DNA barcoding and morpho-anatomical characters of two forms of Convolvulus arvensis L. (Convolvulaceae) grown in Asir region, Saudi Arabia

DNA barcoding and morpho-anatomical characters of two forms of Convolvulus arvensis L. growing in Asir region, Saudi Arabia were studied. We applied three molecular markers, ITS, rbcL and matK for phylogenetic reconstruction in combination with 24 morpho-anatomical characters to provide proposal for the recognition of its population under two forms. Plant samples were collected from ten various populations of C. arvensis supposed to be of two different forms, and finally these have been found to be of two new forms (small leaf) and form (large leaf) for the flora of Saudi Arabia. TWINSPAN results showed that there was negative group with twenty one objects including stem length, stem thickness, seed shapes, petiole length, leaf apex, leaf length, petal circumference, leaf width, capsule length, capsule width, stamen length, anther width, anther length, carpel length, venations type, pollen grains shape, T.S of stem, root and leaf, stigma shape and anther base surface and positive group including hairs, stomata and anther shape. Dendrogram upon morpho-anatomical characters resolved two main clusters, one cluster of form (large leaf) and the other of form (small leaf). Amplified gene region for the form (small leaf)-rbcL showed 100% identity with C. lineatus; form (small leaf)-matK and form (large leaf)-matK had 99% identity with Calystegia sepium (L.) R. Br. Form of (small leaf)–ITS had 100% identity with that of C. arvensis; form (large leaf)-rbcL had 100% identity with those of C. arvensis and C. lineatus. Form (large leaf)–ITS had 99% identity with that of C. arvensis. The results of molecular phylogenetic analyses based on certain morphological and rbcL data support two main clusters within C. arvensis which is consistent with two forms, form (small leaf) and form (large leaf).


Introduction
Convolvulus arvensis L. is one of the most common weeds all over the world found mostly in about 54 countries distributed in 32 different economic crops (Holm et al., 1997).It was named as a field-bind weed due to its extensive deep root system together with long-term seed bank that were considered as key features to the noxious weed status (Gianoli, 2001).C. arvensis plants are native to Europe and grows extensively in Mediterranean climates, temperate and tropical regions and in many other climatic zones (Lyons, 1998).Field bindweed was considered as a perennial vine of the glorious family Convolvulaceae which spreads by rhizome or/and seed and its prostrate weak -stem, often twine forming tangled dense mats (Wiese and Phillips, 1976;Gleason and

Statistical analysis
The multivariate statistical analysis for morpho-anatomical characters was carried out using Community Analysis Package (CAP) Statistics program, version 5. Cluster analysis for the morpho-anatomical characters including (Euclidean Ward's (Nath et al., 2014;Ward, 1963), TWINSPAN (Two Way Indicator Species Analysis, Manhas et al., 2009) and Jaccard's Similarity Coefficient were performed.
Resulted sequence from primer was compared with each other by using Bio edit version 7.1.11and Clustal W multiple sequence alignment to investigate molecular phylogenetic analysis (Tamura et al., 2011;Grimm et al., 2006).Obtained sequences from the matK, rbcL and ITS regions for the two taxa were compared by other relevant sequences using BLAST (Basic Local Alignment Search Tool).A split is a bipartition in a species set, that divides all dataset of species into two groups: a functional out-group and a functional in-group (Huson et al., 2010).

