Gene 554 (2015) 205–214

Contents lists available at ScienceDirect

Gene

j ourna l homepage: www.e lsev ie r .com/ locate /gene

Validation of reference genes for accurate normalization of geneexpression for real time-quantitative PCR in strawberry fruits usingdifferent cultivars and osmotic stresses☆

Vanessa Galli a,b,⁎,1, Joyce Moura Borowski a,b,1, Ellen Cristina Perin a,b,1, Rafael da Silva Messias a,b,Julia Labonde c, Ivan dos Santos Pereira a, Sérgio Delmar dos Anjos Silva a, Cesar Valmor Rombaldi b

a Embrapa Clima Temperado, Rodovia BR 396, Km 78 Caixa Postal 403, CEP 96001-970 Pelotas, RS, Brazilb Universidade Federal de Pelotas, Faculdade de Agronomia Eliseu Maciel, Campus Universitário s/n, Caixa Postal 354, CEP 96010-900 Pelotas, RS, Brazilc Universidade Federal de Pelotas, Centro de Ciências Biológicas, Campus Universitário s/n, Caixa Postal 354, CEP 96010-900 Pelotas, RS, Brazil

Abbreviations:DBP, DNAbindingprotein;HISTH4, histo3-phosphate dehydrogenase; 18S, 18S ribosomal RNA; PALNCED1, 9-cis epoxycarotenoiddioxygenase;RT-qPCR, revetitativepolymerasechainreaction;ACT,actin;TUB, tubulin;al RNA; ABA, abscisic acid; NaCl, sodium chloride; UBQ11decarboxylase; Tm,melting temperature; Cq, quantificativalue; V, pairwise variation.☆ Key message: The stability of candidate reference gencultivars and strawberry submitted to osmotic stresses tosion by RT-qPCR.⁎ Corresponding author at: Universidade Federal de Pe

Eliseu Maciel, Campus Universitário s/n, Caixa Postal 35Brazil.

E-mail address: vane.galli@yahoo.com.br (V. Galli).1 These authors contributed equally to this study.

http://dx.doi.org/10.1016/j.gene.2014.10.0490378-1119/© 2014 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 29 June 2014Received in revised form 18 October 2014Accepted 27 October 2014Available online 29 October 2014

Keywords:Fragaria × ananassaStrawberry cultivarsAbiotic stressesReference genesRefFinder

The increasing demand of strawberry (Fragaria × ananassa Duch) fruits is associatedmainly with their sensorialcharacteristics and the content of antioxidant compounds. Nevertheless, the strawberry production has beenhampered due to its sensitivity to abiotic stresses. Therefore, to understand themolecularmechanisms highlight-ing stress response is of great importance to enable genetic engineering approaches aiming to improve strawber-ry tolerance. However, the study of expression of genes in strawberry requires the use of suitable reference genes.In the present study, seven traditional and novel candidate reference geneswere evaluated for transcript normal-ization in fruits of ten strawberry cultivars and two abiotic stresses, using RefFinder, which integrates the fourmajor currently available software programs: geNorm, NormFinder, BestKeeper and the comparative delta-Ctmethod. The results indicate that the expression stability is dependent on the experimental conditions. The can-didate reference geneDBP (DNA binding protein)was considered themost suitable to normalize expression datain samples of strawberry cultivars and under drought stress condition, and the candidate reference gene HISTH4(histone H4) was the most stable under osmotic stresses and salt stress. The traditional genes GAPDH (glyceral-dehyde-3-phosphate dehydrogenase) and 18S (18S ribosomal RNA) were considered themost unstable genes inall conditions. The expression of phenylalanine ammonia lyase (PAL) and 9-cis epoxycarotenoid dioxygenase(NCED1) genes were used to further confirm the validated candidate reference genes, showing that the use ofan inappropriate reference genemay induce erroneous results. This study is thefirst survey on the stability of ref-erence genes in strawberry cultivars and osmotic stresses and provides guidelines to obtain more accurate RT-qPCR results for future breeding efforts.

© 2014 Elsevier B.V. All rights reserved.

neH4;GAPDH, glyceraldehyde-, phenylalanine ammonia lyase;rse transcription real timequan-UBI,ubiquitin;40S,40Sribosom-, ubiquitin 11; PIRUV, pyruvateon cycle;M, expression stability

es was evaluated in strawberryfurther studies of gene expres-

lotas, Faculdade de Agronomia4, CEP 96010-900 Pelotas, RS,

1. Introduction

Strawberry (Fragaria × ananassa Duch) is one of the most economi-cally important fruit crops worldwide. The increasing cultivation areasand the rising consumption of strawberry fruits are associated with itspleasant flavor, taste and texture, as well as with their essential nutri-ents, minerals, vitamins, and antioxidant compounds. The antioxidantproperties of strawberry fruits are related to the high content of L-ascorbic acid (vitamin C), anthocyanins and phenolic compounds(Erkan et al., 2008; Pineli et al., 2011),whichhavebeenmedically recog-nized as having positive influences on protecting against the risk ofmany diseases (Zhu et al., 2013).

Therefore, tomeet increased strawberry fruit demands in our chang-ingworld, it is important to expandour agricultural systems to drier andsaline lands, especially in developing countries (Van den Ende and El-Esawe, 2014; Huang et al., 2012; Krasensky and Jonak, 2012). However,osmotic stresses are among the factors most limiting crop productivity

206 V. Galli et al. / Gene 554 (2015) 205–214

(Orsini et al., 2012). Even in not arid or semiarid regions, which aremore exposed to salinization, this phenomenon is also increasinglyexpanding because it is tightly associatedwith the practice of irrigation,the most used cultivation system for strawberry crop (Orsini et al.,2012). Therefore, several efforts have been made to breed this speciesfor higher abiotic stress tolerance (Husaini and Abdin, 2008; Christouet al., 2013, 2014).

The plant engineering strategies for abiotic stress tolerance rely onthe expression of genes involved in the stress-mediated response,since it induces a cascade of events in plant cell that trigger gene expres-sion regulation, leading to biochemical and physiological changes. How-ever, the information of gene expression and adaptation capability ofstrawberry is limited and the strategy employed by plant to survivesuch adverse conditions is dependent on the abiotic stress and the ge-netic background.

