Introduction

Table 1 Accessions of C. quinoa and C. album studied in our present investigation.
Full size table

Chenopodium is one of the taxonomically most complex genus belongs to the subfamily Chenopodioideae within the family Chenopodeace. The highly polymorphic habit of species within this genus has caused more difficulties in their proper taxonomic identification. Taxonomic identification of Chenopodium has been controversial because of the highly polymorphic leaf shape, floral structure and seed morphology (La Duke and Crawford 1979; Kurashige and Agrawal 2005). While Wilson and Manhart (1993) have described the genus Chenopodium as a ‘taxonomic receptacle’, Rahiminejad and Gornall (2004) have described it as a complex group which lacked good morphological characteristics to distinguish between species. Most of the work on genetic diversity and phylogeny in Chenopodium has focussed on C. quinoa and C. berlandieri subsp. Nuttalliae (Ruas et al. 1999; Gangopadhyay et al. 2002) and only a few studies have been carried out on its other important species including C. album. Earlier studies carried out to resolve the taxonomic complexities in this genus clearly indicated the existence of C. album as the most polymorphic species of the genus Chenopodium. (Mehra and Malik 1963; Mukherjee 1986), karyotypic analysis (Bhargava et al. 2006; Kolano et al. 2008), flavonoids (Rahiminejad and Gornall 2004), RAPD profiles (Ruas et al. 1999; Gangopadhyay et al. 2002; Rana et al. 2010), direct amplification of minisatellite DNA (DAMD) (Rana et al. 2010) and intersimple sequence repeats (ISSR) markers (Rana et al. 2011) Bhargava et al. (2006) and Rana et al. (2010) have suggested it to be an aggregate taxon, thereby confusing its identity compared to Linnaean C. album. C. quinoa is an allotetraploid (\(2n=4x=36\)) with an estimated genome size of \(\sim \)1.5 Gbp (Palomino et al. 2008; Kolano et al. 2016). In addition, C. album is known as a complex of diploid \((2n=18)\), tetraploid \((2n=36)\) or hexaploid (\(2n=54\)) cytotypes with endopolyploidy and autopolyploidy as the origin of polyploidy (Kolano et al. 2008). Limited genetic resources have long been considered as a major factor hindering molecular marker-assisted breeding in quinoa (Jarvis et al. 2008). While the subfamily Chenopodiaceae has been considered to be monophyletic on the basis of sequence data of chloroplast rbcL (Kadereit et al. 2003) and matK/trnK (Müller and Borsch 2005) genes, Fuentes-Bazan et al. (2012b) have argued that Chenopodium, as tradtionally recognized, consists of six independent lineages. The arguement was based on sequence diversity in plastid trnL-F and nuclear internal transcribed spacer (ITS) regions of different species of this genus. Walsh et al. (2015) have however, identified two distinct polyploid lineages of which one lineage comprised of American tetraploid species and the other of Eastern Hemisphere hexaploid species. Thus, phylogenetic relationships among major lineages within Chenopodiaceae still remains poorly understood.

Materials and methods

Plant materials

Seeds of 19 accessions of Chenopodium (table 1) were procured from National Bureau of Plant Genetic resources, NBPGR, Shimla, India. The accessions were chosen on the basis of variations in colour, shape and arrangement of their leaves; colour, shape and surface morphology of their seeds; texture and type of pollen grains. The selection of accessions was based on the study by Devi and Chrungoo (2015). Plants of each accession were raised to full maturity in the experimental garden of the North Eastern Hill University, Shillong.

