Journal of Genetics (2019) 98:37 © Indian Academy of Sciences https://doi.org/10.1007/s12041-019-1079-0

RESEARCH ARTICLE

Deciphering species relationships and evolution in Chenopodium through sequence variations in nuclear internal transcribed spacer region and amplified fragment-length polymorphism in nuclear DNA

NIKHIL K. CHRUNGOO1∗, RAJKUMARI JASHMI DEVI1,3, SHAILENDRA GOEL2 and KAMAL DAS2

1Department of Botany, Centre for Advanced Studies, North-Eastern Hill University, Shillong 793022, India 2Department of Botany, University of Delhi, Delhi 110 007, India 3Present address: Institute of Bioresources and Sustainable Development, Takyelpat, Imphal 795 001, India *For correspondence. E-mail: [email protected], [email protected].

Received 22 June 2018; revised 30 September 2018; accepted 2 November 2018; published online 16 April 2019

Abstract. Evaluation of sequence variations in the internal transcribed spacer (ITS) region of 19 accessions, comprising of 11 accessions of Chenopodium quinoa, eight accessions of Chenopodium album and 165 retrieved sequences of different species of Chenopodium belonging to subfamily Chenopodioideae revealed a higher intraspecific genetic diversity in Himalayan C. album than that in C. quinoa. ITS and amplified fragment-length profiles of the accessions suggest the existence of accessions of Himalayan C. album as heteromorphs of the same species rather than a heterogenous assemblage of taxa. While the evolutionary relationship reconstructed from variations in 184 sequences of ITS region from species belonging to Chenopodiaceae, Amaranthaceae, Polygonaceae and Nelumbonaceae established a paraphyletic evolution of family Chenopodiaceae, it also revealed a monophyletic evolution of Chenopodieae I. The reconstruction also established five independent lineages of the subfamily Chenopodioideae with C. album as a sister clade of C. quinoa within the tribe Chenopodieae I. The results also indicate a much younger age for Himalayan chenopods (C. album) than the reported crown age of Chenopodieae I.

Keywords. Chenopodioideae; internal transcribed spacer; amplified fragment-length profile; time-measured phylogenetic tree; evolutionary divergence; Chenopodium.

Introduction have described it as a complex group which lacked good morphological characteristics to distinguish between Chenopodium is one of the taxonomically most complex species. Most of the work on genetic diversity and phy- genus belongs to the subfamily Chenopodioideae within logeny in Chenopodium has focussed on C. quinoa and the family Chenopodeace. The highly polymorphic habit C. berlandieri subsp. Nuttalliae (Ruas et al. 1999; Gan- of species within this genus has caused more difficulties in gopadhyay et al. 2002) and only a few studies have been their proper taxonomic identification. Taxonomic identi- carried out on its other important species including C. fication of Chenopodium has been controversial because album. Earlier studies carried out to resolve the taxonomic of the highly polymorphic leaf shape, floral structure complexities in this genus clearly indicated the existence and seed morphology (La Duke and Crawford 1979; of C. album as the most polymorphic species of the Kurashige and Agrawal 2005). While Wilson and Man- genus Chenopodium. (Mehra and Malik 1963; Mukherjee hart (1993) have described the genus Chenopodium as a 1986), karyotypic analysis (Bhargava et al. 2006; Kolano ‘taxonomic receptacle’, Rahiminejad and Gornall (2004) et al. 2008), flavonoids (Rahiminejad and Gornall 2004),

Electronic supplementary material: The online version of this article (https://doi.org/10.1007/s12041-019-1079-0) contains supplemen- tary material, which is available to authorized users.

1 37 Page 2 of 11 Nikhil K. Chrungoo et al.

RAPD profiles (Ruas et al. 1999; Gangopadhyay et al. Materials and methods 2002; Rana et al. 2010), direct amplification of minisatel- lite DNA (DAMD) (Rana et al. 2010) and intersimple Plant materials sequence repeats (ISSR) markers (Rana et al. 2011) Bhar- gava et al. (2006) and Rana et al. (2010) have suggested Seeds of 19 accessions of Chenopodium (table 1) were it to be an aggregate taxon, thereby confusing its identity procured from National Bureau of Plant Genetic resour- compared to Linnaean C. album. C. quinoa is an allote- ces, NBPGR, Shimla, India. The accessions were chosen traploid (2n = 4x = 36) with an estimated genome size on the basis of variations in colour,shape and arrangement of ∼1.5 Gbp (Palomino et al. 2008; Kolano et al. 2016). of their leaves; colour, shape and surface morphology of In addition, C. album is known as a complex of diploid their seeds; texture and type of pollen grains. The selection (2n = 18), tetraploid (2n = 36) or hexaploid (2n = 54) of accessions was based on the study by Devi and Chrun- cytotypes with endopolyploidy and autopolyploidy as the goo (2015). Plants of each accession were raised to full origin of polyploidy (Kolano et al. 2008). Limited genetic maturity in the experimental garden of the North Eastern resources have long been considered as a major factor Hill University, Shillong. hindering molecular marker-assisted breeding in quinoa PCR amplification of ITS region: Genomic DNA, (Jarvis et al. 2008). While the subfamily Chenopodiaceae to be used as a template for PCR, was isolated from has been considered to be monophyletic on the basis of 10-day-old aetiolated seedlings of all the accessions fol- sequence data of chloroplast rbcL (Kadereit et al. 2003) lowing the method of Murray and Thompson (1980). and matK/trnK (Müller and Borsch 2005) genes, Fuentes- PCR amplification of ITS region from genomic DNA of Bazan et al. (2012b) have argued that Chenopodium,as each accession was carried out with the forward (ITS5:   tradtionally recognized, consists of six independent lin- 5 GGAAGTAAAAGTCGTAACAAGG3 ) and reverse   eages. The arguement was based on sequence diversity (ITS4: 5 TCCTCCGCTTATTGATATGC3 ) primers in plastid trnL-F and nuclear internal transcribed spacer designed by White et al. (1990). The reaction mixtures (ITS) regions of different species of this genus. Walsh for amplification were optimized with 1 mM MgCl2,0.4 et al. (2015) have however, identified two distinct poly- mM dNTPs, 0.03 U Taq polymerase, 0.2 pmols primers ploid lineages of which one lineage comprised of American and 50 ng of the template DNA. Amplification was per- tetraploid species and the other of Eastern Hemisphere formed in a thermal cycler programmed to one cycle of ‘hot ◦ ◦ hexaploid species. Thus, phylogenetic relationships among start’at 94 C; 35 cycles of denaturation at 94 C for 1.0 min; ◦ ◦ major lineages within Chenopodiaceae still remains poorly annealing at 50 C for 1 min and extension at 72 C for 1 ◦ understood. min and one final step of chain elongation at 72 Cfor10

