Identification of reference genes for gene expression normalization in safflower (Carthamus tinctorius)

ABSTRACT

Safflower (Carthamus tinctorius L., Asteraceae) is an important oil crop and medicinal plant. Gene expression analysis is gaining importance in the research of safflower. Quantitative PCR has become a powerful method for gene study. Reference genes are one of the major qualification requirements of qPCR because they can reduce the variability. To identify the reference genes in safflower, nine candidate genes of the housekeeping genes were selected from the EST library of safflower constructed by our lab: CtACT (actin), CtGAPDH (glyceraldehyde 3-phosphate dehydrogenase), CtE1F4A (elongation factor 1 alpha), CtTUA (alpha-tubulin), CtTUB (beta-tubulin), CtPP2A (serine/threonine-protein phosphatase), CtE1F4A (eukaryotic initiation factor 4A), CtUBI (Ubiquitin), and Ct60S (60S acidic ribosomal protein). Expression stability was examined by qPCR across 54 samples, representing tissues at different flowering stages and two chemotype of safflower lines. We assessed the expression stability of these candidate genes by employing four different algorithms (geNorm, NormFinder, ΔCt approach, and BestKeeper) and found that CtUBI and Ct60S were the highly ranked candidate genes. CtUBI and Ct60S were used as reference genes to evaluate the expression of CtFAD2-10 and CtKASII. Our data suggest CtUBI and Ct60S could be used as internal controls to normalize gene expression in safflower.

Keywords:
Safflower; Reference gene; RT-qPCR

Introduction

Safflower (Carthamus tinctorius L., Asteraceae) is a thistle-like, self-compatible, annual, diploid (2n = 24) herbaceous crop that thrives in hot, dry climates and survives on minimal surface moisture. The safflower cultivars are distributed from the Mediterranean to the Pacific Ocean at latitudes between 20ºS and 40ºN, wherever a hot, dry climate suits the crop. In some countries, safflower has become an important crop due to the rich content of edible oil, which has the highest polyunsaturated/saturated ratios of any oil available (Gecgel et al., 2007; Yeilaghi et al., 2012). Safflower is also a valuable medicinal plant. The flowers of safflower can be used for the treatment of cardiovascular and cerebrovascular diseases (Tian et al., 2010; Asgarpanah and Kazemivash, 2013) and the extracts from safflower can be used as anti-inflammatory agents as well (Jun et al., 2011). There are several studies about the genetic variation of safflower cultivars using various molecular DNA markers, such as SNP (Chapman and Burke, 2007), SRAP (Peng et al., 2008), ISSR (Chapman et al., 2009; Golkar et al., 2011), and AFLP (Zhang et al., 2009; Feng et al., 2010; Li et al., 2010). Meanwhile, in the previous study of our lab, we have found that obvious differentiation has occurred in safflower populations from exterior appearance to inner chemical constituent, due to the long natural and artificial selection (Peng et al., 2008; Zhang et al., 2009; Feng et al., 2010; Li et al., 2010). From our conclusion, hydroxysafflor yellow A (HYSA), as the active compound, is the main factor to determine the diversity of safflower (Yang et al., 2011).

As secondary metabolites in safflower, the flavonoids are major components of the extracts from the flowers with medicinal function (Andersen and Markham, 2010; Asgarpanah and Kazemivash, 2013). HSYA, one of the most important flavonoids with a unique presence in the flower petals of safflower, plays a major role in the pharmacological effects of flavonoids (Feng et al., 2013; Sun et al., 2013; Wang et al., 2013). The flowering process of safflower is a complex development associated with the biosynthesis of flavonoids, particularly the color change (Tanaka et al., 2010). In safflower seeds, the identification and initial characterization of the FAD2 gene family with eleven members provides an insight into the principal determinants of synthesis of linoleic acid in safflower seed oil (Cao et al., 2013; Liu et al., 2013). And the CtFAD3 enzyme activity is important for fatty acid desaturation in safflower flower (Guan et al., 2014). The understanding of the expression of the key genes will help investigate the mechanism involved in the biosynthesis of flavonoids and fatty acids.

Quantitative real-time PCR (qPCR) is a routine tool with high sensitivity and specificity for the quantification of gene expression (Gachon et al., 2004). When we study the gene expression, the reference genes can be used in the normalization of experimental variations, such as the amount of material, extraction of RNA, efficiency of reverse transcription, and so on (Silver et al., 2006; Gutierrez et al., 2008; Bustin et al., 2009; Guénin et al., 2009). It is important to normalize the expression of the target gene in order to obtain reliable and accurate results by using the reference genes. Housekeeping genes are often considered as being constantly expressed and are always used as internal standards. But many researches showed that the expression of housekeeping genes varies in different materials and even in the same materials during different treatments (Thellin et al., 1999; Schmidt and Delaney, 2010). So the evaluation of reference genes in different plant tissues during biotic or abiotic stress is necessary for expression studies.

