Characterization of reference genes for quantitative real-timePCR analysis in various tissues of Anoectochilus roxburghii

Gang Zhang • Mingming Zhao • Chao Song •

Anxiong Luo • Jianfa Bai • Shunxing Guo

Received: 3 July 2011 / Accepted: 17 December 2011 / Published online: 27 December 2011

� Springer Science+Business Media B.V. 2011

Abstract Accurate quantification of transcript profiling

with quantitative real time polymerase chain reaction

(qRT-PCR) relies on the reliable normalization of an

appropriate reference gene. This study reported the iden-

tification and validation of nine reference genes, including

b-tubulin (b-TUB), elongation factor 1 alpha (EF-1a),

elongation factor 1 beta (EF-1b), glyceraldehyde-3-phos-

phate dehydrogenase (GAPDH), ubiquitin (UBQ), actin

1/2(ACT-1 and ACT-2), 18S rRNA, and 26S rRNA, from

Anoectochilus roxburghii (Wall.) Lindl., a valuable herb

remedy widely used for various diseases treatment in tra-

ditional Chinese medicine. Transcriptional levels of the

candidate reference genes were examined using qRT-PCR

analysis and revealed differential expression of the genes in

the leaf, stem, root, flower, and peduncle tissues. The rel-

ative quantities data were subjected to geNorm software for

ranking the expression stability of the reference genes and

the results showed that EF-1b and ACT-2 were the two

best stable genes whereas GAPDH and 26S rRNA did not

favor normalization of qRT-PCR in these tissues. The

expression pattern of a squalene synthase encoding gene

(SS) was also determined in parallel. The analyses were in

great consistency when the qRT-PCR data was normalized

to the expression of each or both of EF-1b and ACT-2 as

the internal control, further confirming the reliability of

EF-1b and ACT-2 as the best internal control. The present

study provided the first important clues for accurate data

normalization in transcript profiling in A. roxburghii,

which will be essential to further functional genomics

study in the valuable medicinal plant.

Keywords Anoectochilus roxburghii � Gene expression �Quantitative real time PCR � Internal control gene �Squalene synthase

Abbreviations

ACT-1 Actin 1

ACT-2 Actin 2

CT Threshold value

E PCR efficiency

EF-1a Elongation factor 1 alpha

EF-1b Elongation factor 1 beta

GAPDH Glyceraldehyde-3-phosphate dehydrogenase

Q Relative quantities

qRT-PCR Quantitative reverse transcription PCR

RT-PCR Reverse transcription polymerase chain

reaction

sqPCR Semi-quantitative PCR

UBQ Ubiquitin

18S rRNA 18S ribosomal RNA

26S rRNA 26S ribosomal RNA

b-TUB b-Tubulin

Electronic supplementary material The online version of thisarticle (doi:10.1007/s11033-011-1402-1) contains supplementarymaterial, which is available to authorized users.

G. Zhang � M. Zhao � C. Song � A. Luo � S. Guo (&)

Institute of Medicinal Plant Development, Chinese Academy

of Medical Sciences and Peking Union Medical College,

Beijing 100193, China

e-mail: sxguo2006@yahoo.com.cn

G. Zhang

College of Pharmacy and Shaanxi Provincial Key Laboratory

for Chinese Medicine Basis & New Drugs Research, Shaanxi

University of Chinese Medicine, Xi’an, Shaanxi 712046, China

J. Bai

College of Veterinary Medicine, Kansas State University,

L-222 Mosier Hall, Manhattan, KS 66506, USA

123

Mol Biol Rep (2012) 39:5905–5912

DOI 10.1007/s11033-011-1402-1

Introduction

Gene expression profiling serves one of the cornerstones of

modern molecular biology and contributes to the funda-

mental interpretation of molecular and genetic mechanisms

under certain environmental and developmental conditions

[1]. Compared with conventional analytical methodologies,

like Northern blotting, quantitative competitive polymerase

chain reaction (PCR), and reverse transcription (RT) semi-

quantitative (sq) PCR, etc. for instance, quantitative real-

time PCR (qRT-PCR) is a robust means for rapid and

reliable quantification of transcript levels with high sensi-

tivity, specificity and broad dynamic range, and thus is

widely used in quantification of gene expression, validation

of microarray data, and molecular diagnostics [2, 3].

The accuracy of qRT-PCR, however, is inevitably

affected by a variety of variables including the initial

amounts of samples, integrity of RNA, efficiencies of RT,

PCR reaction variations, etc. [4]. To reduce such variations

during qRT-PCR analysis, appropriate endogenous refer-

ence genes (refer also as internal control genes), which

ideally have a stable expression at a certain level inde-

pendent of cell or tissue types or experimental conditions,

are commonly used for accurate and reliable normalization

[2]. Nevertheless, several studies have also revealed

unstable expression with considerable variations for some

reference genes analyzed [5, 6]. Therefore, the stability of

candidate reference genes must be systematically deter-

mined before utilization in qRT-PCR normalization [7].

