Identification and molecular study of medicinal Plectranthus species (Lamiaceae) from Saudi Arabia using plastid DNA regions and ITS2 of the nrDNA gene

2.1 Study area

The study area included different locations in the Jazan-Fifa Mountains, Baha, and Taif. These areas are located in the Tihama mountain range, which stretches across the southwest of the KSA, with height ranging from approximately 2000 to 12,000 m. The Tihama range is located in the Afro-Alpine vegetation zone, featuring a high plant diversity. Details of sample names, dates, and collection areas are presented in Table 1.

2.2 Plant material

Samples (leaves, flowers, and stems) of six Plectranthus species were collected between 2014 and 2018. Voucher specimens were prepared and identified by experts in taxonomy according to Collenette (1999) and Chaudhary (2000) and were deposited at the Herbaria of King Saud University/KSA, Kew and Edinburgh herbarium/UK, and Biology Department, Faculty of Applied Science, Umm Al-Qura University, Mecca/KSA after comparison with voucher specimens. At least three accessions per species were selected for molecular analysis from these voucher specimens.

2.3 DNA extraction

Initially, the following DNA extraction techniques were tested to select the best method: DNeasy Plant Mini Kit extraction (Qiagen/USA) following commercial protocols, traditional CTAB method (Doyle and Doyle, 1987), and modified DNA extraction method of Sahu et al. (2012), as described in Manikandan et al., (2017). The quantity of the extracted DNA was estimated using a NanoDrop spectrophotometer (Thermo Scientific, Wilmington, DE, USA). DNA was tested via 1% agarose gel electrophoresis and ultraviolet light.

2.4 Primer sequences

Different DNA barcodes were amplified using the following primers: a-F (5′-ATGTCACCACAAACAGAGACTAAAGC-3′) and a-R (5′-GTAAAATCAAGTCCACCRCG-3′) for rbcL; KIM3F (5′-ACCCAGTCCATCTGGAAATCTTGGTTC-3′) and KIM1R (5′-CGTACAGTACTTTTGTGTTTACGAG3′) for matK, according to CBOL Plant Working Group (2009); 2F (5′-CGTAGCTACTTCTTCGCAGC-3′) and 5R (5-CCTTATCATTTAGAGGAAGGAG-3′) for ITS2, according to Chen et al., (2010); for ITS1; 2F (5′-ATGCGATACTTGGTGTGAAT-3′) and 3R (5′-GACGCTTCTCCAGACTACAAT-3′) and F (5′-ACTGCCTTGATCCACTTGGC-3′) and R (5′-CGAAGCTCCATCTACAAATGG-3′) for trnH-psbA, according to (Kress and Erickson (2007).

2.5 DNA amplification and sequencing

DNA regions were amplified according to the method described by Maloukh et al., (2017), using 25 µL reaction volumes containing 12.5 µL Master Mix, 8.5 µL nano-pure water, 1 µL of each primer, and 2 µL DNA. The thermal cycle included an initial denaturation level of 94 °C for 5 min, followed by 94 °C for 45 s. An annealing temperature dependent on the type of primer (ITS1, 50 °C; ITS2, 56 °C; matK, rbcL, and trnH-psbA, 55 °C) was applied for 45 s, followed by 72 °C for 1.5 min, and a final extension step at 72 °C for 10 min, in 40 cycles, using a Mastercycler (Eppendorf Vapo Protect, NY, USA). The quality of the PCR products was confirmed by electrophoresis on 1.5% agarose gel, and products were purified with a 1.0% agarose gel using the QIAquick purification Kit (Qiagen, USA). Sequencing was performed using a genetic analyzer (Applied Biosystem, CA, USA).

2.6 Bioinformatic analysis

Forward and reverse sequences were assembled and edited per sample and per gene in CodonCode Aligner v.8.0.2 (Condon Code Co., USA). The ITS2 region of each specimen was detected using a hidden Markov model to remove the 5.8S and 28S regions, which may overlap with the ITS2 region (Keller et al., 2009). A multiple sequence alignment was run for each gene using the Muscle algorithm tool in MEGA 7.0.27 (Kumar et al., 2016). The sequences were manually adjusted, and ambiguous regions were removed.

BLAST was run against known specimens in GenBank. The identification process took place at three levels: family, genus, and species. The query sequences were detected by selecting the highest maximum score and the lowest E-value. Outcomes were classified into three levels, as described in (Meier et al., 2006).

