Biologia 68/6: 1211—1220, 2013Section ZoologyDOI: 10.2478/s11756-013-0267-2

Identification and validation of reference genes for real time PCRexpression studies in heads of nurse honeybees

Lenka Kohútová1, Jaroslav Klaudiny1*, Róbert Nádašdy2, Mária Šedivá1,Ján Kopernický2 & Juraj Majtán3

1Institute of Chemistry, Slovak Academy of Sciences, Dúbravská cesta 9, 84538 Bratislava, Slovakia;e-mail: jaroslav.klaudiny@savba.sk2Animal Production Research Centre, Institute of Apiculture, Gašperíkova 599, 03308, Liptovský Hrádok, Slovakia3Institute of Zoology, Slovak Academy of Sciences, Dúbravská cesta 9, 84506 Bratislava, Slovakia

Abstract: The primary aim of this study was to identify reference genes and workers of particular role and ages that wouldbe suitable for exploring genetic/epigenetic variations in constitutive expression of a gene encoding antimicrobial peptidedefensin1 in worker heads using real-time PCR. This peptide is an integral component of larval food and honey and haspotential to act against some brood pathogens. Expression levels of distinct genes may vary in worker heads due to geneticfactors, age of bee, and particular role of a worker that depends on its age or colony needs. Prerequisite for exploring thevariations in defensin1 expression was therefore to identify such workers in which correlated expression of defensin1 andsuitable reference genes occurs. Selection process was done by carefully designed quantitative real-time PCR procedurein two colonies showing different age-related division of labor. Expression of ten candidate reference genes, defensin1 andamylase, as a marker of forager bees, was assessed in pooled head samples of workers aged 2 to 30 days. Correlatedand moreover stable expression of defensin1 and six candidate genes was detected in nursing bees in both colonies. Thesuitable reference genes were therefore selected on the basis of their expression stability. This was evaluated by geNormand NormFinder algorithms in pooled head samples and through plotted Cq data in head samples of individual nurse bees.As the best reference genes were selected: psa1, tctp1, cyclophilin, gapdh and mrjp4 (in this order). They are suitable foraforementioned defensin1 expression studies and also for studies of other genes expressed in heads of nurses. In addition,an amylase expression-based procedure for reliable distinguishing nurses from foragers was elaborated.

Key words: Quantitative real-time PCR; Apis mellifera; age-related division of labor; gene expression; defensin1; larvaljelly; honey

Introduction

The vast majority of the genetically heterogeneous pop-ulation of a honeybee hive consists of workers that per-form most tasks in the colony. A particular task is usu-ally performed by workers of certain age: cell cleaningand capping by the youngest bees, nursing (brood andqueen care/feeding) by 4–16 days old bees, food han-dling and comb building by middle aged bees and ven-tilating, guarding and foraging by older bees between15–30 days (Winston 1987). This age-related divisionof labor in colony is not a rigid system. A worker isable to modify its role in response to colony needs.Role changes can be triggered by specific conditions,such as amount and type of available forage as well asa shortage of workers performing a specific task in thecolony (Page et al. 1989; Robinson 1992; Robinson etal. 1992; Huang & Robinson 1996). Genetic differencesbetween bee races and strains were also described as afactor that can cause differences in foraging age (Win-

ston & Katz 1982; Calderone & Page 1987). Transi-tions between tasks are accompanied with changes inexpression of various genes participating in exocrine,endocrine and brain activities (Whitfield et al. 2003;Cash et al. 2005; Liu et al. 2011). A good example ofexpression changes are hypopharyngeal glands. Maturehypopharyngeal glands in nurse heads synthesize largeamounts of nutritive major royal jelly proteins (Kuboet al. 1996; Ohashi et al. 1997; Schmitzova et al. 1998),whereas in foragers, the glands shrink and predomi-nantly synthesize carbohydrate-metabolizing enzymes,such as α-glucosidase, α-amylase and glucose oxidase(Kubo et al. 1996; Ohashi et al. 1997, 1999). Expres-sion of some genes can change not only with differentroles but also at maturation of workers for a partic-ular role what can comprise a period of several days.Such examples are changing expression levels of genesencoding royal jelly proteins MRJP1 and MRJP3 in de-veloping hypopharyngeal glands of nurses (Feng et al.2009).

