Advertisement

Genetic Resources and Crop Evolution

, Volume 63, Issue 1, pp 79–96 | Cite as

Molecular phylogenetics and systematics of Trigonella L. (Fabaceae) based on nuclear ribosomal ITS and chloroplast trnL intron sequences

  • Rakhee Dangi
  • Shubhada Tamhankar
  • Ritesh Kumar Choudhary
  • Suryaprakasa Rao
Research Article

Abstract

The genus Trigonella includes the widely cultivated T. foenum-graecum and a number of other medicinal and aromatic plant species distributed around the Mediterranean region. Sectional delimitation of Trigonella species is still based on morphology and interspecific relationships within the genus are not yet sufficiently resolved. Phylogenetic relationships in 22 species representing 11 of the 12 sections recognized within the genus Trigonella were analyzed using nuclear ITS and plastid trnL-F DNA sequences. Within nuclear ITS region, ITS-II was found to evolve faster compared to ITS-I. Maximal phylogenetic resolution and support was obtained in the combined analysis of the two selected regions. Trees resulting from maximum parsimony as well as Bayesian inference of combined data provided evidence for Trigonella being monophyletic with high support but did not agree with the traditional subgeneric division. Phylogenetic reconstructions indicated three major lineages supported by apomorphies in inflorescence and stipule. The phylogram supported the sectional delimitation of sections Cylindricae, Vérae, Samaroideae, Pectinatae, Erosae and Callicerates. There is strong support to combine monotypic sections Pectinatae and Erosae into one. However, species belonging to section Foenum-graecum and Falcatulae clustered in different subclades, contrary to their current classification. Inflorescence type appears to be a useful character with taxonomic potential for phenetic implications at subgeneric and sectional level within the genus. Moreover, some legume and seed characteristics, considered important in existing taxonomy, appear to have arisen more than once in Trigonella and are homoplastic.

Keywords

ITS region Molecular phylogeny Trigonella trnL intron trnL-F intergenic spacer 

Introduction

The genus Trigonella L. belongs to the family Fabaceae, subfamily Papilionoideae and tribe Trifolieae. T. foenum-graecum (fenugreek) is the most economically important species of the genus. Traditionally, consumed as fresh vegetable and as spice to add flavor to the Indian cuisines, fenugreek is gaining importance around the world due to its rare medicinal properties (Dangi et al. 2004). Fenugreek contains three important chemical constituents with medicinal value; i.e. steroidal sapogenins; galactomannans and isoleucine. These constituents seem to work in a synergistic way to produce health effects and have placed fenugreek among the most commonly recognized “nutraceutical” or health food products. Because of its high protein content and favorable amino-acid composition, fenugreek seed is considered equal in nutritive value with soy. Wild species of Trigonella, with protein content of the seeds higher than that of T. foenum-graecum, can also be considered as new natural protein sources (Niknam et al. 2004). Besides fenugreek, other species of Trigonella are used as food, medicine and have potential as new species for pasture legume production in phase farming system (Petropoulos 2002; Dangi 2013).

Trigonella contains mostly annual or perennial plants that are often scented and rarely more than 50 cm tall. They are characterized by a campanulate or tubular calyx with two large and three small equal lobes, diadelphous stamens, uniform anthers, terminal stigma and ovary with numerous ovules (Širjaev 1928−1932; Sinskaya 1961; Hutchinson 1964). Legumes of the genus vary greatly in size and are cylindrical or compressed, linear or oblong, straight or curved, indehiscent or dehiscing with a pronounced short or long mucro (beak). The genus is mainly distributed in the Mediterranean region with some species extending to Macaronesia, Central Europe and Asia. The main centre of diversity for some of the species is located in Turkey and adjoining Syria (Dangi et al. 2004; Dangi 2013).

Taxonomically, the tribe Trifolieae includes four genera, Trifolium, Medicago, Melilotus and Trigonella. Species of this group are mainly characterized by the presence of digitately trifoliolate leaves with stipules adnate to the stem, but not encircling it entirely (Small 1987a, b). Trifolieae is a member of a large clade of legumes lacking one copy of the chloroplast inverted repeat (IRLC, Lavin et al. 1990; Liston 1995). Molecular phylogenetic studies have identified a strongly supported monophyletic “vicioid clade” within the IRLC comprising tribes Trifolieae and Fabeae (Liston and Wheeler 1994; Sanderson and Wojciechowski 1996; Wojciechowski et al. 2000, 2004). Phylogenetic analysis within the “vicioid clade” conducted by Steele and Wojciechowski (2003) strongly supported the monophyly of Trifolium but this genus was resolved as a sister lineage to Fabeae (with moderate bootstrap support), making Trifolieae paraphyletic. A more recent phylogenetic analysis within the “vicioid clade” conducted by Ellison et al. (2006) resolved Trifolium as a sister group to Trigonella + Melilotus clade but with a weak support. Thus a close relationship of Trifolium to other genera in Trifolieae is questioned indicating that the phylogenetic relationship of genera among Trifolieae still needs to be tested.

Sequence analysis of a variety of genes and genic regions from both nucleus and chloroplast strongly supported the monophyly of Medicago and Trifolium (Watson et al. 2000; Bena 2001; Steele and Wojciechowski 2003; Wojciechowski et al. 2004; Ellison et al. 2006; Steele et al. 2010). However, in all these phylogenetic studies the monophyly of Trigonella as delimited by Small (1987a, b) was in question. Trigonella was always resolved paraphyletic with respect to Melilotus indicating the need for a critical evaluation of the monophyly of the genus.

The comprehensive monograph of Širjaev (1928–1932) summarizes the extensive taxonomic history of Trigonella and provides detailed descriptions and illustrations of all recognized species. Based on morphological characters he divided the genus into three subgenera and further 15 sections. On the basis of 54 morphological characters, Small (1987a) transferred 23 “medicagoids” Trigonella species to the genus Medicago and delimited the genus into 12 sections recognized by Širjaev (1928–1932, Table 1). In the phylogenetic analysis conducted by Bena (2001) “medicagoid” species, described earlier as transition between Trigonella and Medicago, joined to the Medicago clade with a very good support rather than the Trigonella/Melilotus clade; supporting the morphology based taxonomic transfer of the “medicagoids” Trigonella species to the genus Medicago. Although a number of morphology based taxonomic studies have been conducted in Trigonella (Vasil’chenko 1953; Lashin 2006; Small 1987a, b), Širjaev’s (1928–1932) taxonomic concept has remained unchanged.
Table 1

Details of species used in phylogenetic analysis

No.

