Calendula L. (marigold) is the type genus of the small tribe Calenduleae (Asteraceae). While all other genera of the Calenduleae are native to southern Africa, Calendula is distributed in the Northern Hemisphere. Calendula species occur mainly in the Mediterranean area, from Morocco and Spain to Iran, southward to the Hoggar Mountains (Algeria) and Yemen (Norlindh, 1946), and northward to Germany and Poland. The center of distribution is northwestern Africa; eight species are listed in the Flora of northern Morocco (Valdés et al., 2002). The genus Calendula consists of 12 annual or perennial species, which are regarded as taxonomically complicated due to hybridizations (Norlindh, 1977; Heyn and Joel, 1983). Within the genus, C. officinalis L. (common marigold) is of special importance due to its use as an economic crop. Calendula officinalis flowers are used for pharmaceutical purposes (EDQM, 2014), in skin care products because of their anti-inflammatory activity (Talhouk et al., 2007), and as feed additives to improve the color of food because of their orange color (carotenoids) (Mukherjee et al., 2011). Florets of orange cultivars are also used as an adulterant of the expensive spice saffron (Marieschi et al., 2012). The fruits of C. officinalis are rich in fatty oil that has, because of its unusual composition, numerous technical applications (Zanetti et al., 2013). Common marigold is also an important ornamental plant with many cultivars. The flower heads are up to 5 cm in diameter, which is relatively large compared to other species of the genus. The flower heads vary from pastel yellow to deep orange, and several cultivars are double flowered.

At present, the identification of C. officinalis is often performed by (high-performance) thin-layer chromatography (TLC) or by using morphological characters (EDQM, 2014; AHPA, 2015). To the best of our knowledge, DNA-based methods do not yet exist. It can be assumed that TLC is not able to distinguish all Calendula species, and that processed plant material (e.g., fine-cut or ground flowers) cannot be identified to species level by morphology. Therefore, a DNA-based method to identify this species has the potential to complement existing methods in quality control. High-resolution melting curve analysis (HRM) is based on the melting behavior of relatively short, double-stranded DNA fragments and is a fast and reliable post-PCR method to detect mutations like single-nucleotide polymorphisms (SNPs) or indels. With a slow, stepwise increase of temperature, a fluorescent dye incorporated between the two DNA strands is released depending on sequence, GC content, and length of PCR products, resulting in a specific melting curve (Ririe et al., 1997; Liew et al., 2004).

Compared to sequencing standard barcode markers, the designed assay is much faster, less labor-intensive, and hence much cheaper. After only 2 h of PCR and subsequent HRM analysis, results are available. Furthermore, the short amplification products facilitate analysis of degraded DNA, as is often present in finely powdered material. Marieschi et al. (2012) developed sequence-characterized amplified region (SCAR) markers for the discrimination of saffron from several adulterants (including C. officinalis) and were able to detect adulterations of as little as 1%. Jiang et al. (2014) reported on a barcode melting curve analysis using general psbA-trnH primers for the same purpose. According to their methodology and results (extensively overlapping peaks of Calendula and saffron), we would suppose that the detection limit of Calendula adulterations is considerably higher than 1%. Both assays were not tested for the species-specificity of C. officinalis.

The aim of this study was to develop a DNA-based assay to identify the economically important plant C. officinalis and to distinguish it from other species of the genus. The analysis of outgroup samples should demonstrate the specificity of the assay and improve the reliability of the results. Several outgroup species grow wild in Central Europe and are therefore potential contaminants as “weeds,” but frequent adulterations are not reported. Additionally, we tested whether the assay is able to detect C. officinalis as an adulterant in saffron samples.

METHODS AND RESULTS

DNA extraction—The sample set included dried leaves of 225 Calendula samples of 10 species, 63 outgroup samples of 14 genera (all Asteraceae), and three samples of saffron stigmata (Crocus sativus L., Iridaceae) (Appendix 1). Genomic DNA was extracted using a modified cetyltrimethylammonium bromide (CTAB) protocol (“CTAB method 1”; Schmiderer et al., 2013, based on Doyle and Doyle, 1990). This extraction included a mixture of 1 mL CTAB extraction buffer containing 27.4 mM CTAB, 0.7 M NaCl, 13.5 mM β-mercaptoethanol, 14.4 mM sodium dodecyl sulphate, 4.1 µg Proteinase K, 10 mg polyvinylpyrrolidone K30 (all reagents from Carl Roth GmbH, Karlsruhe, Germany), 1 mM EDTA (pH 8), and 10 mM Tris-HCl (pH 8) (Sigma-Aldrich, Vienna, Austria) per sample. For the DNA extraction of saffron samples, an additional washing step with 70% ethanol was performed.