Nowadays, due to the subsequent crossing efforts with differentFragaria species, there is a great range of cultivars with distinct pheno-typic and adaptation characteristics (Capocasa et al., 2008; Padulaet al., 2013; Vandendriessche et al., 2013; Mazur et al., 2014), whichcould be exploited. Therefore, the study of expression of genes in differ-ent cultivars and under environmental stresses may provide useful in-formation regarding the molecular mechanisms highlighting the stresstolerance enabling for the development of genetic engineering ap-proaches. Furthermore, the strawberry has been proposed as a modelfor functional genomics and transgenic studies within the Rosaceae(Mezzetti, 2009); therefore, the molecular information obtained withthis crop may also be applied to other rosaceous crops.

Currently, one of the most used techniques for determination ofgene expression is the reverse transcription real time quantitative poly-merase chain reaction (RT-qPCR), because of its high sensitivity, repro-ducibility and specificity (Bustin, 2002; Derveaux et al., 2010). However,reliable results of RT-qPCR are dependent on specific experimentalstrategies in order to minimize the variations in the quality, stabilityand input of RNA, and in the efficiency of reverse transcription andPCR steps (Fleige and Pfaffl, 2006; Derveaux et al., 2010). Among thesestrategies, the selection of suitable reference genes to normalize datais of great importance to obtain accurate results. A suitable referencegene should be expressed at a constant level among samples, and its ex-pression is assumed to be unaffected by the experimental conditions(Bustin, 2002). The use of inadequate reference genes may result inquantification errors; consequently, the expression data may bemisinterpreted (Jain et al., 2006; Amil-Ruiz et al., 2013).

Some genes, because of their roles in basic cellular process, primarymetabolism and cell structure maintenance, are often referred to as ref-erence genes (Wong and Medrano, 2005; Czechowski et al., 2005).Thus, the most traditional reference genes currently used in RT-qPCRstudies with plants include actin (ACT) (Maroufi et al., 2010), tubulin(TUB) (Wan et al., 2010), ubiquitin (UBI) (Chen et al., 2011), 18S ribo-somal RNA (18S) (Jain et al., 2006) and 40S ribosomal RNA (40S)(Cruz-Rus et al., 2011). Nevertheless, it is important to highlight thatthe stability among some of these commonly used reference genes isrelative, and there is no single gene having a constant stable expressionunder all experimental conditions (Radonic et al., 2004; Czechowskiet al., 2005). Therefore, the reliability of the results of gene expressionis dependent on the use of suitable reference genes for the crop and con-ditions under study.

In the present study, we evaluated the stability of traditional andnovel candidate reference genes, in order to identify the most suitablereference genes for transcript normalization in ten strawberry cultivars(Fragaria × ananassa Duch) and under salt, drought or osmotic (sam-ples from the salt and drought experiments were analyzed together)stresses. For demonstrating the efficacy of the selected referencegenes, we investigated the expression of phenylalanine ammonialyase (PAL), a key enzyme from the phenylpropanoid pathway, and 9-cis epoxycarotenoid dioxygenase (NCED1), responsible for the synthe-sis of abscisic acid (ABA), both related to abiotic stress response. As far

as we know, this is the first attempt to evaluate the expression stabilityof candidate genes in strawberry under these conditions.

2. Material and methods

2.1. Experimental conditions and plant material

In this study, the expression stability of putative reference geneswasanalyzed in the following experimental conditions: (1) among samplesfrom different cultivars; (2) among samples from strawberry fruits cul-tivated under salt stress; (3) among samples from strawberry fruits cul-tivated under drought stress; and (4) analyzing together the samplesfrom the salt stress experiment (2) and from the drought stress exper-iment (3).

To understand the molecular events involved in abiotic stress, it isimportant to study the expression profile of related genes in plantsfrom strawberry cultivars differing in the ability to tolerate those condi-tions. Therefore, to select the appropriate reference genes to normalizeRT-qPCR data in different strawberry cultivars (Aromas, Albion,Camarosa, Camino Real, Diamante, Festival, Palomar, Portola, SanAndreas andVentana), the seedlingswere transplanted to experimentalfield at Brazilian Agricultural Research Corporation (Embrapa, Pelotas/RS, Brazil). The fertilizer was calculated and applied according to thetechnical recommendations for the crop (CQFS, 2004) and the irrigationwas performed by dripping. Each cultivar was represented by three rep-licates with five plants for each replicate.

To select the appropriate reference genes to normalize RT-qPCR datain strawberries submitted to osmotic stresses, the experiment was con-ducted in a greenhouse where strawberry seedlings (cv. Camarosa)were transplanted to 9 L pots containing a mixture of soil and vermicu-lite (3:1) as substrate. The fertilizer was calculated and applied accord-ing to the technical recommendations for the crop (CQFS, 2004).Irrigationwas performed by dripping in order tomaintain the soil mois-ture content constant (20% ± 5%), as measured by electronic meterHydrofarm (Falker™). Plants were watered according to the evapo-transpiration value of the culture, as proposed by Marouelli et al.(2008). For salt stress treatment, plants were submitted to four applica-tions of 200 mM sodium chloride (NaCl) (one per week) in the soil,starting at 90 days after transplanting. For the drought stress treatment,from the flowering stage, plants were cultivated with 50% of the totalamount of water required by this crop, according to the evapotranspira-tion values (Marouelli et al., 2008). The experiment was in a completelyrandomized design, which consisted of three treatments: non-stressedplants (control), salt stressed plants, and drought stressed plants. Eachtreatment consisted of four replicates with six plants per replicate.

In the first experiment (strawberry cultivars), at least one ripe fruits(stage FR, according to Jia et al., 2011) from each of the five plants perreplicate was sampled. In the second experiment (osmotic stresses), atleast one ripe fruit was sampled from each of the six plants per replicate.The sampled ripe fruits from each replicate were pooled, and immedi-ately frozen in liquid nitrogen and stored at −80 °C until analysis.Therefore, we obtained three RNA samples from each cultivar, totalizing30 samples, and four RNA samples from each of the three treatmentsfrom the osmotic stress experiment, totalizing 12 samples.

2.2. Total RNA isolation and first strand cDNA synthesis

Total RNA of strawberry was isolated using a modified CTAB(hexadecyltrimethylammonium bromide) protocol (Messias et al.,2014). RNAqualitywas evaluated in 1% agarose gel after electrophoresisand by spectrometry, using A260/A280 and A260/A230 ratios. RNA concen-tration was measured in Qubit® fluorometer (Invitrogen). Total RNA(1 μg) was treated with 1 U DNAse I and DNAse 1× reaction buffer(Invitrogen) before cDNA synthesis. For the amplification of the candi-date reference genes, all treated RNA was reverse transcribed using

207V. Galli et al. / Gene 554 (2015) 205–214

the M-MLV enzyme and oligo-dT primers, according to themanufacturer's instructions (Invitrogen).