Subgenus/section

Species

Origin

SA Numbera

EC Numberb

GenBank Numbers

ITS

trnL-F

 

Subgenus I: Trigonella

1

Falcatulae

T. anguina Delile

n/a

40210

583495

JX274190

JX274134

T. balansae Boiss. et Reut.

Pakistan

32204

583507

JX274193

JX274137

Great Britain

32217

583508

JX274194

JX274138

Turkey

32604

583509

JX274195

JX274139

India

PI216049

546586

JX274132

JX274188

T. maritima Delile ex Poir.

Libya

9841

583600

JX274113

JX274170

Libya

9842

583601

JX274114

JX274171

T. stellata Forssk.

Jordan

34597

583621

JX274127

JX274184

T. suavissima Lindl.

Australia

27514

583623

JX274129

JX274186

Australia

35940

583624

JX274130

JX274187

2

Callicerates

T. calliceras Fisch. ex M. Bieb.

Canada

32202

583570

JX274204

JX274148

3

Uncinatae

T. spicata Sm.

Iraq

14495

583614

JX274124

JX274181

Syria

34561

583616

JX274125

JX274182

4

Cylindricae

T. cylindracea Desv.

Iraq

34528

583578

JX274210

JX274154

Syria

35551

583579

JX274211

JX274155

T. filipes Boiss.

Syria

13459

583580

JX274212

JX274156

Iran

16134

583582

JX274213

JX274157

Turkey

33584

583584

JX274214

JX274158

T. kotschyi Boiss.

Turkey

19732

583597

JX274110

JX274167

Turkey

34544

583598

JX274111

JX274168

Jordan

36833

583599

JX274112

JX274169

T. mesopotamica Hub.-Mor.

Syria

14518

583604

JX274115

JX274172

Iran

14519

583605

JX274116

JX274173

Turkey

16864

583606

JX274117

JX274174

Turkey

32847

583607

JX274118

JX274175

T. strangulata Boiss.

Iraq

14514

583622

JX274128

JX274185

5

Samaroideae

T. cretica (L.) Boiss.

Sweden

5048

583575

JX274207

JX274151

Canada

32208

583576

JX274208

JX274152

Switzerland

32209

583577

JX274209

JX274153

6

Pectinatae

T. arabica Delile

Jordan

5040

583496

JX274191

JX274135

Egypt

34593

583498

JX274192

JX274136

7

Erosae

T. schlumbergeri Boiss.

Jordan

5044

583609

JX274120

JX274177

Turkey

12521

583510

JX274121

JX274178

8

Verae

T. caelesyriaca Boiss.

Syria

5043

583557

JX274197

JX274141

Israel

16848

583558

JX274198

JX274142

Jordan

34612

583565

JX274200

JX274144

T. grandiflora Bunge

Russia

6142

583595

JX274109

JX274166

9

Spinosae

T. spinosa L.

Syria

34549

583620

JX274183

 

Subgenus II: Trifoliastrum

10

Capitatae

T. caerulea (L.) Ser.

Australia

32197

583567

JX274201

JX274145

Turkey

32198

583568

JX274202

JX274146

Spain

32200

583569

JX274203

JX274147

 

Subgenus III: Foenum-graecum

11

Foenum-graecum

T. coerulescens (M. Bieb.) Halácsy

Turkey

14939

583573

JX274205

JX274149

 

Russia

32203

583574

JX274206

JX274150

T. foenum-graecum L.

Iran

35684

583588

JX274215

JX274159

Jordan

35685

583589

JX274216

JX274160

Egypt

35686

583590

JX274217

JX274161

Syria

35687

583591

JX274218

JX274162

Saudi Arabia

37117

583592

JX274219

JX274163

T. gladiata Steven ex M. Bieb.

Spain

32220

583593

JX274107

JX274164

UK

36477

583594

JX274108

JX274165

Medicagoids

M. brachycarpa (Fischer ex M. Bieb.) Moris

Turkey

12543

583556

JX274196

JX274140

M. pamphylica (Huber-Mor. et Širj.) E. Smallc

Turkey

34530

583562

JX274199

JX274143

M. rostrata (Boiss. et Balansa) E. Smallc

Turkey

16846

583611

JX274122

JX274179

M. lunata Rchb.c

Turkey

32224

583612

JX274123

JX274180

M. plicata (Boiss.) Širj.

n/a

5056

583608

JX274119

JX274176

Medicago

M. sativa L.

   

AF053142

DQ311711

M. lupulina L.

   

DQ311980

DQ311708

Melilotus

M. albus Medik.

   

DQ311984

DQ311713

M. officinalis (L.) Pall.

   

DQ311985

DQ311714

Trifolium

T. polyphyllum (C.A. Mey.) Latsch.

   

DQ312135

DQ311885

T. lupinaster L.

   

DQ312083

DQ311827

T. pseudostriatum Baker f.

   

DQ312139

DQ311889

T. acaule A. Rich.

   

DQ311989

DQ311718

n/a not available

aAccession identity number—The Genetic Resource Centre, South Australian Research and Development Institute (SARDI), Waite Research Precinct, Waite Institute, Urrbrae, South Australia

bAccession identity number—National Bureau of Plant Genetic Resources (NBPGR), New Delhi, India

cReclassified

The exact number of species that comprise the genus Trigonella has been debated. Linnaeus suggested as many as 260 species of Trigonella (Petropoulos 2002) whereas about 128 species were reported by Vasil’chenko (1953), 97 by Fazli (1967) and 70 by Hector (1938), Rouk and Mangesha (1963), Hutchinson (1964) and Tutin et al. (1969). Currently the genus comprises 62 species according to pertinent literature (Small 1987a, b).