* Corresponding author

c©2013 Institute of Zoology, Slovak Academy of Sciences

1212 L. Kohútová et al.

Both nurses and foragers produce antimicrobialpeptide defensin1 (Liu et al. 2011; Klaudiny et al.2012; Majtan et al. 2012). Nurses synthesize the pep-tide in their hypopharyngeal and mandibular glands(located in head) from which it is secreted into lar-val food – larval jelly (LJ) (Klaudiny et al. 2005).In foragers, it is possible to assume that at least hy-popharyngeal glands synthesize the peptide and secreteit into honey. Defensin1 is active in vitro against thepathogen Paenibacillus larvae indicating that it couldcontribute to colony resistance against a serious bacte-rial brood disease, American foulbrood (Bíliková et al.2001; Bachanová et al. 2002). This disease causes sub-stantial financial losses in beekeeping and in agricul-tural production dependent on pollination. Our recentinvestigation of defensin1 expression in colonies, at theprotein and transcript levels, showed that defensin1 wasa constitutive component of LJs, but at the same timeintercolonial and intracolonial differences were observedin defensin1 contents in number of analyzed LJ samplesas well as in defensin1 levels in RNA samples isolatedfrom nurse heads. The obtained data indicated that de-fensin1 levels in LJs could be determined by genetic fac-tors regulating defensin1 transcriptional and/or post-transcriptional processes in heads of nurses. Moreover,the data suggested that it should be possible to breednovel honeybee lines expressing higher amounts of de-fensin1 in LJs (Klaudiny et al. 2012).For practical breeding and general knowledge, it is

interesting to explore genetic (epigenetic) factors deter-mining high constitutive defensin1 expression in headsof some workers in bee population. Such an investiga-tion involves identification of bees expressing increasedconstitutive levels of defensin1 in colonies. A quanti-tative real-time PCR (qPCR) is the method of choicefor this purpose. Genetic (epigenetic) factors are how-ever only one of factors that may influence constitu-tive defensin1 expression in a worker head. Other fac-tors are: age of bee, role of bee, maturation state ofbee within a role and colony needs as was mentionedabove on examples of expression of some genes. Theinfluence of single factors on the expression of head de-fensin1 during worker life is difficult to assess. In thiscase, for intended defensin1 expression analyses, it isnecessary besides suitable reference genes to identifyand use specific workers in which the reference genesand defensin1 have the same expression profile (cor-related expression). At analyses under such conditions,the normalization of defensin1 expression with the suit-able reference genes enables to eliminate a contributionof “not genetic” factors to expression and to measurevariations in constitutive defensin1 expression levels inworkers caused only by genetic (epigenetic) factors.In this work, expression of defensin1 and another

11 genes in heads of workers was analyzed across abroad age range of workers in two colonies to find refer-ence genes and workers in which correlated expressionof suitable reference genes and defensin1 occurs. Theprofiles of transcription levels of all genes were com-pared in samples of pooled worker heads in one colony.

The correlated and also stable expression of six genesand defensin1 was found to occur in nursing bees ofvarious ages. The same expression profile of these genesoccurred in nurses in the second examined colony hav-ing different age-related division of labor. Therefore, thereference quality of the genes was assessed on the basisof their expression stability. Based on obtained results,the most suitable reference genes for investigation ofexpression not only of defensin1 but also of other nursehead genes were selected. In addition, some referencegenes and amylase expression were validated for distin-guishing nurses from foragers.

Material and methods

Honeybee samplesApis mellifera carnica workers were obtained from naturallymated queen colonies in the apiary of the Institute of Api-culture in Liptovský Hrádok, Slovakia. They were collectedfrom two hives in July 2012. The age of bees was determinedby color labeling of 11–19 h old workers which emerged onisolated combs maintained in an incubator at 34◦C. Workersof particular ages were gathered into plastic tubes, frozen inliquid nitrogen and stored at –80◦C.

RNA isolation and cDNA synthesisTotal RNA was isolated either from pool of heads or in-dividual heads of workers using the SiMAX total RNAcolumns isolation kit (Ecoli, Slovakia, www.ecoli.sk) em-ploying guanidine thiocyanate-acid phenol extraction in thefirst step. In the first case, heads were cut off from frozenbees and pooled samples, each containing 4 heads, were pul-verized in liquid nitrogen with a mortar and a pestle, placedinto microcentrifuge tubes and kept at –80◦C before iso-lation. In the case of single heads, each frozen head washalved between the eyes, immediately ground first in 80 µlof cold denaturing solution (4◦C) using a plastic pestle andthen immediately processed. Powdered pooled and groundhead samples were each homogenized in a denaturing so-lution (850 µl and 250 µl, respectively) using an ULTRA-TURRAX followed by further procedure according to isola-tion kit manual. In addition to the manufacturer’s recom-mendations, RNA eluted from a column was incubated at58◦C for 12 min to ensure dissolution of RNA aggregatesand then centrifuged at 15,000 g for 3 min to remove anymaterial released from columns. RNA purity and concen-tration were estimated spectrometrically and its integritydetermined by agarose gel electrophoresis.

First strand cDNA was synthesized from 1.6 µg to-tal RNA using the Thermo reverse transcription cDNA kit(Ecoli, Slovakia, www.ecoli.sk) with oligo (dT)18 as primer.The reaction (20 µl) was performed in a thermal cycler at50◦C for one hour.

Gene selection, primer design, amplification conditions andstrategySix genes encoding proteins secreted by nurses into LJ:mrjp3, mrjp4, mrjp5 (Klaudiny et al. 1994; Schmitzova etal. 1998; Albert et al. 1999a, b), apisimin (Bíliková et al.2002), regucalcin and tctp1 (Han et al. 2011) were selectedas candidate reference genes for qPCR with an expecta-tion that some of them have correlated expression with de-fensin1 gene at least during the period of nursing.Amylase isabundantly expressed in foragers and very weakly in nurses(Ohashi et al. 1999; Liu et al. 2011) and it was selected as

Reference genes for qPCR in nurse bee heads 1213

Table 1. Genes and primers used for real-time PCR.