Housekeeping genes, such as actin (Bas et al., 2004), elongation factor 1 alpha (Schmidt and Delaney, 2010; Li et al., 2012), alpha-tubulin (Jarosova and Kundu, 2010; Wan et al., 2011), beta-tubulin (Jarosova and Kundu, 2010), serine/threonine-protein phosphatase (Liu et al., 2012), eukaryotic initiation factor 4A (Silveira et al., 2009), ubiquitin (Infante et al., 2008), and 60S acidic ribosomal protein (Le et al., 2012), are commonly used as reference genes for gene expression studies. There are studies which have used reference genes for the normalization of gene expression in safflower (Li et al., 2012; Cao et al., 2013). But these reference genes were not evaluated for qPCR data in safflower tissues. In this study, nine housekeeping genes namely CtACT (actin), CtGAPDH (glyceraldehyde 3-phosphate dehydrogenase), CtE1F4A (elongation factor 1 alpha), CtTUA (alpha-tubulin), CtTUB (beta-tubulin), CtPP2A (serine/threonine-protein phosphatase), CtE1F4A (eukaryotic initiation factor 4A), CtUBI (Ubiquitin), and Ct60S (60S acidic ribosomal protein) were selected as candidate reference genes from our constructeded transcriptome data of safflower. At the same time, two representative chemotype of safflower varieties, with-HSYA (yellow flower) and without-HSYA (white flower) were selected to identify reference genes. The expression of these genes in the bract, ovary, stem, leaf, calyx, petal (I–IV, four different stages of development during flowering) of two safflower lines were analyzed using four different algorithms, that is, geNorm (Vandesompele et al., 2002), NormFinder (Andersen et al., 2004), ΔCt approach (Silver et al., 2006), and BestKeeper (Pfaffl et al., 2004). Our study will help to achieve more accurate and reliable results in a wide variety of safflower samples.

Materials and methods

Plant material

The seeds of two lines of safflower, Carthamus tinctorius L., Asteraceae (ZHH0082 with yellow flower and Xin Honghua NO.7 with white flower) from the Chinese populations were cultivated in the field at the medicinal botanical garden of Second Military Medical University. The samples (bract, ovary, stem, leaf, calyx, petal (I–IV)) were collected from three different flowering plants (biological triplicates) and immediately frozen in liquid nitrogen. All 54 samples were then stored at -70 ºC for RNA extraction.

Total RNA extraction

Total RNA was isolated using the RNA Extraction Plant Mini Kit (LifeFeng, Shanghai, China) according to the protocol provided by the manufacturer. RNA concentration and quality were measured with the NanoDrop spectrophotometer (NanoDrop Technologies) and agarose gel electrophoresis. Only the RNA samples with A₂₆₀/A₂₈₀ ratios between 1.9 and 2.1 and A₂₆₀/A₂₃₀ ratios greater than 2.0 were used for cDNA synthesis.

First strand cDNAs synthesis

First strand cDNAs were synthesized according to the manufacturer's instructions of TransScript^® One-Step gDNA Removal and cDNA Synthesis SuperMix (TransGen Biotech, Beijing, China). Total RNA (1 µg), suitable volumes of H₂O, and 1 µl anchored Oligo (dT) 18 primer (0.5 µg/µl) were mixed and incubated at 65 ºC for 5 min followed by cooling on ice. The reverse transcriptase reactions were started after adding 10 µl 2× TS reaction mix, 1 µl Enzyme Mix, and 1 µl gDNA Remover at 42 ºC for 45 min. And then the mixture was heated for 5 min at 85 ºC for inactivating the enzymes. All cDNA samples were diluted 1:10 with RNase-free water before being used as templates in the qPCR analysis.

Q-PCR

Nine sequences were selected as candidate reference genes from the petal EST libraries of C. tinctorius (Table 1). The qPCR primers were designed using the Beacon Designer v8.0 software. Amplicon lengths varied from 75 to 200 bp, with melting temperatures (Tm) varying between 52 ºC and 60 ºC and primer lengths between 18 and 22 bp. Primer pairs were tested for specificity by qPCR, followed by a dissociation curve and agarose gel electrophoresis (^{Fig. S1}). PCR reactions were performed in 96-well plates with the CFX96 Touch^TM Real-Time PCR Detection System (Bio-Rad Laboratories, CA, USA) and ABI 7500 Real-Time PCR System (Applied Biosystems, Foster City, USA). The CFX96 was used to obtain the qPCR data of nine candidate genes and ABI 7500 was used to obtain the qPCR data of CtFAD2-10 (KC257456, F-CTTTACCGTATGGCTTTAG, R-GTGTGTTGAAGGTATGTG), CtKASII (KC257458, F-GACAGGTTTATGCTCTAC, R-CAATCAGAACTCCACATC), and two stable reference genes. Each reaction comprising 20 µl of the following was prepared as follows: 0.3 µm forward primer, 0.3 µm reverse primer, 10 µl 2× TranStartTM Top Green qPCR SuperMix (TransGen Biotech, Beijing, China), 2 µl cDNA, 0.4 µl passive reference dye II (only used in ABI system), and suitable volumes of H₂O. The thermocycling conditions were set at 95 ºC for 30 s, followed by 40 cycles of 10 s at 95 ºC for template denaturation, 15 s at Tm for annealing, and 20 s (30 s for ABI system) at 72 ºC for extension and fluorescence measurement. Afterwards, the dissociation curve was obtained by heating the amplicon from 60 ºC to 95 ºC and reading at each 0.5 ºC increase (0.2 ºC for ABI system). Three technical replicates were performed for each PCR reaction.

Table 1
Nine candidate reference genes and their primer sequences for qPCR.

Gene identification/gene description	E value ID (%)	Primer sequence	Amplicon size	Amplification efficiency	GeneBank Accession Number	Tm
CtACT	0.0	F-ACTGGTGTTATGGTAGGA	89	1.85	KJ634809	60
actin	100	R-GGATACTTCAAGGTAAGGATA
CtGAPDH	0.0	F-GTTGTGGACTTAACCGTAA	78	1.83	KJ634805	53
glyceraldehyde 3-phosphate dehydrogenase	93	R-TTCTGATTCCTCCTTAATAGC
CtEF1	0.0	F-CCAAGAGACCATCAGACAA	76	1.88	KJ634806	58
elongation factor 1 alpha	98	R-GGCACAGTTCCAATACCA
CtTUA	0.0	F-CTACACCAACCTCAATCG	91	1.89	KJ634803	58
α-tubulin	99	R-AGTCACATCCACATTCAAG
CtTUB	0.0	F-GGAAGAGGAGTATGATGA	93	1.88	KJ634802	54
β-tubulin	99	R-AATGGCAGTTGAGATTAC
CtPP2A	0.0	F-GAGAACCTGATGTAACGAGAC	80	1.89	KJ634804	58
serine/threonine-protein phosphatase	95	R-ACCACCAAGCAAGCAATC
CtE1F4A	0.0	F-CGCTGATTACATTAAGATG	90	1.85	KJ634807	54
eukaryotic initiation factor 4A	98	R-AATTGGAAGATATCGTAGAT
CtUBI	3e−87	F-TCACTTATGTTTACCAGAA	150	1.85	KJ634808	54
Ubiquitin	98	R-GCTTTCAATTTCAACTCA
Ct60S	9e−165	F-CATCCATTATCCAACAATC	92	1.86	KJ634810	54
60S acidic ribosomal protein	88	R-AAGAGTAATCAGTCTCCA