2.3. Primer design and RT-qPCR conditions

Seven candidate reference genes were used in the present study.They included the traditional reference genes ACT, GAPDH (glyceralde-hyde-3-phosphate dehydrogenase), UBQ11 (ubiquitin 11) and 18S;the gene DBP (DNA binding protein) and the gene HISTH4 (histoneH4), which were previously reported as reference gene in studies withstrawberry (Mehli et al., 2004; Lunkenbein et al. 2006; Landi et al.,2014); and the reference gene PIRUV (pyruvate decarboxylase), whichwas not previously described for this crop. The last two referencegenes were selected based on a preliminary in silico evaluation of geneexpression stability using in house RNA-Seq libraries from strawberryfruits under control, salt stress and drought stress conditions.

These primers were designed based on the sequences ofFragaria × ananassa extracted fromGenBank, using Vector NTI10 soft-ware (Invitrogen), according to the following parameters:melting tem-perature (Tm) of 58–62 °C and GC content of 45–55%. All primersequences and relevant information about the genes are presented inTable 1.

The cDNAs, in the concentration of 20 ng·μL−1, were amplified byRT-qPCR in a final volume of 20 μL containing 1 μL cDNA, 10 μL of Plati-num SYBR green UDG (Invitrogen), and 3–5 pmol of each primer. Am-plification 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 at60 °C. The specificity of the amplicons was verified by the presence ofa single peak in RT-qPCR melting curve products and a single band ofthe expected size in the 3% agarose gel after electrophoresis. The meltcurve analysis ranged from 60 to 95 °C increasing the temperature step-wise by 1%. No-template controls were included to ensure that therewas no reagent or genomic DNA contamination.

2.4. Data analysis

In order to estimate the expression stability of the seven candidatereference genes, all amplification plots were analyzed with a thresholdfluorescence value of 0.1 to obtain amplification cycle (Cq) valuesusing the SDS version 1.1 software (Applied Biosystems). The efficiencyof the PCR was estimated using the LinReg PCR program (Ramakers

Table 1Candidate reference gene description and parameters derived from RT-qPCR analysis.

Genesymbol

Primer sequence (5′–3′) (forward/reverse) GenBank accession

18S TGTGAAACTGCGAATGGCTCATTAA/GAAGTCGGGATTTGTTGCACGTATT

X15590.1

ACT GGGATGACATGGAGAAGATTTGGC/TCTCACGATTAGCCTTGGGATTCAG

AB116565.1

DBP TTGGCAGCGGGACTTTACC/CGGTTGTGTGACGCTGTCAT

–c

GAPDH CCAAGGCTGTCGGAAAGGTT/CAACATCATCTTCGGTGTAACCC

AB363963.1

HISTH4 GTGGCGTCAAGCGTATCTCC/TGTCCTTCCCTGCCTCTTGA

AB197150.1

PIRUV AGGTGCGTTGCGAAGAGGA/CTAAATCTGTGAATGCGAATGAGG

AF141016.2

UBQ11 CAGACCAGCAGAGGCTTATCTT/TTCTGGATATTGTAGTCTGCTAGGG

–d

FaPAL AACCACGACATTTCCAACGAGGC/GCCCTACCATTGATTTCAGCGAC

HM641823.1; AB3603AB360391.1; AB36039

FaNCED1 TCGCCATTGCTGAACCGT/TCCACCTTCAAACTCACGGC JX013944.1

a The melting temperature was calculated by SDS version 1.1 software in a 7500 Real time fb The RT-qPCR efficiency and correlation coefficients (R2) were determined by LinReg PCR pc According to Schaart et al. (2002).d According to Hytönen et al. (2009).e PAL primers were designed based on a consensus sequence of these accesses.

et al., 2003). There aremany tools to evaluate the stability of a referencegene, such as geNorm (Vandesompele et al., 2002), NormFinder(Andersen et al., 2004), BestKeeper (Pfaffl et al., 2004), and the compar-ative ΔCt method (Silver et al., 2006). The ΔCt method compares rela-tive expression of pairs of genes within each sample to identify usefulreference genes (Silver et al., 2006). The Bestkeeper is an Excel basedtool that uses raw data (Cq values) and PCR efficiency to determinethe best suited references and combines them into an index by the coef-ficient of determination and the P value (Pfaffl et al., 2004). In this case,any studied genewith the standard deviation (SD) higher than 1 can beconsidered inconsistent. NormFinder provides a ranking from the mostto the least stable gene, based on the stability value of each gene whichis a direct measure for the estimated expression variation enabling theuser to evaluate the systematic error introduced when using the genefor normalization (Andersen et al., 2004). GeNorm algorithm firstcalculates an expression stability value (M) for each gene and then com-pares the pairwise variation (V) of this gene with the others. Referencegenes are ranked according to their expression stability by a repeatedprocess of stepwise exclusion of the least stably expressed genes(Vandesompele et al., 2002). Genes with highly variable results have ahigh M value, which indicates a low stability of expression, and viceversa, when using a cutoff of 1.5. A user-friendly web-based compre-hensive tool, RefFinder (http://www.leonxie.com/referencegene.php?type=reference) was used, integrating the four currently availablemajor software programs, geNorm, NormFinder, BestKeeper, and thecomparative delta-Ctmethod, to compare and ranking the tested candi-date reference genes. RT-qPCR data were exported into an Exceldatasheet (Microsoft Excel 2010) and Cq valueswere converted accord-ing to the requirements of the software. The figures for relative expres-sion profiles were generated using the program GraphPad Prism 6(GraphPad Prism Software).

2.5. Validation of reference genes

To confirm the reliability of the potential reference genes, the relativeexpression profiles of FaPAL, a key enzyme from the phenylpropanoidmetabolic pathway, and FaNCED1, a gene involved in the synthesis ofABA, were measured and normalized with the most stable and least sta-ble reference genes, as determined by RefFinder. The RT-qPCR amplifica-tion conditionswere the same as described above. The relative expressiondata was calculated according to the 2−ΔΔCq method and presented asfold change (Livak and Schmittgen, 2001). Samples from non-stressed

Ampliconlength

RT-qPCRefficiencyb

R2 Product Tma

(°C)

109 1.962 0.994 80.74

117 1.872 0.990 85.40

–c 1.938 0.999 79.54

209 1.851 0.998 86.44

167 1.845 0.997 88.27

219 1.943 0.999 85.24

–d 1.884 0.998 83.43

94.1; AB360393.1;0.1

115e 2.01 0.998 84.39

148 1.906 0.997 81.02

ast thermocycler (Applied Biosystems).rogram.