AmpliconGene name and symbol Accession number Primer sequence (5’→3’) size (bp) E (%) a R2 b

actin related protein 1(actin)

NM 001185145.1NM 001185146.1 c

GAAGCTTGCGGTATTCATGAGACGTGGATGGTGCTAGGGCAGTG

164 104.1 0.999

alpha-amylase(amylase)

NM 001011598.1 GATAATTGGTGGGATAACGGAAGCTATTGGCCAGGAGGTAGACAGAC

128 96.8 1.0

apisimin NM 001011582.1 CCATGTTGGTCAGCGATGTGTCAGGAGTACTGCTGACACGTTGG

133 93.5 1.0

cyclophilin XM 393381.4 GTATTCTTTCCATGGCTAATGCTGGCCTCCAATTTTCTAACAACATCC

145 90.8 1.0

defensin 1 NM 001011616.2 CTTCTCTTCATGGCTATGGTTGCCAAGTTACTCTTCTATGTCTGTCG

113 95.6 1.0

glyceraldehyde-3-phosphatedehydrogenase (gapdh)

XM 393605.4 TTGTTGACTTAACAGTTAGACTTGGTTTAGCATCAAAAATACTGGCATGG

170 98 1.0

major royal jelly protein 3(mrjp3)

NM 001011601.1 TGGGACTCGTGGGTAATTCAGGTGCCTTCAGGATCATTAATTCTGC

168 95.2 1.0

major royal jelly protein 4(mrjp4)

NM 001011610.1 TTCCCACACTTTGAATCACAACTCGAGAATTGTAATAAAGATTATGCGTC

131 95.3 1.0

major royal jelly protein 5(mrjp5)

NM 001011599.1 CATTCATGAAATCAGAATATGGAGCACCAATAGCTGAATTATTCATGAGG

140 95.3 0.999

proteasome subunit alphatype-1 (psa1)

XM 625029.3 ATGATGATCTATTACCTGTGAGTCGGTGGTCCTTGATCATCATATCCAG

132 100.3 1.0

regucalcin XM 001121327.2 CTTTGGATAGCTTTATATGGTGGAGCTAATAGAGGTCCTCCAAATGTGC

127 97.5 1.0

translationally-controlledtumor protein 1 (tctp1)

XM 395299.4 GTTCTATATGAAGTATATGGCAAGGATGATTCATAACTATATCAACACCAG

141 91.1 1.0

Explanations: a Real-time PCR reaction efficiency calculated by the standard curve method. b Coefficient of determination for thestandard curve. c Primers were designed in the sequence region common for both transcript variants.

a marker for foragers. Gapdh, actin and cyclophilin were se-lected as commonly used reference genes (Scharlaken et al.2008; Lourenco et al. 2008; Pfaffl 2004), whereas psa1 dueto its presumed stable expression in workers of various age(Feng et al. 2009).

Primers with higher Tm (60–70◦C) (Table 1) were de-signed within open reading frames; in most cases, the ampli-cons span an intron or overlap exon/intron junction (in oneprimer of a pair) to distinguish genomic DNA contamina-tion in RNA/cDNA samples. Amplification conditions em-ploying primers with increased Tm and a reduced annealingtemperature were used in order to reduce a possible impacton qPCR of sequence variations within primer binding sites(Boyle et al. 2009). This strategy reflected our experiencewith a negative effect of a SNP on mrjp4 amplification (un-published results).

In pooled head samples, selected genes were ampli-fied in sets of samples, each set contained samples originat-ing from workers of different age (2–30 days). Samples ofeach set were processed in parallel during RNA preparationand cDNA synthesis. Subsequent qPCR amplifications weredone gradually with cDNA samples of each set in 96 wellplates as follows. All examined genes were amplified fromeach cDNA sample in one run on one plate to minimizeqPCR variations caused by manipulations of the sample indifferent experiments; thus, the methodology resembles amultiplex qPCR.

Real-time PCRThe qPCR reactions were performed with HOT FIRE-Pol EvaGreen qPCR mix plus (no ROX) (Ecoli, Slovakia,www.ecoli.sk). Each sample was analyzed in duplicate in areaction of 15 µl consisting of 5 µl of diluted cDNA (1:50in water) and 10 µl of a reaction mix containing 3 µl qPCR

mix and primers (0.4 µM final concentrations in reaction).The qPCRs were carried out on CFX96 thermocycler (Bio-Rad, www.bio-rad.com) using the following conditions: 95◦Cfor 15 min and 40 cycles at 95◦C for 15 s, 58◦C for 20 s,69◦C for 20 s. The specificity of amplification was moni-tored by melting curve analysis, electrophoretic determina-tion of amplicon sizes on 2.5% agarose gel and sequencingof PCR products (reactions without fluorescent dye) aftertheir purification by Wizard SV gel and PCR clean-up sys-tem (Promega, www.promega.com). Control reactions withRNA templates (no RT control) and primers for actin genewere performed to evaluate DNA contamination in RNApreparations (none or not significant contaminations werefound). No-template control reactions were performed foreach primer pair. Standard curves using five or four pointsof template cDNA 10-fold dilutions were constructed to de-termine amplification efficiencies for each gene.