PCR amplification of ITS region: Genomic DNA, to be used as a template for PCR, was isolated from 10-day-old aetiolated seedlings of all the accessions following the method of Murray and Thompson (1980). PCR amplification of ITS region from genomic DNA of each accession was carried out with the forward (ITS5: 5\(^{\prime }\)GGAAGTAAAAGTCGTAACAAGG3\(^{\prime })\) and reverse (ITS4: 5\(^{\prime }\)TCCTCCGCTTATTGATATGC3\(^{\prime })\) primers designed by White et al. (1990). The reaction mixtures for amplification were optimized with 1 mM MgCl\(_{2}\), 0.4 mM dNTPs, 0.03 U Taq polymerase, 0.2 pmols primers and 50 ng of the template DNA. Amplification was performed in a thermal cycler programmed to one cycle of ‘hot start’at 94\(^{\circ }\)C; 35 cycles of denaturation at 94\(^{\circ }\)C for 1.0 min; annealing at 50\(^{\circ }\)C for 1 min and extension at 72\(^{\circ }\)C for 1 min and one final step of chain elongation at 72\(^{\circ }\)C for 10 min. Each reaction mixture was electrophoresed on 1.2% agarose gel. The amplicons were visualized under UV light in ChemiDoc XRS+ system (Biorad, California, USA) after staining the gels with 0.5 \(\mu \)g/mL ethidium bromide. The amplified DNA fragments were eluted from the gel and purified with QIA quick Gel Extraction kit (Qiagen, Hilden, Germany) following the manufacturer’s protocol. The precipitated DNA was pelletted by centrifuging at 13,000 rpm for 20 min at 4\(^{\circ }\)C, washed twice with 70% alcohol and vacuum dried. The dried samples were dissolved in sterile distilled water.

Table 2 Primers used in selective amplification of AFLP analysis.
Full size table

Nucleotide sequencing

Nucleotide sequencing of each amplified DNA was carried out by capillary gel electrophoresis in ABI 3130 automated sequencer (Applied Biosystems, California, USA) using performance optimized polymer-7 (POP-7) as the resolving matrix. Prior to sequencing, each amplicon was subjected to cycle sequencing with individual primers using the BigDye terminator v3.1 cycle sequencing kit (Thermo Fischer Scientific, Waltham, USA) as per the manufacturer’s protocol. The cycle sequencing protocol comprised of one cycle of hot start at 96\(^{\circ }\)C for 1 min, 35 cycles of denaturation at 96\(^{\circ }\)C for 10 s, annealing (\(T_{\mathrm{a}})\) for 2 min and chain extension at 60\(^{\circ }\)C for 3 min. DNA was precipitated from the amplified reaction mixture by adding 25 \(\mu \)L of absolute alcohol and 1.0 \(\mu \)L of 125 mM EDTA. The precipitated DNA was pelleted by centrifugation at 13,000 rpm for 20 min at 4\(^{\circ }\)C, washed twice with cold 70% alcohol and vacuum dried. The samples were denatured by addition of 10 \(\mu \)L of formamide and heating at 96\(^{\circ }\)C for 2 min prior to nucleotide sequencing.

Sequence analysis

The nucleotide sequences were analysed by BLAST (http://www.ncbi.nlm.nih.gov/Blast) to determine their homology with other known sequences in the GenBank databases. Statistical analyses for determining interspecific/intraspecific sequence diversity were carried out using SeqState v.1.21(Müller 2005). Multiple sequence alignments to determine conserved sites (CS), variable sites (VS), parsimony informative sites (PIS), singleton sites (SS), and transition/transversion ratio and nucleotide pair frequencies, namely, identical pair (ip), transitional pair (Ts) and transversional pair (Tv) were generated using MEGA 6.06 (Tamura et al. 2011).

Compensatory base changes in predicted secondary structure of ITS2

The secondary structure of ITS2 was predicted using Vienna RNA web servers (Vienna RNA package, v2.1.9) at default energy and temperature (37\(^{\circ }\)C) with no manual modification and optimization in sequences. Compensatory base changes (CBCs), which are defined as base changes in a paired region of a primary RNA transcript when both nucleotides of a paired site mutate while the pairing itself is maintained (Gutell et al. 1994), were analysed in the predicted secondary structure of ITS2 using CBC analyser 1.1 (Wolf et al. 2005).