Despite the increasing use of molecular markers in phylogenetic and systematic studies (Hillis 1995), the sectional delimitation in Trigonella is still based on morphology. Analysis of seed protein electrophoresis profiles in the taxonomy of Trigonella species has indicated the need for reassessment of nomenclature of Trigonella (Niknam et al. 2004). A very recent classification (Çeter et al. 2012) based on seed characteristics did not support the sub generic classification proposed by Širjaev (1928–1932). Since no significant datasets with taxonomic relevance (other than morphological characters) were available in Trigonella, the aim of the present study was to develop a phylogenetic framework for Trigonella and assess its generic affinities in the tribe Trifolieae using regions from both nuclear and chloroplast genome. Internal transcribed spacer (ITS) region of nuclear ribosomal DNA was chosen because it’s utility for examining interspecific relationships in many genera within Fabaceae (Wojciechowski 2003) and especially the closely related Medicago and Trifolium (Bena 2001; Ellison et al. 2006) has been well documented. The chloroplast trnL-F region was selected because these are universally useful markers for application in a broad spectrum of phylogenetic questions (Gielly and Taberlet 1994; Kores et al. 2001). The major objectives were to test the monophyly of the genus and to assess phylogenetic relationships within and between sections and subsections in view of Širjaev’s systematic treatment (1928–1932). Further, inflorescence, pod and seed character evolution were reconsidered in the light of molecular data.

Materials and methods

Plant material

Fifty-five accessions representing 22 Trigonella species (Table 1) and 2 Medicago species (medicagoids—M. plicata and M. brachycarpa) were procured from South Australian Research and Development Institute’s (SARDI) Australian Medicago Genetic Resource Centre (AMGRC), Urrbrae, South Australia through National Bureau of Plant Genetic Resources (NBPGR), New Delhi, India. The details of each accession along with their original country of collection are listed in Table 1. The Trigonella species represented 11 of the 12 sections of the genus recognized (Širjaev 1928−1932; Small 1987a, b). Species belonging to section Ellipticae could not be sampled. Within Trigonella, six species (T. anguina, T. calliceras, T. grandiflora, T. spinosa, T. stellata and T. strangulata) were represented by a single accession each while remaining 16 species had multiple accessions from different geographic areas (Table 1). Prior to phylogenetic analysis, species identification was confirmed using the relevant literature and type specimens (Širjaev 1928−1932; Small 1987a, b). Three misidentified accessions (EC 583562, EC 583611 and EC 583612) were reclassified using the “key to medicagoid species of Trigonella” specified by Small (1987a), as Medicago pamphylica, M. rostrata and M. lunata respectively (Dangi 2013). The final data set considered for phylogenetic analysis consisted of total 55 accessions representing 22 Trigonella and 5 Medicago (medicagoids—M. plicata, M. brachycarpa, M. pamphylica, M. lunata and M. rostrata) species.

DNA extraction, amplification and sequencing

Total genomic DNA was extracted from one gram of young leaf tissue using CTAB method (Rogers and Bendich 1985). The entire nuclear ribosomal DNA spacer region including ITS-I, ITS-II and the 5.8S cistron was amplified by PCR with primers ITS-F (5′CGTAACAAGGTTTCCGTAGGTGAACC3′) and ITS-R (5′TTATTGATATGCTTAAACTCAGCGGG3′) which differed ITS1 and ITS4 of White et al. (1990) by a few bases. The amplifications were carried out in GeneAmp PCR system 9700 (Applied Biosystems, Foster City, CA, USA). The thermal cycling conditions were as follows: Initial denaturation of 4 min at 94 °C, 35 cycles of 30 s denaturation (94 °C), 30 s annealing (50 °C) and 1 min 30 s elongation (72 °C) followed by final extension of 5 min at 72 °C.

The trnL-F region was amplified either as a single fragment using the primers c/f or in two shorter fragments using the primers c/d and e/f of Taberlet et al. (1991). The thermal cycling conditions were as follows: Initial denaturation of 3 min at 94 °C, 35 cycles of 30 s denaturation (94 °C), 45 s annealing (50 °C), 1 min 30 s elongation (72 °C) followed by final extension of 5 min at 72 °C. Amplified products were purified by polyethylene glycol (PEG) precipitation (Sambrook and Russell 2001) for sequencing. Sequencing reactions were carried out using the Big Dye®Terminator v3.1 Cycle sequencing Kit (Applied Biosystems, Foster City, CA, USA) as per the manufacturer’s protocol. Both ITS and trnL-F regions were sequenced bi directionally using the same primer pair used in amplification. For ITS region, additional internal primers ITS 2 and ITS 3 (White et al. 1990) were used in some accessions. Sequencing reactions were purified by ethanol/EDTA precipitation according to manufacturer’s protocol. Dried pellets were suspended in 10 µL of Hi-Di formamide and run on ABI 3100 Avant Genetic Analyzer as per the recommended protocol. Sequences of taxa from allied genera in Trifolieae namely Melilotus, Medicago and Trifolium were retrieved from GenBank (Table 1). Since most of the retrieved sequences were available only for trnL intron region, for combined phylogenetic analysis only trnL intron region was considered.

Phylogenetic analysis

Sequences were aligned using Clustal W with default settings (Thompson et al. 1994). Maximum parsimony and Bayesian analyses were performed on separate and combined data sets using PAUP* 4.0 Beta version 10 and MrBayes 3.0b4 respectively (Swofford 2001; Ronquist and Huelsenbeck 2003). Only maximum parsimony analysis was performed for separate ITS-I and ITS-II datasets. Gaps were treated as missing data. All parsimony analyses were simultaneous and unconstrained (Nixon and Carpenter 1996) with character state changes unordered and weighed equally. Analysis was conducted using an initial heuristic search comprising 1000 replicates of random stepwise addition using tree bisection and reconnection (TBR) branch swapping with MULTREES option on, but saving only one tree per replicate. Multiple most parsimonious trees resulting from this analysis were used to compute a strict consensus tree, which was then used as a constraint for another round of heuristic searches. The consistency index (CI, Kluge and Farris 1969) and retention index (RI, Farris 1989) were calculated. The robustness of the clade in the strict consensus tree was evaluated by non-parametric bootstrap analysis (Felsenstein 1985) and by computing decay values (Bremer 1994). The following general descriptions for categories of bootstrap support were used: poor <50 %, weak, 50–74 %, moderate, 75–84 %, strong, 85–100 % (Chase et al. 2000). Decay indices (DI) were obtained using the programs PAUP* and TreeRot.v2 (Sorenson 1999) and clades having DI greater than or equal to 4 were considered well supported (Marcilla et al. 2001).