Data analysisQuantification cycle (Cq) values were determined using theBiorad CFX Manager 2.1 software of CFX96 that calcu-lated the arithmetic mean of two replicates for each sam-ple and each gene. This software was also employed for theestimation of (1) relative normalized transcriptional levelof an examined gene in samples of all sets of one colonyrelated to the lowest value, and (2) relative quantity of atarget gene in samples of all sets of one colony related tothe highest expressed sample (having the lowest Cq value).The second type of data and the raw Cq values were usedfor the evaluation of expression stability of genes acrosssome samples using respectively the freely available pro-grams NormFinder (version 0.953, Andersen et al. 2004) andgeNormPLUS (part of qbasePLUS software, version 2.4, Van-desompele et al. 2002; Hellemans et al. 2007). Statistical

1214 L. Kohútová et al.

Fig. 1. Raw Cq values for defensin1 gene in pooled head samples of workers of various ages from colony A and colony B. Three setsof samples from each colony were analyzed.

analyses were performed with one-way ANOVA and cor-responding graphs were generated using GraphPad Prism(www.graphpad.com). Some data were expressed as meanvalues ± standard error of the mean (SEM).

Results

Amplification of genesAll genes were amplified under optimized amplificationconditions with high amplification efficiencies consis-tent with the standard requirements for qPCR (Ta-ble 1). Primer specificities were confirmed by the se-quencing of PCR products and determination of am-plicon sizes. Amplification specificities of each qPCRreaction were routinely determined by analyzing melt-ing temperatures of the products.

Transcription profiles of all examined genes across dif-ferent ages of workers from the colony ATranscript levels of defensin1, amylase (the marker genefor foragers) and ten candidate reference genes wereassessed in samples consisting of pooled heads, eachsample prepared from four workers of a particular age(2–30 days). Three sets of samples comprising this agerange were analyzed. Similar levels of defensin1 mRNA(similar Cq values) were detected in the analyzed sam-ples, although slightly increased levels (lower Cq val-ues) were observed in the youngest workers (< 8 days)and older workers (>20 days) (Fig. 1). To find amongstcandidate reference genes those with expression corre-lating with that of defensin1, their transcription levelsand also those of amylase gene were normalized to de-fensin1 and compared in all sets of samples of differentage (Fig. 2). Basically two types of profiles were found:(i) those having a peak form, demonstrated by uncorre-lated expression (mrjp3, mrjp5, regucalcin, apisimin),and (ii) those displaying correlated expression over a pe-riod 6–20 days (psa1, tctp1, gapdh, cyclophilin, mrjp4,actin). The correlated expression resulted from the factthat defensin1 (Fig. 1) and corresponding six genes (Cqdata not shown) were stably expressed in the work-ers of these ages. The 6–20 day age corresponds to theperiod of low expression of amylase, but excludes theyoungest bees. This indicated that correlated and sta-

ble expression occurred in nurses. During the periodof 20–30 days the correlation was lost, the normalizedtranscription levels of the genes initially decreased andsubsequently increased in the oldest workers for somegenes (psa1, gapdh, cyclophilin, actin). This period wascharacterized by high amylase transcription, a charac-teristic feature of foragers.

Transcription profiles of eight genes in workers from thecolony BTranscriptional levels of those six candidate referencegenes shown in colony A to have expression correlatedwith that of defensin1, as well as of the defensin1 andthe amylase genes were further assessed by an identicalapproach with samples obtained from the colony B. Theobtained results confirmed correlation between the ex-pression of these genes and defensin1 as well as the sta-ble expression of defensin1 (Fig. 1) and these genes (Cqdata not shown) in nurses. The correlations extendedover shorter age range (4–12 days) in this colony (Fig. 3)as compared to the colony A (6–20 days; Fig. 2). Theloss of the correlation was associated with an increaseof amylase transcription. This result confirmed amylasemRNA as a suitable marker for differentiation betweennurses and foragers. Important observation was thatthe earlier transition to foraging in workers of colony Bdid not affect the pattern of normalized transcriptionlevels of the genes observed in the colony A; higher val-ues in the youngest bees, similar values in nurses, lowerin “younger” foragers and then progressively increasingwith aging of foragers.