Amplified fragment length polymorphism analysis

Amplified fragment length polymorphism (AFLP) analysis of DNA from the accessions of Chenopodium was carried out with LI-COR4300 DNA analyser following the procedure described by Vos et al. (1995). Genomic DNA from each accession was digested with MseI and EcoRI and ligated with respective adaptors at 20\(^{\circ }\)C for 18 h. Preamplification of the ligated fragments was performed using adaptor primers with one selective base at \(3'\) end. The preamplification products were subsequently used as templates for selective amplification. A total of 24 random primer combinations were used in selective amplification from a set of 13MseI and fourIR700 labelled EcoRI primers (table 2). The amplification products were size fractionated on 6.5% polyacrylamide gel using LI-COR 4300 automated DNA analyser and the bands were scored as 0 (absent) / 1 (present). The data matrix on presence or absence of AFLP bands was subjected to cluster analysis to generate Jaccard’s similarity coefficient using NTSYS-pc v2.1 (Rohlf 2000). The polymorphic information content (PIC) of each AFLP marker was calculated by the formula PIC\(_{i}=2f_{i}\)(1\(-f_{i})\), where \(f_{i}\) is the frequency of the amplified allele (band present), and 1–\(f_{i}\) is the frequency of the null allele (Roldan-Ruiz et al. 2000). The marker index (MI) was calculated according to the formula MI=PIC\(\times \)EMR, where effective multiplex ratio (EMR)\(= n\beta \), in which ‘n’ represents total number of loci per fragment per primer and ‘\(\beta \)’ represents the fraction of polymorphic loci. The resolving power (RP) of each primer was calculated as RP=\(\Sigma I_{\mathrm{b}}\), in which \(I_{\mathrm{b}}\) represents fragment informativeness.

Phylogenetic analysis and age estimation

Taxon sampling

The dataset for phylogenetic analysis included nucleotide sequences of nuclear ITS region of Chenopodium species belonging to Chenopodieae I (including the Rhagodia and Einadia), Chenopodieae II (Chenopodiastrummurale lineage) and Chenopodieae III (Chenopodiastrumrubrum lineage) of subfamily Chenopodioideae. Taxa included in the sampling also included representatives from various tribes of Chenopodiodeae, namely, Dysphanieae, Anserineae, Atripliceae, Axyrideae and the families Amaranthaceae, Polygonaceae and Nelumbonaceae.

Age calibration

The estimated age of different accessions of C. quinoa and C. album were determined from nucleotide sequences of whole ITS sequence using the calibrated ages of Amaranthaceae (87–47 million years ago (mya); Kadereit et al. 2003), Polygonaceae (90.7–125 mya; Schuster et al. 2013) and Nelumbonaceae (125–137 mya; Wikström et al. 2001) as reference markers. Representative taxa of Nelumbonaceae, namely, Nelumbolutea, N. nucifera and N. pentapetala were used as the out group.

BEAST analyses

The dataset for Bayesian evolutionary analysis sampling trees (BEAST) analysis comprised of 184 sequences of the ITS region including 165 sequences retrieved from GenBank databases and 19 sequences from the accessions of Chenopodium investigated in the present study. The sequences were aligned using MEGA 6.06 software. The output file of the alignment matrix was loaded on BEAUti v1.8.0 in nexus format for setting model parameters for BEAST. The ‘HKY substitution model’ with ‘estimated’ base frequencies and ‘gamma+invariant’ distributed rate variation among sites, ‘SRD06’ model with two partitions: position (1+2), three with unlinked substitution rate parameters and unlinked rate heterogeneity across codon position and the ‘Clock’ model with log normal relaxed clock were used for data analysis. ‘Yule tree prior’ was used to construct the tree with the ucld.mean adjusted to a uniform prior of 10-0.000001 to reflect reasonable substitution rates per site. The output file generated with BEAUti for 100 million generations was loaded in BEAST v1.8.0 for generating the time-measured phylogenetic tree. The maximum credibility tree was generated from the output file of BEAST using Tree Annotator v1.8.0 (beast.bio.ed.ac.uk/tree annotator) with a burn-in of 25%. Posterior probability values of 1.00–0.90, 0.89–0.70 and 0.69–0.50 indicated strong, moderate and weak clade support, respectively.