Sequencing and sequence analysis—The nuclear internal transcribed spacer region (ITS), the chloroplast rbcL gene, and part of the matK gene, all commonly used DNA barcoding regions (Fazekas et al., 2012), and the trnK 5′ intron, trnL-trnF intergenic spacer, and the psbA-trnH intergenic spacer were sequenced from 22 samples of 10 Calendula species and two Dimorphotheca pluvialis (L.) Moench samples (GenBank accession no.: KM356075–KM356196, KM668487). For a 15-µL PCR reaction, 1 µL of genomic DNA (1:50 dilution of the original DNA extract, equivalent to approx. 1–50 ng) was added to a master mix containing 1× PCR buffer B, 2.5 mM MgCl₂, 133 µM dNTP mix, 0.6 units Taq HOT FIREPol DNA Polymerase (all reagents from Solis BioDyne, Tartu, Estonia), and 0.6 µM forward and reverse primer (Life Technologies, Vienna, Austria). The PCR cycle profile included a denaturation step at 95°C for 15 min, followed by 45 cycles at 95/55/72°C for 45/45/90 s, with a final elongation step of 9 min at 72°C. PCR products were checked on 1.4% agarose gels and purified with ExoI and SAP (Fermentas, St. Leon-Roth, Germany) according to the manufacturer's instructions. Sequencing was performed by Microsynth (Vienna, Austria) using the same primers as for the original amplification (Table 1). The obtained sequences were edited using Chromas version 2.24 (Technelysium, Tewantin, Australia) and aligned with MEGA6 (Tamura et al., 2013). The sequence analysis involved an alignment of 37 ITS sequences with a total of 641 positions ( Appendix S1 (apps.1500069_s1.doc)) and an alignment of 23 chloroplast sequences with a total of 2413 positions ( Appendix S2 (apps.1500069_s2.doc)). Each chloroplast sequence was a combination of the trnK 5′ intron, part of matK, trnL-trnF, psbA-trnH, and rbcL sequences of one sample. Candidate diagnostic nucleotides were identified using nucDiag from the R package Spider 1.3-0 (Brown et al., 2012).

Table 1.

Base composition of PCR, sequencing^(*), and HRM primers used in this study.

Primer design and HRM—HRM-suitable primers were designed based on the chloroplast trnK 5′ intron and trnL-trnF intergenic spacer alignments. Primers with an optimum melting temperature ranging from 56°C to 58°C were designed using Primer Express 2.0 (Applied Biosystems, Foster City, California, USA) (Table 1). HRM with preamplification was performed with a Rotor-Gene 6000 (QIAGEN, Hilden, Germany). For a 10-µL PCR reaction, 1 µL of genomic DNA (1:50 dilution of the original DNA extract, equivalent to approx. 1–50 ng) was added to a master mix containing 1× HOT FIREPol EvaGreen HRM Mix (no ROX) (Solis BioDyne) and 0.15 µM forward and reverse primers (Life Technologies). The PCR cycle profile included a denaturation step at 95°C for 15 min, followed by 45 cycles at 95/58/72°C for 10/20/20 s. The melting analysis was performed by increasing the temperature from 68°C to 82°C by 0.1°C/s. All reactions were done in duplicates. In each HRM run, reference samples for each expected curve type were included. The melting curves were analyzed using Rotor-Gene 6000 Series software (QIAGEN). The PCR efficiency (E) was calculated with a 10-fold dilution series following the formula E = 10ˆ(−1/slope) − 1. The straight calibration line included five measuring points for each primer combination. The efficiency of the trnK primers was 93.0% (R² = 0.9994), and the efficiency of the trnL-trnF primers was 78.5% (R² = 0.9981).

Identification of C. officinalis—For C. officinalis, only one species-specific mutation could be found in all sequenced loci, located at position 211 of the trnK-matK alignment (Table 2). The confirmation of this diagnostic nucleotide was performed by developing HRM-suitable primers and testing an extensive sample set (Appendix 1). The primer pair Cal_trnK_2F&R was designed to amplify 71 bp of the trnK 5′ intron including this SNP (A/C transversion), which divided all Calendula samples into two groups. Group 1 consisted only of C. officinalis samples, and group 2 consisted of samples of all other Calendula species (Fig. 1A). One outgroup sample of Senecio L. sp. grouped with C. officinalis, whereas Tagetes patula L. and a part of the Anthemis tinctoria L. samples showed melting curves of group 2. The other outgroup samples formed three further groups with higher melting temperatures (Fig. 1B). The Helianthus L. samples showed poor amplification due to an indel in the primer-binding site and unspecific HRM curves. The primer pair Cal_trnL-F_1F&R amplifies 126 bp of the trnL-trnF intergenic spacer. Several SNPs divided the Calendula samples in three groups. Group I consisted of samples of C. maroccana (Ball) B. D. Jacks. and C. lanzae Maire, group II consisted of samples of C. eckerleinii Ohle and C. meuselii Ohle, and group III consisted of samples of C. officinalis and all other Calendula species (Fig. 1C). The tested outgroup samples showed many different melting curves, but all of them with higher melting temperatures than the Calendula samples, except Petasites Mill. spp. The latter showed melting curves very similar to C. officinalis but distinguishable from our target species by the trnK primers (Fig. 1D). The Tagetes L. samples showed an insufficient amplification resulting in unspecific HRM curves. With the application of both primer pairs, all samples of C. officinalis were reliably identified.