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

GenBank accession number Accession no. Species Source of nr ITS sequence

IC341704 C. album NBPGRa KC577850 NIC22517 C. album NBPGR KC577845 IC341700 C. album NBPGR KC577846 IC447575 C. album Chaura, Kinnaur, Himachal Pradesh, India KC577847 EC359447 C. album NBPGR KC577849 EC359451 C. album NBPGR KC577848 IC411825 C. album Thakma, Leh Jammu and Kashmir, India KC577851 IC411824 C. album Thakma, Leh Jammu and Kashmir, India KC577836 EC507744 C. quinoa NBPGR KC577834 EC507742 C. quinoa Chile KC577835 EC507738 C. quinoa Peru KC577837 EC507739 C. quinoa Ecuador KC577838 EC5077391 C. quinoa NBPGR KC577839 EC507740 C. quinoa USA KC577840 EC5077401 C. quinoa USA KC577841 EC5077402 C. quinoa USA KC577842 EC507741 C. quinoa USA KC577843 EC507747 C. quinoa USA KF709219 EC507748 C. quinoa Argentina KC577844 aNational Bureau of Plant Genetic Resources, Shimla station, India. Species relationships and evolution in Chenopodium Page 3 of 11 37 min. Each reaction mixture was electrophoresed on 1.2% no manual modification and optimization in sequences. agarose gel. The amplicons were visualized under UV light Compensatory base changes (CBCs), which are defined in ChemiDoc XRS+ system (Biorad, California, USA) as base changes in a paired region of a primary RNA after staining the gels with 0.5 μg/mL ethidium bromide. transcript when both nucleotides of a paired site mutate The amplified DNA fragments were eluted from the gel while the pairing itself is maintained (Gutell et al. 1994), and purified with QIA quick Gel Extraction kit (Qiagen, were analysed in the predicted secondary structure of ITS2 Hilden, Germany) following the manufacturer’s protocol. using CBC analyser 1.1 (Wolf et al. 2005). The precipitated DNA was pelletted by centrifuging at 13,000 rpm for 20 min at 4◦C, washed twice with 70% alco- hol and vacuum dried. The dried samples were dissolved Amplified fragment length polymorphism analysis in sterile distilled water. Amplified fragment length polymorphism (AFLP) Nucleotide sequencing analysis of DNA from the accessions of Chenopodium was carried out with LI-COR4300 DNA analyser following Nucleotide sequencing of each amplified DNA was carried the procedure described by Vos et al. (1995). Genomic out by capillary gel electrophoresis in ABI 3130 auto- DNA from each accession was digested with MseIand mated sequencer (Applied Biosystems, California, USA) EcoRI and ligated with respective adaptors at 20◦Cfor using performance optimized polymer-7 (POP-7) as the 18 h. Preamplification of the ligated fragments was per- resolving matrix. Prior to sequencing, each amplicon was formed using adaptor primers with one selective base at subjected to cycle sequencing with individual primers 3 end. The preamplification products were subsequently using the BigDye terminator v3.1 cycle sequencing kit used as templates for selective amplification. A total of (Thermo Fischer Scientific, Waltham, USA) as per the 24 random primer combinations were used in selective manufacturer’s protocol. The cycle sequencing protocol amplification from a set of 13MseI and fourIR700 labelled ◦ comprised of one cycle of hot start at 96 C for 1 min, 35 EcoRI primers (table 2). The amplification products were ◦ cycles of denaturation at 96 C for 10 s, annealing (Ta) for size fractionated on 6.5% polyacrylamide gel using LI- ◦ 2 min and chain extension at 60 C for 3 min. DNA was COR 4300 automated DNA analyser and the bands were precipitated from the amplified reaction mixture by adding scored as 0 (absent) / 1 (present). The data matrix on pres- 25 μL of absolute alcohol and 1.0 μL of 125 mM EDTA. ence or absence of AFLP bands was subjected to cluster The precipitated DNA was pelleted by centrifugation at analysis to generate Jaccard’s similarity coefficient using ◦ 13,000 rpm for 20 min at 4 C, washed twice with cold 70% NTSYS-pc v2.1 (Rohlf 2000). The polymorphic informa- alcohol and vacuum dried. The samples were denatured tion content (PIC) of each AFLP marker was calculated μ ◦ by addition of 10 L of formamide and heating at 96 C by the formula PICi = 2fi(1−fi),wherefi is the fre- for 2 min prior to nucleotide sequencing. quency of the amplified allele (band present), and 1–fi is the frequency of the null allele (Roldan-Ruiz et al. 2000). Sequence analysis The marker index (MI) was calculated according to the formula MI=PIC×EMR, where effective multiplex ratio The nucleotide sequences were analysed by BLAST (http:// (EMR)= nβ, in which ‘n’ represents total number of loci www.ncbi.nlm.nih.gov/Blast) to determine their homology per fragment per primer and ‘β’ represents the fraction with other known sequences in the GenBank databases. of polymorphic loci. The resolving power (RP) of each  Statistical analyses for determining interspecific/ primer was calculated as RP= Ib, in which Ib represents intraspecific sequence diversity were carried out using fragment informativeness. SeqState v.1.21(Müller 2005). Multiple sequence align- ments to determine conserved sites (CS), variable sites Phylogenetic analysis and age estimation (VS), parsimony informative sites (PIS), singleton sites (SS), and transition/transversion ratio and nucleotide pair Taxon sampling: The dataset for phylogenetic analysis frequencies, namely, identical pair (ip), transitional pair included nucleotide sequences of nuclear ITS region (Ts) and transversional pair (Tv) were generated using of Chenopodium species belonging to Chenopodieae I MEGA 6.06 (Tamura et al. 2011). (including the Rhagodia and Einadia), Chenopodieae II (Chenopodiastrummurale lineage) and Chenopodieae III Compensatory base changes in predicted secondary structure of (Chenopodiastrumrubrum lineage) of subfamily Chenopo- ITS2 dioideae. Taxa included in the sampling also included representatives from various tribes of Chenopodiodeae, The secondary structure of ITS2 was predicted using namely, Dysphanieae, Anserineae, Atripliceae, Axyrideae Vienna RNA web servers (Vienna RNA package, v2.1.9) and the families Amaranthaceae, Polygonaceae and at default energy and temperature (37◦C) with Nelumbonaceae. 37 Page 4 of 11 Nikhil K. Chrungoo et al.