Table 3 Indels, sequence statistics and nucleotide pair frequencies analysis of sequences representing nuclear ITS region of different accessions belonging to C. quinoa and eight C. album.
Full size table

Results

PCR amplification of ITS and its sequence analysis

PCR with primer pair ITS5-ITS4 successfully amplified the nuclear ITS region from all the accessions of Chenopodium investigated in the present study. The amplicons generated with DNA template from each accession showed an apparent molecular mass of 0.7 kb. The length of ITS1, 5.8S and ITS2 ranged from 263 to 270, 169 to 172 and 229 to 248 bases, respectively. The average G+C% for ITS1, 5.8S and ITS2 regions was 54.5, 51.90 and 60.1, respectively. BLASTn analysis of the nucleotide sequences of entire ITS region, showed 88–99% homology with the ITS of other species of Chenopodium available in the database. All sequences have been deposited in GenBank (table 1). A comparative analysis of the interspecific sequence divergence in ITS1, ITS2 and 5.8S regions among the 19 accessions of Chenopodium revealed highest sequence divergence of 10.15% in ITS2 (table 3). The number of indels (including gaps) recorded in ITS1, 5.8S and ITS2 regions of the ITS region were 14, 4 and 28, respectively. Sequence analysis revealed 18.55, 13.70 and 2.34% parsimony informative sites in ITS2, ITS1 and 5.8S regions, respectively. With an R value of 1.02, ITS1 showed a higher ratio of transition to transversion than ITS2, which had an R value of 0.55. While the ITS1, 5.8S and ITS2 regions of accessions belonging to C. album, respectively, showed 7.42, 3.45 and 14.95% intraspecific sequence diversity those accessions belonging to C. quinoa showed 5.89, 1.05 and 3.48% sequence diversity, respectively (table 3). Clustal multiple alignment of the sequences revealed the presence of 24 SNPs in ITS1 and 18 SNPs in ITS2, which could discriminate between C. quinoa and C. album (table 1, a&b in electronic supplementary material). The 50% majority-rule consensus tree constructed from alignment matrix of ITS region of the 19 accessions of chenopods investigated in the present study as well as the sequences available in the GenBank databases revealed clustering of the accessions into six broad groups (figure 1). Species belonging to section Blitum, namely, C. bonus-henricus, C. californicum, C. capitatum and C. foliosum formed a single cluster (cluster I). While five accessions of C. album, namely, EC-359451, EC-359447, IC-341700, IC-22517 and IC-447575 grouped into one cluster (cluster II) along with sequences bearing accession numbers FN561549 (C. album), HE577468 (C. iljinii), FN561556 (C. giganteum), HM005835 (C. murale) and HE577413 (C. gigantospermum), other three accessions of C. album, namely, IC-341704, IC-411824 and IC-411825 clustered together with sequences from 11 accessions of C. quinoa (cluster III). Cluster IV comprised of sequences from C. leptophyllum, C. pallescens, C. standleyanum, C. nevadense, C. pallidicaule, C. cycloides, C. desertorum, C. atrovirens, C. petiolare, C. incanum, C. dessicatum, C. subglabrum, C. fremontii, C. pratericola, C. hians, C. neomaxicanum and C. watsonii. While C. coronopus, C. hybridum, C. rubrum, C. urbicum of section Pseudoblitum and C. glaucum of section Glauca formed the fifth cluster, sequences from species of section Dysphania, namely, C. graveolans, C. ambrosioides and C. schraderianum formed the sixth cluster (figure 1).