208 V. Galli et al. / Gene 554 (2015) 205–214

plants were used as reference samples. Statistical analyses were per-formed using the computer program SAS System for Windows version9.1.3 (SAS, 2000). Data were subjected to variance analysis (p ≤ 0.05).In case of statistical significance, the relative quantitation results werecompared from the more stable and more unstable reference genes by ttest (p ≤ 0.05).

3. Results and discussion

3.1. Description and expression profiling of the candidate reference genes

Gene expression studies by RT-qPCR require the knowledge of stablyexpressed reference genes for data normalization of target genes underspecific experimental conditions, because the use of inadequate non-validated reference genes may result in doubtful interpretation of thedata (Bustin et al., 2009; Derveaux et al., 2010). Therefore, to allowthe evaluation of the expression of genes related to stress response instrawberry cultivars it is of primary importance to determine the mostsuitable reference genes to normalize expression data.

High RT-qPCRefficiency is usually correlatedwith robust andpreciseresults of gene expression (Bustin et al., 2009). In this study, the efficien-cy of RT-qPCR was calculated for all candidate reference genes as themean values obtained from the technical and biological replicates andit varied from 1.84 (HISTH4) to 1.93 (PIRUV), indicating high efficiency.The Tm for all PCR products ranged from 79.54 °C (DBP) to 88.27 °C(HISTH4), which was in the rage expected for these amplicons basedon the percentage of CG composition and on the length used as criteriafor the primer design (Table 1). The specificity of candidate referencegenes and target genes were confirmed by the existence of a singlepeak in the melting curves and a single band in the 3% agarose gel,after electrophoresis (Online Resource 1). Additionally, no RT-qPCR de-tection signals were observed in the no-template controls and reversetranscription negative control reactions. These results allow the use ofthe designed primers for gene expression analysis by RT-qPCR.

A suitable reference gene shows a constant expression level amongsamples evaluated; therefore, we used the quantification cycle (Cq)values to determine the expression levels of the traditional and novelcandidate reference genes evaluated in this study. We considered twosubsets of samples: one composed by the strawberry cultivars(Fig. 1a), and the other composed by samples from osmotic (salt anddrought) stresses (Fig. 1b). The results show that the ACT was theleast expressed gene, with the highestmean Cq value in the two subsetsof samples (24.54 and 28.18, respectively) (Fig. 1). On the other hand,18S was the most expressed gene in both subsets, with the lowestmean Cq value (13.24 and 15.24, respectively) (Fig. 1). Additionally,GAPDH showed themost variation in expression levels among the eval-uated genes in the subset of osmotic stresses, while 18S showed themost variation in expression levels in the subset of strawberry cultivars,

Fig. 1. Expression levels of different candidate reference genes. Expression data displayed as RT(a) and twoosmotic stresses (salt anddrought stress) (b). The line across the box is depicted as tmaximum and minimum values. The higher the boxes and whisker are, the greater the variati

as showed by the larger whisker taps and boxes in Fig. 1 compared tothose for the other genes, suggesting its low stability. Similar resultswere demonstrated by Wei et al. (2013), which found 18S gene as themost expressed, evaluating ten candidate reference genes in Sesamesamples submitted to different stresses and hormone treatments.More-over, in a study identifying themost suitable reference genes for abioticstress in soybean, thirteen candidate genes collected from literaturewere evaluated for stability of expression under dehydration, high salin-ity, cold and abscisic acid treatments using delta Ct and geNorm ap-proaches. The 18S gene was excluded of the candidates due to itsextremely high abundance. Template cDNA for the samples analyzedwith 18S primers had to be diluted at least 1000-fold relative to allother candidate genes to avoid the signal saturation obtained at lowCT values. This dilution introduces a random element of variabilitywhich can potentially alter the apparent expression levels of targetgenes normalized with 18S (Le et al., 2012).

3.2. Expression stability of candidate reference genes in strawberry fruits

The main characteristic of a reference gene is the expression stabili-ty, which is not affected by tissue type, developmental periods, or phys-iological conditions. In this study, we used RefFinder, a user-friendly,web-based comprehensive tool that was developed for evaluating andscreening reference genes, which integrates the comparativeΔCtmeth-od, BestKeeper, NormFinder, and geNorm approaches. Overall, the re-sults of these methods were similar in the subset of strawberrycultivars: themost unstable genewas 18S and the secondmost unstablegene was GAPDH, suggesting that the traditional housekeeping genesare not suitable to normalize RT-qPCR data using these samples(Fig. 2A–D). For the most stable candidate reference genes, accordingto the ΔCt method, the UBQ11 gene was the first gene of the rank inthe samples of strawberry cultivars, indicating that it is the most stablegene (Fig. 2A), followed byDBP gene. However, according to Bestkeeper(Fig. 2B), none of the genes should be considered as reference gene,since all genes showed SD higher than 1. Although only 18S showedM value higher than the cut-off (M N 1.5), and thus considered unstableaccording to geNorm (Fig. 2B), this method pointed DBP and UBQ11 asthe most stable genes in these samples; and ACT, followed by DBP, ac-cording to NormFinder (Fig. 2D). The small differences observed be-tween stability estimation rankings determined by those methodswere expected based on the different designs of the algorithms in the al-ternative approaches (Andersen et al., 2004), and because the stabilityamong most of the genes is similar.

Besides integrating the four above methods, the RefFinder softwareuses the Cq value to assign an appropriate weight to an individualgene and calculates the geometric mean of their weights for the overallfinal ranking, based on the rankings from each of the four programs. Al-though the RefFinder do not indicate a cut-off value to classify a

-qPCR quantification cycle (Cq) values for each reference gene in ten strawberry cultivarshemedian. The box indicates the 25th and 75th percentiles andwhisker caps represent theons.

Fig. 2. Expression stability of the candidate reference genes in ten strawberry cultivars as calculated by the comparative delta-Ctmethod (A), BestKeeper (B), NormFinder (C), geNorm(D),and RefFinder (E).