Several statistical approaches with different algorithms

have been proposed for reference genes selection such as

geNorm [8], Normfinder [9], and BestKeeper [10]. Among

the algorithms, geNorm, the most widely used software,

offers a measure of gene expression stability (M) by cal-

culating the mean pairwise variation between an individual

gene and all other tested control genes, and thus plays an

important role in systematic validation of reference genes

selection [1, 11]. Overall, these algorithms significantly

facilitate research on reference genes selection. Studies

have reported the evaluation of a number of reference

genes to identify the best suited internal control genes for

normalization of real-time PCR data under specific condi-

tions in various organisms, including human [12], animals

[13, 14], microbes [3], and a variety of plants [5, 6, 11, 15–

17, 33–35]. The reference genes generally fall into the

following categories based on cellular processes impli-

cated, including protein synthesis and degradation (EF-1a,

EF-1b and UBQ), cell structure (ACT and TUB), ribo-

somal component (18S rRNA), and glycolysis (GAPDH),

etc., and do play essential roles in transcript profiling

analysis for offering accurate normalization.

Anoectochilus roxburghii (Wall.) Lindl. (Orchidaceae)

is one of the most valuable medicinal plants used in

traditional Chinese medicine for various diseases’ treat-

ment including diabetes, tumor, hyperliposis, and hepatitis

[18]. Phytochemical studies demonstrated various biolog-

ical active constituents from the herb, such as aliphatic

compounds, flavonoids, glucosides and steroids [19]. Kin-

senoside, an active compound from A. roxburghii, exhibits

diverse pharmacological effects, including antihypergly-

cemia [18], antiosteoporosis [20], antiadiposity [21], anti-

fatigue [22] and hepatoprotection [23]. Given the

importance of A. roxburghii and its valuable active com-

pounds, there is an urgent need for the functional genomic

research into this medicinal plant. However, no transcrip-

tomic or genomic data has been available yet. The objec-

tive of this work is to establish a basis for comparative

expression profiling by validating reference genes suitable

for transcript analysis in A. roxburghii.

Materials and methods

Plant materials and samplings

Anoectochilus roxburghii (Wall.) Lindl. plants with the

height of 15–20 cm were collected at Hongtian County,

Yong’an City, Fujian Province, China (September 23, 2010).

The whole plants were carefully transferred to the 20 cm

diameter pots, and kept in a high-humidity (90–100%)

growth chamber at 25 ± 2�C under a 16-h photoperiod with

the light density of 400–1,000 lux till blossomed. Then, these

plants were harvested from soil and the roots were rinsed

with distilled water. Precaution has to be mandatory to

exclude any mechanical damage. Intact root, leaf, stem,

flower and peduncle tissues of the plants were sampled with a

pair of sterilized scissors, immediately frozen in liquid

nitrogen and stored at -80�C prior to total RNA extraction.

Three independent biological replications were performed.

RNA isolation, quality control and cDNA synthesis

Strict RNA quality control and precise RNA concentration

determination were performed. Extraction of total RNA

was carried out using the RNeasy plant kit (BioTeke,

Beijing, China). Genomic DNA contamination was

removed by RQ1 RNase-free DNase (Promega, Madison,

WI, USA). Quality and integrity of the total RNA were

determined using a formamide denaturing gel along with

an RNA ladder (Invitrogen, Carlsbad, CA, USA) for

comparison and quantity of total RNA was examined using

a NanoDropTM 2000 spectrophotometer (Thermo Fisher

Scientific, Waltham, MA, USA). The first strand cDNA

was synthesized with 2 lg of total RNA using the RT-PCR

system (Promega, Madison, WI, USA) according to the

manufacturer’s protocol.

5906 Mol Biol Rep (2012) 39:5905–5912

123

Primer design and RT-PCR verification

Nine commonly used reference genes representing distinct

functional classes and gene families and a gene of interest

were selected in this study. The reference genes included

ubiquitin, actin, tubulin, elongation factor, glyceraldehyde-

3-phosphate dehydrogenase, and ribosomal RNA

(Table 1). The qRT-PCR primers for 18S rRNA, 26S

rRNA, ACT-1, and b-TUB were identical to those previ-

ously used for normalization in Salvia miltiorrhiza [24],

Oncidium [25], Dendrobium thyrsiflorum (Reichb. f.) [26],

and Triticum aestivum [27], respectively. The cDNA

fragments for EF-1a, GAPDH, and ACT-2 were firstly

cloned from A. roxburghii by amplification with the

reported primers [28–30], followed by the qRT-PCR

primers design. The qRT-PCR primers for EF-1b, UBQ,

and SS were, respectively, designed based on their cDNA

sequences previously identified (unpublished data). All

primers were designed with Primer 3 [31] and synthesized

by Sangon Biotech (Shanghai, China).