Fig. 1
figure 1

Fifty per cent majority-rule consensus tree based on the alignment matrix of ITS sequences for inferring relationship between 38 species of the genus Chenopodium. Numbers at each node represent bootstrap values. Numbers prefixed with IC/ EC/NIC at each branch indicate the accession number of Chenopodium species studied in the present investigation. Tree rooted on Fagopyrumesculentum.

Full size image

RNA secondary structure prediction

The predicted RNA secondary structures of ITS2 for the 19 accessions of Chenopodium investigated in the present study showed four common helices comprising of helix III, a U–U mismatch in helix II, conserved motif AAA between helix II and III, and an UGGGU/UGGU/GGU motif near the apex of helix III (figure 2).While the secondary structure of ITS2 of six accessions of C. quinoa, namely, EC-507738, EC-507740, EC-507741, EC-507748, EC-5077391 and EC-5077401 showed additional small loops between helix II and III, helix III and IV, and helix IV and I, the additional loop between helix III and IV was not detected in four accessions, namely, EC-507739, EC-507742, EC-507744 and EC-507747 of this species. Further, the secondary structure of ITS2 of two accessions of C. album, namely, IC-447575 and NIC-22517 showed an additional branched loop on the \(5'\) side of the Y-shaped helix III (figure 2). While the predicted secondary structure of consensus ITS2 sequence of C. album and C. quinoa revealed additional loops between helix II–III and helix IV–I, the sequences from C. quinoa showed an additional loop between helix III–IV. CBC analysis in the predicted RNA secondary structures of ITS2 of six accessions of Chenopodium, representing RNA secondary structures of ITS2 of C. quinoa and C. album, did not reveal any CBC between EC-507740 and EC507747, and between IC-411824 and EC-359451. However, we could detect three CBC in the predicted RNA secondary structures of ITS2 between IC-411824 on one hand and the accessions IC-447575 and IC-341704 on the other.

Fig. 2
figure 2

Graphical visualization (2Dplot) of the secondary structure of ITS2 spacer region of accessions belonging to C. quinoa (1, EC-507738; 2, EC-507739; 3, EC-507740; 4, EC-507741; 5, EC-507742; 6, EC-507744; 7, EC-507747; 8, EC-507748; 9, EC-5077391; 10, EC-5077401; 11, EC-5077402) and C. album (12, EC-359451; 13, EC-359447; 14, IC-341700; 15, IC-341704; 16, IC-411824; 17, IC-411825; 18, IC-447575; 19, NIC-22517); 20, consensus structure of C. quinoa; 21, consensus structure of C. album. The four common helices of conserved structure of ITS2 are numbered I–IV. U–U mismatched in helix II, AAA motif in between helices II–III and UGGU motif in helix III are highlighted.

Full size image

AFLP

Of the 24 random primer combinations used in selective amplification for AFLP analysis, eight primer combinations generated best polymorphic profiles (figure 3). The eight primer pairs generated a total of 373 bands of which 291 (78.01%) were polymorphic and 82 (21.99%) were monomorphic. The highest number of polymorphic bands was generated by primer pair E-ACG+M-CAG. The highest and lowest PIC values of 0.35 and 0.23 were observed for the profiles generated with primer pairs E-ACG+M-CTGC and E-ACT+ M-CTAG, respectively. The RP of the selected eight primer pairs ranged from 0.42 for the primer pair E-AGG+M-CAT to 0.52 for primer pair E-ACG+M-CTGC. The EMR for the amplicons generated with different primer combinations ranged from 19.84 for the primer pair E-AGC+M-CAA to 42.67 for the primer pair E-ACG+M-CAG. For determining utility of the marker system, we calculated the MI for all primer combinations. The MI value ranged from 5.35 for the primer combination E-AGC+M-CAA to 10.67 for the primer combination E-ACG+M-CAG (table 4). The highest correlation with \(r^{2}\) value of 0.7725 at \(P_{0.01}\) was observed between PIC and RP. The concatenated unweighted pair group method with arithmetic mean (UPGMA) dendrogram generated from the combined scoring profiles of all eight primer combinations resolved the accessions into two clades. While the accessions IC-341700, IC-447575, EC-359451 and EC-359447, all of which are hexaploid C. album, formed a single clade, the accessions EC-507744, EC-507741, EC-507742, EC-507738, EC-5077402, EC-5077401, EC-507748, EC-507739, EC-507747 all of which are hexaploid C. quinoa as well as IC-411825, IC-411824, which are tetraploid and have been reported as C. album, formed the other clade. The accession IC341704, which is a hexaploid species of C. album, emerged as a separate group independent of the other two clades (figure 4).

