR. BBRC 343:646–651
Khraiwesh B, Zhu J, Zhu J (2012) Role of miRNAs and siRNAs in
biotic and abiotic stress responses of plants. Biochim Biophys
Acta 1819:137–148
Kulcheski FR, Marcelino-Guimaraes FC, Nepomuceno AL, Abdel-
noor RV, Margis R (2010) The use of microRNAs as reference
genes for quantitative polymerase chain reaction in soybean. Anal
Biochem 406:185–192
Lai Z, Kane NC, Kozik A et al (2012) Genomics of Compositae
weeds: EST libraries, microarrays, and evidence of introgression.
Am J Bot 99:209–218
Lee H, Aedin C (2006) A review of the health care potential of bioac-
tive compounds. J Sci Food Agric 86:1805–1813
Libault M, Thibivilliers S, Bilgin D et al (2008) Identification of four
soybean reference genes for gene expression normalization. Plant
Genome 1:44–54
Livak KJ, Schmittgen TD (2001) Analysis of relative gene expression
data using real-time quantitative PCR and the 2-Delta Delta CT
method. Methods 25:402–408
Llorach R, Martínez-Sánchez A, Tomás-Barberán FA, Gil MI, Fer-
reres F (2008) Characterisation of polyphenols and antioxidant
properties of five lettuce varieties and escarole. Food Chem
108:1028–1038
Mahmoudi H, Nasri MB, Baâtour O et al (2012) Differential response
to elevated NaCl by antioxidant enzymes and gene transcripts in
two contrasting lettuce genotypes. Austr J Crop Sci 6:632–640
Maltseva DV, Khaustova NA, Fedotov NN et al (2013) High-through-
put identification of reference genes for research and clinical RT-
qPCR analysis of breast cancer samples. J Clin Bioinform 3:13.
doi:10.1186/2043-9113-3-13
Manoli A, Sturaro A, Trevisan S, Quaggiotti S, Nonis A (2012) Evalu-
ation of candidate reference genes for qPCR in maize. J Plant
Physiol 169:807–815
Marouelli WA, Silva WLC, Silva HR (2008) Irrigação por aspersão
em hortaliças: qualidade da água, aspectos do sistema e método
prático de manejo. Embrapa Informação Tecnológica, Embrapa
Hortaliças, Brasília
Maroufi A, Van Bockstaele E, De Loose M (2010) Validation of ref-
erence genes for gene expression analysis in chicory (Cichorium
intybus) using quantitative real-time PCR. BMC Mol Biol 11:15
Messias R, Galli V, Silva S, Schirmer M, Rombaldi C (2013) Micro-
nutrient and functional compounds biofortification of maize
grains. Crit Rev Food Sci Nutr. doi:10.1080/10408398.2011.649
314
Migocka M, Papierniak A (2011) Identification of suitable reference
genes for studying gene expression in cucumber plants subjected
to abiotic stress and growth regulators. Mol Breed 28:343–357
Nicolle C, Cardinault N, Gueux E et al (2004) Health effect of vegeta-
ble-based diet: lettuce consumption improves cholesterol metabo-
lism and antioxidant status in the rat. Clin Nutr 23:605–614
Peltier HJ, Latham GJ (2008) Normalization of microRNA expression
levels in quantitative RT-PCR assays: identification of suitable
reference RNA targets in normal and cancerous human solid tis-
sues. RNA 14:844–852
Qin F, Shinozaki K, Yamaguchi-Shinozaki K (2011) Achievements
and challenges in understanding plant abiotic stress responses
and tolerance. Plant Cell Physiol 52:1569–1582
Rabbani MA, Maruyama K, Abe H et al (2003) Monitoring expres-
sion profiles of rice genes under cold, drought, and high-salinity
stresses and abscisic acid application using cDNA microarray and
RNA gel-blot analyses. Plant Physiol 133:1755–1767
Ramakers C, Ruijter JM, Deprez RH, Moorman AF (2003) Assump-
tion-free analysis of quantitative real-time polymerase chain reac-
tion (PCR) data. Neurosci Lett 339:62–66
Ren W, Zhao L, Zhang L, Wang Y, Cui L, Tang Y, Sun X, Tang K
(2011) Molecular cloning and characterization of 4-hydroxyphe-
nylpyruvate dioxygenase gene from Lactuca sativa. J Plant Phys-
iol 168:1076–1083
Soós V, Juhás A, Light ME, Staden JV, Balázs E (2009) Smoke water
induced changes of expression pattern in grand rapids lettuce
achenes. Seed Sci Res 19:37–49
Sunkar R, Li YF, Jagadeeswaran G (2012) Functions of microRNAs
in plant stress responses. Trends Plant Sci 17:196–302
Tiecher A, Paula LA, Chaves FC, Rombaldi CV (2013) UV-C effect
on ethylene, polyamines and the regulation of tomatofruit ripen-
ing. Postharvest Biol Technol 86:230–239
Vandesompele J, De Preter K, Pattyn F et al (2002) Accurate nor-
malization of real-time quantitative RT-PCR data by geomet-
ric averaging of multiple internal control genes. Genome Biol
3:research0034
Wan H, Zhao Z, Qian C et al (2010) Selection of appropriate reference
genes for gene expression studies by quantitative real-time poly-
merase chain reaction in cucumber. Anal Biochem 399:257–261
Wang Y, Frei M (2011) Stressed food—The impact of abiotic envi-
ronmental stresses on crop quality. Agric Ecosyst Environ
141:271–286
Wei L, Miao H, Zhao R et al (2013) Identification and testing of refer-
ence genes for Sesame gene expression analysis by quantitative
real-time PCR. Planta 237:873–889
Wheeler S, Loveys B, Ford C, Davies C (2009) The relationship
between the expression of abscisic acid biosynthesis genes,
accumulation of abscisic acid and the promotion of Vitis vinif-
era L. berry ripening by abscisic acid. Austr J Grape Wine Res
15:195–204
Wu Y, Kuzma J, Marechal E et al (1997) Abscisic acid signaling
through cyclic ADP-ribose in plants. Science 278:2126–2130
13

1200
Wu Y, Thorne ET, Sharp RE, Cosgrove DJ (2001) Modification of
expansin transcript levels in the maize primary root at low water
potentials. Plant Physiol 126:1471–1479
Xu F, Deng G, Cheng S et al (2012) Molecular cloning, characteri-
zation and expression of the phenylalanine ammonia-lyase gene
from Juglans regia. Molecules 17:7810–7823
Yuan G, Wang X, Guo R, Wang Q (2010) Effect of salt stress on phe-
nolic compounds, glucosinolates, myrosinase and antioxidant
activity in radish sprouts. Food Chem 121:1014–1019
Zdravković JM, Aćamović-Đoković GS, Mladenović JD, Pavlović
RM, Zdravković MS (2014) Antioxidant capacity and contents of
13
Planta (2014) 239:1187–1200
phenols, ascorbic acid, β-carotene and lycopene in lettuce. Hemi-
jska industrija. doi:10.2298/HEMIND130222043Z
Zhu J, Li W, Yang W, Qi L, Han S (2013) Identification of microR-
NAs in Caragana intermedia by high throughput sequencing and
expression analysis of 12 microRNAs and their targets under salt
stress. Plant Cell Rep 32:1339–1349
Zou JJ, Wei FJ, Wang C et al (2010) Arabidopsis calcium-dependent
protein kinase CPK10 functions in abscisic acid- and Ca2+-medi-
ated stomatal regulation in response to drought stress. Plant Phys-
iol 154:1232–1243