Data analysis

The PCR efficiency shown in Table 1 was calculated for each candidate gene with LinRegPCR program (Ruijter et al., 2009). The raw fluorescence data, Ct values, and the relative quantity (ΔCt) values of candidate genes were generated with the CFX manager software (BioRad). The Ct values of CtFAD2-10, CtKASII, and two stable reference genes were calculated using the 7500 software v2.06 (ABI). The data obtained were converted into correct input files and analyzed using geNorm, NormFinder, ΔCt approach, and BestKeeper. In Fig. 1, normalized CtFAD2-10 and CtKASII expressions were calculated using the geometric means of CtUBI and Ct60S Ct values of each sample.

Fig. 1
Relative quantification of CtFAD2-10 and CtKASII expression using different internal controls analyzed by the 2^-ΔΔCt method in all samples. Relative quantification of CtFAD2-10 and CtKASII expression were detected using CtUBI and Ct60S both individually and in combination. The relative expression of petal I was set to 1. (A) CtUBI, Ct60S and the geometric average of CtUBI and Ct60S were used as internal controls for CtFAD2-10 expression. (B) CtUBI, Ct60S and the geometric average of CtUBI and Ct60S were used as internal controls for CtKASII expression. B, bract; C, calyx; L, leaf; O, ovary; S, stem; I–IV, petal I–IV.

Results

Selection of putative reference genes for qPCR experiments

In order to find the best reference genes in safflower, nine housekeeping genes, CtACT, CtGAPDH, CtE1F4A, CtTUA, CtTUB, CtPP2A, CtE1F4A, CtUBI, and Ct60S, were selected as putative reference genes (Table 1). These genes were used to BLAST search against a C. tinctorius EST (expressed sequence tag) library constructed by our laboratory from the petal during flowering. The housekeeping gene commonly used as reference gene is the 18S rRNA. In this study, however, the Ct values of 18S in the safflower tissues were at least 10 cycles higher than the Ct values of the other candidate genes (data not shown). Such high Ct values make 18S unsuitable for use as a reference gene (Li et al., 2012). The dissociation cures of amplicons of these nine candidate genes exhibited a single peak showed that the primer pairs were specificity, which were also verified by agarose gel electrophoresis results (^{Fig. S1}). The expression of these genes in bract, ovary, stem, leaf, calyx, petal (I–IV, four different stages of development during flowering, Fig. 2, ^{Table S1}) was analyzed. Since flowering begins in the outer circle of florets and progresses centripetally toward the center of the capitulum, the four stages of petals can be found in one capitulum. The reason we differentiate the four stages of petals is that the compounds have significant variation during the development of the C. tinctorius flowers (Salem et al., 2011) and the gene expressions are different as well (Mallona et al., 2010). In our study, a total of 54 samples (two lines of safflower (ZHH0082 and Xin Honghua NO.7), nine organs, three biological replicates) were used to evaluate the stability of putative reference genes. The significant difference between the ZHH0082 and Xin Honghua NO.7 was the color of flowers. The flowers of ZHH0082 were yellow and the flowers of Xin Honghua NO.7 were white.

Fig. 2
Two lines of safflower. (1) A, B, C are the ZHH0082 (Yellow); (2) D, E, F are the Xin Honghua NO.7 (White); (A, D) capitulum of safflower; (B, E) longitudinal section of capitulum; (C, F) four different stages of the petal during flowering. Black arrows point to the samples collected from safflower. B, bract; O, ovary; T, stem; L, leaf; C, calyx; P, petal.

Expression stability of putative reference genes via differential statistical analyses

The cycle threshold (Ct) values profiling of candidate genes in different organs and safflower lines are shown in Fig. 3. The mean of each of the nine candidate genes in 54 samples are shown in Table S2. In this study, geNorm, NormFinder, ΔCt approach and BestKeeper were used to analyze the stability of gene expression.

Fig. 3
Ct values of candidate reference genes. Expression data showed the Ct values (black spots) of each gene. The vertical boxes represent the 25th and 75th percentiles with error bars (whiskers). The lines across the boxes are depicted as the median. (A) All samples of safflower, (B) samples of ZHH0082 (Yellow), and (C) samples of Xin Honghua NO.7 (White).

GeNorm calculated the gene expression stability value M (the average pairwise variation of a particular gene with all other control genes) for a reference gene. The lower the M value, the more stably expressed the gene is. NormFinder used a model-based approach for identifying the optimal normalization gene(s). The intra- and intergroup variations were calculated and included in the gene expression stability values. Genes with the lowest values had the most stable expression. GeNorm and NormFinder used the linear scale expression quantities, which can be calculated from the Ct values by using a standard curve or comparative CT method (Schmittgen and Livak, 2008). ΔCt approach compared the ΔCt variation of pairs of housekeeping genes within individual samples to identify the reference genes. The genes with lower ΔCt variation were stably expressed. BestKeeper, an Excel-based tool, determined the most stably expressed genes based on the variation (SD and SV values) and repeated pairwise correlation analysis. The rank-ordered genes calculated by these four algorithms were further analyzed by RankAggreg (Pihur et al., 2009), which calculated footrule distances and obtained the consensus rank list of genes by means of Cross-Entropy Monte Carlo algorithm. Six groups (AS, Y, W, F, YF, WF) were analyzed by these algorithms and the results are shown in Tables 2, 3 and Fig. S2.