Standard RT-PCR amplification was carried out to

amplify the gene fragments with ExTaqTM DNA polymerase

(TaKaRa, Dalian, China). The procedure was as follows:

94�C for 3 min; followed by 34 cycles of 94�C 30 s, 60�C

30 s, 72�C 45 s; 72�C 7 min for extension; hold at 4�C. The

PCR products were loaded on 1.8% agarose gel with EB

(Sigma-Aldrich, St. Louis, MO, USA) staining, and purified

using the TIANgel Midi purification kit (Tiangen, Beijing,

China) according to the manufacturer’s instructions, fol-

lowed by cloning into the pMD18-T vector (TaKaRa,

Dalian, China). The positive clones were obtained after

being transformed into the E. coli JM109 competent cells

(TaKaRa, Dalian, China). Triplicate recombinant plasmids

were sequenced on an ABI PRISM 3130XL genetic analyzer

(Applied Biosystems, Foster City, CA, USA) using the ABI

PRISMTM Big DyeTM Terminator Cycle (Applied Biosys-

tems, Foster City, CA, USA). The sequences were finally

confirmed using BLAST analysis against the non-redundant

nucleotide database (nr) and the ESTs database (non-mouse

and non-human ESTs collections) in GenBank.

qRT-PCR analyses

The leaf, stem, root, flower, and peduncle tissues of A. rox-

burghii were harvested and three biological replicates were

performed independently. RT was performed on 2 lg of total

RNA using the MMLV Reverse Transcriptase (Promega,

Madison, WI, USA) with the oligo(dT)18 and random hex-

amer pd(N)6 primers. Templates for qRT-PCR analysis were

40 9 diluted cDNAs from each 2 lg RNA sample. The

qRT-PCRs were performed on the ABI 7500 Real-Time PCR

System (Applied Biosystems, Foster City, CA, USA) using

Table 1 The qRT-PCR primer sequences, amplicon size (bp), PCR efficiency (E), and GenBank accession of the candidate genes in A.roxburghii

Symbol Gene name Primer sequence (50–30) Accession E Size

EF-1a Elongation factor 1 alpha F: TCAGGCTGACTGTGCTGTCCT JF825419 2.02 156

R: GTGGTGGCGTCCATCTTGTT

EF-1b Elongation factor 1 beta F: GGATGTGAAGCCCTGGGA JF825420 1.96 296

R: CCTTGGCAGTGCAGCTCTT

GAPDH Glyceraldehyde-3-phosphate dehydrogenase F: CAAGGACTGGAGAGGTGGAAGA JF825421 2.18 297

R: GACCTGCTGTCACCCAAGAAGT

18S rRNA 18S ribosomal RNA F: CCAGGTCCAGACATAGTAAG JF825422 2.07 429

R: GTACAAAGGGCAGGGACGTA

26S rRNA 26S ribosomal RNA F: CTGATTTCCAGTGCGAATACGA Unavaa 2.13 69

R: TCCGAACGACTAAAGGATCGA

UBQ Ubiquitin F: TGGTCGCACCTTAGCGGATT JF825423 1.82 110

R: TGGTTTTCCCAGTGAGGGTCTT

ACT-1 Actin 1 F: AAGCTGTTCTTTCCCTATATGCTAGTGG JF825424 2.15 236

R: CTTCTCCTTGATGTCCCTGACAATTT

ACT-2 Actin 2 F: CGGGCATTCACGAGACCAC JF825425 1.91 221

R: AATAGACCCTCCAATCCAGACACT

SS Squalene synthase F: TGGCAGGTTTAGTTGGCTTGG JF825426 2.13 360

R: TGATTTGCGGTATGGCACAGAA

b-TUB Tubulin beta F: TGTGCCCCGTGCTGTTCTTATG Unavaa 135

R: CCCTTGGCCCAGTTGTTACCC

a Unava indicates that the data is unavailable

Mol Biol Rep (2012) 39:5905–5912 5907

123

the SYBR� Premix Ex TaqTM (Perfect Real Time) (TaKaRa,

Dalian, China). The reactions were set up at 20 ll with 2 ll

cDNA template, 10 ll SYBR� Premix Ex TaqTM Master

Mix, 0.4 ll each primer (10 lM), 0.4 ll ROX Reference

Dye and 6.8 ll double distilled water. The thermal condi-

tions were: 95�C for 30 s; followed by 40 cycles of 95�C

15 s, 60�C 45 s. Dissociation curves were generated for each

reaction to ensure specific amplification. All reactions,

including non-template controls, were carried out in three

times. Cycle threshold (CT) values were generated from the

ABI PRISM 7500 SDS Software v 1.4 (Applied Biosystems,

Foster City, CA, USA).

PCR efficiencies (E) of candidate genes were deter-

mined by generating standard curves. Triplicate qRT-PCR

assays were applied on a dilution series of cDNAs (bulked

cDNA of analyzed samples) and the data were subjected to

analysis using the ABI PRISM 7500 SDS Software v 1.4

(Applied Biosystems, Foster City, CA, USA) to establish

the standard curve for each candidate gene. Amplification

efficiency was then calculated based on the slope of

the standard curve for each gene using the formula

E = (10-1/slope). To pass validation, each primer pair must

demonstrate high amplification efficiency above 80%.