Genetic distances were calculated using the Kimura two-parameter (K2P) model in MEGA v.7.0.27 (Kumar et al., 2016). Inter- and intraspecific distances were calculated as the barcoding gaps by using TaxonDNA v.1.7.8 (Meier et al., 2006). The secondary structures of the ITS2 sequences were predicted using the ITS2 ribosomal RNA database ( http://its2.bioapps.biozentrum.uni-wuerzburg.de/) (Schultz et al., 2005).

Two phylogenetic trees were built in MEGA 7.0.27 (Kumar et al., 2016): a neighbor‐joining (NJ) tree and a maximum likelihood (ML) tree. Both trees were created by running 1,000 bootstrap replicates and K2P model of selection. The gamma distributed invariant site (G 1) nucleotide model was used for for ML tree. The number of discrete gamma categories was set as five (Liu et al., 2012; Kumar et al., 2016).

3.1 Evaluation Efficiency of DNA barcoding

The preliminary results obtained from the DNA isolation experiments showed that the method of (Manikandan et al., 2017) allowed for the extraction of relatively more and better quality DNA than other methods, and therefore, DNA was extracted from all samples using this method.

The 21 plant specimens for each gene yielded 21 (100%) PCR products for ITS2 and rbcL (Table 2). The lowest percentages of samples yielding PCR products occurred using the matK and trnH-psbA genes. All 21 (100%) specimens were successfully sequenced for their rbcL genes and 20 (95%) specimens for their ITS2 loci. The ITS2 region had the highest guanine-cytosine content (Table 2). Thus, two rbcL and ITS2 DNA barcodes were selected for the following analyses.

3.2 Species resolution and barcode analysis

Based on BLAST, 100%, 85.7%, and 28.5% of the ITS2 sequences were correctly identified at the family, genus, and species levels, respectively (Fig. 1). Furthermore, 100%, 71.4%, and 14.2% of the rbcL sequences were successfully identified at the family, genus, and species levels, respectively. E-values were (0) for each gene.

Close

Fig. 1

Successful identification of Plectranthus spp., local to the Kingdom of Saudi Arabia at three levels (family, genus, and species) using BLAST and matching sequences drawn from the NCBI database. Values listed in the columns are estimated in percentages (%).

Export to PPT

3.3 Inter- and intraspecific distance and barcode gap

The interspecific distances for the rbcL and ITS2 barcode loci are shown in Table 3. The ITS2 locus showed a minimum interspecific distance of 0.139, while the maximum interspecific distance was 0.000. The rbcL loci exhibited a minimum interspecific value of 0.018 and a maximum intraspecific distance of 0.000. Fig. 2 shows that the ITS2 loci have a wider barcode gap than the rbcL loci.

Close

Fig. 2

Barcode gaps for DNA barcodes, a-ITS2 and b-rbcL, based on the K2P distance and six Plectranthus spp. local in the Kingdom of Saudi Arabia. Total barcoded sequences for ITS2 = 21, rbcL = 21, three specimens per species.

Export to PPT

3.4 ITS2 secondary structure

Fig. 3 shows the predicted secondary structures of the ITS2 sequences used to identify the Plectranthus spp. The six species shared a similar ITS2 morphological structure, with a central ring and four similar helices, namely, I, II, III, and IV. However, there were variations in the stem loop number, size, and position and degree of angle from the center of the spiral arm of the ITS2.

Close

Fig. 3

Observed variation in the secondary structure of ITS2 barcodes for the identification of Plectranthus spp. native to the Kingdom of Saudi Arabia, using an ITS2 workbench. a: Plectranthus arabicus. b: Plectranthus barbatus. c: Plectranthus hijazensis. d: Plectranthus pseudomarrubioides. e: Plectranthus asirensis. f: Plectranthus tenuiflorus.

Export to PPT

3.5 Phylogenetic trees

The rbcL ML tree (Fig. 4) was constructed using 21 Saudi Arabian Plectranthus spp. sequences, with accessions numbers MN381814–MN381834. The ITS2 ML tree (Fig. 5) was constructed using 20 Saudi Arabian Plectranthus spp. sequences, with accessions numbers MN82135–MN382154. In addition, 36 Plectranthus spp. sequences were downloaded from GenBank representing 8 different countries in Asia, Africa, Australia, and USA.

