Collection of local medicinal plants of Kingdom of Bahrain covered eight different regions as shown in Figure 1.

Plant samples were collected during their vegetative and reproductive growth stage in the two successive year 2016-2017 (January 2016 to April 2016, December 2016-April 2017) by several field trips in the different area, as the most of the annual desert plants, start growing after winter rainfall and perennial plants enter reproductive phase. Following the methodology of Barcode of Life Database (BOLD), whole plants with root system (if small) and part of the plant with branches, leaves, flowers showing maximum morphological characteristics useful for identification were collected for voucher preparation and plant identification. Plants were collected with all detail information of locality, plant habitat, road number, GPS coordinate, elevation level, field photograph, collector name, date, time etc.

Plant vouchers were prepared with the proper sample ID after collection. Plant parts cleaned, and dried using spacers between layers of blotting papers and finally pasted on the Herbarium sheet with proper labelling. Field/Herbarium photograph taken and uploaded in the BOLD specimen submission portal and the records of each plant with pictures are traceable online in Plant Taxomony portal in BOLD system (http://www.boldsystems.org). Plants identified and botanical names assigned based on reference books of Flora of Bahrain by ourselves. In some cases, we resolved the identification problem by consulting with taxonomists. Plants identified using standard identification method by comparing plant habitat, nodal characteristics of stem, leaf shape, leaf type, leaf arrangement, floral type, floral appearance, flower shape (sepal and petal), fruit characteristics etc.

Young leaves from the collected plant samples used for sample tissue preparation. Clean leaves properly dried either room temperature (thin and small) or heat dried at 37˚C (in case of succulents and thick leaves of desert plants). The leaves stored in sealable plastic packs and kept at room temperature [25].

As DNA was extracted using the plate based method, plant tissue sample of particular ID, around 2 - 3 pieces of plant leaves (0.5 cm) were subsampled in 96-well tube strips placed in plant box following the specific guideline of BOLD systems.

DNA extraction, PCR and sequencing work performed at Canadian Center for DNA Barcoding (CCDB), Canada using their standardized protocols. DNA extracted following the Glass fiber plate DNA extraction protocol, using a small amount of dry leaf samples [26] [27]. The leaf tissues were homogenized into fine powder using Tissue Lyser (Qiagen, USA) with rotated rack adopter at 28 Hz for 30 seconds by two times. DNA was extracted using 2X CTAB (250 µl) buffer and by incubating at 65˚C for 90 min. Cell lysates (50 µl) were transferred into 96-well Eppendrpf plate and Plant Binding Buffer (PBB, 100 µl) was added and incubated for 5 min at RT. 96-well Glass Fiber plate (PALL1) used and semi-automated glass fiber filtration method followed, while the corresponding DNA was bound to GF membrane [25]. GF Membrane rinsed by using Protein Wash Buffer (PWB) and the DNA eluted from GF plate to the collection microplate by adding pre-warmed double distilled water (60 µl), covered with cap strip and stored for PCR.