Expression of selected genes in individual workersThe expression stabilities of eight genes analyzed incolonies A and B were further examined in individualnurses and forages. For this purpose, the expression ofthe genes was analyzed in individual workers from bothcolonies aged 10, 20 or 22 days (11 workers per eachgroup). The highest expression stability was observedin nurses for psa1, mrjp4, and tctp1 (the difference be-tween max. and min. Cq values within both colonies≤1.66) followed by gapdh and cyclophilin (≤1.88). Thelowest stability was observed for defensin1, actin andamylase (Fig. 4). The expression stability was evaluated

Reference genes for qPCR in nurse bee heads 1215

Fig. 2. Relative transcription levels of ten candidate reference genes and of the amylase gene normalized to those of defensin1 in pooledhead samples of workers of various ages from colony A. The scales of transcription levels of each gene were individually adjusted tosee optimally the expression differences.

only in nurses in colony A as it was found that 20 days-old workers were not foragers as had been thought attheir collection. Nevertheless, this fact enabled us to as-sess the expression stability on larger number of nurses.The expression stability in foragers was thus evaluatedonly in colony B and it was lower than in nurses in thecase of the examined genes.

Selection of the best reference genes for exploring tran-scription of defensin1 and other genes expressed innurse headsThe analyses above revealed that six candidate ref-

erence genes (psa1, tctp1, gapdh, cyclophilin, mrjp4,actin) had transcription profiles reasonably correlatingwith that of defensin1 in the heads of nurses of differentage. They also revealed that most of these genes wererelatively stably expressed in nurses of two examinedcolonies. The ranks of their expression stabilities in thepooled nurse heads were determined with geNormPLUS

and NormFinder programs. Two algorithms used in theprograms ranked these genes differently in colonies Aand B. Actin was ranked as the least stable in bothcolonies (Table 2). The NormFinder determined alsothe most suitable combinations of two reference genes

1216 L. Kohútová et al.

Fig. 3. Relative transcription levels of six selected candidate reference genes and of the amylase gene normalized to those of defensin1in pooled head samples of workers of various ages from colony B. The scales of transcription levels of each gene were adjusted asdescribed for Fig. 2.

Table 2. Ranking of selected candidate reference genes based on their expression stability.

Colony A Colony B

geNormPLUS NormFinder geNormPLUS NormFinder

Stability value M Rank Stability value Rank Stability value M Rank Stability value Rank

psa1 0.167 1 0.098 3 0.086 1/2 0.082 3gapdh 0.179 2 0.096 2 0.177 5 0.133 5cyclophilin 0.184 3 0.116 4 0.09 3 0.052 1/2tctp1 0.2 4 0.089 1 0.086 1/2 0.052 1/2mrjp4 0.279 5 0.138 5 0.135 4 0.088 4actin 0.367 6 0.224 6 0.227 6 0.171 6

Explanations: Expression stabilities were determined with geNormPLUS and NormFinder programs in pooled head samples preparedfrom nurses of various ages 6–20 days from colony A and 4–10 days from colony B. A low numerical value corresponds to high expressionstability.

to be tctp1-gapdh and tctp1-cyclophilin in the colonyA and B, respectively (Table 3). Considering the ranksobtained by the programs and the rank of expressionstability in individual nurses (see above) we ranked themost stable reference genes in both colonies in the fol-lowing order: psa1, tctp1, cyclophilin, gapdh and mrjp4.

The fact that these reference genes possess high ex-pression stability in workers during the period of nurs-ing validates these genes as generally suitable refer-ence genes for qPCR studies in heads of nurses, includ-ing studies focused on identification of individuals withhigh constitutive expression of defensin1.

Reference genes for qPCR in nurse bee heads 1217

Table 3. The best combination of two reference genes recom-mended by NormFinder.

Colony Best combination Stability value

A tctp1 and gapdh 0.076B tctp1 and cyclophilin 0.054

Fig. 4. Expression stability of the defensin1, amylase and six se-lected candidate reference genes in the heads of individual work-ers. Raw Cq values of nurses (10 and 20 days old) from the colonyA as well as nurses and foragers (10 and 22 days old, respectively)from the colony B (11 workers per group) are presented as fol-lows: Cq medians as lines, 25th to 75th percentile within boxesand ranges by whiskers.

Validation of different reference genes and amylase ex-pression for distinguishing nurses from foragersAn important issue for the study of gene expression innurse heads is the selection of true nurses in colonies.High expression of amylase versus defensin1 was deter-mined in our expression analyses to be a marker fordistinguishing foragers from nurses and the youngestbees. However, use of defensin1 as a reference gene forassessment of amylase transcription does not seem tobe reasonable. It could lead to a false identification ofsome foragers as nurses in the case of high defensin1expression. In such foragers, normalized amylase levels

could be decreased to low values typical for nurses. Useof some of the selected reference genes as reference foramylase transcription in workers of various ages seemedto be more appropriate. Four these genes were exam-ined. As the most suitable for the identification of theage of the nurse-foragers transition were verified gapdh,tctp1 and cyclophilin; the relative transcription levels ofamylase normalized to these genes showed the biggestdifferences between nurses and foragers (Fig. 5). A cleardividing line existed in the colony A; low amylase ex-pression until day 20 and very high at day 22. The situ-ation was not so clear in the colony B because here thechange of relatively young nurses to foragers occurredover longer interval (between 10–14 days). This is doc-umented by gradually increasing amylase transcriptionwithin this period. The observed gradual increase intranscription was presumably caused by the fact thatanalyzed pooled samples contained increasing numberof forager heads. Based on this assumption and the lowvalues of relative amylase transcription levels detectedin the pooled samples derived from nurses until day 20and days 8–10 in colony A and B, respectively (verylikely these samples contained only nurse heads), weconcluded that normalized relative amylase transcrip-tion levels in individual nurses in these pooled samplesdid not exceed the value of ∼35. The values in nurseswere lower than those supposed to belong to individualforagers in pooled samples (> 150).