Possible conflicts between the ITS and trnL intron data sets were evaluated with an incongruence length difference test (ILD) (Farris et al. 1994, 1995) prior to combining the data. This test, implemented as the partition homogeneity test in PAUP* (Swofford 2001), determines whether the original data partitions differ significantly from randomly shuffled partitions of the combined data sets. One hundred replicates were performed on parsimony informative characters using TBR branch swapping, simple sequence addition, MULTREE on, Steepest Descent in effect, and MaxTrees set at 100. The trees obtained for each gene region were also examined for “hard” or “soft” incongruences (Seelanan et al. 1997) and the data was combined following the suggestion of Liu and Miyamoto (1999).

Bayesian analysis of the separate and the combined data sets were conducted with MrBayes 3.0b4 (Huelsenbeck et al. 2002; Ronquist and Huelsenbeck 2003). The best fit model of sequence evolution was chosen using the hierarchical likelihood ratio test (hLRT) and alkali information criterion (AIC), calculated with Mr. Model Test 2.3 (Posoda and Crandall 1998). These models were applied to their respective partitions in the separate and combined analysis. In each analysis, a single run of 3,000,000 generations were conducted. In each run trees were sampled every 100 generation and burn-in was determined by inspection of the log-likelihoods of the sample trees. Branch length information was recorded and averaged across all retained trees, and majority rule consensus tree were computed to obtain posterior probabilities (PP). Clades with >0.90 posterior probabilities were considered strongly supported.

Results

Sequence characteristics

A total of 55 accessions, including 50 representing 22 Trigonella species and 5 “medicagoids” were sequenced for both ITS and trnL-F regions and nucleotide sequences have been deposited in GenBank (Table 1). For ITS region, length varied from 609 to 730 bp (including the out-group) while that in trnL intron and trnL-F regions the length varied from 150 to 288 bp and 341 to 514 bp respectively. Clean ITS sequence could not be obtained for T. spinosa.

The intra specific sequence variation between accessions of species used in the present study was very low, ranging from no variation for most species to 1.1 % in T. arabica with the exception of T. cylindracea which showed a sequence variation of 2.5 % between the two accessions (Online resource Table 1). Within ITS region, the low levels of intra specific difference observed was due to point mutations rather than length variation and was associated with ITS-II region. For trnL-F sequences also very low intraspecific sequence variation was observed ranging from no variation for most species to 1.1 % in T. cylindracea. Whenever present, intra species nucleotide polymorphism was associated with length variation resulting from insertions/deletions within the trnL-F region. All sequences obtained were included in a series of parsimony analysis. Multiple accessions of a given species clustered together, and therefore, a single accession was selected for inclusion in the phylogenetic analysis (Online Resource Fig. 1). The two accessions of T. cylindracea, although clustered in the same clade showed divergent ITS and trnL-F sequences. Since a single accession could not be determined as a “representative” of the species, both the accessions were included in the phylogenetic analysis.

Within Trigonella, (excluding the out-group) the ITS-I region varied from 248 for T. caerulea to 233 for T. maritima. Out of the 237 aligned characters 45 were variable and parsimony uninformative and 64 were potentially informative for parsimony analysis. ITS-II varied from 226 for T. caerulea to 218 for T. stellata, T. caelesyriaca, T. maritima and T. suavissima. Of the 220 aligned characters 27 were variable and 60 were potentially informative for parsimony analysis. ITS-I was longer as compared to ITS-II. The sequence statistics for each dataset are summarized in Table 2. The complete ITS region (excluding the out-group) varied from 609 to 730 bp. For the final alignment (including the out-groups), 94 bps corresponding to regions of ambiguous alignment were removed from the data set. This resulted in a final data set of 674 aligned characters of which 207 were variable but uninformative and 127 were potentially parsimony informative. The mean GC content of the ITS region was 48.7 %. The trnL intron region (out-groups included) varied in length from 341 to 514. The aligned sequences however, included only 230 sites, of which 11 were variable but uninformative and seven were potentially parsimony informative. The mean GC content of the trnL intron region was 35.4 %. The trnL intron data set was characterized by numerous indels.
Table 2

Sequence statistics for separate and combined, ITS (internal transcribed spacer), chloroplast trnL intron and combined (ITS + trnL) data sets used in the phylogenetic analysis

Description

ITS

trnL intron

ITS + trnL

Number of taxa included- ingroup

22

23

22

Number of taxa included- outgroup included

35

36

33

Length range (bp)- ingroup

609–730

341–514

1307–1421

Aligned length (bp)- ingroup

698

311

1166

Aligned length (bp)- outgroup included

674

230

997

G + C content mean (%)- ingroup

48.6

35.7

40.0

G + C content mean (%)- outgroup included

48.7

35.1

42.1

Parsimony uninformative sites- ingroup

87

19

123

Parsimony uninformative sites- outgroup included

207

11

104

Potentially informative characters- ingroup

79

5

91

Potentially informative characters- outgroup included

127

7

220

CI of MPTs

0.69

0.95

0.77

RI of MPTs

0.83

0.93

0.83

Number of MPTs

112

28

108

Length of MPTs

383

22

511

CI Consistency Index, RI Retention Index, MPTs Maximum Parsimonious Tree

The combined ITS and trnL intron sequences included 862 aligned sites among the 35 taxa. Of these, 96 were variable but uninformative and 157 were potentially parsimony informative. The mean GC content was 42.1 %. For the combined ITS + trnL sequence data without Melilotus, of the 997 aligned sites among the 33 terminal taxa 104 were variable but uninformative and 220 were potentially parsimony informative. The mean GC content was 42.1 %. The number of parsimony informative characters was higher for the combined analysis without Melilotus species.