Three gene regions (rbcLa, matK, ITS2) were amplified using CCDB plant protocol [25] [26]. Amplification of the different plant markers required different primers and PCR recipes in PCR plates shown in Table 1 and Table 2. PCR plate placed into the thermo-cycling block and different PCR thermocycle programs were used for different primer set of three different barcode genes. Amplification of rbcLa: 94˚C for 4 min; 35 cycles of 94˚C for 30 sec, 55˚C for 30 sec; 72˚C for 1 min; final extension 72˚C for 10 min; hold at 4˚C. For ITS2: 94˚C for 5 min; 35 cycles of 94˚C for 30 sec, 56˚C for 30 sec, 72˚C for 45 sec; final extension 72˚C for 10 min; hold at 4˚C. Amplification of matK: 98˚C for 45 s; 35 cycles of 98˚C for 10 s, 54˚C for 30 s, 72˚C for 40 s; final extension 72˚C for 10 min; hold at 4˚C. PCR products visualized and analyzed on 2% Agarose E-gel ^R 96 system (Invitogen) using recommended program and software [27]. Strong amplification of rbcL and ITS2 were obtained using low concentration of primers, dNTPs and Taq polymerase with one primer set rbcLa-F [28] , rbcLa-R [29] and another primer set IITS-S2F [30] ITS4 [31]. In case of matK amplification, higher concentration of forward primer matK-xF [32] and reverse primer matK-MALPR1 [33] , dNTPs and Taq polymerase were used (Table 2). According to CCDB protocol, diluted PCR replicons used directly for sequencing [26]. PCR products were sequenced on an ABI 3730Xl DNA analyzer (Applied Biosystem, California, USA) following standard procedure. For bidirectional sequencing of rbcLa and ITS2, same PCR primer set used separately for bidirectional sequencing at 96˚C for 2 min; 30 cycles of 96˚C for 30 s, 55˚C for 15 s, 60˚C for 4 min; hold at 4˚C. For bidirectional sequencing of matK, three different primers used separately depending on PCR products used for sequencing at 96˚C for 2 min; 30 cycles of 96˚C for 30 s, 50˚C for 15 s, 60˚C for 4 min; hold at 4˚C.

Chromatographs of all the three markers of those plants recovered, trimmed, edited and aligned by using Codon Code Aligner version 3.7.1-6.0.2 (CodonCode

*Bidirectional Sequencing.

*Platimum DNA Polymerase (Invitrogen, Carlsbad, California, USA) used for PCR reactions.

Co., Massachusetts, USA) by CCDB and received DNA sequences for further analysis and study. The nucleotide sequences of different plants aligned using Basic Local Alignment Search Tool (BLAST) algorithm provided by National Centre for Biotechnology Information (NCBI) and European Bioinformatics Institute (EMBL-EBI). The level of similarity between the samples studied by BLAST algorithm using blastclust (http://www.ncbi.nih.gov), which automatically and systematically clusters the nucleotide and protein sequences based on pairwise distance. It is a query method of genus and species of the samples from which sequences separated by the smallest genetic distances in the matrix [18]. The FASTA files of barcode sequences of rbcL, matk and ITS2 prepared for comparison with reference data sites. Different BLAST matching tools used widely to optimize the molecular data mining and for the identification of plants at genus as well as species level. This comparison brought out the genus identification and in some cases species identification. FASTA files of nucleotide sequences explored MUSCLE program [34]. The nucleotide sequences were aligned; analyzed and phylogenetic tree was constructed using various algorithm programs in Molecular Evolutionary Genetics Analysis (MEGA 7) software [35]. Sequence editing and alignment were generated by MUSCLE [34] to construct phylogenetic tree and evolutionary distance measurement by Kimura 2-parameter of studied plants [36]. The nucleotide sequences of coding region of chloroplast gene rbcLa, matK and nuclear ITS2 regions deposited in GenBank (Table 4).