Discussion

In this work, data are presented from qPCR analysesof twelve genes in worker heads of various ages. Tran-scription levels of the genes were explored in workersfrom two randomly selected colonies that accidentallyshowed a different age-related division of labor. Ourprimary aim was to find reference genes and workers ofcertain role and ages in which correlated expression ofsuitable reference genes and defensin1 occurs. The anal-yses showed that correlated expression of some genesand defensin1 occurs in nursing bees. The correlationwas associated with the fact that defensin1 and theidentified genes were stably expressed in nurses. Eval-uating the expression stability of these genes in pooledhead samples during period of nursing by geNorm andNormFinder (recommended as the most reliable evalu-ation approach by Teng et al. 2011) and in individualnurses of certain age, led to selection of best referencegenes. We ranked these reference genes according totheir expression stabilities in two colonies (psa1, tctp1,cyclophilin, gapdh and mrjp4). The reference genes canbe employed for studies focused on characterization ofgenetic (epigenetic) factors determining constitutive ex-pression levels of defensin1 and other stably expressedgenes in the heads of nurses as well as generally forstudies of gene expression in nurses.An experimental procedure for reliable identifica-

tion of foragers was elaborated in this study on thebasis of evaluation of amylase transcription in heads ofworkers of various ages. The use of any of three selected

1218 L. Kohútová et al.

Fig. 5. Relative amylase transcription levels normalized to the selected best reference genes in pooled head samples of workers ofvarious ages from colony A and B. Note a sharp increase of amylase expression between days 20 and 22 in colony A and gradualincrease increase between days 10 and 16 in colony B.

reference genes (gapdh, tctp1 and cyclophilin) was val-idated for assessment of amylase expression in work-ers. The procedure can be universally used for iden-tification of foragers in colonies and for distinguish-ing foragers from nurses. Foragers collected by mis-take together with nurses particularly in colonies withshifted age-related labor division can be identified bythis way and eliminated from the dataset. This is im-portant for correct identification of workers with in-creased constitutive expression of defensin1. Amylase

expression is capable of discriminating reliably foragersfrom nurses but not nurses from the youngest work-ers, cleaners. In this study, no effort was made to finda marker specific only for nurses or youngest work-ers; further research is required for the exact identifica-tion of these workers. Nevertheless, possibly sufficientfor many experiments can be the careful selection ofnurses in colonies and elimination of cleaners based oncollection of nurses on brood comb during feeding ofyoung larvae and a consecutive evaluation of abdomen

Reference genes for qPCR in nurse bee heads 1219

size, which is bigger in nurses as compared to clean-ers.In the primary screening of all genes, several genes

with expression profile having a peak form in workersof different age were revealed. Such profiles were foundfor mrjp3, mrjp5, apisimin and regucalcin, which allencode LJ proteins. The profile was known for mrjp3(Feng et al. 2009) but it was novel for the other genes.These genes can be potentially examined as candidatereference genes for genetic (epigenetic) expression stud-ies of genes encoding other LJ proteins possessing thesame transcription profile in nurse head.In comparison to other published data, our analy-

ses documented around 2-fold higher transcription lev-els of defensin1 in younger foragers than in nurses,which is congruent with results of Liu et al. (2011). Sim-ilarly, our results correlate with these of Liu regardinghigh mrjp5 transcription in both nurses and foragers,but they do not agree with results of Drapeau et al.(2006) that did not detect mrjp5 mRNA in bees olderthan 12 days. Furthermore, actin was found here topossess the lowest expression stability among the sixstably expressed genes in nurse heads. Other studiesfocused on the evaluation of reference genes for expres-sion studies during brood development and comparisonof tissues and heads of bacterially infected honeybeesranked actin as moderately suitable amongst the genestested (Lourenco et al. 2008; Scharlaken et al. 2008).Finally, mrjp4 gene was found here as one of the ref-erence genes suitable for analysis of defensin1 expres-sion in nurses; thereby, this result supports our dataobtained in previous defensin1 expression studies, inwhich mrjp4 was used as a reference gene for analysesof defensin1 expression in heads of 10 and 14 days oldworkers. These studies documented variability of de-fensin1 expression levels in workers within and amongcolonies; this is assumed to be associated with geneticdiversity of bees (Klaudiny et al. 2012).In conclusion, this work identified several genes

stably expressed in the heads of workers during theperiod of nursing. They can be used as suitable ref-erence genes for expression studies in honeybees andtheir colonies dealing with genetic or epigenetic ques-tions concerning some genes and also for other studiesexamining gene expression in nurses in experimentalconditions that will not influence their expression sta-bility.