Table 2. Primers used in selective amplification of AFLP analysis.

    MseI primers (5 –3 ) EcoRI primers (5 –3 )

GATGAGTCCTGAGTAA CAA A1. GACTGCGTACCAATTC ACG GATGAGTCCTGAGTAA CAC A2. GACTGCGTACCAATTC ACT GATGAGTCCTGAGTAA CAG A3. GACTGCGTACCAATTC AGC GATGAGTCCTGAGTAA CAT A4. GACTGCGTACCAATTC AGG GATGAGTCCTGAGTAA CTA GATGAGTCCTGAGTAA CTC GATGAGTCCTGAGTAA CTG GATGAGTCCTGAGTAA CTT GATGAGTCCTGAGTAA CAGT GATGAGTCCTGAGTAA CTAG GATGAGTCCTGAGTAA CTGC GATGAGTCCTGAGTAA CGAT GATGAGTCCTGAGTAA CACG

EcoRI primers are infrared 700 dye labelled.

Age calibration: The estimated age of different accessions of Results C. quinoa and C. album were determined from nucleotide sequences of whole ITS sequence using the calibrated PCR amplification of ITS and its sequence analysis ages of Amaranthaceae (87–47 million years ago (mya); Kadereit et al. 2003), Polygonaceae (90.7–125 mya; Schus- PCR with primer pair ITS5-ITS4 successfully ter et al. 2013) and Nelumbonaceae (125–137 mya; Wik- amplified the nuclear ITS region from all the accessions ström et al. 2001) as reference markers. Representative taxa of Chenopodium investigated in the present study. The of Nelumbonaceae, namely, Nelumbolutea, N. nucifera and amplicons generated with DNA template from each acces- N. pentapetala were used as the out group. sion 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 BEAST analyses: The dataset for Bayesian evolutionary G+C% for ITS1, 5.8S and ITS2 regions was 54.5, 51.90 analysis sampling trees (BEAST) analysis comprised of and 60.1, respectively. BLASTn analysis of the nucleotide 184 sequences of the ITS region including 165 sequences sequences of entire ITS region, showed 88–99% homology retrieved from GenBank databases and 19 sequences from with the ITS of other species of Chenopodium available the accessions of Chenopodium investigated in the present in the database. All sequences have been deposited in study. The sequences were aligned using MEGA 6.06 GenBank (table 1). A comparative analysis of the interspe- software. The output file of the alignment matrix was cific sequence divergence in ITS1, ITS2 and 5.8S regions loaded on BEAUti v1.8.0 in nexus format for setting model among the 19 accessions of Chenopodium revealed high- parameters for BEAST. The ‘HKY substitution model’ est sequence divergence of 10.15% in ITS2 (table 3). The with ‘estimated’ base frequencies and ‘gamma+invariant’ number of indels (including gaps) recorded in ITS1, 5.8S distributed rate variation among sites, ‘SRD06’ model with and ITS2 regions of the ITS region were 14, 4 and 28, two partitions: position (1+2), three with unlinked sub- respectively. Sequence analysis revealed 18.55, 13.70 and stitution rate parameters and unlinked rate heterogeneity 2.34% parsimony informative sites in ITS2, ITS1 and across codon position and the ‘Clock’ model with log nor- 5.8S regions, respectively. With an R value of 1.02, ITS1 mal relaxed clock were used for data analysis. ‘Yule tree showed a higher ratio of transition to transversion than prior’ was used to construct the tree with the ucld.mean ITS2, which had an R value of 0.55. While the ITS1, 5.8S adjusted to a uniform prior of 10-0.000001 to reflect and ITS2 regions of accessions belonging to C. album, reasonable substitution rates per site. The output file gen- respectively, showed 7.42, 3.45 and 14.95% intraspecific erated with BEAUtifor 100 million generations was loaded sequence diversity those accessions belonging to C. quinoa in BEAST v1.8.0 for generating the time-measured phylo- showed 5.89, 1.05 and 3.48% sequence diversity, respec- genetic tree. The maximum credibility tree was generated tively (table 3). Clustal multiple alignment of the sequences from the output file of BEAST using Tree Annotator v1.8.0 revealed the presence of 24 SNPs in ITS1 and 18 SNPs in (beast.bio.ed.ac.uk/tree annotator) with a burn-in of 25%. ITS2, which could discriminate between C. quinoa and Posterior probability values of 1.00–0.90, 0.89–0.70 and C. album (table 1, a&b in electronic supplementary 0.69–0.50 indicated strong, moderate and weak clade sup- material). The 50% majority-rule consensus tree con- port, respectively. structed from alignment matrix of ITS region of the 19 Species relationships and evolution in Chenopodium Page 5 of 11 37