Morpho-anatomical features
The list of morphological and anatomical characters that were investigated together with their character states is shown in the Figs 1.1 -1.5 and Table 1.The resulted dendrogram based on these characters was resolved in two main discrete clusters of form (small leaf) and form (large leaf) for the ten populations of which were separated at linkage distance of 52.9.While five populations of form (small leaf) were further resolved in to two sub-clusters at 1.2 (Fig. 2).
The Jaccard's pairwise similarity coefficient values for ten populations of C. arvensis ranged from 0.0833 to 1.0, whereas those among the populations of form (small leaf) varied from 0.96 to 1.00 and of form (large leaf) didn't vary, rather remained always highest (1.00) (Table 2).These values between the populations of two forms (small leaf) and form (large leaf) ranged from 0.0833 to 0.125.Whereas, this coefficient was 0.0833 between P2 and P4 of forms (small leaf) and P1-P5 of forms (large leaf) and it was 0.125 between P1, P3 and P5 of forms (small leaf) and P1-P5 of forms (large leaf, Table 2).
TWINSPAN output for the classification of ten populations revealed that there was negative group with twenty one characters including stem length, stem thickness, seed shapes, petiole length, leaf apex, leaf length, petal circumference, leaf width, capsule length, capsule width, stamen length, anther width, anther length, carpel length, venations type, pollen grains shape, transverse section (TS) in stem, in root, in leaf, stigma shape and anther base surface.Also it showed that there was positive group of three characters comprising hairs, stomata and anther shape.As the morpho-anatomical characters of C. arvensis from Saudi Arabia showed distinct variation between the two forms, namely form (small leaf) and form (large leaf).Therefore, these variations can be used to clearly distinguish between some closely related subspecies, varieties, and forms.5.58 ml (Max 5.9 ml ± Min 5.1 ml) 9 ml (Max 9.7 ml ± Min 9.4 ml) Capsule width 3.26 ml (Max 3.4 ml ± Min 3.10 ml) 5.26 ml (Max 5.50 ml ± Min 5.10 ml) Stamen length 9.25 ml (Max 9.4 ml ± Min 9.1 ml (Fig. 1.3.A2) 11.49 ml (Max 13.0 ml ± Min 10.1 ml (Fig. 1.3.B2) Anther width 0.09 ml (Max 0.10 ml ± Min 0.08 ml) (Fig. 1.3.A1) 0.12 ml (Max 0.14 ml ± Min 0.12 ml) (Fig. 1.3.B1) Anther length 2.16 ml (Max 2.22 ml ± Min 2.12 ml) (Fig  This conclusion is consistent with Sa`ad (1967) who distinguished between the morphologically very similar species C. valentinusis and C. supinus on the light of the minor variations in their leaf shapes and the presence or absence of hairs on both sides of their leaves .
Cluster 1 formed at the branch length of 0.8 was consisted of form (small leaf)-rbcL + ITS and form (large leaf)-ITS.Cluster 2 at branch length of 0.3 was consisted of the clade of form (large leaf)-matK and form (small leaf)-matK.The branch length of the out-group form (large leaf)-rbcL was 0.8.The nucleotide composition resulted from rbcL, matK and ITS regions of two forms of C. arvensis are shown in (Table 5).The average of all the three nucleotide sequences had a total of 724 positions, in the final data set, for Pos #1 (241), for Pos #2 and for Pos #3 (240).It revealed that form (small leaf)-rbcL had the highest Guanine (G) content (28.7%) and lowest Adenine (A) (20.6%), while form (large leaf)-rbcL had the highest Thymine (T) and Adenine contents of 27.4% and 27.2% respectively and lowest Cytosine (C) content of 22.0.Form (small leaf)-matK and form (large leaf)-matK had the highest Thymine (37.7%) and lowest Guanine (15.9%).Form (small leaf)-ITS and form (large leaf)-ITS had high amount of Guanine (287%) and low amount of Adenine (20.6%).The degree of bias showed avraiation depends upon the codon composition.At the first codon position for form (small leaf)-rbcL, the usage of G was 33.3%, and those of the other bases were 25.4% (C), 19.3% (A) and 22.0% (T).At the second codon position, the content of T was 23.00% and those of the other bases were 28.6% (C), 22.00% (A) and 26.4% (G).At the third codon position, the base usage was T (26.00%),C (27.6%),A (20.4%) and G (26.2%) (Shrivastava et al., 2013).The detailed numbers of codon composition for form (large leaf)-rbcL, form (small leaf)-matK, form (large leaf)-matK, form (small leaf)-ITS, form (large leaf)-ITS are shown in Table 5. Phylogenetic relationships based on molecular and morphological properties have been considered in many taxa (Bernardi and Crane, 2005;Ward et al., 2005).Also, combinations between DNA sequencing and morphological characters including leaf anatomy and macromorphology had been conducted to investigate the generic and subgeneric relations (Soh and Parnell, 2011).Interestingly in this work, we found concordant results from the analyses of molecular data and morphological and anatomical characters.
We found variations in morphological and anatomical characters accompanied with nucleotide sequences of two forms of C. arvensis.Such variations may be due to the differences in the GC percentage, nucleotide length, nucleotide bases differences, and number of gaps in each nucleotide region (Singh et al., 2016).These differences in each form of C. arvensis can be mainly ascribed to the varying numbers of repeat, alignment gaps, deletions, copy number sand base substitutions and/or additions.Base additions and substitutions are featured by very high C content that might form pure poly C structures (Shrivastava et al., 2013).Therefore, all of these factors individually or altogether will lead to a consequence of mutations in the lineage of C. arvensis forms according the prevailing environmental conditions.Variation in the amount of GC content and other nucleotides A, T and C in the first, second and third position is same to that of some other plant groups, for example, Triticeae tribe (Bieniek et al., 2015) and Pisum sativum varieties (Moustafa et al., 2019).
Comparison the obtained sequences from rbcL, matK and ITS regions with the sequences in the GenBank gave various similarity percentage with other species of convolvulus plants.For example, form (small leaf)-rbcL and form (large leaf)-rbcL gave a 100% similarity with C. lineatus and C. arvensis.There was no 100% similarity was obtained from form (small leaf)-matK, and form (large leaf)-matK, though form (small leaf)-ITS gave 100% similarity with C. arvensis.Also, the molecular phylogenetic analyses showed that form (large leaf) -rbcL clustered alone than other nucleotide sequences generated from matK and ITS regions.Therefore, rbcL can be applied to address the problems of plant phylogeny in lower taxa, even at the level of form, since it could separate form (large leaf) from form (small leaf) and in concomitant with morphoanatomical characters of these two forms.However Baldwin et al. (1995), found that ITS could reconstruct phylogenetic relationships precisely among plant species.Li et al. (2011) found that ITS effectively resolved the problems of species by sorting out 6,286 samples from 1,757 of seed plant species.An application of ITS was also questioned.The sequencing and amplification rate of ITS primer was relatively low (i.e.71.00% and 86.20% respectively) which might be as a result from some second-level structures on it (Huang et al., 2015;DeSalle, 2007;Waugh, 2007).Additionally, Xing et al. (2018) found that rbcL and matK were not appropriated for characterization the plant for tropical forest at the species level.
In conclusion, our results have established two different forms of C. arvensis (small leaf) and (large leaf) based on DNA barcoding and morpho-anatomical characters in Saudi Arabia.In addition, our results conclude that rbcL sequences are suitable for characterization of C. arvensis at the form level.

Fig. 4 .
Fig. 4. ML tree generated through maximum likelihood analysis of rbcL, matK and ITS sequence data of two forms (small leaf (sl)) and form (large leaf (ll)) of C. arvensis.