Acknowledgements

This work was supported by the Scientific grant Agency ofthe Ministry of Education of the Slovak Republic and theSlovak Academy of Sciences, the project VEGA 2/0178/12.This contribution is also the result of the project imple-mentation: Centre of excellence for white-green biotechnol-ogy (ITMS 26220120054), supported by the Research &Development Operational Programme funded by the Euro-pean Regional Development Fund (ERDF). We thank Pro-fesor Iain B.H.Wilson for critical reading of the manuscript

and also anonymous reviewers for suggestions improving themanuscript.

References

Albert Š., Klaudiny J. & Šimúth J. 1999a. Molecular character-ization of MRJP3, highly polymorphic protein of honeybee(Apis mellifera) royal jelly. Insect Biochem. Mol. Biol. 29(5): 427–434. PMID: 10380654

Albert Š., Bhattacharya D., Klaudiny J., Schmitzová J. & ŠimúthJ. 1999b. The family of major royal jelly proteins and itsevolution. J. Mol. Evol. 49 (2): 290–297. PMID: 10441680

Andersen C.L., Jensen J.L. & Orntoft T.F. 2004. Normalizationof real-time quantitative reverse transcription-PCR data: amodel-based variance estimation approach to identify genessuited for normalization, applied to bladder and colon cancerdata sets. Cancer Res. 64 (15): 5245–5250. PMID: 15289330

Bachanová K., Klaudiny J., Kopernický J. & Šimúth J. 2002.Identification of honeybee peptide active against Paenibacil-lus larvae larvae through bacterial growth-inhibition assayon polyacrylamide gel. Apidologie 33 (3): 259–269. DOI:10.1051/apido:2002015

Bíliková K., Gusui W. & Šimúth J. 2001. Isolation of apeptide fraction from honeybee royal jelly as a potentialantifoulbrood factor. Apidologie 32 (3): 275–283. DOI:10.1051/apido:2001129

Bíliková K., Hanes J., Nordhoff E., Saenger W., Klaudiny J. &Šimúth J. 2002. Apisimin, a new serine-valine rich peptidefrom honeybee (Apis mellifera L.) royal jelly: purification andmolecular characterization. FEBS Lett. 528: 125–129.

Boyle B., Dallaire N. & MacKay J. 2009. Evaluation of theimpact of single nucleotide polymorphisms and primer mis-matches on quantitative PCR. BMC Biotech. 9: 75. DOI:10.1186/1472-6750-9-75.

Calderone N.W. & Page Jr. R.E. 1987. Genotypic variability inage polyethism and task specialization in the honeybee, Apismellifera (Hymenoptera: Apidae). Behav. Ecol. Sociobiol. 22(1): 17–25. DOI: 10.1007/BF00395694

Cash A.C., Whitfield C.W., Ismail N. & Robinson G.E. 2005.Behavior and the limits of genomic plasticity: power andreplicability in microarray analysis of honeybee brains.Genes Brain Behav. 4 (4): 267–271. DOI: 10.1111/j.1601-183X.2005.00131.x

Drapeau M.D., Albert S., Kucharski R., Prusko C. & MaleszkaR. 2006. Evolution of the Yellow/Major Royal Jelly Proteinfamily and the emergence of social behavioral in honey bees.Genome Res. 16 (11): 1385–1394. DOI: 10.1101/gr.5012006

Feng M., Fang Y. & Li J. 2009. Proteomic analysis of honeybeeworker (Apis mellifera) hypopharyngeal gland development.BMC Genom. 10: 645. DOI: 10.1186/1471-2164-10-645

Han B., Li C., Zhang L., Fang Y., Feng M. & Li J. 2011. Novelroyal jelly proteins identified by gel-based and gel-free pro-teomics. J. Agric. Food Chem. 59 (18): 10346–10355. DOI:10.1021/jf202355n.

Hellemans J., Mortier G., Paepe A.D., Speleman F. & Vandes-ompele J. 2007. qBase relative quantification framework andsoftware for management and automated analysis of real-time quantitative PCR data. Genome Biol. 8 (2): R19. DOI:10.1186/gb-2007-8-2-r19

Huang Z.Y. & Robinson G.E. 1996. Regulation of honey bee di-vision of labor by colony age demography. Behav. Ecol. So-ciobiol. 39 (3): 147–158.

Klaudiny J., Albert Š., Bachanová K., Kopernický J. & Šimúth J.2005. Two structurally different defensin genes, one of themencoding a novel defensin isoform, are expressed in honeybeeApis mellifera. Insect Biochem. Mol. Biol. 35 (1): 11–22.PMID: 15607651

Klaudiny J., Bachanová K., Kohútová L., Dzúrová M., Koper-nický J. & Majtán J. 2012. Expression of larval jelly antimi-crobial peptide defensin1 in Apis mellifera colonies. Biologia67 (1): 200–211. DOI: 10.2478/s11756-011-0153-8

Klaudiny J., Hanes J., Kulifajová J., Albert Š. & Šimúth J. 1994.Molecular cloning of two cDNAs from the head of the nurse

1220 L. Kohútová et al.

honey bee (Apis mellifera L.) coding for related proteins ofroyal jelly. J. Apic. Res. 33 (2): 105–111.