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 R (Si/Sv) and eight (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 C. quinoa accessions of C. album, namely, EC-359451, EC-359447, Sv (transver sional pairs) IC-341700, IC-22517 and IC-447575 grouped into one ing to cluster (cluster II) along with sequences bearing acces- sion numbers FN561549 (C. album), HE577468 (C. iljinii), FN561556 (C. giganteum), HM005835 (C. murale)and HE577413 (C. gigantospermum), other three accessions of Si (transi tional pairs) 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, II (identi cal pairs) 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,

CpG (100 coverage) C. hybridum, C. rubrum, C. urbicum of section Pseudobli- tum and C. glaucum of section Glauca formed the fifth clus- ter, sequences from species of section Dysphania,namely, SS (%) C. graveolans, C. ambrosioides and C. schraderianum formed the sixth cluster (figure 1). (%) PIS RNA secondary structure prediction

The predicted RNA secondary structures of ITS2 for the VS (%) 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 (%) CS motif near the apex of helix III (figure 2).While the sec- ondary structure of ITS2 of six accessions of C. quinoa,

Sequence statisticsnamely, Nucleotide pair frequencies EC-507738, EC-507740, EC-507741, EC-507748, (%) G+C 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

Sequence was not detected in four accessions, namely, EC-507739, divergence EC-507742, EC-507744 and EC-507747 of this species. Further, the secondary structure of ITS2 of two acces- sions 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 sec- ondary structure of consensus ITS2 sequence of C. album Mean

length (nt) Indels 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 range Length

266–269263–270 268263–270 268 268 4172–172 11169–172 14 172169–172 7.42 171 5.89 171 9.97 0 54.1233–248 54.8 223 4 (82.89)229–269 198 (73.33) 46 54.5 (17.10) 4 237229–248 72 168 (26.67) 31 (62.22) (11.52) 3.45 103 237 (38.14) 15 7 37 (5.58) (2.59) (13.70) 1.05 235 65 (24.07) 64 15 2.32 (23.70) 51.70 149 (86.63) 16 52.00 23 56 164 (13.37) 32/0 (95.90) 28the 51.90 14.95 60 145 1 (84.80) 8 (0.58) (4.68) 27predicted 3.48 (15.79) 22 (12.79) 10.15 59.1 244.00 1 248.00 4 (0.58) (2.34) 141 (56.85) 249.00 23 60.9 97 (13.45) (39.11) 7 (4.09) 60.1 58 204 (75.84) 39 (15.72) 122 (49.19) 35 58 11.00 (13.01) (23.39) 117RNA (47.18) 9.00 46 50 (18.55) 7 (2.60) 70 7.00 62 (28.23) 166 27 (10.04) 26 secondary 11.00 10.00 20 168.00 56 169.00 8.00 200.00 1.02 0.89 2 2.00 210.00 structures 1.00 224.00 0.84 13.00 8.00 2.00 3.00 0.00 4 23.00 of ITS2 1.25 15.00 0.56 3.90 5.00 0.69 of 0.55 six 0.65 accessions of Chenopodium, representing RNA secondary and and and Indels, sequence statistics and nucleotide pair frequencies analysis of sequences representing nuclear ITS region of different accessions belong 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 C. album C. quinoa C. album C. quinoa C. album C. quinoa C. album C. quinoa C. album C. quinoa C. album C. quinoa Table 3. Taxa ITS1 CS, conserved sites; VS, variable sites; PIS, parsimony informative sites; SS, singleton sites. 5.8S ITS2 C. album. three CBC in the predicted RNA secondary structures of 37 Page 6 of 11 Nikhil K. Chrungoo et al.

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 Fagopy- rumesculentum.