The success of sequencing of barcode markers evaluated based on the number of plant species that successfully generated a sequence for a particular marker. The sequence recovery and analysis of three DNA barcodes of 29 medicinal plants collected from different parts of Kingdom of Bahrain showed (Figure 1). DNA barcodes genes of rbcLa, matK and ITS2 of 29 different medicinal plants representing 21 different families analyzed. The primers of rbcLa, matK, and ITS2 are used for PCR amplification, sequencing (Table 1) and PCR reagents (Table 2). For PCR amplification and sequencing of rbcLa and ITS2 same primers worked but in bidirectional sequencing of matK PCR products, 70% PCR products sequenced by choice I combination while 30% sequencing worked choice II (Table 1). The detail sequencing results of 29 medicinal plants of 21 different families using rbcLa, matK and ITS2 represented (Table 3). The success of sequence obtained of the 29 plants of three different barcodes genes rbcLa, matK and ITS2 were 97%, 79% and 75% respectively (Table 3). In bidirectional sequencing for rbcLa, forward primer gave high quality sequencing result in 92% plants while reverse primer gave 78%. In case of Limonium axillare in Plumbaginaceae and Mesembryanthemum nodiflorum of Aizoaceae, we got moderate quality sequencing product from forward primer and no product from reverse primer. High quality sequencing in matK and ITS2 obtained by 68% and 72% plants respectively using forward primer and 65% and 68% using reverse primer respectively. The high sequencing success rate of rbcLa (90%) than matK (70%) and ITS2 (41%) was observed in Amazonian trees while eight different DNA markers were tested [14]. In another DNA barcode study to build a community phylogeny of tropical trees, sequencing success was higher in rbcL region (90%) in comparing matK (68%) [37]. Sequencing success of rbcL (95%), matK (76%) and ITS2 (89%) from fresh specimen of vascular plants from Churchill was reported [38]. Similarly, the low success rate of matK in compare to rbcL reported in the speedy assessment of species abundance in taxonomically poorly known area or in cryptic population [39] , while, the sequence success of 88% and 90% in case of matK region is reported respectively by [20] [23]. The sequencing success of matK around 85% observed in standard and nested multiplex-tandem PCR [32]. In general, the lower success of matK sequence recovery also reflected due to difficulties in primer selection and binding of primers as explained [38] [39]. The low success of in sequence recovery in some genera may be relatively thick leaves of desert plants, slower desiccation and consequent DNA degradation [38]. It has been noted that rbcLa is much easier to sequence than matK and others, so the selection of rbcLa is an important option in barcoding study [40]. In several comparative studies, the variation in the success of sequencing results of different DNA may depend on the nature of plants, their habitat and ecological behavior [41]. The sequences of three different barcode genes submitted in GenBank (Table 4).

The performance of the three-barcode markers for identification of plant genus and species level compared with Global multiple sequence alignment (MSA) and

Primers used for sequencing were 1) rbcLa-F, 2) rbcLa-R, 3) ITS-S2F, 4) ITS4, 5) matK-xF, 6) matK-MALPR1, 7) matK-1RKIM-f and level of sequencing results mentioned as HQ: high quality sequencing and MQ: moderate quality sequencing.

presented (Table 5). The obtained sequence of the three barcode genes, rbcLa, matK and ITS2 analyzed following BLAST to confirm the taxonomic identification of the plants at the molecular level, using online database sites of International Nucleotide Sequence Database Collaboration (INSDC). The genetic distance (GD) between sequence pair and percent identity value (PI) after global multiple sequence alignment is helpful for comparative study, species identification and to score relatedness and distance. BLAST search data is helpful for

genus and species identification of plants at the molecular level, which is one of the main objectives of this study. This is helping in the process of plant taxonomical query in the regional and global accomplishment. Using NCBI database in BLAST analysis, 100%, 92.85% and 86.95% plant genus of the analyzed medicinal plants confirmed by obtained barcode sequences of ITS2, rbcLa and matK sequences respectively (Table 5). The use of rbcLa and ITS2 barcodes worked fine in compare to matK. Overall, 97% plants species were correctly

The percentage of plant identification at species and genus level calculated on the number of identified plants (genus/species level) based on total plants considered for the BLAST study.