Kubo T., Sasaki M., Nakamura J., Sasagawa H., Ohashi K.,Takeuchi H. & Natori S. 1996. Change in the expressionof hypopharyngeal-gland proteins of the worker honeybees(Apis mellifera L.) with the age and /or role. J. Biochem.119 (2): 291–295. PMID: 8882720

Liu F., Li W., Li Z., Zhang S., Chen S. & Su S. 2011. High-abundance mRNA in Apis mellifera: Comparison betweennurses and foragers. J. Insect Physiol. 57 (2): 274–279. DOI:10.1016/j.jinsphys.2010.11.015.

Lourenco A.P., Mackert A., Cristino A.S. & Simoes Z.L.P. 2008.Validation of reference genes for gene expression studies inhoney bee, Apis mellifera, by quantitative real-time RT-PCR.Apidologie 39 (3): 372–385. DOI: 10.1051/apido:2008015

Majtan J., Klaudiny J., Bohova J., Kohutova L., Dzurova M.,Sediva M., Bartosova M. & Majtan V. 2012. Methylglyoxal-induced modifications of significant honeybee proteinouscomponents in manuka honey: possible therapeutic impli-cations. Fitoterapia 83 (4): 671–677. DOI: 10.1016/j.fitote.2012.02.002.

Ohashi K., Natori S. & Kubo T. 1997. Change in the mode ofgene expression of the hypopharyngeal gland cells with age-dependent role change of the worker honeybee Apis melliferaL. Eur. J. Biochem. 249 (3): 797–802. PMID: 9395329

Ohashi K., Natori S. L. & Kubo T. 1999. Expression of amylaseand glucose oxidase in the hypopharyngeal gland with an age-dependent role change of the worker honeybee (Apis melliferaL.). Eur. J. Biochem. 265 (1): 127–133. PMID: 10491166

Pfaffl M.W. 2004. Quantification strategies in real-time PCR,Chapter 3, pp. 87–112. In: Bustin S.A. (ed.), A-Z of quan-titative PCR, International University Line, La Jolla, CA.910 pp. ISBN-10: 0963681788, ISBN-13: 978-0963681782

Page R.E., Robinson G.E., Calderone N.W. & RothenbuhlerW.C. 1989. Genetic structure, division of labor, and theevolution of insect societies, pp. 15–29. In: Breed M.D. &Page R.E. (eds), The Genetics of Social Evolution, West-view Press, Boulder, CO, USA, 213 pp. ISBN: 0813376181,9780813376189

Robinson G.E. 1992. Regulation of division of labor in insect so-cieties. Annu. Rev. Entomol. 37: 637–665. DOI: 10.1146/an-nurev.en.37.010192.003225

Robinson G.E., Page R.E., Strambi C. & Strambi A. 1992.Colony integration in honey bees: mechanisms of behaviouralreversion. Ethology 90 (4): 336–350. DOI: 10.1111/j.1439-0310.1992.tb00844.x

Scharlaken B., de Graaf D.C., Goossens K., Brunain M., Peel-man L.J. & Jacobs F.J. 2008. Reference gene selectionfor insect expression studies using quantitative real-timePCR: The head of the honeybee, Apis mellifera, aftera bacterial challenge. J. Insect Sci. 8 (33): 1–10. DOI:http://dx.doi.org/10.1673/031.008.3301

Schmitzová J., Klaudiny J., Albert Š., Schroder W., Schrecken-gost W., Hanes J., Júdová J. & Šimúth J. 1998. A family ofmajor royal jelly proteins of the honeybee (Apis mellifera L.).Cell. Mol. Life Sci. 54 (9): 1020–1030. PMID: 9791542

Teng X., Zhang Z., He G., Yang L. & Li F. 2011. Validation ofreference genes for quantitative expression analysis by real-time RT-PCR in four lepidopteran insects. J. Insect Sci. 12(60): 1–17. DOI: 10.1673/031.012.6001.

Vandesompele J., De Preter K., Pattyn F., Poppe B., Van RoyN., De Paepe A. & Speleman F. 2002. Accurate normalizationof real-time quantitative RT-PCR data by geometric averag-ing of multiple internal control genes. Genome Biol. 3(7):Research0034. DOI: 10.1186/gb-2002-3-7-research0034

Whitfield C.W., Cziko A.M. & Robinson G.E. 2003. Gene ex-pression profiles in the brain predict behavior in individualhoney bees. Science 302 (5643): 296–299. DOI: 10.1126/science.1086807

Winston M.L. 1987. The Biology of Honey Bees. Harvard Uni-versity Press. Cambridge, MA, USA, 294 pp. ISBN-10:0674074092, ISBN-13: 978-0674074095

Winston M.L. & Katz S.J. 1982. Foraging differences betweencross-fostered honeybee workers (Apis mellifera L.) of Euro-pean and Africanized races. Behav. Ecol. Sociobiol. 10 (2):125–129. DOI: 10.1007/BF00300172

Received April 6, 2013Accepted May 29, 2013