ITS2 between IC-411824 on one hand and the accessions combinations generated best polymorphic profiles IC-447575 and IC-341704 on the other. (figure 3). The eight primer pairs generated a total of 373 bands of which 291 (78.01%) were polymorphic and AFLP 82 (21.99%) were monomorphic. The highest number of polymorphic bands was generated by primer pair E- Of the 24 random primer combinations used in selective ACG+M-CAG. The highest and lowest PIC values of 0.35 amplification for AFLP analysis, eight primer and 0.23 were observed for the profiles generated with Species relationships and evolution in Chenopodium Page 7 of 11 37

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. primer pairs E-ACG+M-CTGC and E-ACT+ M-CTAG, formed the other clade. The accession IC341704, which respectively. The RP of the selected eight primer pairs is a hexaploid species of C. album, emerged as a separate ranged from 0.42 for the primer pair E-AGG+M-CAT to group independent of the other two clades (figure 4). 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 Phylogeny and evolutionary divergence 42.67 for the primer pair E-ACG+M-CAG. For determin- ing utility of the marker system, we calculated the MI for all The phylogenetic tree inferred using maximum likelihood primer combinations. The MI value ranged from 5.35 for from 184 nucleotide sequences of nuclear ITS region the primer combination E-AGC+M-CAA to 10.67 for the representing species belonging to Chenopodiaceae, Ama- primer combination E-ACG+M-CAG (table 4). The high- ranthaceae, Polygonaceae and Nelumbonaceae clustered 2 est correlation with r value of 0.7725 at P0.01 was observed into 10 broad clusters within which species of the genus between PIC and RP. The concatenated unweighted pair Chenopodium were distributed in five clades, namely, group method with arithmetic mean (UPGMA) dendro- Chenopodieae I, Chenopodieae II, Chenopodieae III, gram generated from the combined scoring profiles of all Dysphanieae and Anserineae (figure 1 in electronic supple- eight primer combinations resolved the accessions into mentary material). While clade IV comprised of the tribe two clades. While the accessions IC-341700, IC-447575, Anserineae that included Blitumcalifornicum, B. capita- EC-359451 and EC-359447, all of which are hexaploid C. tum, B. virgatum, Monolepis nuttaliana, Spinaceaoleracea album, formed a single clade, the accessions EC-507744, and Chenopodiumfoliosum with strong clade value of 0.91, EC-507741, EC-507742, EC-507738, EC-5077402, EC- clade V was formed by the representative taxa of tribe 5077401, EC-507748, EC-507739, EC-507747 all of which Axyrideae along with three species belonging to genus are hexaploid C. quinoa as well as IC-411825, IC-411824, Krascheninnikovia,namely,K. ceratoides, K. eversmanni- which are tetraploid and have been reported as C. album, ana, K. lanata with strong clade support value of 0.98. 37 Page 8 of 11 Nikhil K. Chrungoo et al.

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.

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.

Primer Total Band size Total band Monomorphic Polymorphic combinations individual in range scored bands (%) bands (%) PIC Rp EMR MI

1 E-AGC+M-CAA 16 91.4–513.0 48 17 (35.42) 31 (64.58) 0.27 0.44 19.84 5.35 2 E-ACT+M-GAC 16 116.0–561.0 47 14 (29.79) 33 (70.21) 0.32 0.51 23.17 7.34 3 E-AGG+M-GAC 16 114.0–588.0 51 10 (19.61) 41 (80.39) 0.29 0.43 32.96 9.56 4 E-ACG+M-CAG 16 79.0–570.0 54 6 (11.11) 48 (88.89) 0.26 0.43 42.67 10.67 5 E-AGG+M-CAT 16 100.0–459.0 55 12 (21.82) 43 (78.18) 0.26 0.42 33.62 8.74 6 E-ACT+M-CTAG 16 110.0–446.0 34 7 (20.59) 27 (79.41) 0.29 0.48 21.44 6.21 7 E-ACT+M-CTAG 16 85.0–576.0 43 8 (18.60) 35 (81.40) 0.23 0.43 28.49 6.55 8 E-ACG+M-CTGC 16 85.0–629.0 41 8 (19.51) 33 (80.49) 0.35 0.52 26.56 9.29 Total 373 82 (21.99%) 291 (78.01%)

Clade VI comprised of species belonging to the tribe Dys- lentiformis, A. watsonii, A. peruviana, A. phyllostegia, A. phanieae, namely, Dysphaniaambrosioides, D. multifida, patagonica, A.cinerea, A. amnicola and A. prostrate with D. cristata, D. pumilio, D. graveolens, D. aristata, clade support of 0.94. Clade X was an assemblage of Chenopodiumambrosioides, C. botrys and C. schraderi- species belonging to Chenopodieae I including the genera anum with clade support value of 0.86. While species Rhagodia and Einadia with a posterior probability value belonging to Chenopodieae II (Chenopodiastrummurale of 0.53. The representative taxa of the family Amaran- lineage), namely, C. murale, C. coronopus and C. hybridum thaceae, namely, Amaranthusblitoides, A. californicus, A. formed clade VII, clade VIII comprised of species belong- lividus, A. palmeri, A. standleyanus, A. tuberculatus along ing to Chenopodieae III (Chenopodiastrumrubrum lin- with Beta vulgaris formed clade III with clade support eage), namely, C. rubrum, C. ambigum, C. glaucum value of 0.91. The representative taxa belonging to the gen- and Oxybasisurbica. Clade IX comprised of the tribe era Fagopyrum and Polygonum of the family Polygonaceae Atripliceae represented by species belonging to the genus emerged as clade II with a clade value of 0.99. The tree was Atriplex,namely,A. asiatica, A. rosea, A. suberecta, A. rooted to Nelumbonucifera of nelumbonaceae (clade I). Species relationships and evolution in Chenopodium Page 9 of 11 37

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.