Close

Fig. 4

Maximum likelihood tree of ITS2 barcodes showing the phylogenetic relationships among Plectranthus spp. from the Kingdom of Saudi Arabi and other species of Plectranthus originating from different geographic regions. Bootstrap = 1000 replications.

Export to PPT

Close

Fig. 5

Maximum likelihood tree of rbcL barcodes showing the phylogenetic relationships among Plectranthus spp. from the Kingdom of Saudi Arabia and other species of Plectranthus originating from different geographic regions. Bootstrap = 1000 replications.

Export to PPT

Overall, ML produced the best trees in terms of forming monophyletic groups for each species and nodes with high support values. Further, the ML-ITS2 tree showed clades with high support values and branches and good separation of species. The ML tree based on the ITS2 locus data (Fig. 4) divided the sequences into two main clades: Ⅰ and II. Clade I, with a high supporting value (89%), was divided in three main sub-clades, and clade II was also divided into three main branches. The phylogenetic tree constructed from rbcL gene data (Fig. 5) was also split into two clades (I and II). Clade I, with a high support value (89%), was divided into two main branches. Clade II was also divided into two branches. In general, Figs. 4 and 5 show that some species appeared to be highly related to each other as they always appeared in the same clade. For example, P. barbatus P. caninus, P. asirensis, P. otostegioides, and P. lanuginosus appeared to be highly related, as well as P. pseudomarrubioides, P. ornatus, P. montanus, P. amboinicus, P. parviflorus, and P. edulis. Furthermore, P. tenuiflorus, P. hijazensis, and P. arabicus always appeared together in one clade.

4.1 DNA extraction and PCR challenges

In the Plectranthus genus, there are many issues surrounding DNA isolation and purification, which hinder the acquisition of high-quality nrDNA. These include degradation of DNA owing to endonucleases, co-isolation of highly viscous polysaccharides, inhibitor compounds such as polyphenols, and presence of secondary compounds that directly or indirectly interfere with enzymatic reactions (Manikandan et al., 2017). In the present study, three methods of DNA extraction were compared, and the method described in (Sahu et al., 2012) was found to show the best performance in terms quantity and quality of DNA extracted and successful PCR product collection, particularly with the ITS2 and rbcL loci. This finding is consistent with that of (Manikandan et al., 2017).

It was observed that in dried, old, or degraded samples, fewer PCR products and less successful sequencing were observed, especially for the ITS1, matK, and trnH-psbA loci. The findings of the present study showed that the ITS2 and rbcL loci represented the highest PCR success rates (100%) as well as the highest sequencing success rates (95% and 100%, respectively; Table 2). These observations are consistent with those of (Al-Juhani, 2019), who reported that the rbcL and ITS2 regions had higher rates of amplification and sequencing success, even when using old, dry, or degraded specimens.

4.2 DNA barcodes and species resolution

Al-Qurainy et al., (2014) studied Saudi Arabian P. asirensis and reported the possibility of using nrDNA-ITS and the rbcL and rpoC1 chloroplast loci for DNA barcoding in Plectranthus species. The present study showed that, in Plectranthus, ITS2 exhibited a higher discrimination rate at the genus level (85.7%) and species resolution (28.5%) than rbcL, which showed a discrimination rate of 71.4% at the genus level and species resolution of 14.2% (Fig. 1). The ITS2 locus is, therefore, recommended to identify herbal medicinal plants (Chen et al., 2010), for the following reasons: (i) it is located within the nuclear genome, which has a different rate of evolution than the plastid genome, (ii) it provides higher species resolution, and (iii) it has a much shorter sequence, allowing for higher recovery from processed plant materials found within herbal products.

Moreover, a possible explanation for the relatively low species-level identification in the present results is the lack of sufficient reference data; several species of Plectranthus do not have reference sequence data available in GenBank. There were no reference sequences available for P. hijazensis, P. arabicus, or P. tenuiflorus, for most of the markers selected for examination. Therefore, BLAST revealed closely related species but not specific species and was only able to successfully identify to the genus level, with low species resolution. Thus, the voucher specimens of the present study are the first ever sequences for the aforementioned species deposited in GenBank and are publicly available. (Hollingsworth, 2011) argued that species with a single sample are potentially distinguishable (i.e., for successful species identification) if the sequence is unique.