Phylogenetic analysis

Separate analyses

Parsimony analysis of the ITS data set resulted in 112 equally parsimonious trees of length 383 (CI = 0.694, RI = 0.830, HI = 0.305). The strict consensus of these trees with bootstrap (BP) and decay values (DI) is presented in Fig. 1. The Bayesian analysis of the ITS data set using an AIC selected SYM + G substitution model generated trees with a topology highly similar to that produced with MP analysis.
Fig. 1

Phylogenetic relationship in Trigonella based on the maximum parsimony analysis of ITS sequences. Numbers along the branch indicate bootstrap percentage above 50 % followed by the decay index. Numbers below the branch indicate Bayesian posterior probabilities. Sections of Trigonella according to Small (1987a, b) are indicated on the right

The Bayesian posterior probability values (PP) and the MP bootstrap values (BP) were well correlated with the PP values consistently higher. In the MP analysis ten clades had BP ≥ 90 %, DI ≥ 4 while one clade had BP ≥ 70 %, DI ≥ 1. Analysis of the ITS sequence data confirmed the monophyly of Medicago and Trifolium and the clustering of “medicagoids” with the monophyletic Medicago with a BP value of 99 % and DI of 7 (Bena 2001). The two species of Melilotus formed a part of a basal polytome within the clade of Trigonella species (Fig. 1).

Within Trigonella, section Falcatulae was rendered paraphyletic by the position of the strongly supported T. balansae and T. anguina cluster (BP = 95 %, DI = 4) outside a strongly supported clade comprising the remaining three representatives of this section. T. maritima, T. stellata and T. suavissima (BP = 95 %, DI = 5). Section Cylindricae was monophyletic with a moderate support (BP = 68 %, DI = 2) with a sister taxon relationship of T. spicata (section Uncinatae) with T. strangulata weakly resolved. There was a strong bootstrap support for T. arabica (section Pectinatae, BP = 94 %, DI = 4) as a sister group to T. schlumbergeri (section Erosae) and to the strongly supported clade comprising of T. maritima, T. stellata and T. suavissima. In the tree depicted (Fig. 1) T. foenum-graecum was sister to clade comprising T. gladiata (section Foenum-graecum), with a weak support. T. coerulescens, the other representative from section Foenum-graecum was sister to T. caerulea (BP = 53 %). Despite the lack of resolution in this part of the tree, results indicated that section Foenum-graecum is paraphyletic. Remaining species within Trigonella were largely unresolved with little BP support.

Parsimony analysis of the trnL intron data set resulted in 28 equally parsimonious trees of length 22 (Online Resource Fig. 2, CI = 0.95, RI = 0.93, HI = 0.07). Due to low number of parsimony informative characters, the trnL intron data set resulted in trees that are poorly resolved and weakly supported.

Combined analyses

Based on the ILD test, the two partitions were significantly different (p = 0.01). Removal of non Trigonella sequences, where several topological discrepancies were observed, still resulted in significantly different partitions (p = 0.01). Incongruences between ITS and trnL intron data have been reported previously in sub tribe Trigonellinae (Ellison et al. 2006). It has been reported earlier that lack of resolution should not be interpreted as lack of evidence for combining data; however it may simply be evidence of insufficient information and signals (Cunningham 1997a, b). The latter may be the case with the trnL intron data set, in which there is an obvious deficit of discrete characters suitable for parsimony analysis. Moreover, comparison of the poorly resolved trnL intron with ITS tree did not show strongly supported topological incongruence, hence ITS and trnL intron data sets were analyzed simultaneously.

Parsimony analysis of the combined data set (ITS + trnL) resulted in 328 equally parsimonious trees of length 448 (CI = 0.70, RI = 0.820, HI = 0.381, Online Resource Fig. 3). In the MP analysis 14 clades have a BP ≥ 90 %, DI ≥ 4 while three clades have BP ≥ 70 %, DI ≥ 1. Apart from the position of T. cretica (which is poorly supported difference) the MP strict consensus tree and the Bayesian tree were consistent. Combined analysis resolved Trifolium and Trigonella + Melilotus of Trifolieae as sister groups. The monophyly of Trigonella and Melilotus resolved in the combined analysis was not resolved by the separate ITS and trnL intron data sets.

Parsimony analysis of the combined ITS + trnL intron (without Melilotus) resulted in 108 equally parsimonious trees of length 511 (CI = 0.77, RI = 0.873, HI = 0.291). The strict consensus of these trees is presented in Fig. 2 with BP and DI. The Bayesian analysis of the combined data set using an AIC selected GTR + G+I substitution model for both the regions generated trees with a topology highly similar to that produced with MP. The topology of the Bayesian tree was in conflict with the strict consensus tree for MP analysis of the combined data with respect to the position of T. spicata and T. strangulata. Apart from the position of these species in the parsimony strict consensus tree, the MP strict consensus tree and the Bayesian tree were consistent.
Fig. 2

Phylogenetic relationships in Trigonella based on maximum parsimony analysis of ITS + trnL intron sequences. Numbers along the branch indicate bootstrap percentage above 50 % followed by the decay index. Numbers below the branch indicate Bayesian posterior probabilities. Sections of Trigonella according to Small (1987a, b) are indicated on the right

Despite the conflicting signals in the two data sets, the combined analysis was better resolved and the BP support increased for some of the clades. The ITS sequences made a much greater contribution than the trnL intron sequences. The topology of the combined analysis indicated that Trigonella is monophyletic and consisted of three major lineages referred to as Clade I and II and III.
  • Clade I: This clade contained 5 subclades a-e. Section Falcatulae was rendered paraphyletic by the position of the strongly supported T. balansae and T. anguina cluster (subclade I-a, BP = 98, DI = 4) outside a strongly supported subclade (BP = 93 %, DI = 5) comprised of the remaining three representatives of this section, T. maritima, T, stellata and T. suavissima. Unlike ITS, in the combined data set T. foenum-graecum formed a strongly supported clade (subclade I-b BP = 85 %, DI = 2) with T. gladiata (section Foenum-graecum). However, section Foenum-graecum was again rendered paraphyletic by the position of T. coerulescens outside the clade comprising of the remaining two representatives of this section. Like ITS, sectional delimitation of section Cylindricae was well supported (subclade I-c BP 99 %, DI 10) with a sister taxon relationship of T. spicata and T. strangulata weakly supported (subclade I-d). The relationship of T. coerulescens with T. caerulea (subclade I-e) was weakly supported.