209V. Galli et al. / Gene 554 (2015) 205–214

reference gene as stably or unstably expressed, this software providea ranking of the candidate genes according to their stability. In thiscase, candidate genes with the lower ranking were considered tobe most stably expressed under the evaluated experimental condi-tions, and thus could be selected as ideal reference genes. Therefore,we used RefFinder as a confirmatory tool to finally select the mostsuitable reference genes, since small differences in the rank of ex-pression stability was observed among the evaluated approaches.As observed in Fig. 2E, the recommended comprehensive rankingof stability in the subset of strawberry cultivars was determined asDBP N UBQ11 N ACT N HISTH4 N PIRUV N GAPDH N 18S. Therefore, theoverall results indicate the novel candidate reference genes as themost suitable to normalize gene expression data with strawberrycultivars.

Considering the subset of osmotic stresses, the most stable gene ac-cording to theΔCtmethod,wasHISTH4, followedUBQ11 (Fig. 3A). How-ever, Bestkeeper indicated DBP, followed by PIRUV (Fig. 3B), and only

these genes showed SD values below the cut-off; geNorm suggestedPIRUV and UBQ11 (Fig. 3C), although all genes showed M value belowthe cut-off; and NormFinder pointed ACT, followed by HISTH4, as themost suitable reference genes (Fig. 3D). Based on these results, the rec-ommended comprehensive ranking of stability in the subset of osmoticstresses, according to the integrative RefFinder analysis was:HISTH4 N UBQ11 N PIRUV N ACT N DBP N 18S N GAPDH. The HISTH4 andtheACTwere also suggested as themost stable reference genes in straw-berry fruits treatedwith benzothiadiazole (BTH) resistance inducers, al-though 18S-RNA and GAPDH2 showed the best expression stability infruits treated with COA (a mixture of calcium and organic acids),while 18S-RNA and ACTweremore stable in fruits treatedwith chitosan(Landi et al., 2014). Similarly, the gene DBP, which was the RefFinderrecommended gene to be used in samples of strawberry cultivar,was one of the worst genes in samples of strawberry under osmoticstresses, according to NormFinder, RefFinder and the comparativeΔCt method. These results further confirm the importance of

Fig. 3. Expression stability of the candidate reference genes in strawberries submitted to osmotic stresses as calculated by the comparative delta-Ct method (A), BestKeeper (B),NormFinder (C), geNorm (D), and RefFinder (E).

210 V. Galli et al. / Gene 554 (2015) 205–214

validating reference genes according to the experimental condi-tions under evaluation.

Different environmental stresses, such as drought and soil sa-linity may result in different metabolic responses in the cell(Wang and Frei, 2011). Therefore, we also evaluated the stabilityof the candidate reference genes in the salt stress (Fig. 4) anddrought stress (Fig. 5) treatments separately. For salt stress, theUBQ11 gene was the best gene according to the ΔCt method(Fig. 4A). The novel candidate reference genes proposed in thisstudy PIRUV, DBP/PIRUV and HISTH4 were the most suitable genesaccording to BestKeeper, geNorm and NormFinder, respectively(Fig. 4B, C, D). Therefore, the RefFinder recommended ranking ofstability was HISTH4 N PIRUV N UBQ11 N DBP N ACT N 18S N GAPDH(Fig. 4E). For drought stress, the ΔCt method, the BestKeeper soft-ware and the NormFinder software indicated DBP as the most sta-ble gene (Fig. 5A, B, D), while geNorm suggested PIRUV andUBQ11 genes (Fig. 5C). According to RefFinder, the recommended

comprehensive ranking of stability under drought stress wasDBP N UBQ11 N PIRUV N HISTH4 N ACT N 18S N GAPDH (Fig. 5E).Therefore, although the most unstable genes were similar betweenboth stresses (salt and drought), the most suitable reference geneswere not the same, confirming that even similar experimental con-dition may affect the cell metabolism in a different way.

Overall, the traditional genes commonly used to normalize RT-qPCRdata, 18S and GAPDH were the most unstable candidate reference genefor both subset of samples and for individual osmotic stress conditions.These candidate reference genes also demonstrated high variation inthe expression levels (Fig. 1). These results indicate that 18S andGAPDH are not suitable to normalize RT-qPCR data in samples fromthe present study. Similarly, in the study performed by Clancy et al.(2013), both NormFinder and geNorm ranked GAPDH as relatively un-stable throughout Fragaria × ananassa fruit development, althoughthis gene was the top candidate in Fragaria vesca tissues according toNormFinder and ranked third by geNorm. GAPDH and 18S were also

Fig. 4. Expression stability of the candidate reference genes in strawberries submitted to salt stress as calculated by the comparative delta-Ctmethod (A), BestKeeper (B), NormFinder (C),geNorm (D), and RefFinder (E).

211V. Galli et al. / Gene 554 (2015) 205–214

unsuitable as reference genes in samples frommaize grains (Galli et al.,2013), lettuce (Borowski et al., 2014) and radish (Xu et al., 2012),among others. Moreover, the ACT gene, another traditional referencegene, showed substantial variation in the present study, as well as be-tween strawberry tissues (Clancy et al., 2013), in potato (Nicot et al.,2005), poplar (Brunner et al., 2004) and rice (Jain et al., 2006). These re-sults further confirm that there is typically no universally applicable ref-erence gene for all experimental conditions, and that the traditionalreference transcripts may not always show constitutive expression ashas often been assumed.

3.3. Minimum number of reference genes

The geNorm program also indicates the minimum number of refer-ence genes for accurate normalization by calculating the pairwise vari-ation Vn/Vn + 1 between two sequential normalization factors to

determine the necessity of adding the next more consistent referencegene. A large variation means that the added gene had a significant ef-fect on the normalization and should preferably be included for calcula-tion (Vandesompele et al., 2002). According to this criterion, the V2/3value was lower than the cut-off value of 0.15 in the subset of strawber-ry cultivars (0.027) and in the subset of osmotic stresses (0.132), sug-gesting that two reference genes are sufficient to normalize geneexpression data in these samples (Fig. 6). However, to perform RT-qPCR analysis using strawberry fruits submitted to salt stress, three ref-erence genes are suggested, since the V3/4 value was 0.142. Althoughonly the V4/5 value in the drought stress treatment reaches the valuelower than 0.15, the use of five reference genes may increase consider-ably the experimental costs. Therefore we suggest that three referencegenes are sufficient to normalize those samples, since the V3/4 valuewas relatively close to 0.15 and the stepwise addition of referencegene does not reduce significantly the V value.