Table 2
Ranking of the candidate reference genes according to their values calculated by geNorm, NormFinder, ΔCt approach and BestKeeper in AS, Y and W.

Rank	All samples (AS)					Yellow (Y)					White (W)
	geNorm (M)	NormFinder (SV)	ΔCt approach	BestKeeper		geNorm (M)	NormFinder (SV)	ΔCt approach	BestKeeper		geNorm (M)	NormFinder (SV)	ΔCt approach	BestKeeper
1	CtUBI	Ct60S	Ct60S	Ct PP2A		Ct60S	CtUBI	Ct60s	CtPP2A		CtACT	CtPP2A	CtUBI	CtPP2A
	0.776	0.061	0.671			0.774	0.063	0.670			0.756	0.062	0.643
2	Ct60S	CtACT	CtACT	Ct UBI		CtUBI	Ct60s	CtACT	CtUBI		CtTUB	CtEF1	CtTUB	CtUBI
	0.777	0.066	0.673			0.781	0.070	0.672			0.759	0.072	0.649
3	CtACT	CtUBI	CtUBI	Ct 60S		CtTUB	CtACT	CtTUB	CtE1F4A		CtGAPDH	CtACT	Ct60S	Ct60S
	0.788	0.068	0.681			0.788	0.081	0.683			0.767	0.092	0.651
4	CtTUB	CtGAPDH	CtTUB	Ct TUB		CtACT	CtE1F4A	CtUBI	Ct60S		CtUBI	Ct60S	CtACT	CtTUB
	0.837	0.076	0.724			0.804	0.081	0.689			0.770	0.113	0.660
5	CtE1F4A	CtEF1	CtE1F4A	Ct E1F4A		CtEF1	CtTUB	CtGAPDH	CtTUB		Ct60S	CtGAPDH	CtE1F4A	CtE1F4A
	0.839	0.078	0.736			0.814	0.085	0.748			0.773	0.123	0.699
6	CtGAPDH	CtE1F4A	CtGAPDH	Ct A CT		CtE1F4A	CtGAPDH	CtEF1	CtACT		CtE1F4A	CtTUA	CtGAPDH	CtGAPDH
	0.845	0.085	0.748			0.815	0.088	0.772			0.780	0.137	0.700
7	CtEF1	CtTUA	CtEF1	Ct GAPDH		CtGAPDH	CtEF1	CtPP2A	CtEF1		CtPP2A	CtUBI	CtPP2A	CtACT
	0.851	0.127	0.785			0.878	0.103	0.863			0.856	0.144	0.766
8	CtPP2A	CtTUB	CtPP2A	Ct TUA		CtPP2A	CtPP2A	CtE1F4A	CtGAPDH		CtEF1	CtE1F4A	CtEF1	CtTUA
	0.975	0.127	0.818			1.021	0.148	0.869			0.880	0.146	0.781
9	CtTUA	CtPP2A	CtTUA	Ct EF1		CtTUA	CtTUA	CtTUA	CtTUA		CtTUA	CtTUB	CtTUA	CtEF1
	1.104	0.138	0.957			1.130	0.168	0.974			1.077	0.154	0.926
RankAggrega	Ct60S/CtUBI/CtACT					CtUBI/Ct60S/CtTUB					CtUBI/Ct60S/CtTUB

• ^a
  Top three ranked genes were shown.

Table 3
Ranking of the candidate reference genes according to their values calculated by geNorm, NormFinder, ΔCt approach and BestKeeper in F, YF and WF.

Rank	All flowers (F)					Yellow-Flower (YF)					White-Flower (WF)
	geNorm (M)	NormFinder (SV)	ΔCt approach	BestKeeper		geNorm (M)	NormFinder (SV)	ΔCt approach	BestKeeper		geNorm (M)	NormFinder (SV)	ΔCt approach	BestKeeper
1	Ct60S	CtACT	Ct60S	CtPP2A		CtE1F4A	CtE1F4A	CtE1F4A	CtPP2A		Ct60S	CtEF1	Ct60S	CtPP2A
	0.705	0.073	0.625			0.727	0.061	0.635			0.604	0.039	0.550
2	CtE1F4A	CtEF1	CtACT	CtUBI		Ct60S	CtACT	Ct60S	CtUBI		CtE1F4A	Ct60S	CtACT	CtUBI
	0.742	0.086	0.650			0.755	0.071	0.665			0.657	0.050	0.583
3	CtUBI	CtUBI	CtE1F4A	Ct60S		CtTUB	CtGAPDH	CtTUB	Ct60S		CtACT	CtPP2A	CtTUB	Ct60S
	0.749	0.097	0.661			0.764	0.079	0.665			0.671	0.060	0.591
4	CtACT	CtE1F4A	CtUBI	CtTUB		CtUBI	Ct60S	CtACT	CtE1F4A		CtTUB	CtUBI	CtE1F4A	CtTUB
	0.777	0.097	0.665			0.768	0.098	0.687			0.678	0.060	0.593
5	CtTUA	Ct60S	CtGAPDH	CtE1F4A		CtGAPDH	CtTUB	CtUBI	CtTUB		CtTUA	CtACT	CtUBI	CtE1F4A
	0.806	0.100	0.707			0.819	0.109	0.692			0.683	0.091	0.594
6	CtGAPDH	CtGAPDH	CtTUB	CtACT		CtACT	CtUBI	CtGAPDH	CtACT		CtUBI	CtTUA	CtTUA	CtGAPDH
	0.808	0.112	0.709			0.829	0.110	0.697			0.687	0.098	0.637
7	CtTUB	CtTUB	CtTUA	CtGAPDH		CtEF1	CtEF1	CtEF1	CtEF1		CtPP2A	CtE1F4A	CtPP2A	CtACT
	0.809	0.123	0.754			0.836	0.130	0.767			0.745	0.126	0.666
8	CtEF1	CtPP2A	CtPP2A	CtTUA		CtTUA	CtTUA	CtPP2A	CtGAPDH		CtGAPDH	CtGAPDH	CtGAPDH	CtTUA
	0.875	0.127	0.755			0.888	0.146	0.817			0.755	0.158	0.677
9	CtPP2A	CtTUA	CtEF1	CtEF1		CtPP2A	CtPP2A	CtTUA	CtTUA		CtEF1	CtTUB	CtEF1	CtEF1
	0.885	0.128	0.760			0.981	0.183	0.826			0.900	0.164	0.737
RankAggrega	CtUBI/CtE1F4A/Ct60S			CtE1F4A/Ct60S/CtTUB				Ct60S/CtUBI/CtACT