Discussion complex. A significant feature of ITS1 and ITS2 spacer region was the ability to discriminate accessions belong- All the accessions of Chenopodium studied in the present ing to C. quinoa and C. album. Even though Joshi (1991) investigation had yellow flowers, alternate leaf had suggested the Himalayan chenopod (C. album)tobe arrangement, obtuse leaf base, lenticular seeds positioned an assemblage of more than one species, Gangopadhyay vertically in the flower and sunken pollen. Variations were, et al. (2002) and Bhargava et al. (2006) have proposed it to however, observed in leaf colour, leaf shape, leaf mar- be an assemblage of heteromorphic and heterocytotic (2x, gins, type of pollen grains, colour and texture of seed 4x,6x) forms. Our results also suggest a higher intraspe- coat and morphology of seed edges (Devi and Chrun- cific genetic diversity in C. album than that in C. quinoa. goo 2015). While the accessions IC-411824 and IC-411825 Analysis of the predicted RNA secondary structures were tetraploid with 2n = 4x = 36, all other accessions of ITS2 of the 19 accessions of Chenopodium investi- belonging to C. album were hexaploid with 2n = 6x = 54. gated in the present study revealed a typical four helical All the accessions of C. quinoa had a chromosome num- structure in all accessions. A significant feature of the ber of 2n = 4x = 36 (Chrungoo and Babita, personal analysis was the similarity in the secondary structure of communication). While C. quinoa (2n = 4x = 36) is ITS2 between IC-341704, IC-411824, EC-507739, EC- reported as a tetraploid of putative allopolyploid origin 507742 and EC-507744 all of which had green-coloured (Wilson 1980), C. album is known as a complex of diploid leaves. Absence of any CBC in the secondary structure (2n = 18), tetraploid (2n = 36) or hexaploid (2n = 54) of ITS2 between IC-411824 and EC-359451 indicates that cytotypes with endopolyploidy and autopolyploidy as the the two accessions presumably belong to the same species. origin of polyploidy (Kolano et al. 2008). Our results on Existence of CBC between IC-411824 on the one hand the extent of sequence variations in the ITS region indicate and two accessions of C. album, namely, IC-447575 and ITS1 to be more conserved than ITS2. This is in contrast IC-341704 on the other is indicative of taxonomic differ- with the observations of Singh (2010), who has reported a ences between IC-411824 on the one hand and IC-447575 higher percentage of conserved residues in ITS2 than that and IC-341704 on the other. Whereas all the accessions of ITS1. Our results are, however, in agreement with Alice of C. album had rhombic-shaped leaves with an acute and Campbell (1999), who have reported a higher mean apex and dentate margins, the accession IC-341704 had pair-wise divergence of sequences in ITS2 than in ITS1 in lanceolate leaves with an obtuse apex and entire mar- Rubus and Dalibarda. The nucleotide sequences of whole gin. The morphological and genetic variations exhibited ITS of different accessions of Chenopods examined in the by the accession IC-341704 calls for further investigation present study revealed lesser parsimony informative sites into the taxonomic identity of this hexaploid Chenopod. than the previous report of Singh (2010) on 12 different Muller et al. (2007) have suggested a 93% probability species of Chenopodium and Rana et al. (2011) on C. album of two sequences belonging to different species if they 37 Page 10 of 11 Nikhil K. Chrungoo et al. show any CBC in the predicted secondary structures of with the existence of common ancestral genes in the their ITS2 but only a 76% probability of they being of crop. the same species if they do not show any CBC in the On the basis of sequence variations in ITS and chloro- predicted secondary structures of their ITS2. Since the plast trnL-F regions, Fuentes-Bazan et al. (2012a) have 76% probability of being the same species in case of reported five lineages comprising of the tribes Chenopo- absence of CBC is lower than the 93% probability to be dieae I, Chenopodieae II, Chenopodieae III, Anserineae different species when there is CBCs and the secondary and Dysphanieae within sub-family Chenopodioideae. structure of ITS2 of IC-411824 is similar to the sec- The evolutionary relationship reconstructed in the present ondary structure of ITS of other accessions belonging to study confirms five independent well-supported lineages C. quinoa, we suggest that the accession IC-411824 belongs of the sub-family Chenopodioideae. The ITS and AFLP to C. quinoa rather than C. album. This suggestion is also profiles observed in the present study indicate that the supported by the similarity in the chromosome number of accessions of