  • Clade II: This clade moderately supported the monophyly of section Vérae with a sister taxon relationship with section Callicerates (T. calliceras).

  • Clade III: This well supported clade contained T. arabica (section Pectinatae) and T. schlumbergeri forming a strongly supported subclade (III-a, BP = 99 %, DI = 4) with a sister group relationship with T. maritima, T. stellata and T. suavissima (III-b)

Discussion

Intraspecific sequence variation

The ITS sequences obtained in the present study were compared with the previous reports in T. foenum-graecum, T. caerulea, T. cretica, T. arabica, T. calliceras, T. stellata, T. spicata and T. kotschyi (Bena 2001; Kakani et al. 2011) for nucleotide variation, if any. The nucleotide variation between the accessions of the same species was very low and ranged from no variation for T. foenum-graecum and T. cretica to three nucleotides for T. arabica. In the present analysis, T. stellata and T. calliceras were represented by a single accession. In these species also, when compared with reported accessions, very low nucleotide variation was observed and this supported the inclusion of a “representative” of these species. The intra specific sequence variation did not affect the overall phylogenetic position of that species. Whenever present, intraspecific sequence variation was observed in ITS-II, suggesting that in Trigonella, ITS-II changes faster as compared to ITS-I.

Monophyly of Trigonella and Melilotus

Various phylogenetic studies in tribe Trifolieae have reported that Trigonella is paraphyletic with regard to Melilotus. Phylogenetic analysis of tribe Trifolieae and Fabeae based on the sequence of matK gene revealed that Medicago and Trigonella are sister taxa but Melilotus was nested with Trigonella (Steele and Wojciechowski 2003). The ITS + ETS combined data positioned Melilotus as a sister group within the Trigonella clade (Bena 2001). In the nuclear GA3ox1 sequence analysis Melilotus species formed a basal polytome within the clade of all Trigonella species (Steele et al. 2010). The trnK/matK analysis also placed Melilotus species as a weakly supported group within Trigonella clade (Steele et al. 2010). In morphological tree also Melilotus was nested within Trigonella and Trifolium was basal in the tree (Gazara et al. 2001). Similar to these reports, in the present study too, in separate ITS analysis Melilotus was nested within the Trigonella clade. However, combined ITS and trnL data set resolved the monophyly of Trigonella and Melilotus with a high BP support (Online resource Fig. 3). Generalized morphological distances of the genera of Trifolieae based on 54 morphological characters (Small 1987a, b) showed that Trigonella is somewhat intermediate between Medicago and Melilotus. Although sampling within Melilotus is limited in the present study, results strongly indicate that the closest relative of Trigonella is Melilotus. The close relationship between the two genera is also supported by a number of morphological characters viz. incised stipules margin, notched apex of standard, style longer than ovary and smooth surface of seed coat. Moreover, some species of Trigonella and nearly all species of Melilotus release coumarins upon maceration of leaf tissue while species of Medicago and Trifolium are coumarin negative (Ingham 1981). Similarity in pollen grain morphology of Trigonella and Melilotus further supports their position together (Lashin 2006).

Phylogeny of tribe Trifolieae

Taxonomically, Medicago along with Melilotus (sweet clovers) and Trigonella were included in the tribe Trigonellinae, first recognized by Schultz (1901), but as circumscribed this tribe was not accepted by most taxonomists. Instead, most authors recognized the tribe Trifolieae, which included Trifolium along with these three genera (Rechinger 1984). The monophyly of Trifolium was strongly supported in the matK analysis (Steele and Wojciechowski 2003) and was apparent in the supertree which incorporated nrDNA ITS results from many Old World (Watson et al. 2000) and New World (Liston et al. 2001) species. Surprisingly the genus was resolved (with moderate bootstrap support) as a sister lineage to the Fabeae, making Trifolieae paraphyletic although the position was only weekly supported (Wojciechowski et al. 2000, 2004). Based on these results a close relationship of Trifolium to other genera in Trifolieae was questioned. A more recent phylogenetic analysis among the genera of the “vicioid clade” (Ellison et al. 2006) using combined nrDNA ITS and trnL resolved Trifolium and Trigonella + Melilotus of Trifolieae as sister groups with weak support. Combined analysis of nrDNA ITS and trnL intron in the present study also resolved Trifolium and Trigonella + Melilotus of Trifolieae as sister groups with a high bootstrap support (BP 95 %, DI 8, Online resource Fig. 3) which was in agreement with the traditional classification (Heyn 1981) and also supported the placement of Trifolium within Trifolieae as suggested by Ellison et al. (2006).

Medicago and Trigonella, as delimited by Small and Jomphe (1989) have been strongly supported as sister genera by nrDNA ITS and the flanking external transcribed spacer region (Bena 2001) as well as the plastid-encoded matK gene analyses (Steele and Wojciechowski 2003; Wojciechowski et al. 2004). However, in the present combined ITS + trnL intron analysis Medicago is resolved as a sister group to Trifolium + Trigonella clade. These results are in accordance with study of Ellison et al. (2006) where Medicago was resolved as a sister group to Trifolium and Trigonella + Melilotus. Although sampling within Medicago and Trifolium is limited in the present study, results indicate that Trifolium is the closest relative of Trigonella after Melilotus.

Phylogenetic relationships and classification of the genus

The phylogeny derived from the combined data sets provided strong support for the monophyly of the genus Trigonella as delimited by Small (1987a, b). The Trigonella species studied here represent 10 of the 12 recognized sections. Clean ITS sequence could not be obtained for T. spinosa and hence this species representing the monotypic section Spinosae could not be included in ITS and the combined analysis. Species belonging to section Ellipticae could not be included because of unavailability of the fresh as well as herbarium material which are rare to find (Steele and Wojciechowski 2003).