identified either anyone of the barcodes (rbcLa, matK or ITS2) using BLAST survey. 28.57% plant species were confirmed by the analysis of all the three barcode sequences (rbcLa, matK and ITS2). 21.42% plant species were confirmed by the analysis of sequences of rbcLa and matK while 28.57% plant species were confirmed by the analysis of sequences of rbcLa and ITS2 combinations. 10.71% plant species were identified only by rbcLa (Limonium axillare, Herniaria hemistemon, Dipcadi erythraeum) and only by ITS2 sequences separately (Cynanchum variance, Cyanomorium coccineum, Cistanche coccineum). Only one barcode gene specifically confirmed the species identification using GenBank BLAST tool of those plants. BLAST analysis confirmed 100% medicinal plant genus and 89.28%, 86.32% and 60.86% plant species of medicinal plants by rbcLa, ITS2 and matK barcodes respectively. Similarly using BLAST analysis of sequence in the genetic distance analysis (GD), 99% success rate of genus identification and 73% and 75% species identification for rbcL and matK respectively observed in the identification of African rainforest trees was reported [18]. Similarly, the selection of rbcL and matK as core plant barcode marker reported by several workers and they have universal discriminatory power [14] [17] [42]. For identification of medicinal plant, the importance ITS2 together with rbcL, matK reported and justified that ITS2 is a valuable DNA barcode gene for the identification of closely related species [16] [30] [43]. Moreover, DNA polymorphism is typically low in coding plastid genes such as rbcL and matK but is frequent in noncoding ITS2 region and is helpful for species identification [18].

The comparative analysis of molecular sequence data is essentially important for reconstructing evolutionary history of plants. The discriminatory influence of three barcode genes was considered in pairwise distance matrix to construct phylogenetic tree using in MEGA7 software [34]. The evolutionary history was inferred using pairwise distances in Maximum Composite Likelihood (MCL) [35]. The phylogenetic relatedness and evolutionary history drown by means of neighbor-Join (NJ) method [44] using nucleotide sequence alignment of rbcLa and ITS2 (Figure 2 and Figure 3). The evolutionary distance analysis conducted using Kimura 2-parameter [36] method of base substitution per site and mentioned

by units. The pairwise distance matrix value indicates their closeness during evolution and ultimately helps to draw their phylogenetic tree and the history of evolution. The clades formed in the trees were mostly mixture of several plant species and the big and small branches showed their relatedness and distances clearly among the plant species. Cynanchum varians and Leptadenia pyrotechnica showed their closeness and common ancestry as they belong to the same family confirmed by taxonomic observation too. Similarly Herniaria hemistemon and Spargularia marina are closely related in their evolutionary history and belong to the same family Caryophyllaceae. In a double-locus (rbcLa and ITS2) phylogenetic trees, tree nodes and close branches supported the taxonomic clarity, relatedness and clade phylogeny. For each species, the higher interspecific distance obtained by ITS2 (33.17%) in compare to rbcLa (8.72%), so the value of closeness differed in phylogenetic tree of ITS2 (Figure 2) and rbcLa (Figure 3). Phylogenetic tree analysis of ITS2 and rbcL provided species resolution in a better way by using pairwise distance matrix value and indicated their closeness and distance with each other. Similarly, in the building of community parsimony, rbcL and ITS2 sequence alignment in constructing phylogenetic tress nodes showed support value in plant phylogeny [18] [29]. We noticed that the matK region of chloroplast genome is little problematic in sequence retrieving, phylogenetic data analysis and identification of plants in comparing to rbcLa and ITS2 barcodes, similar to other studies [17] [32] [42]. In the present study, rbcLa and ITS2 demonstrated excellent reliability for species authentication and less genetic diversity, due to the limited dispersal capacity in the plant communities of the arid and semiarid regions in comparison to tropics where it is frequent factor. So, the DNA barcode analysis of those plants is of a little bit challenging for species differentiation [18]. We noticed that the matK region of chloroplast genome is little problematic in sequence retrieving, phylogenetic data analysis and identification of plants in comparing to rbcLa and ITS2 barcodes, similar to other studies [17] [32] [42].

In conclusion, primarily, rbcLa and ITS2 are very useful barcode region and can be of wider application in the study of desert plant identification, ecology, and plant diversity study. Moreover, our sequences submitted to the GenBank will be of very helpful data for future study in various aspects. As, DNA barcoding has potentials to transfigure the systems of taxonomists work by giving considerable momentum during this era as a helpful tool for plant identification at molecular level.