Data analyses

Following qRT-PCR analysis and data collection, raw

expression levels of CT values for each gene in different

samples were transformed to relative quantities Q using

Q = E (minCT - sampleCT) equation. The Q data were con-

verted into the correct input files as required by geNorm

software and analyzed for ranking the expression variability of

reference genes [8]. The stability of each reference gene was

shown as M value calculated by geNorm analysis, and a larger

M value indicated the less stability while a smaller suggesting

the more stability. For validation of the selected reference

genes, tissue specific expression of a squalene synthase

encoding gene SS, a key regulator involving terpenoid bio-

synthesis, was analyzed by qRT-PCR using the comparative

DDCT method of relative gene quantification [32]. A proba-

bility (P) value B0.05 was used to determine the significance

of difference between samples, or when relative quantity of

RNA was at least two fold higher or lower than that of the

calibrator tissue sample.

Results

Sequences analyses and establishment of qRT-PCR

assays

Using gene specific primers, each RT-PCR assay yielded

a unique specific band. After recovering, cloning and

sequencing analyses, all candidate gene fragments were

identified from A. roxburghii. The cDNA sequences of EF-

1a, EF-1b, GAPDH, b-TUB, UBQ, ACT-1, and ACT-2

were, respectively, 496, 296, 297, 135, 643, 236, and

221 bp in length. BLASTX analyses revealed that they had

85–100% identities with corresponding homologues in

other plant species. For example, A. roxburghii EF-1bshowed 85% identity with that of Pisum sativum (GenBank

accession AAR15081); EF-1a was 98% homologous to that

of Solanum tuberosum (GenBank accession ABB55388);

And ACT-1 exhibited 98% identity with the Picea abies

Actin 1 (GenBank accession ACP19072). BLASTN anal-

yses indicated that the A. roxburghii 18S rRNA and 26S

rRNA, 429 and 69 bp in length, was 97% similar to that

of Oncidium sphacelatum and Glycyrrhiza uralensis

(GenBank accession U59939 and EF571299), respectively.

Moreover, a 360 bp SS gene was obtained and had 89%

identity with the Glycyrrhiza uralensis SS (GenBank

accession ADG36707). Except b-TUB and 26S rRNA

sequences (see Supplementary Data), the other eight can-

didate genes ([200 bp) were deposited in GenBank under

the accessions from JF825419 to JF825426 (Table 1).

Before proceeding with qRT-PCR, PCR efficiencies for

the candidate genes were determined by amplifying a serial

dilution of cDNA pools. The qRT-PCR assays yielded

specific products and generated reliable standard curves

(see Supplementary Data), except that using the b-TUB

primer pair (data not shown), which did not generate an

ideal standard curve and was therefore excluded from

further analysis. The coefficients of determination (R2) for

all standard curves were mostly above 0.990 except for that

of 18S rRNA was 0.989578. The primer efficiencies for

each candidate gene calculated from the slope of the

standard curve ranged from 82 to 118%. These results

indicated good establishment of qRT-PCR assays for our

circumstances. The qRT-PCR primers, PCR efficiency, and

amplicon size of the candidate genes were documented in

Table 1.

Expression of reference genes and selection of internal

controls

The optimal qRT-PCR assays established were then per-

formed to examine the transcription levels of the eight

reference genes in the five tissues and the raw CT values

were shown in Fig. 1. Overall, the eight reference genes

displayed a wide expression range with CT values from

7.31 to 29.7. 18S rRNA transcripts maintained the most

abundant with the lowest CT values below 10 cycles and

showed the relatively stable expression among the tissues.

Higher expressed genes with CT values blow 20 cycles

were 26S rRNA and EF-1a, respectively. 26S rRNA

exhibited the significant expression change with CT values

5908 Mol Biol Rep (2012) 39:5905–5912

123

from 15.0 to 17.8 in the leaf and flower samples compared

with the other three tissues. By contrast, EF-1a expression

was relatively stable. GAPDH was the lowest expressed

gene with CT values from 26.4 to 29.7 and showed dra-

matic change in leaf and stem tissues. The other four

intermediately expressed genes including EF-1b, UBQ,

ACT-1, and ACT-2, with CT values from 20 to 25, dem-

onstrated relatively stable expression.