In the separate ITS and combined analyses, section Falcatulae appeared paraphyletic. Širjaev (1928–1932) divided this section into two subsections: Leves (smooth seeds) and Tuberculatae (tuberculate seeds). Subsection Leves was further divided into four series—Hamosae (T. hamosa, T. falcata, T. media and T. uncata), Stellatae (T. stellata and T. maritima), Anguinae (T. anguina and T. suavissima) and Lacinatae (T. laciniata and T. occulta). Subsection Tuberculatae includes two species, T. corniculata and T. balansae. Species belonging to section Falcatulae formed two separate well supported clades. The first clade clusters T. balansae and T. anguina together with a strong support (BP = 99 %, DI = 4). The presence of shared chloroplast haplotypes also confirmed the very close relationship between these species (Dangi 2013). The second clade clusters T. maritima, T. stellata and T. suavissima with strong support (BP = 95 %, DI = 5) with morphological feature like umbellate inflorescence synapomorphic for this clade.

The close relationship of T. anguina and T. suavissima suggested by Širjaev (1928–1932) is not supported by the present phylogeny. T. anguina and T. suavissima were placed in series Anguinae due to presence of linear legumes which are plicated. However, the plicated nature of the pods is a polymorphic character state in T. suavissima as agreed by Širjaev (1928–1932) and also revealed by the different accessions (EC 583623 with non-plicated legumes and EC 583624 with plicated legumes) used in the present analysis. Apparently this trait is not of much importance in determining phylogenetic relationships. T. anguina is atypical when compared with other species of section Falcatulae because it possesses an umbellate sessile inflorescence. It has an ascending to erect shrubby habit with a considerably hardened stem at the base. In addition to above mentioned synapomorphic character for clade I-a, T. suavissima differs from T. anguina in possessing an ascending to prostrate herbaceous habit (Širjaev 1928−1932). Moreover, the two species have drastically different distribution (T. suavissima in salty and grassy plains of South Australia and T. anguina in desert and semi-desert regions of South Africa with clayey soil (Širjaev 1928−1932).

One of the important morphological traits used by Širjaev (1928–1932) for subsectional delimitation of section Falcatulae is the smooth or tuberculate nature of seed surface. In spite of having smooth seeds, T. suavissima and T. anguina clustered at different positions in the phylogram pointing out that this feature may have arisen in parallel and might lack taxonomic significance. Investigations by Çeter et al. (2012) have revealed that although the great variation in seed sculptures, shapes and colours is useful in species delimitation in Trigonella, unique seed ornamentation and combination of other seed characteristics did not provide considerable information that could be used to distinguish sections or subsections of this genus.

T. corniculata and T. balansae which belong to subsection Tuberculatae are morphologically similar but readily distinguishable by the shape of capitulum and acute/obtuse legume apex. These species are more closely related to species of section Ellipticae in having somewhat deeper roots, with an erect to ascending shrubby habit with considerably hardened and thicker stem along with other morphological features related to the inflorescence, flowers, legumes and legume striations (Širjaev 1928–1932). Section Ellipticae includes all the known perennial species of Trigonella and if T. balansae was a perennial it would have been placed in section Ellipticae by Širjaev (1928–1932). Both molecular data and morphology suggest that species of sections Falcatulae belong to different strongly supported subclades and its current circumscription should be reconsidered. Although remaining species of section Falcatulae not sampled here share morphological features synapomorphic for species of clade I-a, a redefinition of taxonomic status of section Falcatulae would require further in depth molecular and morphological analyses of all the species.

With a moderate support the circumscription of section Vérae was in agreement with the morphological similarities between T. grandiflora and T. caelesyriaca (Širjaev 1928−1932; Small 1987a, b). The monotypic section Callicerates (T. calliceras) was a sister taxon to species of section Vérae. Phylogenetic analysis along with morphological features like the presence of semisagittate stipules with dentate base and entire upper part, umbellate inflorescence, distinct legume sutures and longitudinal/obliquely longitudinal venation shared by species of clade II (Fig. 2) prove that section Callicerates is more closely related to section Vérae and not to section Cylindricae which are characterized by lanceolate/triangular oblong, always entire stipules, capitulum inflorescence, indistinct legume suture and longitudinal venation or section Foenum-graecum where the stipules are ovate/semi cone shaped, entire rarely dentate, sessile flowers, capitulum inflorescence (if present), legume sutures thickened at both poles and legume venation longitudinal/obliquely longitudinal (Širjaev 1928−1932).

Five of the ten species reported in section Cylindricae are sampled (Small 1987a, b). In the combined analysis species belonging to subsection Boissierianae characterized by non-strangulated seeds (Fig. 2, subclade I-c, BP = 99 %, DI = 10) form a strongly supported cluster with a sister taxon relationship with subsection Strangulatae (seeds strangulated) supported by a high Bayesian PP value. The clear synapomorphic characters that the species of these subsections as well as the remaining species of section Cylindricae not sampled here share provide adequate justification to support the sectional delimitation of section Cylindricae.

T. spicata and T. cephalotes are two species of section Uncinatae. Principal coordinate analysis by Small (1987a, b) showed that Trigonella section Uncinatae was close to Melilotus, as postulated by Širjaev (1928–1932). However, this view is not supported by the present phylogeny. Separate ITS analysis placed T. spicata with the core species of clade I-c (Fig. 2) with a moderate support (Fig. 1, BP = 68 %, DI = 2) while the combined analysis supports this relationship only by a high Bayesian PP value. Section Uncinatae has a number of unique features indicating the need for a more extensive sampling of molecular sequence data for a better phylogenetic resolution.

T. cretica and T. graeca from section Samaroideae were assigned to Melilotus by Lassen (in Greuter and Raus 1987). In cluster analysis by Small (1987a, b), this section was intermediate between the two genera, while in the principal coordinate analysis it was closer to Trigonella than to Melilotus. In the maximum parsimony analysis of trnK/matK sequence data T. cretica was sister to the clade comprising four Melilotus species, M. alba, M. segetalis, M. sulcatus and M. indica. The present analysis supports the sectional delimitation of section Samaroideae and provides a strong support for its inclusion within Trigonella.