Table 2.

Diagnostic nucleotide candidates to distinguish individual species.^a

Detection of C. officinalis as an adulterant of saffron—For the detection of Calendula in saffron, artificial DNA admixture series of 0%, 0.0001%, 0.001%, 0.01%, 0.1%, 1%, 10%, and 100% C. officinalis DNA in Crocus sativus DNA were prepared and standardized to 10 ng/µL. Concentrations of the DNA extracts were determined using a NanoDrop ND-2000c (Peqlab Biotechnologie GmbH, Erlagen, Germany). For the mixture series, two different samples of saffron (Cal139 and Cal142) were used; each mixture series was prepared and tested twice. The amplification ability of the admixture series and pure saffron DNA was tested with both primer combinations. The homology of primer-binding sites in saffron was tested in silico with the most closely related, published sequences (trnK: Crocus banaticus Heuff. [GenBank accession no. JX903623.1], Crocus cartwrightianus Herb. [JX903624.1], Iris pseudacorus L. [KC118962.1]; trnL-trnF: Iris laevigata Fisch. [DQ286792.1]). Several mismatches in the primer-binding sites led to no or very poor, unspecific amplification products of saffron DNA. The analysis of the admixture series revealed that with both primer combinations, admixtures of above 0.01% C. officinalis (equivalent to 1 pg DNA, = limit of detection) were consistently identified as C. officinalis (Fig. 2A, C). In the qPCR, the admixtures showed an increase of the Cq value according to the decrease of the Calendula DNA concentration (Fig. 2B, D), while the HRM curves of samples containing between 1 pg and 100 ng DNA (introduced to PCR) were equal. Lower admixtures were amplified only randomly but showed, if properly amplified, in most cases an HRM curve like that of higher admixtures.

Fig. 1.

HRM analysis based on two chloroplast markers. A. = Anthemis. Ad. = Adenostyles, C. = Calendula, Ci. = Cichorium, L. = Leucanthemum, T. perfor. = Tripleurospermum perforatum. (A) HRM analysis with the primer pair Cal_trnK_2F&R amplifying one species-specific SNP (A/C) located in the trnK 5′ intron, distinguishing Calendula officinalis samples from all other analyzed samples of the genus. (B) HRM analysis of outgroup samples with the primers Cal_trnK_2F&R. (C) HRM analysis with the primer pair Cal_trnL-F_1F&R of a 126-bp part of the trnL-trnF intergenic spacer including several SNPs. The Calendula samples were divided in three groups. Group I: C. maroccana and C. lanzae, group II: C. eckerleinii and C. meuselii, group III: C. officinalis and all other Calendula samples. (D) HRM analysis of outgroup samples with the primers Cal_trnL-F_1F&R. Group IV: Adenostyles glabra, Eupatorium cannabinum, E. perfoliatum, Matricaria nigellifolia, Scorzonera sp., Senecio sp. Group V: E. purpureum, Helianthus annuus. Group VI: Tanacetum vulgare. Group VII: Anthemis spp., Ci. intybus, Dimorphotheca pluvialis, Helianthus tuberosus, Leucanthemum vulgare, Matricaria spp., Tanacetum parthenium, Tripleurospermum perforatum. HRM curves of other Tanacetum samples appeared between V and VI (data not shown).

DISCUSSION

DNA barcoding has become an important technique for taxonomy, as well as in applications like quality (i.e., identity) control of food or herbal raw materials. Although genetic differences in the chloroplast set as well as in ITS were relatively small, one SNP was detected that distinguished the economically important target species C. officinalis from all other Calendula species. Testing our HRM assay with an extensive set of Asteraceae species revealed that one sample of Senecio sp. gave the same result as C. officinalis in the trnK primer combination. Therefore, a second assay in the trnL-trnF intergenic spacer was applied, to distinguish this Senecio sample from C. officinalis. The combination of both analyses had greater discriminatory power than just the trnK assay, although all closely related species could be distinguished with the trnK primers only. Additionally, this assay can be used to detect adulterations of saffron with Calendula flowers. Due to the high specificity of the used Calendula primers, even traces of marigold would be detected.

Fig. 2.

Analysis of artificial admixtures of Calendula officinalis in saffron. All properly amplified admixture samples showed an equivalent HRM curve like the C. officinalis references. (A) HRM analysis with the primer pair Cal_trnK_2F&R. (B) Amplification plot of the qPCR corresponding to A. (C) HRM analysis with the primer pair Cal_trnL-F_1F&R. (D) Amplification plot of the qPCR corresponding to B. %-Values mean proportion of C. officinalis DNA in saffron DNA of each sample. NTC = no template control.