C. album constitute heteromorphs of the IC-411824 with other accessions of C. quinoa. same species rather than a heterogenous assemblage of The phylogenetic tree generated from the combined taxa as suggested by Rana et al. (2011). The results AFLP scoring profiles of all primer combinations resolved also establish C. album as a sister clade of C. quinoa the accessions into two clades. While four hexaploid (2n = within the tribe Chenopodieae I. Even though Müller and 6x = 54) C. album formed a single clade, nine tetraploid Borsch (2005) have suggested a monophyletic evolution of (2n = 4x = 36) accessions belonging to C. quinoa clus- Chenopodiaceae and Amaranthaceae families on the basis tered together into a single clade with three sub-groups. of variations in chloroplast matK/trnK sequences, nesting The accessions IC-411825 and IC-411824, which are of Amaranthaceae within sub-family Chenopodioideae, as tetraploid and have been reported as C. album, clustered observed in the present study, indicates a paraphyletic evo- together with the tetraploid C. quinoa. Similar results on lution of family Chenopodiaceae. This observation is in clustering of different accessions of Chenopodium,onthe conformity with many others who have also reported nest- basis of RAPD and DAMD markers have been reported by ing of Amaranthaceae within Chenopodiaceae (Manhart Rana et al. (2010). Here, the accessions clustered together and Rettig 1994; Downie et al. 1997) as well as Fuentes- on the basis of their ploidy level. The higher EMR and MI Bazan et al. (2012a, b), who classified Chenopodium sensu- for primer pair E-ACG+M-CAG reveals the higher effi- lato as paraphyletic. Our results suggesting a monophyletic ciency of this primer set in divulging the genetic diversity evolution of Chenopodieae I are in conformity with those of the collection. of Fuentes-Bazan et al. (2012b), who have suggested a The accessions IC-411824 and IC-411825, showed the monophyletic evolution for Chenopodium sensu strict. Our same ITS and AFLP profiles as other accessions of C. results indicate that the Himalayan chenopods evolved quinoa, thereby indicating the closeness of these acces- 10.3-5.9 mya, which is much younger than the reported sions of C. quinoa. The variations exhibited by IC341704 crown age of Chenopodieae I. This compares well with calls for further investigation on the taxonomic iden- the age of tetraploidization in Chenopodium suggested by tity of this hexaploid species. Our observations are also Jarvis et al. (2017). supported by the time-measured phylogenetic tree that clearly revealed IC-411824 and IC-411825 as components Acknowledgements of the clade comprised of accessions belonging to C. quinoa. Anabalon-Rodiguez and Thomet-Isla (2009) were Financial support received from Department of Biotechnology, able to correlate geographic distribution, grain colour, Govt. of India vide grant no. BT/PR-8953/BCE/08/533/2007 and panicle colour and phenology of C. quinoa with AFLP- grant no BT/04/NE/2009 under the Biotech Hub programme based marker fragments. Their results clustered the acces- is gratefully acknowledged. RJD gratefully acknowledges the receipt of financial support from Department of Science & Tech- sions into three groups where group-I comprised of eight nology, Govt. of India in the form of a research fellowship under local accessions from northern Chile having yellow to INSPIRE brown grains, yellow but compact to intermediate den- sity panicle, agglomerated form and precocity. Group-II References comprised of highland accessions of later phenology, yel- low to red coloured grains, compact and amarantiform Alice L. A. and Campbell C. S. 1999 Phylogeny of Rubus panicle from precordillera sector and group III included (Rosaceae) based on nuclear ribosomal DNA internal tran- the species C. album and C. ambrosioides which were scribed spacer region sequences. Am. J. Bot. 86, 81–97. denominated as outgroup controls. While the morpho- Anabalon-Rodiguez L. and Thomet-Isla M. 2009 Comparative logical analysis of the Andean cultivars placed them as analysis of genetic and morphologic diversity among quinoa an independent Group (group-II), the AFLP data inte- accessions (Chenopodium quinoa Willd.) of the south of Chile and highland accessions. J. Plant Breed. Crop. Sci. 1, 210–216. grated the Andean cultivars with group-I, indicating the Bhargava A., Shukla S. and Ohri D. 2006 Karyotypic studies on existence of similarity in the genetic material of acces- some cultivated and wild species of Chenopodium (Chenopo- sions collected from the region. This result is consistent diaceae). Genet. Resour. Crop Evol. 53, 1309–1320. Species relationships and evolution in Chenopodium Page 11 of 11 37