Next, triplicate relative quantities Q for each gene in every

sample transformed from the qRT-PCR data were subjected to

geNorm analysis to evaluate the best internal control among the

eight reference genes. As showed in Fig. 2a, the expression

stability M values for GAPDH, 26S rRNA, ACT-1, EF-1a,

UBQ, 18S rRNA, ACT-2, and EF-1b in the five tissues were

1.056, 0.943, 0.784, 0.702, 0.647, 0.592, 0.393, and 0.393,

respectively. ACT-2 and EF-1b with the lowest M value were

ranked the most stable genes, while GAPDH and 26S rRNA

with the highest M value therefore had the lowest expression

stabilities. And, the expression stabilities among these genes

were ranked as follows: ACT-2/EF-1b[18S rRNA[UBQ [EF-1a[ACT-1[26S rRNA[ GAPDH. In addi-

tion, the optimal number of reference genes for accurate nor-

malization was also analyzed by geNorm (Fig. 2b). Based on

the default cut-off V value of 0.15 (the point at which it is

unnecessary to include additional genes in a normalization

strategy) [8], the V4/5 value (0.14) would suggest the use of four

of the five most stable genes for normalization.

Validation of SS expression using the best reference

gene

For validation of the best stable reference genes identified

from geNorm analysis, we tested the tissue specific

expression of SS by qRT-PCR analysis normalizing to

ACT-2 and EF-1b (Fig. 3). As presented in Fig. 3, the

expression pattern of SS normalized to the ACT-2

expression levels showed similar in the five tissues with

that normalized to the EF-1b expression levels. When

using a combination of ACT-2 and EF-1b as the reference

genes, the results were in good agreement with that using

each ACT-2 or EF-1b alone as the unique internal control.

The transcription levels of SS were the highest in the roots,

followed by flower, peduncle, and stem tissues, whereas

the lowest in leaf tissues.

Discussion

Accurate normalization of gene expression against an

adequate set of reference genes is a pre-requisite for

accurate gene expression analysis. Accordingly, the num-

ber of studies on reference genes selection in different plant

Fig. 1 Transcription levels of the eight reference genes presented as

absolute CT values in five tissue samples of A. roxburghii. Templates

were the 40 9 diluted cDNA reverse transcribed from each 2 lg

RNA sample. The absolute quantification assays measured in three

replicates were employed. Data are analyzed using the ABI PRISM

7500 SDS Software v 1.4 (Applied Biosystems, Foster City, CA,

USA) and are represented by mean from three independent replicates

Fig. 2 Selection of the most suitable reference genes for qRT-PCR

normalization using geNorm analysis. a Average expression stability

(M) of reference genes during stepwise exclusion of the least stable

gene. Data are from three biological replicates and three technical

replicates. The X-axis from left to right indicates the ranking of the

genes according to their expression stability. A larger M value

indicates the less stability while a smaller M value suggested the more

stability. b Determination of the optimal number of reference genes

for normalization. Vn/Vn?1 is the pairwise variation between normal-

ization factors of n and n ? 1 genes. The geNorm algorithm proposes

a default cut-off V value of 0.15, below which the inclusion of an

additional reference gene is not required

Mol Biol Rep (2012) 39:5905–5912 5909

123

species has increased over the last years [5, 6, 11, 15–17,

33–35]. Here, we attempted the current research on

validation of nice candidate reference genes in various

A. roxburghii tissues for qRT-PCR normalization after a

serial of strict quality controls by well harvested samplings,

RNA sample quality evaluation, DNase I digestion, and

careful manipulation to largely decrease variability in all

experimental replicates. The data obtained substantially

increased the credibility of geNorm analysis. The tissue

specific expression of the target SS gene using qRT-PCR

passed validation using ACT-2 and EF-1b as internal con-

trol, ensuring the validity of expression analysis in this study.

Reference genes are ordinarily regulated differently with

differential expression patterns in different plant species.

In the present study, we analyzed the expression stability

of eight reference genes and found that most of the genes

exhibited differential expression in A. roxburghii tissues

(Fig. 1). GAPDH displayed the most considerable expres-

sion variability and was discouraged its use as an internal

control, the same as that in Petunia hybrida [11] and in

coffee [17]. In tomato, various abiotic stress elicitors

including nitrogen starvation, low temperatures, and

suboptimal light conditions induced a high expression

variability of GAPDH in the leaf tissues [6]. ACT-1 and

ACT-2, two members of actin gene family in A. roxburghii,

showed apparently distinct expression stability. ACT-2 was

ranked as one of the most stable genes whereas ACT-1 had

relatively poor stability (Fig. 2a), limiting its use as inter-

nal control. Similar results were obtained that the expres-

sion of ACT2/7 had the highest lowest stability index and

was less stable than ACT11 in different tissues of soybean

[16]. The ACT case implies that the expression stability of

different reference genes from the same gene family vary

in the same experimental condition.

EF-1a and UBQ, involving in protein metabolism, are

widely used as good internal control for transcript profiling

[6, 11, 33]. However, our analysis revealed EF-1a was less

stable in various A. roxburghii tissues. Other reports evi-

dently showed that EF-1a had constant stability in Petunia

hybrida tissues [11] and in tomato tissues even under

cold stress [6]. The expressions of UBQ orthologues show

discrepancy in diverse plant species. UBQ10 gene, for

instance, was remarkably stable expressed in Arabidopsis

[5], whereas it was unsuitable for normalization of different

tissues at different developmental stages in rice and soy-

bean [15, 16]. In our case, UBQ was ranked as the forth

stable expressed gene with moderately high abundance

implying the inapplicability of UBQ as a good internal

control. Taken together, these findings collectively suggest

the expressions of reference genes can be varied in the

specific experimental conditions and thus need specific

consideration for data normalization. The fact that refer-

ence genes are not only involved in basal cell metabolism,

but are also involved in specific cellular functions may be

partly explained at this point [36].