Fig. 3
figure 3

AFLP profiles of genomic DNA from different accessions of Chenopodium studied in the present investigation using primer pairs: 1, EAGC+MCAA; 2, EACT+MGAC; 3, EAGG+MGAC; 4, EACG+MCAG; 5, EAGG+MCAT; 6, EACT+MCTG; 7, EACG+MCTGC; 8, EACT+MCGAT. For each primer combination the sequence of accessions in the lanes is IC-341704, IC-341700, IC-447575, EC-359447, EC-359451, EC-507744, EC-507742, IC-411825, IC-411824, EC-507738, EC-5077391, EC-5077401, EC-5077402, EC-507741, EC-507747, EC-507748. Marker: IR-700 labelled 100-bp ladder.

Full size image
Table 4 Band size, total number of band, monomorphic bands (%), polymorphic bands (%), PIC, RP, EMR and MIscored from AFLP banding analysis in different accessions of Chenopodium investigated in the present study.
Full size table
Fig. 4
figure 4

Concatenated UPGMA dendrogram generated using NTSYS-pc v2.1 from the AFLP scoring profiles of eight primer combinations. The numbers prefixed with EC/IC/NIC at the tip of the branch indicate the accession numbers of the plants. Figures at branch points represent bootstrap values.

Full size image

Phylogeny and evolutionary divergence

The phylogenetic tree inferred using maximum likelihood from 184 nucleotide sequences of nuclear ITS region representing species belonging to Chenopodiaceae, Amaranthaceae, Polygonaceae and Nelumbonaceae clustered into 10 broad clusters within which species of the genus Chenopodium were distributed in five clades, namely, Chenopodieae I, Chenopodieae II, Chenopodieae III, Dysphanieae and Anserineae (figure 1 in electronic supplementary material). While clade IV comprised of the tribe Anserineae that included Blitumcalifornicum, B. capitatum, B. virgatum, Monolepis nuttaliana, Spinaceaoleracea and Chenopodiumfoliosum with strong clade value of 0.91, clade V was formed by the representative taxa of tribe Axyrideae along with three species belonging to genus Krascheninnikovia, namely, K. ceratoides, K. eversmanniana, K. lanata with strong clade support value of 0.98. Clade VI comprised of species belonging to the tribe Dysphanieae, namely, Dysphaniaambrosioides, D. multifida, D. cristata, D. pumilio, D. graveolens, D. aristata, Chenopodiumambrosioides, C. botrys and C. schraderianum with clade support value of 0.86. While species belonging to Chenopodieae II (Chenopodiastrummurale lineage), namely, C. murale, C. coronopus and C. hybridum formed clade VII, clade VIII comprised of species belonging to Chenopodieae III (Chenopodiastrumrubrum lineage), namely, C. rubrum,   C. ambigum, C. glaucum and Oxybasisurbica. Clade IX comprised of the tribe Atripliceae represented by species belonging to the genus Atriplex, namely, A. asiatica, A. rosea, A. suberecta, A. lentiformis, A. watsonii, A. peruviana, A. phyllostegia, A. patagonica, A.cinerea, A. amnicola and A. prostrate with clade support of 0.94. Clade X was an assemblage of species belonging to Chenopodieae I including the genera Rhagodia and Einadia with a posterior probability value of 0.53. The representative taxa of the family Amaranthaceae, namely, Amaranthusblitoides, A. californicus, A. lividus, A. palmeri, A. standleyanus, A. tuberculatus along with Beta vulgaris formed clade III with clade support value of 0.91. The representative taxa belonging to the genera Fagopyrum and Polygonum of the family Polygonaceae emerged as clade II with a clade value of 0.99. The tree was rooted to Nelumbonucifera of nelumbonaceae (clade I).