Fig. 5. Expression stability of the candidate reference genes in strawberries submitted to drought stress as calculated by the comparative delta-Ctmethod (A), BestKeeper (B), NormFinder(C), geNorm (D), and RefFinder (E).

212 V. Galli et al. / Gene 554 (2015) 205–214

3.4. Validation of reference genes

The expression of NCED genes (involved in the synthesis of ABA) isinduced during drought stress in several plants, such as Arabidopsis(Iuchi et al., 2001), citrus (Agusti et al., 2007), sweet cherry (Ren et al.,2010), and tomato (Sun et al., 2011), among others. Abiotic stressesalso induce the expression of PAL gene (Jeong et al., 2012; Borowskiet al., 2014), related to the synthesis of several antioxidant compoundsfrom the phenylpropanoid metabolic pathway, which act as radicalscavenger, thus helping to reduce cell damages caused by the stress.Therefore, to detect the effect of using different reference genes in thedata normalization, we evaluated the relative expression of FaNCED1and FaPAL in strawberry fruits subjected to drought stress. Themost sta-ble (DBP, UBQ11 and PIRUV) and the least stable (GAPDH) referencegenes, according to RefFinder, were used as internal controls.

As shown in Fig. 7, when we used the set of three most stable refer-ence genes to normalize the expression of FaNCED1 and FaPAL in

strawberry fruits under drought stress, we observe that both genes areindeed upregulated. However, when we used the most unstable refer-ence gene in the analysis, their expression is not regulated comparedto control. Therefore, the use of unsuitable references leads to differ-ences in the relative expression profile. These results further confirmedthe importance of validating reference genes prior to experimentalapplications.

4. Conclusions

In this study, we evaluated seven candidate reference genes for thenormalization of gene expression in strawberry cultivars and under os-motic stresses (salt and drought). Our results suggested that differentsuitable reference genes should be selected case by case, according tothe experimental condition under evaluation, since the most stable ref-erence genes for the subset of strawberry cultivars were not the same asfor the subset of osmotic stresses. In our work, the traditional genes

Fig. 6. Pairwise variation (V) calculated by geNorm to determine theminimumnumber of reference genes for accurate normalization in samples from strawberry cultivars (A), and straw-berry under osmotic stresses (B), salt stress (C) and drought stress (D).

213V. Galli et al. / Gene 554 (2015) 205–214

GAPDH and 18S were considered the most unstable genes in both sub-sets of samples. Otherwise, the novel candidate reference gene DBPwas considered the most suitable to normalize expression data in fruitsamples of strawberry cultivars and under drought stress, and thenovel candidate reference gene HISTH4 was the most stable under os-motic stresses and salt stress, according to RefFinder. The use ofRefFinder as a confirmatory toll clearly improved the decision for themost stable genes to be selected in further studies. The expression anal-ysis of FaPAL and FaNCED1 confirmed the importance of validating refer-ence genes to achieve accurate RT-qPCR results. This study will be ofgreat importance in further analyses of gene expression in order to un-derstand the mechanisms highlighting stress tolerance in strawberryplants for future breeding efforts.

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.gene.2014.10.049.

Fig. 7. Expression profile of FaPAL and FaNCED1 in strawberry fruits cultivated under drought ststable reference gene (GAPDH) were used to normalize the expression data. Samples from the cstandard deviation calculated from three biological replicates.

Author's contributions

For this study, authors V.G., J.M.B. and E.C.P. carried out the experi-ment and executed the analyses; I.S.P. and J.L. carried out parts of the ex-periment. R.S.M., S.D.A.S. and C.V.R. designed the experiment anddirected the work. All authors contributed to the writing of the article.

Conflict of interest

No conflict of interest declared.

Acknowledgments

The authors gratefully acknowledge the technical and financial sup-port of the Fundação de Amparo à Pesquisa do Estado do Rio Grande do

ress. The best stable combination of reference genes (DBP, UBQ11 and PIRUV) and the leastontrol treatment were used as reference samples. The data show themean expression ±

214 V. Galli et al. / Gene 554 (2015) 205–214

Sul (FAPERGS) (grant number 1636-2551/13-4) and Conselho Nacionalde Desenvolvimento Científico e Tecnológico (CNPq)/ Embrapa ClimaTemperado (grant number 484951/2013-0 and 150078/2014-5).

References

Agusti, J., Zapater, M., Iglesias, D.J., Cercós, M., Tadeo, F.R., Talón, M., 2007. Differential expres-sion of putative 9-cis-epoxycarotenoid dioxygenases and abscisic acid accumulation inwater stressed vegetative and reproductive tissues of citrus. Plant Sci. 172, 85–94.

Amil-Ruiz, F., Garrido-Gala, J., Blanco-Portales, R., Folta, K.M., Muñoz-Blanco, J., Caballero,J.L., 2013. Identification and validation of reference genes for transcript normalizationin strawberry (Fragaria × ananassa) defense responses. PLoS One http://dx.doi.org/10.1371/journal.pone.0070603.

Andersen, C.L., Jensen, J.L., Orntoft, T.F., 2004. Normalization of real-time quantitative re-verse transcription-PCR data: a model-based variance estimation approach to identi-fy genes suited for normalization, applied to bladder and colon cancer data sets.Cancer Res. 64, 5245–5250.

Borowski, J.M., Galli, V., Messias, RdS, Perin, E.C., Buss, J.H., Anjos, E., Silva, SDd, Rombaldi,C.V., 2014. Selection of candidate reference genes for real-time PCR studies in lettuceunder abiotic stress. Planta http://dx.doi.org/10.1007/s00425-014-2041-2.

Brunner, A.M., Busov, V.B., Strauss, S.H., 2004. Poplar genome sequence: functional geno-mics in an ecologically dominant plant species. Trends Plant Sci. 9, 49–56.

Bustin, S.A., 2002. Quantification of mRNA using real-time reverse transcription PCR (RT-PCR): trends and problems. J. Mol. Endocrinol. 29, 23–39.

Bustin, S.A., Benes, V., Garson, J.A., et al., 2009. The MIQE guidelines: minimum informationfor publication of quantitative real-time PCR experiments. Clin. Chem. 55, 611–622.

Capocasa, F., Scalzo, J., Mezzeti, B., Battino, M., 2008. Combining quality and antioxidantattributes in the strawberry: the role of genotype. Food Chem. 111, 872–878.