The rRNA genes, especially 18S rRNA, are mostly

commonly employed reference genes for normalization of

real-time PCR data. Our results indicated 18S rRNA had

much better expression stability than 26S rRNA (Fig. 2a).

26S rRNA was firstly validated for the use of internal

control in seven plant species under various developmental

stages and external stimuli by sq-PCR, proposing 26S

rRNA as a suitable internal control in expression studies

[37]. The expression pattern of an Oncidium orchid HDR

gene was examined normalized to the 26S rRNA expres-

sion without validation [25]. In A. roxburghii, 26S rRNA

showed fluctuated expression and exhibited inappropriate

stability in the tissues analyzed by geNorm (Fig. 2a),

which was not in agreement with the previous reports [36].

The finding is not surprising because the expression sta-

bility of a same reference gene may be varied in different

species. Hence, it is likely that 26S rRNA may not serve as

a good internal control for gene profiling in A. roxburghii

and the potential utilization of 26S rRNA as internal con-

trol in other organisms is still needed to be further inves-

tigated using qRT-PCR analysis combined statistical

calculation, like geNorm.

Most research regarding gene expression using qRT-

PCR analysis normally used only one single reference gene

for normalization. Though, normalization with two or more

reference genes for qRT-PCR studies might be more reli-

able [8]. In this study, geNorm analysis demonstrated that

four genes would be required for accurate normalization

(Fig 2b). However, the number of reference genes needs

Fig. 3 Validation of the expression pattern of a squalene synthase

encoding gene SS in A. roxburghii tissues by qRT-PCR analysis.

Relative quantitation is calculated using the comparative CT method.

The raw qRT-PCR data of SS transcripts were normalized to the

expression levels of ACT-2 (light dark column), EF-1b (deep darkcolumn), and a combination of ACT-2 and EF-1b (blank column) as

the internal control, respectively. Leaf sample was considered as

calibrator. Mean and error bars are calculated from three independent

replicates. The mean relative expression values are presented in the

columns

5910 Mol Biol Rep (2012) 39:5905–5912

123

balancing between accuracy and practical consideration,

thus the default cut-off V value of 0.15 proposed by geN-

orm should not be taken too strictly, as suggested by the

geNorm manual itself. The V3/4 value of 0.156 was slightly

higher than the geNorm cut-off value and thus three

of the four most stable genes should be suggested for

normalization. However, 18S rRNA was extremely highly

expressed with the lowest CT values across the samples

(Fig. 1), much more abundant than that in other plant

species [33]. It is therefore not recommended the use of

18S rRNA as an internal control for normalization of

weakly or moderately expressed genes. In addition, the

imbalances in rRNA and mRNA fractions between differ-

ent samples also made 18S rRNA far from ideal [2]. Taken

together, we propose ACT-2 and EF-1b as the suitable

internal control for qRT-PCR normalization in A. rox-

burghii tissues.

For validation, we further examined the tissue expres-

sion of SS gene in A. roxburghii using qRT-PCR with

ACT-2 and EF-1b as the internal control. The results were

as expected with great consistency, indicating reliability of

ACT-2 and EF-1b as the suitable internal control for nor-

malization in expression analysis in A. roxburghii tissues

(Fig. 3). Squalene synthase catalyzes the first committed

enzymatic step from the central isoprenoid pathway toward

sterol and triterpenoid biosynthesis. Regulation of SS

contributes to either enhancement or decrease of the active

constituents involving this pathway [38, 39]. A recent study

reported the improved amount of artemisinin, an effective

anti-malarial drug, by the knockdown of SS transcripts in

transgenic Artemisia annua (Chinese wormwood) plants

using gene silencing [40]. The findings promote us priorly

to validate the expression of SS in A. roxburghii tissues.

High level expression of SS transcripts in the roots suggests

more active molecular regulation of SS. However, unveil-

ing the molecular regulation networks regarding the

pathways of pharmacological effective constituents, like

kinsenoside, in A. roxburghii needs extensive and intensive

investigation at the transcriptomic and biochemical levels.

The present study preferably provides the first important

clues benefiting such research in this precious medicinal

plant in future.

Acknowledgments This research was financially supported by

the National Natural Sciences Foundation of China (No. 31101608

and 31070300), and the China Postdoctoral Science Foundation

(No. 20100470244).