Discussion

All the accessions of Chenopodium studied in the present investigation had yellow flowers, alternate leaf arrangement, obtuse leaf base, lenticular seeds positioned vertically in the flower and sunken pollen. Variations were, however, observed in leaf colour, leaf shape, leaf margins, type of pollen grains, colour and texture of seed coat and morphology of seed edges (Devi and Chrungoo 2015). While the accessions IC-411824 and IC-411825 were tetraploid with \(2n=4x=36\), all other accessions belonging to C. album were hexaploid with \(2n=6x=54\). All the accessions of C. quinoa had a chromosome number of \(2n=4x=36\) (Chrungoo and Babita, personal communication). While C. quinoa \((2n=4x=36)\) is reported as a tetraploid of putative allopolyploid origin (Wilson 1980), C. album is known as a complex of diploid (2\(n=18\)), tetraploid (2\(n=36\)) or hexaploid (2\(n=54\)) cytotypes with endopolyploidy and autopolyploidy as the origin of polyploidy (Kolano et al. 2008). Our results on the extent of sequence variations in the ITS region indicate ITS1 to be more conserved than ITS2. This is in contrast with the observations of Singh (2010), who has reported a higher percentage of conserved residues in ITS2 than that of ITS1. Our results are, however, in agreement with Alice and Campbell (1999), who have reported a higher mean pair-wise divergence of sequences in ITS2 than in ITS1 in Rubus and Dalibarda. The nucleotide sequences of whole ITS of different accessions of Chenopods examined in the present study revealed lesser parsimony informative sites than the previous report of Singh (2010) on 12 different species of Chenopodium and Rana et al. (2011) on C. album complex. A significant feature of ITS1 and ITS2 spacer region was the ability to discriminate accessions belonging to C. quinoa and C. album. Even though Joshi (1991) had suggested the Himalayan chenopod (C. album) to be an assemblage of more than one species, Gangopadhyay et al. (2002) and Bhargava et al. (2006) have proposed it to be an assemblage of heteromorphic and heterocytotic (2x, 4x, 6x) forms. Our results also suggest a higher intraspecific genetic diversity in C. album than that in C. quinoa.

Analysis of the predicted RNA secondary structures of ITS2 of the 19 accessions of Chenopodium investigated in the present study revealed a typical four helical structure in all accessions. A significant feature of the analysis was the similarity in the secondary structure of ITS2 between IC-341704, IC-411824, EC-507739, EC-507742 and EC-507744 all of which had green-coloured leaves. Absence of any CBC in the secondary structure of ITS2 between IC-411824 and EC-359451 indicates that the two accessions presumably belong to the same species. Existence of CBC between IC-411824 on the one hand and two accessions of C. album, namely, IC-447575 and IC-341704 on the other is indicative of taxonomic differences between IC-411824 on the one hand and IC-447575 and IC-341704 on the other. Whereas all the accessions of C. album  had rhombic-shaped leaves with an acute apex and dentate margins, the accession IC-341704 had lanceolate leaves with an obtuse apex and entire margin. The morphological and genetic variations exhibited by the accession IC-341704 calls for further investigation into the taxonomic identity of this hexaploid Chenopod. Muller et al. (2007) have suggested a 93% probability of two sequences belonging to different species if they show any CBC in the predicted secondary structures of their ITS2 but only a 76% probability of they being of the same species if they do not show any CBC in the predicted secondary structures of their ITS2. Since the 76% probability of being the same species in case of absence of CBC is lower than the 93% probability to be different species when there is CBCs and the secondary structure of ITS2 of IC-411824 is similar to the secondary structure of ITS of other accessions belonging to C. quinoa, we suggest that the accession IC-411824 belongs to C. quinoa rather than C. album. This suggestion is also supported by the similarity in the chromosome number of IC-411824 with other accessions of C. quinoa.