• ^a
  Top three ranked genes were shown.

In AS (all 54 samples of the two lines of safflower), CtUBI, Ct60S, and CtACT were the three most stably expressed candidate genes calculated by geNorm, NormFinder and ΔCt. In the list of BestKeeper, the top three were CtPP2A, CtUBI, and Ct60S. The most stably expressed gene identified by BestKeeper was CtPP2A, which had the highest variation in the results from other methods (geNorm ranked 8, NormFinder ranked 9, and ΔCt ranked 8). The top three aggregate order calculated by RankAggreg was Ct60S/CtUBI/CtACT in AS (Table 2).

The results of AS were analyzed by segregating into two subgroups, Y and W. The Y group included 27 samples collected from ZHH0082 (Yellow) and W contained 27 samples from Xin Honghua NO.7 (White). The four algorithms produced similar ranking lists in AS and Y. But there were more differences between two lines, Y and W. In Y, Ct60S was the first, second, and the first best-ranked gene in the geNorm, NormFinder, and ΔCt approach, respectively. However, Ct60S was ranked at fifth, fourth, and third in W. In the analysis of NormFinder, the position of CtPP2A had remarkable variation between Y and W, which was the eighth in Y but the first in W. Although there were discrepancies among the orders produced by these algorithms in the two lines, the aggregate orders were the same, namely CtUBI/Ct60S/CtTUB (Table 2).

The results analyzed by algorithms were different in all Flowers (F), Yellow-Flower (YF, flowers of ZHH0082), and White-Flower (WF, flowers of Xin Honghua NO.7). According to the results of NormFinder, the CtE1F4A was the second most stably expressed gene in F, seventh in YF, and first in WF. In particular, CtE1F4A was the most stably expressed gene in YF calculated by the geNorm, NormFinder, and ΔCt approach. The upgraded positions of CtE1F4A resulted in changing of the aggregate orders in SA, F, and YF. But the order in WF was Ct60S/CtUBI/CtACT, which was the same as that in AS (Table 3).

It has been suggested that the use of two or more reference genes for RT-qPCR studies might generate more reliable results (Vandesompele et al., 2002). To determine the optimal number of genes required for accurate normalization, pairwise variations V_n/V_n+1 were calculated based on the normalization factor (NF_n and NF_n+1) values according to the geNorm algorithm. If the V_n/V_n+1 of n genes were below the cut-off value, the additional housekeeping gene (n + 1) was not necessary for reliable normalization. The cut-off value of pairwise variations V_n/V_n+1 was set at 0.15. In the results of the six groups analyzed, at least four reference genes can be used for accurate normalization in AS (V₂/V₃ = 0.173, V₃/V₄ = 0.176), Y (V₂/V₃ = 0.193, V₃/V₄ = 0.175) and F (V₂/V₃ = 0.193, V₃/V₄ = 0.183). In W (V₂/V₃ = 0.22) and YF (V₂/V₃ = 0.203), the best number was three. Two genes were sufficient for expression normalization in WF (V₂/V₃ < 1.5) (Fig. 4).

Fig. 4
Optimal number of reference genes for normalization according to geNorm results. Pairwise variation (V_n/V_n+1) was measured between the normalization factors NF_n and NF_n+1. The inclusion of an additional reference gene was not required below the cut-off value of 0.15 (the chartreuse line). AS, all samples; Y, ZHH0082 (Yellow); W, Xin Honghua NO.7 (White). F, flower; YF, yellow flower; WF, white flower.

Quantification of CtFAD2-10 and CtKASII expression with stable reference genes

The expression of CtFAD2-10, one of the FAD2 gene family in safflower, was normalized using KASII in an early study (Cao et al., 2013). Keto acyl-acyl carrier protein synthase II (KASII) also plays a role in the fatty acid biosynthesis in plants. The expression of KASII was not evaluated for using as an internal control for normalization. When compared with organs obtained from different individuals, single housekeeping gene RNA levels were not appropriate to be used for normalization of RNA levels (Tricarico et al., 2002). To further verify the suitability of reference genes selected in the present study, CtFAD2-10 and CtKASII expression levels were detected in safflower (Fig. 1). The relative expression data were calculated using 2^-ΔΔCt method. The internal control genes were the CtUBI and Ct60S and the geometric average of CtUBI and Ct60S.