Chen, I.-C., Hill, J.K., Ohlemüller, R., Roy, D.B., Thomas, C.D., 2011. Rapid range shifts of spe-cies associated with high levels of climate warming. Science 333, 1024–1026.

Christou, A., Manganaris, G.A., Papadopoulos, I., Fotopoulos, V., 2013. Hydrogen sulfide in-duces systemic tolerance to salinity and non-ionic osmotic stress in strawberry plantsthrough modification of reactive species biosynthesis and transcriptional regulationof multiple defence pathways. J. Exp. Bot. 64, 1953–1966.

Christou, A., Manganaris, G.A., Fotopoulos, V., 2014. Systemic mitigation of salt stress by hy-drogen peroxide and sodium nitroprusside in strawberry plants via transcriptional reg-ulation of enzymatic and non-enzymatic antioxidants. Environ. Exp. Bot. 107, 46–54.

Clancy, M.A., Rosli, H.G., Chamala, S., Barbazuk, W.B., Civello, P.M., Folta, K.M., 2013. Vali-dation of reference transcripts in strawberry (Fragaria spp.). Mol. Gen. Genomics.288, 671–681.

Cqfs. Comissão de química e fertilidade do solo- RS/SC, 2004. Manual de adubação ecalagem para os estados do Rio Grande do Sul e Santa Catarina, 10th ed. SociedadeBrasileira de Ciência do Solo/Núcleo Regional Sul/UFRGS, Porto Alegre (400 pp.).

Cruz-Rus, E., Amaya, I., Sánchez-Sevilla, J.F., Botella, M.A., Valpuesta, V., 2011. Regulationof L-ascorbic acid content in strawberry fruits. J. Exp. Bot. 62 (2), 4191–4201.

Czechowski, T., Stitt, M., Altmann, T., Udvardi, M.K., Scheible, W.-R., 2005. Genome-wideidentification and testing of superior reference genes for transcript normalization inArabidopsis. Plant Physiol. 139, 5–17.

Derveaux, S., Vandesompele, J., Hellemens, J., 2010. How to do successful gene expressionanalysis using real-time PCR. Methods 50, 227–230.

Erkan, M., Wang, S.Y., Wang, C.Y., 2008. Effect of UV treatment on antioxidant capacity,antioxidant enzyme and decay in strawberries fruit. Postharvest Biol. Technol. 48(2), 163–171.

Fleige, S., Pfaffl, M.W., 2006. RNA integrity and the effect on the real-time qRT-PCR perfor-mance. Mol. Asp. Med. 7, 126–139.

Galli, V., Silva, M.R., Anjos e Silva, S.D., Rombaldi, C.V., 2013. Selection of reliable referencegenes for quantitative real-time polymerase chain reaction studies in maize grains.Plant Cell Rep. 32, 1869–1877.

Huang, G.T., Ma, S.L., Bai, L.P., Zhang, L., Ma, H., Jia, P., Liu, J., Zhong, M., Guo, Z.F., 2012. Sig-nal transduction during cold salt, and drought stresses in plants. Molecular BiologyReports 39, 969–987.

Husaini, A.M., Abdin, M.Z., 2008. Development of transgenic strawberry(Fragaria × ananassa Duch.) plants tolerant to salt stress. Plant Sci. 174, 446–455.

Hytönen, T., Elomaa, P., Moritz, T., Junttila, O., 2009. Gibberellin mediates daylength-con-trolled differentiation of vegetative meristems in strawberry (Fragaria × ananassaDuch). BMC Plant Biology 9, 18.

Iuchi, S., Kobayashi, M., Taji, T., Naramoto, M., Seki, M., Kato, T., Tabata, S., Kakubari, Y.,Yamaguchi-Shinozaki, K., Shinozaki, K., 2001. Regulation of drought tolerance bygene manipulation of 9-cis-epoxycarotenoid dioxygenase, a key enzyme in abscisicacid biosynthesis in Arabidopsis. Plant J. 27, 325–333.

Jain, M., Nijhawan, A., Tyagi, A.K., Khurana, J.P., 2006. Validation of housekeeping genes asinternal control for studying gene expression in rice by quantitative real-time PCR.Biochem. Biophys. Res. Commun. 345, 646–651.

Jeong, M.J., Choi, B.S., Bae, D.W., Shin, S.C., Un, P.S., Hyoun-Sub, L., Kim, J., Kim, J.B., Cho,B.K., Hanhong, B., 2012. Differential expression of kenaf phenylalanine ammonia-lyase (PAL) ortholog during developmental stages and in response to abiotic stresses.Plant OMICS 5, 392–399.

Jia, H., Chai, Y., Li, C., Lu, D., Luo, J., Qin, L., Shen, Y., 2011. Abscisic acid plays an importantrole in the regulation of strawberry fruit ripening. Plant Physiol 157, 188–199.

Krasensky, J., Jonak, C., 2012. Drought, salt, and temperature stress-inducedmetabolic re-arrangements and regulatory networks. Journal of Experimental Botany 63,1593–1608.

Landi, L., Feliziani, E., Romanazzi, G., 2014. Expression of defense genes in strawberryfruits treated with different resistance inducers. J. Agric. Food Chem. 62, 3047–3056.

Le, D.T., Aldrich, D.L., Valliyodan, B., Watanabe, Y., Ha, C.V., 2012. Evaluation of candidate ref-erence genes for normalization of quantitative RT-PCR in soybean tissues under variousabiotic stress conditions. PLoS One http://dx.doi.org/10.1371/journal.pone.0046487.

Livak, K.J., Schmittgen, T.D., 2001. Analysis of relative gene expression data using real-time quantitative PCR and the 2−ΔΔCt method. Methods 25, 402–408.

Lunkenbein, S., Bellido, M.L., Aharoni, A., Salentijn, E.M.J., Kaldenhoff, R., Coiner, H.A.,Muñoz-Blanco, J., Schwab, W., 2006. Cinnamate Metabolism in Ripening Fruit. Char-acterization of a UDP-Glucose:Cinnamate Glucosyltransferase from Strawberry. PlantPhysiology 140, 1047–1058.

Marouelli, W.A., Silva, W.L.C., Silva, H.R., 2008. Irrigação por aspersão em hortaliças:qualidade da água, aspectos do sistema e método prático de manejo, 2nd ed.Embrapa Informação Tecnológica: Embrapa Hortaliças, Brasília, DF, p. 150.

Maroufi, A., Van Bockstaele, E., De Loose, M., 2010. Validation of reference genes for geneexpression analysis in chicory (Cichorium intybus) using quantitative real-time PCR.BMC Mol. Biol. 11, 15.