The phylogenetic tree generated from the combined AFLP scoring profiles of all primer combinations resolved the accessions into two clades. While four hexaploid \((2n=6x=54)\) C. album formed a single clade, nine tetraploid (\(2n=4x=36\)) accessions belonging to C. quinoa clustered together into a single clade with three sub-groups. The accessions IC-411825 and IC-411824,  which are tetraploid and have been reported as C. album, clustered together with the tetraploid C. quinoa. Similar results on clustering of different accessions of Chenopodium, on the basis of RAPD and DAMD markers have been reported by Rana et al. (2010). Here, the accessions clustered together on the basis of their ploidy level. The higher EMR and MI for primer pair E-ACG+M-CAG reveals the higher efficiency of this primer set in divulging the genetic diversity of the collection.

The accessions IC-411824 and IC-411825, showed the same ITS and AFLP profiles as other accessions of C. quinoa, thereby indicating the closeness of these accessions of C. quinoa. The variations exhibited by IC341704 calls for further investigation on the taxonomic identity of this hexaploid species. Our observations are also supported by the time-measured phylogenetic tree that clearly revealed IC-411824 and IC-411825 as components of the clade comprised of accessions belonging to C. quinoa.Anabalon-Rodiguez and Thomet-Isla (2009) were able to correlate geographic distribution, grain colour, panicle colour and phenology of C. quinoa with AFLP-based marker fragments. Their results clustered the accessions into three groups where group-I comprised of eight local accessions from northern Chile having yellow to brown grains, yellow but compact to intermediate density panicle, agglomerated form and precocity. Group-II comprised of highland accessions of later phenology, yellow to red coloured grains, compact and amarantiform panicle from precordillera sector and group III included the species C. album and C. ambrosioides which were denominated as outgroup controls. While the morphological analysis of the Andean cultivars placed them as an independent Group (group-II), the AFLP data integrated the Andean cultivars with group-I, indicating the existence of similarity in the genetic material of accessions collected from the region. This result is consistent with the existence of common ancestral genes in the crop.

On the basis of sequence variations in ITS and chloroplast trnL-F regions, Fuentes-Bazan et al. (2012a) have reported five lineages comprising of the tribes Chenopodieae I, Chenopodieae II, Chenopodieae III, Anserineae and Dysphanieae within sub-family Chenopodioideae. The evolutionary relationship reconstructed in the present study confirms five independent well-supported lineages of the sub-family Chenopodioideae. The ITS and AFLP profiles observed in the present study indicate that the accessions of C. album constitute heteromorphs of the same species rather than a heterogenous assemblage of taxa as suggested by Rana et al. (2011). The results also establish C. album as a sister clade of C. quinoa within the tribe Chenopodieae I. Even though Müller and Borsch (2005) have suggested a monophyletic evolution of Chenopodiaceae and Amaranthaceae families on the basis of variations in chloroplast matK/trnK sequences, nesting of Amaranthaceae within sub-family Chenopodioideae, as observed in the present study, indicates a paraphyletic evolution of family Chenopodiaceae. This observation is in conformity with many others who have also reported nesting of Amaranthaceae within Chenopodiaceae (Manhart and Rettig 1994; Downie et al. 1997) as well as Fuentes-Bazan et al. (2012a, b), who classified Chenopodium sensulato as paraphyletic. Our results suggesting a monophyletic evolution of Chenopodieae I are in conformity with those of Fuentes-Bazan et al. (2012b), who have suggested a monophyletic evolution for Chenopodium sensu strict. Our results indicate that the Himalayan chenopods evolved 10.3-5.9 mya, which is much younger than the reported crown age of Chenopodieae I. This compares well with the age of tetraploidization in Chenopodium suggested by Jarvis et al. (2017).