References

1. Gutierrez L (2008) Towards a systematic validation of references

in real-time RT-PCR. Plant Cell 20:1734–1735

2. Bustin SA (2002) Quantification of mRNA using real-time

reverse transcription PCR (RT-PCR): trends and problems. J Mol

Endocrinol 29:23–39

3. Hacquard S, Veneault-Fourrey C, Delaruelle C, Frey P, Martin F,

Duplessis S (2011) Validation of Melampsora larici-populinareference genes for in planta RT-quantitative PCR expression

profiling during time-course infection of poplar leaves. Physiol

Mol Plant Pathol 75:106–112

4. Gachon C, Mingam A, Charrier B (2004) Real-time PCR: what

relevance to plant studies? J Exp Bot 55:1445–1454

5. Czechowski T, Stitt M, Altmann T, Udvardi MK, Scheible W

(2005) Genome-wide identification and testing of superior ref-

erence genes for transcript normalization in Arabidopsis. Plant

Physiol 139:5–17

6. Løvdal T, Lillo C (2009) Reference gene selection for quantita-

tive real-time PCR normalization in tomato subjected to nitrogen,

cold, and light stress. Anal Biochem 387:238–242

7. Guenin S, Mauriat M, Pelloux J, Wuytswinkel OV, Bellini C,

Gutierrez L (2009) Normalization of qRT-PCR data: the neces-

sity of adopting a systematic, experimental conditions specific,

validation of references. J Exp Bot 60:487–493

8. Vandesompele J, De Preter K, Pattyn F, Poppe B, Van Roy N, De

Paepe A, Speleman F (2002) Accurate normalization of real-time

quantitative RT-PCR data by geometric averaging of multiple

internal control genes. Genome Biol 3(7):0034.1–0034.11

9. Andersen CL, Jensen JL, Ørntoft TF (2004) Normalization of

real-time quantitative reverse transcription-PCR data: a model-

based variance estimation approach to identify genes suited for

normalization, applied to bladder and colon cancer data sets.

Cancer Res 64:5245–5250

10. Pfaffl MW, Tichopad A, Prgomet C (2004) Determination of

stable housekeeping genes, differentially regulated target genes

and sample integrity: BestKeeper–Excel-based using pair-wise

correlations. Biotechnol Lett 26:509–515

11. Mallona I, Lischewski S, Weiss J (2010) Validation of reference

genes for quantitative real-time PCR during leaf and flower

development in Petunia hybrida. BMC Plant Biol 10:4

12. Mahoney DJ, Carey K, Fu MH, Snow R, Cameron-Smith D,

Parise G, Tarnopolsky MA (2004) Real-time RT-PCR analysis of

housekeeping genes in human skeletal muscle following acute

exercise. Physiol Genomics 18:226–231

13. Zhong Q, Zhang Q, Wang Z, Qi J, Chen Y, Li S, Sun Y, Li C,

Lan X (2008) Expression profiling and validation of potential

reference genes during Paralichthys olivaceus embryogenesis.

Mar Biotechnol 10:310–318

14. Yuwen Y, Dong Z, Wang Q, Sun X, Shi C, Chen G (2011)

Evaluation of endogenous reference genes for analysis of gene

expression with real-time RT-PCR during planarian regeneration.

Mol Biol Rep 38(7):16–21

15. Jain M, Nijhawan A, Tyagi AK, Khurana JP (2006) Validation of

housekeeping genes as internal control for studying gene

expression in rice by quantitative real-time PCR. Biochem Bio-

phys Res Commun 345:646–651

16. Jian B, Liu B, Bi Y, Hou W, Wu C, Han T (2008) Validation of

internal control for gene expression study in soybean by quanti-

tative real-time PCR. BMC Mol Biol 9:59

17. Cruz F, Kalaoun S, Nobile P, Colombo C, Almeida J, Barros

LMG, Romano E, Grossi-de-Sa MF, Vaslin M, Alves-Ferreira M

(2009) Evaluation of coffee reference genes for relative expres-

sion studies by quantitative real-time RT-PCR. Mol Breeding

23(4):607–616

18. Zhang Y, Cai J, Ruan H, Pi H, Wu J (2007) Antihyperglycemic

activity of kinsenoside, a high yielding constituent from Anoec-tochilus roxburghii in streptozotocin diabetic rats. J Ethnophar-

macol 14(2):141–145

Mol Biol Rep (2012) 39:5905–5912 5911

123

19. He C, Wang C, Guo S, Yang J, Xiao P (2006) A novel flavonoid

glucoside from Anoectochilus roxberghii (Wall.) Lindl. J Integr

Plant Biol 48(3):359–363

20. Surh YJ, Chus KS, Cha HH, Han SS, Keum YS, Park KK, Lee SS

(2001) Molecular mechanisms underling chemopreventive

activities of anti-inflammatory phytochemicals: down-regulation

of COX-2 and iNOS through suppression of NF-jB activation.

Mutat Res 480–481:243–268

21. Du XM, Irino N, Furusho N, Hayashi J, Shoyama Y (2008)

Pharmacologically active compounds in the Anoectochilus and

Goodyera species. J Nat Med 62:132–148

22. Ikeuchi M, Yamaguchi K, Nishimura T, Yazawa K (2005) Effects

of Anoectochilus formosanus on endurance capacity in mice.