Devi R. J. and Chrungoo N. K. 2015 Species relationships in Muller T., Philippi N., Dandekar T, Schltz J. and Wolf M. 2007 Chenopodiumquinoa and Chenopodiumalbum on the basis of Distinguishing species. RNA 13, 1469–1472. morphology and SDS-PAGE profiles of soluble seed proteins. Murray M. G. and Thompson W. F. 1980 Rapid isolation of J. Appl. Biol. Biotech. 3, 29–33. high molecular weight plant DNA. Nucleic Acids Res. 8, Downie S., Katz-Downie D. and Cho K. 1997 Relationships in 4321–4325. the Caryophyllales as suggested by phylogenetic analyses of Palomino G., Hernandez L. T. and Torres E. D. 2008 Nuclear partial chloroplast DNA ORF2280 homolog sequences. Am. genome size and chromosome analysis in Chenopodium J. Bot. 84, 253–273. quinoa and C-berlandieri subsp nuttalliae. Euphytica 164, Fuentes-Bazan S., Mansion G. and Borsch T. 2012a Towards a 221–230. species level tree of the globally diverse genus Chenopodium Rahiminejad M. R. and Gornall R. J. 2004 Flavonoid evidence (Chenopodiaceae). Mol. Phylogenet. Evol. 62, 359–374. for allopolyploidy in the Chenopodium album aggregate (Ama- Fuentes-Bazan S., Uotila P. and Borsch T. 2012b A novel ranthaceae). Plant Syst. Evol. 246, 77–87. phylogeny-based generic classification for Chenopodium sen- Rana T. S., Narzary D. and Ohri D. 2010 Genetic diversity sulato, and a tribal rearrangement of Chenopodioideae and relationships among some wild and cultivated species of (Chenopodiaceae). Willdenowia 42, 5–24. Chenopodium l. (Amaranthaceae) using RAPD and DAMD Gangopadhyay G., Das S. and Mukerjee K. K. 2002 Speciation methods. Curr. Sci. 98, 840–846. in Chenopodium in west Bengal, India. Genet. Resour. Crop Rana T. S., Narzary D. and Ohri D. 2011 Molecular differentia- Evol. 49, 503–510. tion of Chenopodium album complex and some related species Gutell R. R., Larsen N. and Woese C. R. 1994 Lessons from using ISSR profiles and ITS sequences. Gene 495, 29–35. an evolving ribosomal-RNA – 16S and 23S ribosomal-RNA Rohlf F.J. 2000 NTSYS-pc: numerical taxonomy and multivariate structures from a comparative perspective. 2Microbiol Rev. 58, analysis system, version 2.10f. Exeter Publishing, New York. 10–26. Roldan-Ruiz I., Dendauw J., Bockstaele E. V., Depicker A. and Jarvis D. E., Kopp O. R., Jellen E. N., Mallory M. A., Pattee Loose M. D. 2000 AFLP markers reveal high polymorphic J., Bonifacio A. et al. 2008 Simple sequence repeat marker rates in rye grasses (Lolium spp.). Mol. Breed. 6, 125–134. development and genetic mapping in quinoa (Chenopodium Ruas P.M., Bonifacio A., Ruas C. F., Fairbanks D. J. and Ander- quinoa Willd.). J. Genet. 87, 39–51. sen W. R. 1999 Genetic relationships among 19 accessions of Jarvis D. E., Yung S. H, Damien J. L., Sandra M. S., Bo L., six species of Chenopodium l. by random amplified polymor- Theo J. A. B. et al. 2017 The genome of Chenopodium quinoa. phic DNA fragments (RAPD). Euphytica 105, 25–32. Nature 542, 307–312. Schuster T. M., Setaro S. D. and Kron K. A. 2013 Age estimates Joshi B. D. 1991 Genetic resources of leaf and grain Amaranthus for the buckwheat family Polygonaceae based on sequence data and Chenopod. In Biodiversity (ed. M. S. Swam inathan and calibrated by fossils and with a focus on the Amphi-Pacific S. Jana), pp. 121–134. Macmillan India, Chennai. Muehlenbeckia. PLoS One 8, e61261. Kadereit G., Borsch T., Weising K. and Freitag H. 2003 Phy- Singh S. 2010 Understanding the weedy Chenopodium complex logeny of Amaranthaceae and Chenopodiaceae and the evo- in the north central states. Dissertation, University of Illinois lution of C4 photosynthesis. Int. J. Plant Sci. 164, 959–986. at Urbana-Champaign. Kolano B., Plucienniczak A., Kwasniewski M. and Maluszyn- Tamura K., Peterson D., Peterson N., Stecher G., Nei M. and ska J. 2008 Chromosomal localization of a novel repeti- Kumar S. 2011 MEGA5: molecular evolutionary genetics tive sequence in the Chenopodium quinoa genome. J. Appl. analysis using maximum likelihood, evolutionary distance, Genet. 49, 313–320. and maximum parsimony methods. Mol. Biol. Evol. 28, 2731– Kolano B., McCann J., Orzechowska M., Siwinska D., Temsch E. 2739. and Weiss-Schneeweiss H. 2016 Molecular and cytogenetic evi- Vos P.,Hogers R., Bleeker R., Reijans M., Van de Lee T., Hornes dence for an allotetraploid origin of Chenopodium quinoa and M. et al. 1995 AFLP: a new technique for DNA fingerprinting. C. berlandieri (Amaranthaceae). Mol. Phylogenet. Evol. 100, Nucleic Acids Res. 23, 4407–4414. 109–123. Walsh B. M., Adhikary D., Maughan P. J., Emshwiller E. and Kurashige N. S. and Agrawal A. A. 2005 Phenotypic plastic- Jellen E. N. 2015 Chenopodium polyploidy inferences from salt ity to light competition and herbivory in Chenopodium album overly sensitive 1 (sos1)data.Am. J. Bot. 102, 533–543. (Chenopodiaceae). Am. J. Bot. 92, 21–26. White T.J., Bruns T.D., Lee S. and Taylor J.1990 Analysis of phy- La Duke J. and Crawford D. J. 1979 Character compatibility logenetic relationships by amplification and direct sequencing and phyletic relationships in several closely related species of of ribosomal RNA genes. In: PCR protocols: a guide to meth- Chenopodium of the WesternUnited States. Taxon 28, 307–314. ods and applications (ed.M.A.Innis,D.H.Gelfand,J.J. Manhart J. R. and Rettig J. H. 1994 Gene sequence data. In Sninsky and T. J. White), pp. 315–322, Academic Press, New Caryophyllales: evolution and systematics (ed. H.-D. Behnke Yo rk . and T. J. Mabry), pp. 235–246. Springer, Berlin. Wikström, N., Savolainen V. and Chase M. W. 2001 Evolution Mehra P.and Malik C. 1963 Cytology of some Indian Chenopo- of the angiosperms: Calibrating the family tree. Proc.R.Soc. diaceae. Caryologia 16, 67–84. Lond. B, Biol. Sci. 268, 2211–2220. Mukherjee K. K. 1986 A comparative study of two cytotypes Wilson H. D. 1980 Artificial hybridization among species of of Chenopodium album in West Bengal, India. Can. J. Bot. 64, Chenopodium sect. Chenopodium. Sys. Bot. 5, 253–263. 754–759. Wilson H. and Manhart J. 1993 Crop/weed gene flow: Müller K. 2005 Seqstate: primer design and sequence statistics Chenopodium quinoa Willd. and C. berlandieri moq. Theor. for phylogenetic DNA datasets. Appl. Bioinform. 4, 65–69. Appl. Genet. 86, 642–648. Müller K. and Borsch T. 2005 Phylogenetics of Amaranthaceae Wolf M., Achtziger M., Schultz J., Dandekar T. and Mueller T. based on matK/trnK sequence data: evidence from parsimony, 2005 Homology modeling revealed more than 20,000 rRNA likelihood, and Bayesian analyses. Ann. Missouri Bot. Gard. 92, internal transcribed spacer 2 (ITS2) secondary structures. 66–102. RNA 11, 1616–1623.

Corresponding editor: Manoj Prasad