Its pattern of expression was assessed in all samples (bract, ovary, stem, leaf, calyx, petal (I–IV)). Similar expression patterns were generated when either one or two of the most stable genes were used for normalization (Fig. 1A and B). When normalized the geometric average of CtUBI and Ct60S, transcript abundance of CtKASII gradually increased in different developmental stages of flower, peaking at the Stage IV (Fig. 1B). It suggests that the CtKASII was not a stable gene for gene expression normalization. The expression levels of CtFAD2-10 were high in petals, especially in the flowering stage IV (Fig. 1A). When only one reference gene was employed, expression profiles of CtKASII were similar (Fig. 1B), but differences were evident in estimated transcript abundance, which was higher when normalized against CtUBI than against Ct60S.

Discussion

Extensive studies are being carried out on the safflower, because it is both an oil crop and a medicinal plant. Furthermore, research on the biosynthesis of these flavonoids and fatty acid metabolism were also reported recently (Cao et al., 2013; Li et al., 2012). Because it is a sensitive, specific, reproducible, and conventional method, qPCR has become an essential tool for gene expression analysis, especially the gene expression of the secondary metabolites pathway(Al-Ghazi et al., 2009).

Although some reference genes have been used for the normalization of gene expression in safflower, but these reference genes were not evaluated for qPCR data in safflower tissues (Li et al., 2012; Cao et al., 2013). Recently, the safflower reference genes were reported in our process of revising, but their transcriptome data were only from the seed and algorithms for the reference genes evaluation were incomplete (Li et al., 2015). In this study, two representative chemotype of safflower lines, with-HSYA (yellow flower) and without-HSYA (white flower), was chosen to identify the safflower reference genes. The nine housekeeping genes were selected as the candidate genes from transcriptome data of different developmental flowering stage of safflower constructed by our lab. The profiling of these genes was carried out using four algorithms and was ranked by RankAggreg. All 54 samples were separated into six groups in order to evaluate the variation of these candidate genes in different lines, organs, and flowers. Although there were different lists of the most stably expressed genes from the results of geNorm, NormFinder, ΔCt approach and BestKeeper, the CtUBI and Ct60S were in the top three ranked genes according to the RankAggreg analysis in the five groups (except the YF). Our findings were in accordance with the result that UBQ (ubiquitin) showed highly stable expression in Arabidopsis (Czechowski et al., 2005). Similarly, UBQ was the recommended housekeeping gene for normalization in poplar (Brunner et al., 2004) and rice (Jain et al., 2006). However, UBQ was not suggested using as internal controls to normalize gene expression in soybean (Jian et al., 2008). As for 60S, it was found to be the best reference gene in different tissues and under various stress conditions in soybean (Le et al., 2012). Therefore, we used CtUBI and Ct60S as reference genes and evaluated the expression of CtFAD2-10 and CtKASII in safflower. CtKASII had been used as the reference gene to profile the CtFAD2 gene expression. But according to our study, the CtKASII was not suitable for gene expression normalization because its expressions were highly variable.

Our results indicated that the stability of reference gene expression must be validated for each line and the development stages of flowering in safflower. Based on these results, we strongly suggest that CtUBI and Ct60S should be used as reference genes for gene expression in safflower. The identification of these two stable reference genes will enable accurate and reliable gene expression studies related to functional genomics and metabolomics about safflower.

• Ethical disclosures
  All authors declare to our manuscript did not involve any ethical issues.

Acknowledgments

The authors thank the editor and referees for their constructive comments. This work was funded by the National Natural Science Foundation of China (81173484 and 81473300), Shanghai Natural Science Foundation (13ZR1448200) and “863” High Technology Project (2008AA02Z137) grants from the Ministry of Science Technology of P. R. China.

Appendix A Supplementary data

Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.bjp.2016.05.006.