J Nutr Sci Vitaminol 51:40–44

23. Hsieh WT, Tsai CT, Wu JB, Hsiao HB, Yang LC, Lin WC (2011)

Kinsenoside, a high yielding constituent from Anoectochilusformosanus, inhibits carbon tetrachloride induced Kupffer cells

mediated liver damage. J Ethnopharmacol 135(2):440–449

24. Liao P, Zhou W, Zhang L, Wang J, Yan X, Zhang Y, Zhang R,

Zhou G, Kai G (2009) Molecular cloning, characterization and

expression analysis of a new gene encoding 3-hydroxy-3-meth-

ylglutaryl coenzyme A reductase from Salvia miltiorrhiza. Acta

Physiol Plant 31(3):565–572

25. Huang JZ, Cheng TC, Wen PJ, Hsieh MH, Chen FC (2009)

Molecular characterization of the Oncidium orchid HDR gene

encoding 1-hydroxy-2-methyl-2-(E)-butenyl 4-diphosphate

reductase, the last step of the methylerythritol phosphate path-

way. Plant Cell Rep 28(10):1475–1486

26. Skipper M, Pedersen KB, Johansen LB, Frederiksen S, Irish VF,

Johansen BB (2005) Identification and quantification of expres-

sion levels of three FRUITFULL-like MADS-box genes from

the orchid Dendrobium thyrsiflorum (Reichb. f.). Plant Sci

169:579–586

27. Jia XY, Xu CY, Jing RL, Li RZ, Mao XG, Wang JP, Chang XP

(2008) Molecular cloning and characterization of wheat calreti-

culin (CRT) gene involved in drought-stressed responses. J Exp

Bot 59(4):739–751

28. Yang Y, Wu J, Zhu K, Liu L, Chen F, Yu D (2009) Identification

and characterization of two chrysanthemum (Dendronthe-ma 9 moriforlium) DREB genes, belonging to the AP2/EREBP

family. Mol Biol Rep 36:71–81

29. Lu Q, Yang Q (2006) cDNA cloning and expression of antho-

cyanin biosynthetic genes in wild potato (Solanum pinnatisec-tum). Afr J Biotechnol 5(10):811–818

30. Li SH, Kuoh CS, Chen YH, Chen HH, Chen WH (2005) Osmotic

sucrose enhancement of single-cell embryogenesis and transfor-

mation efficiency in Oncidium. Plant Cell Tiss Org 81:183–192

31. Rozen S, Skaletsky H (2000) Primer3 on the WWW for general

users and for biologist programmers. Methods Mol Bio

132:365–386

32. Pfaffl MW (2001) A new mathematical model for relative

quantification in real-time RT-PCR. Nucleic Acids Res 29(9):e45

33. Yang Y, Hou S, Cui G, Chen S, Wei J, Huang L (2010) Char-

acterization of reference genes for quantitative real-time PCR

analysis in various tissues of Salvia miltiorrhiza. Mol Biol Rep

37:507–513

34. Dong L, Sui C, Liu Y, Yang Y, Wei J, Yang Y (2011) Validation

and application of reference genes for quantitative gene expres-

sion analyses in various tissues of Bupleurum chinense. Mol Biol

Rep 38(8):5017–5023

35. Yan J, Yuan F, Long G, Qin L, Deng Z (2011) Selection of

reference genes for quantitative real-time RT-PCR analysis in

citrus. Mol Biol Rep doi:10.1007/s11033-011-0925-9

36. Thellin O, ElMoualij B, Heinen E, Zorzi W (2009) A decade of

improvements in quantification of gene expression and internal

standard selection. Biotechnol Adv 27(4):323–333

37. Singh K, Raizada J, Bhardwaj P, Ghawana S, Rani A, Singh H,

Kaul K, Kumar S (2004) 26S rRNA-based internal control gene

primer pair for reverse transcription-polymerase chain reaction-

based quantitative expression studies in diverse plant species.

Anal Biochem 335:330–333

38. Lee MH, Jeong JH, Seo JW, Shin CG, Kim YS, In JG, Yang DC,

Yi JS, Choi YE (2004) Enhanced triterpene and phytosterol

biosynthesis in Panax ginseng overexpressing squalene synthase

gene. Plant Cell Physiol 45(8):976–984

39. Kim TD, Han JY, Huh GH, Choi YE (2011) Expression and

functional characterization of three squalene synthase genes

associated with saponin biosynthesis in Panax ginseng. Plant Cell

Physiol 52(1):125–137

40. Zhang L, Jing F, Li F, Li M, Wang Y, Wang G, Sun X, Tang K

(2009) Development of transgenic Artemisia annua (Chinese

wormwood) plants with an enhanced content of artemisinin, an

effective anti-malarial drug, by hairpin-RNA-mediated gene

silencing. Biotechnol Appl Biochem 52:199–207

5912 Mol Biol Rep (2012) 39:5905–5912

123