Mazur, S.P., Nes, A., Wold, A., Remberg, S.F., Martinsen, B.K., Aaby, K., 2014. Effects of ripe-ness and cultivar on chemical composition of strawberry (Fragaria× ananassaDuch.)fruits and their suitability for jam production as a stable product at different storagetemperatures. Food Chem. 246, 412–422.

Mehli, L., Schaart, J.G., Kjellsen, T.D., Tran, D.H., Salentijn, E.M.J., Schouten, H.J., Iversen, T.-H., 2004. A gene encoding a polygalacturonase-inhibiting protein (PGIP) shows de-velopmental regulation and pathogen-induced expression in strawberry. New Phytol.163 (Issue 1), 99–110.

Messias, R.S., Galli, V., Borowski, J., Buss, J., Nora, L., Margis, R., Rombaldi, C.V., 2014. Isola-tion of high quality RNA from grains of different maize varieties. Prep. Biochem.Biotechnol. http://dx.doi.org/10.1080/10826068.2013.868355.

Mezzetti, B., 2009. GMO strawberry: methods, risk and benefits. In: Folta, K.M., Gardiner,S.E. (Eds.), Genetics and Genomics of Rosaceae. Springer Science Business Media, LLC.

Nicot, N., Hausman, J.F., Hoffmann, L., Evers, D., 2005. Housekeeping gene selection forreal-time RT-PCR normalization in potato during biotic and abiotic stress. J. Exp.Bot. 56, 2907–2914.

Orsini, F., Alnayef, M., Bona, S., Maggio, A., Gianquinto, G., 2012. Environ. Exp. Bot. 81, 1–10.Padula, M.C., Lepore, L., Milella, L., Ovesna, J., Malafronte, N., Martelli, G., de Tommasi, N.,

2013. Cultivar based selection and genetic analysis of strawberry fruits with highlevels of health promoting compounds. Food Chem. 140, 639–646.

Pfaffl, M.W., Tichopad, A., Prgomet, C., Neuvians, T.P., 2004. Determination of stablehousekeeping genes, differentially regulated target genes and sample integrity:Bestkeeper-Excel-based tool using pair-wise correlations. Biotechnol. Lett. 26,509–515.

Pineli, L. de O., Moretti, C.L., Santos, M.S. dos, Campos, A.B., Brasileiro, A.V., Córdova, A.C.,Chiarello, M.D., 2011. Antioxidants and other chemical and physical characteristics oftwo strawberry cultivars at different ripeness stages. J. Food Compos. Anal. 24, 11–16.

Radonic, A., Thulke, S., Mackay, I.M., Landt, O., Siegert, W., Nitsche, A., 2004. Guideline toreference gene selection for quantitative real-time PCR. Biochem. Biophys. Res.Commun. 313, 856–862.

Ramakers, C., Ruijter, J.M., Deprez, R.H., 2003. Assumption-free analysis of quantitativereal-time polymerase chain reaction (PCR) data. Neurosci. Lett. 339, 62–66.

Ren, J., Sun, L., Wu, J., Zhao, S., Wang, C., Wang, Y., Ji, K., Leng, P., 2010. Cloning and expres-sion analysis of cDNAs for ABA 8′-hydroxylase during sweet cherry fruit maturationand under stress conditions. J. Plant Physiol. 17, 1486–1493.

Schaart, J.G., Salentijn, E.M.J., Krens, F.A., 2002. Tissue-specific expression of the β-glucu-ronidase reporter gene in transgenic strawberry (Fragaria × ananassa) plants. PlantCell Reports 21, 313–319.

Silver, N., Best, S., Jiang, J., Thein, S.L., 2006. Selection of housekeeping genes for gene ex-pression studies in human reticulocytes using real-time PCR. BMC Mol. Biol. 7, 33.

SAS Statistical Analysis System, 2000. SAS Users Guide: Statistics. SAS Institute, Cary, NCUSA.

Sun, L., Wang, Y.P., Chen, P., Ren, J., Ji, K., Li, Q., Li, P., Dai, S.J., Leng, P., 2011. Transcriptionalregulation of SlPYL, SlPP2C, and SlSnRK2 gene families encoding ABA signal core com-ponents during tomato fruit development and drought stress. J. Exp. Bot. 62,5659–5669.

Van den Ende, W., El-Esawe, S.K., 2014. Sucrose signaling pathways leading to fructan andanthocyaninaccumulation: A dual function in abiotic and biotic stress responses? En-vironmental and Experimental Botany 108, 4–13.

Vandendriessche, T., Vermeir, S., Mayayo Martinez, C., Hendrickx, Y., Lammertyn, J.,Nicolay, B.M., Hertog, M. La T.M., 2013. Effect of ripening and inter-cultivar differ-ences on strawberry quality. LWT Food Sci. Technol. 52, 62–70.

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 aver-aging of multiple internal control genes. Genome Biol. 3, 0034.1–0034.11.

Wan, H., Zhao, Z., Qian, C., Shui, Y., Malik, A.A., Chen, J., 2010. Selection of appropriate ref-erence genes for gene expression studies by quantitative real-time polymerase chainreaction in cucumber. Anal. Biochem. 399, 257–261.

Wang, Y., Frei, M., 2011. Stressed food — the impact of abiotic environmental stresses oncrop quality. Agric. Ecosyst. Environ. 141, 271–286.

Wei, L., Miao, H., Zhao, R., et al., 2013. Identification and testing of reference genes for Ses-ame gene expression analysis by quantitative real-time PCR. Planta 237, 873–889.

Wong, M.L., Medrano, J.F., 2005. Real-time PCR for mRNA quantitation. BioTechniques 39, 1.Xu, Y., Zhu, X., Gong, Y., Xu, L., Wang, Y., Liu, L., 2012. Evaluation of reference genes for

gene expression studies in radish (Raphanus sativus L.) using quantitative real-timePCR. Biochem. Biophys. Res. Commun. 424 (3), 398–403.

Zhu, C., Sanahuja, G., Yuan, D., Farré, G., Arjó, G., Berman, J., Zorrilla-López, U., Banakar, R.,Bai, C., Pérez-Massot, E., Bassie, L., Capell, T., Cristou, P., 2013. Biofortification of plantswith altered antioxidant content and composition: genetic engineering strategies.Plant Biotechnol. J. 11 (Issue 2), 129–141.