References

• Al-Ghazi, Y., Bourot, S., Arioli, T., Dennis, E.S., Llewellyn, D.J., 2009. Transcript profiling during fiber development identifies pathways in secondary metabolism and cell wall structure that may contribute to cotton fiber quality. Plant Cell Physiol. 50, 1364-1381.
• Andersen, C.L., Jensen, J.L., Ørntoft, T.F., 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.
• Andersen, O.M., Markham, K.R., 2010. Flavonoids: Chemistry, Biochemistry and Applications. CRC Press.
• Asgarpanah, J., Kazemivash, N., 2013. Phytochemistry, pharmacology and medicinal properties of Carthamus tinctorius L. Chin. J. Integr. Med. 19, 153-159.
• Bas, A., Forsberg, G., Hammarström, S., Hammarström, M.L., 2004. Utility of the housekeeping genes 18S rRNA, β-Actin and glyceraldehyde-3-phosphate-dehydrogenase for normalization in real-time quantitative reverse transcriptase-polymerase chain reaction analysis of gene expression in human T lymphocytes. Scand. J. Immunol. 59, 566-573.
• Brunner, A.M., Yakovlev, I.A., Strauss, S.H., 2004. Validating internal controls for quantitative plant gene expression studies. BMC Plant Biol. 4, 14.
• Bustin, S.A., Benes, V., Garson, J.A., Hellemans, J., Huggett, J., Kubista, M., Mueller, R., Nolan, T., Pfaffl, M.W., Shipley, G.L., Vandesompele, J., Wittwer, C.T., 2009. The MIQE guidelines: minimum information for publication of quantitative real-time PCR experiments. Clin. Chem. 55, 611-622.
• Cao, S., Zhou, X.R., Wood, C.C., Green, A.G., Singh, S.P., Liu, L., Liu, Q., 2013. A large and functionally diverse family of Fad2 genes in safflower (Carthamus tinctorius L.). BMC Plant Biol. 13, 5.
• Chapman, M., Burke, J., 2007. DNA sequence diversity and the origin of cultivated safflower (Carthamus tinctorius L.; Asteraceae). BMC Plant Biol. 7, 60.
• Chapman, M.A., Hvala, J., Strever, J., Matvienko, M., Kozik, A., Michelmore, R.W., Tang, S., Knapp, S.J., Burke, J.M., 2009. Development, polymorphism, and cross-taxon utility of EST–SSR markers from safflower (Carthamus tinctorius L.). Theor. Appl. Genet. 120, 85-91.
• Czechowski, T., Stitt, M., Altmann, T., Udvardi, M.K., Scheible, W.-R., 2005. Genome-wide identification and testing of superior reference genes for transcript normalization in Arabidopsis. Plant Physiol. 139, 5-17.
• Feng, N., Li, Y., Tang, J., Wang, Y., Guo, M., 2010. cDNA-AFLP analysis on transcripts associated with hydroxysafflor yellow A(HSYA) biosynthetic pathway in Carthamus tinctorius Biochem. Syst. Ecol. 38, 971-980.
• Feng, Z.M., He, J., Jiang, J.S., Chen, Z., Yang, Y.N., Zhang, P.C., 2013. NMR solution structure study of the representative component hydroxysafflor yellow A and other quinochalcone C-glycosides from Carthamus tinctorius J. Nat. Prod. 76, 270-274.
• Gachon, C., Mingam, A., Charrier, B., 2004. Real-time PCR: what relevance to plant studies?. J. Exp. Bot. 55, 1445-1454.
• Gecgel, U., Demirci, M., Esendal, E., Tasan, M., 2007. Fatty acid composition of the oil from developing seeds of different varieties of safflower (Carthamus tinctorius L.). J. Am. Oil Chem. Soc. 84, 47-54.
• Golkar, P., Arzani, A., Rezaei, A.M., 2011. Genetic variation in safflower (Carthamus tinctorious L.) for seed quality-related traits and inter-simple sequence repeat (ISSR) markers. Int. J. Mol. Sci. 12, 2664-2677.
• Guénin, S., Mauriat, M., Pelloux, J., Van Wuytswinkel, O., Bellini, C., Gutierrez, L., 2009. Normalization of qRT-PCR data: the necessity of adopting a systematic, experimental conditions-specific, validation of references. J. Exp. Bot. 60, 487-493.
• Guan, L., Wu, W., Hu, B., Li, D., Chen, J., Hou, K., Wang, L., 2014. Devolopmental and growth temperature regulation of omega-3 fatty acid desaturase genes in safflower (Carthamus tinctorius L.). Genet. Mol. Res. 13, 6623.
• Gutierrez, L., Mauriat, M., Guénin, S., Pelloux, J., Lefebvre, J.-F., Louvet, R., Rusterucci, C., Moritz, T., Guerineau, F., Bellini, C., Van Wuytswinkel, O., 2008. The lack of a systematic validation of reference genes: a serious pitfall undervalued in reverse transcription-polymerase chain reaction (RT-PCR) analysis in plants. Plant Biotechnol. J. 6, 609-618.
• Infante, C., Matsuoka, M., Asensio, E., Canavate, J., Reith, M., Manchado, M., 2008. Selection of housekeeping genes for gene expression studies in larvae from flatfish using real-time PCR. BMC Mol. Biol. 9, 28.
• Jain, M., Nijhawan, A., Tyagi, A.K., Khurana, J.P., 2006. Validation of housekeeping genes as internal control for studying gene expression in rice by quantitative real-time PCR. Biochem. Biophys. Res. Commun. 345, 646-651.
• Jarosova, J., Kundu, J., 2010. Validation of reference genes as internal control for studying viral infections in cereals by quantitative real-time RT-PCR. BMC Plant Biol. 10, 146.
• Jian, B., Liu, B., Bi, Y., Hou, W., Wu, C., Han, T., 2008. Validation of internal control for gene expression study in soybean by quantitative real-time PCR. BMC Mol. Biol. 9, 59.
• Jun, M.S., Ha, Y.M., Kim, H.S., Jang, H.J., Kim, Y.M., Lee, Y.S., Kim, H.J., Seo, H.G., Lee, J.H., Lee, S.H., Chang, K.C., 2011. Anti-inflammatory action of methanol extract of Carthamus tinctorius involves in heme oxygenase-1 induction. J. Ethnopharmacol. 133, 524-530.
• Le, D.T., Aldrich, D.L., Valliyodan, B., Watanabe, Y., Van Ha, C., Nishiyama, R., Guttikonda, S.K., Quach, T.N., Gutierrez-Gonzalez, J.J., Tran, L.-S.P., 2012. Evaluation of candidate reference genes for normalization of quantitative RT-PCR in soybean tissues under various abiotic stress conditions. PLoS ONE 7, e46487.