In general, the minimum interspecific distances were greater than the maximum intraspecific distances for the rbcL and ITS2 barcodes (Table 3). The barcoding gap in the ITS2 region was obvious and larger than that in rbcL, where the minimum interspecific distance for ITS2 was 0.139 and the maximum intraspecific distance was 0.000. Species discrimination is considered successful if the minimum interspecific K2P distance is larger than the maximum intraspecific distance for that species (CPBG China Plant BOL Group, 2011); this condition applies to the samples in the current study as well (Fig. 2).

The secondary structures of the ITS2 are shown in Fig. 3. Similar ITS2 secondary structures were observed across all Plectranthus spp. However, differences were observed in the morphology of the four helices among species. Variation in the secondary structure of ITS2 has previously been reported as a tool for identifying species at the molecular morphological characteristics level (Keller et al., 2010). Liu et al. (2019) noted a clear difference between the ITS2 structures of true “Gaoben” samples and the structures of three adulterant species and recommended using ITS2 as a minibarcode to distinguish between closely or distantly related plant species used in Chinese medicine. Hence, the ITS2 secondary structure could be used to validate Plectranthus species identification.

4.3 Phylogeny and taxonomy of Plectranthus spp.

The results of the present study demonstrated that P. asirensis, P. caninus, P. otostegioides, P. barbatus, and P. lanuginosus were grouped together in clade II of the phylogenetic tree constructed using ITS2 data and in clade I of the rbcL phylogenetic tree. This is also consistent with the observations of (Al-Qurainy et al., 2014) and (Musila et al., 2017), who showed that P. asirensis, P. caninus, and P. barbatus are grouped together in trees built based on the nrITS gene, matK, and rbcL.

Furthermore, we showed that three species, P. tenuiflorus, P. hijazensis, and P. arabicus, were always grouped together in the same clade in the trees constructed based on the ITS2 and rbcL loci. This observation broadly supports the findings of (Abdel Khalik and Osman, 2017), who used RAPD and ISSR markers to study Plectranthus species native to Saudi Arabia and found that these species are clustered together.

Comparisons with the results of (Paton et al., 2018), who utilized the trnL-F, rps16, and trnS-G plastid genomes, confirmed the validity of the findings of the present study. The species studied herein were classified into two branches within the Coleus clade. Furthermore, our results showed that P. hijazensis and P. asirensis were separated in two clades with high bootstrap values, which supports the recommendation of (Abdel Khalik and Osman, 2017) to separate these two species into different subgenera.

The findings of the present study support the observations of (Musila et al., 2017), as P. aegyptiacus (P. tenuiflorus) was grouped into the same clade as P. pseudomarrubioides and P. amboinicus in the tree constructed based on the ITS2 locus. Nevertheless, in the tree based on the rbcL data, P. edulis was grouped with P. pseudomarrubioides, P. ornatus, P. montanus, P. amboinicus, P. amicorum, P. madagascariensis, and P. parviflorus.

The present outcomes are consistent with those in previous studies (Frigerio et al., 2019) that have recommended DNA barcoding (trnH-psbA, ITS, and rbcL 1-B) as a useful tool for identifying medicinal and aromatic plants. Our findings also support the observations of Al-Juhani (2019), who showed that ITS2 and rbcL produce species resolutions of up to 77% for 100 species growing on drylands in KSA, most of which are used in traditional medicine. Han et al., (2016) used ITS2 DNA barcoding to verify and detect adulterants of 295 herbal medicine species in China, confirming that the ITS2 region could be used to identify most of the samples (87.7%) and detect adulterants (Han et al., 2016). Furthermore, (Vassou et al., 2016) used the rbcL region to create an API-RDBL library of 374 medicinal plants used in the Indian Pharmacopoeia and mentioned the usefulness of rbcL DNA barcoding in authenticating herbal drugs and detecting adulterants.

Our results are also in line with the results of (Gao et al., 2010), who highlighted the use of the ITS2 region to identify 114 samples of the Fabaceae family, which contains many important medicinal plants, and confirmed the power of the ITS2 region as a tool to authenticate herbal medicines. Furthermore, (Lv et al., 2020) indicated that ITS2 can identify 98% of medicinal specimens from the Apocynaceae family at the species and genus levels.

Finally, (Le et al., 2020) confirmed the effectiveness of ITS2 barcoding in detecting, identifying, and classifying the genetic relationships in agricultural varieties of palm used in botanical gardens and orchards. Together, these findings strongly support the use of the ITS2 region as a tool to distinguish between varieties of Plectranthus, considering that several species of Plectranthus spp. are used ornamentally and in gardens.