• Li, D., Hu, B., Wang, Q., Liu, H., Pan, F., Wu, W., 2015. Identification and evaluation of reference genes for accurate transcription normalization in safflower under different experimental conditions. PLOS ONE 10, e0140218.
• Li, H., Dong, Y., Yang, J., Liu, X., Wang, Y., Yao, N., Guan, L., Wang, N., Wu, J., Li, X., 2012. De novo transcriptome of safflower and the identification of putative genes for oleosin and the biosynthesis of flavonoids. PLoS ONE 7, e30987.
• Li, Y., Wang, Z., Chang, H., Wang, Y., Guo, M., 2010. Expression of CT-wpr, screened by cDNA-AFLP approach, associated with hydroxysafflor yellow A in Carthamus tinctorius L. Biochem. Syst. Ecol. 38, 1148-1155.
• Liu, D., Shi, L., Han, C., Yu, J., Li, D., Zhang, Y., 2012. Validation of reference genes for gene expression studies in virus-infected Nicotiana benthamiana using quantitative real-time PCR. PLoS ONE 7, e46451.
• Liu, Q., Cao, S., Zhou, X.-R., Wood, C., Green, A., Singh, S., 2013. Nonsense-mediated mRNA degradation of CtFAD2-1 and development of a perfect molecular marker for olol mutation in high oleic safflower (Carthamus tinctorius L.). Theor. Appl. Genet. 126, 2219-2231.
• Mallona, I., Lischewski, S., Weiss, J., Hause, B., Egea-Cortines, M., 2010. Validation of reference genes for quantitative real-time PCR during leaf and flower development in Petunia hybrida BMC Plant Biol. 10, 4.
• Peng, S., Feng, N., Guo, M., Chen, Y., Guo, Q., 2008. Genetic variation of Carthamus tinctorius L. and related species revealed by SRAP analysis. Biochem. Syst. Ecol. 36, 531-538.
• Pfaffl, M.W., Tichopad, A., Prgomet, C., Neuvians, T.P., 2004. Determination of stable housekeeping genes, differentially regulated target genes and sample integrity: BestKeeper–Excel-based tool using pair-wise correlations. Biotechnol. Lett. 26, 509-515.
• Pihur, V., Datta, S., Datta, S., 2009. RankAggreg, an R package for weighted rank aggregation. BMC Bioinform. 10, 62.
• Ruijter, J.M., Ramakers, C., Hoogaars, W.M.H., Karlen, Y., Bakker, O., van den Hoff, M.J.B., Moorman, A.F.M., 2009. Amplification efficiency: linking baseline and bias in the analysis of quantitative PCR data. Nucleic Acids Res. 37, e45.
• Salem, N., Msaada, K., Hamdaoui, G., Limam, F., Marzouk, B., 2011. Variation in phenolic composition and antioxidant activity during flower development of safflower (Carthamus tinctorius L.). J. Agric. Food Chem. 59, 4455-4463.
• Schmidt, G.W., Delaney, S.K., 2010. Stable internal reference genes for normalization of real-time RT-PCR in tobacco (Nicotiana tabacum) during development and abiotic stress. Mol. Genet. Genomics 283, 233-241.
• Schmittgen, T.D., Livak, K.J., 2008. Analyzing real-time PCR data by the comparative CT method. Nat. Protoc. 3, 1101-1108.
• Silveira, E., Alves-Ferreira, M., Guimaraes, L., da Silva, F., Carneiro, V., 2009. Selection of reference genes for quantitative real-time PCR expression studies in the apomictic and sexual grass Brachiaria brizantha BMC Plant Biol. 9, 84.
• Silver, N., Best, S., Jiang, J., Thein, S., 2006. Selection of housekeeping genes for gene expression studies in human reticulocytes using real-time PCR. BMC Mol. Biol. 7, 33.
• Sun, L., Yang, L., Fu, Y., Han, J., Xu, Y., Liang, H., Cheng, Y., 2013. Capacity of HSYA to inhibit nitrotyrosine formation induced by focal ischemic brain injury. Nitric Oxide 35, 144-151.
• Tanaka, Y., Brugliera, F., Kalc, G., Senior, M., Dyson, B., Nakamura, N., Katsumoto, Y., Chandler, S., 2010. Flower color modification by engineering of the flavonoid biosynthetic pathway: practical perspectives. Biosci. Biotechnol. Biochem. 74, 1760-1769.
• Thellin, O., Zorzi, W., Lakaye, B., De Borman, B., Coumans, B., Hennen, G., Grisar, T., Igout, A., Heinen, E., 1999. Housekeeping genes as internal standards: use and limits. J. Biotechnol. 75, 291-295.
• Tian, Y., Yang, Z.-F., Li, Y., Qiao, Y., Yang, J., Jia, Y.-Y., Wen, A.-D., 2010. Pharmacokinetic comparisons of hydroxysafflower yellow A in normal and blood stasis syndrome rats. J. Ethnopharmacol. 129, 1-4.
• Tricarico, C., Pinzani, P., Bianchi, S., Paglierani, M., Distante, V., Pazzagli, M., Bustin, S.A., Orlando, C., 2002. Quantitative real-time reverse transcription polymerase chain reaction: normalization to rRNA or single housekeeping genes is inappropriate for human tissue biopsies. Anal. Biochem. 309, 293-300.
• 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, 1-12.
• Wan, H., Yuan, W., Ruan, M., Ye, Q., Wang, R., Li, Z., Zhou, G., Yao, Z., Zhao, J., Liu, S., Yang, Y., 2011. Identification of reference genes for reverse transcription quantitative real-time PCR normalization in pepper (Capsicum annuum L.). Biochem. Biophys. Res. Commun. 416, 24-30.
• Wang, C., Huang, Q., Zhu, X., Duan, Y., Yuan, S., Bai, X., 2013. Hydroxysafflor yellow A suppresses oleic acid-induced acute lung injury via protein kinase A. Toxicol. Appl. Pharmacol. 272, 895-904.
• Yang, J., Wang, Y., Guo, M.-L., 2011. Identification and mapping of a novel hydroxysafflor yellow A (HSYA) biosynthetic gene in Carthamus tinctorius Biochem. Genet. 49, 410-415.
• Yeilaghi, H., Arzani, A., Ghaderian, M., Fotovat, R., Feizi, M., Pourdad, S.S., 2012. Effect of salinity on seed oil content and fatty acid composition of safflower (Carthamus tinctorius L.) genotypes. Food Chem. 130, 618-625.
• Zhang, Z., Guo, M., Zhang, J., 2009. Identification of AFLP fragments linked to hydroxysafflor yellow A in Flos Carthami and conversion to a SCAR marker for rapid selection. Mol. Breed. 23, 229-237.

Publication Dates

• Publication in this collection
  Sep-Oct 2016

History

• Received
  05 June 2015
• Accepted
  24 May 2016