Section Capitatae includes three species namely T. besseriana, T. capitata and T. caerulea characterized by dentate stipules, dense capitulum inflorescence, violet corolla, small lanceolate legumes with thin beak, 1–2(3) ovules and thinly tuberculate seeds. A close affinity to Melilotus, once proposed for section Capitatae by Širjaev (1928–1932) is not supported by the present phylogeny. T. caerulea is sister to T. coerulescens (section Foenum-graecum) in ITS and the combined data. However, this relationship is only weekly supported in both ITS (Fig. 1, BP = 53 %, DI = 1, PP = 0.53) and the combined analysis (Fig. 2, BP = 39 %, DI = 1, PP = 0.52). According to Širjaev (1928–1932), section Foenum-graecum of the genus Trigonella includes eight species arranged in two subsections—Biebersteinianae and Gladiatae. The first subsection includes a single species, T. coerulescens which by its inflorescence, pod and seed morphology seemed to be more related to section Capitatae than to section Foenum-graecum (Širjaev 1928−1932). Although a better phylogenetic resolution at this taxonomic level is required, the phylogeny inferred from this study supports a close relationship between section Capitatae and section Foenum-graecum subsection Biebersteinianae as postulated by Širjaev (1928–1932). However, additional molecular data would be required to conclude whether T. coerulescens should be placed in a new monotypic section within Trigonella or included in section Capitatae.

T. arabica, placed in monotypic section Pectinatae, clustered with a very strong support (Fig. 2, BP = 100 %, DI = 6, PP = 1.00) with T. schlumbergeri (monotypic section Erosae, Clade III-a) indicating closer relationship between the two species and thereby the sections. Numerical taxonomic studies based on morphological characters placed T. arabica and T. schlumbergeri in the same clade (Small 1987a, b) in agreement with the classical monographer, Širjaev (1928–1932). However, Širjaev (1928–1932) and Small (1987a, b) placed these species in separate sections due to differences in legume characteristics. Legumes in T. arabica are oblong and papery with spines on both sutures while in T. schlumbergeri legumes are semi-ovate, papery with winged margin. Flattened and papery pods with thinly tuberculate seeds are synapomorphic characters for clade III-a while the spiny legumes in T. arabica and winged legumes in T. schlumbergeri are apomorphic traits.

Results show that as currently circumscribed section Foenum-graecum is paraphyletic. Section Foenum-graecum subsection Biebersteinianae as circumscribed by Širjaev (1928–1932) is characterized by dense capitulum with long peduncle; bracteolate flowers; long lanceolate and suddenly narrowed legumes with a thin beak, 4–6 seeds; villous, ovate/subtubulate and dentate stipules. On the other hand species belonging to subsection Gladiatae possess flowers that are sessile and ebracteolate, solitary or in groups (up to 4); relatively long flat legumes that contain several seeds, ± a long beak; small, semi-cone shaped and ± hairy stipules. Distinct morphological differences and the separate position of subsection Biebersteinianae and Gladiatae in different regions of the phylogram indicate that the two subsections of section Foenum-graecum should be treated separately.

Seven species in section Foenum-graecum subsection Gladiatae are organized into two series—Compressae (T. gladiata) and Teretes (T. foenum-graecum, T. berythea, T. macrorryncha, T. cariensis, T. cassia and T. raphanina, Širjaev 1928−1932). Of these T. berythea, considered much closer to cultivated T. foenum-graecum and T. macrorryncha are apparently endemic only to South East Turkey (Huber-Morath 1970) while T. cariensis, T. cassia and T. raphanina are known only from herbarium material (Ladizinsky 1979). T. gladiata has the widest distribution among the wild forms of section Foenum-graecum (Ladizinsky 1979). In the separate ITS and trnL data, the sister taxa relationship between T. gladiata and T. foenum-graecum is weakly or moderately supported. However, in the combined analysis, both species clustered with a high BP support (85 %, PP = 1.00). Close relationship between these species was also supported by chloroplast haplotype data revealing a common maternal ancestral linkage (Dangi 2013). Although the remaining species of series Teretes were not sampled, the morphological features that these species share with T. foenum-graecum provide strong support for their placement in clade I-b.

Evolution of inflorescence structure

Molecular data has proven indispensable to sort out the complex phylogenetic relationship in Trigonella but structural characters are needed to support the recircumscription of the taxa into recognizable groups. Inflorescence appears to be an important character with taxonomic potential in Trigonella useful for defining and recognizing taxa at various ranks within the genus. The phylogram suggests that Trigonella is separated into two groups, species of clade II and clade III with umbellate inflorescence and clade I with capitulum inflorescence. Species of clade II and III separated first from the remaining Trigonella species indicating that umbellate inflorescence appears to be a basic structure from which the more evolved capitulum inflorescence originated. Unlike other species of Trigonella, T. anguina is atypical because it possesses a sessile sub-umbellate inflorescence. Included in clade I-a, T. anguina can be considered a transitional species in view of the hypothesis that heads (Inflorescence type in T. balansae) evolved from umbel by the suppression of pedicels (Parkin 1914; Stebbins 1974; Harris 1999). The solitary flowers of Foenum-graecum subsection Gladiatae may have evolved several times from different kinds of inflorescence (Endress and Doyle 2009). However, a detailed comparative study of inflorescence in all Trigonella species combined with phylogeny is needed to prove the hypothesis that umbels and heads evolved from branched inflorescences (in ancestral Trigonella) by the suppression of inflorescence branches to form umbels and by suppression of pedicels in umbellate forms to produce heads (Feng et al. 2011; Harris 1999).

Evolution of pod character

In distinguishing Trigonella species legume characteristics are considered of great taxonomic value. However, in several instances the classification key based on pod characters fails to group phylogenetically related taxa. For example, recognition of sections Uncinatae, Samaroideae and Capitatae and their close proximity to Melilotus was in part attributed to the presence of one or two seeded indehiscent fruits, a character also involved in the delimitation of genus Melilotus (Širjaev 1928−1932). However, the species of section Uncinatae, Samaroideae and Capitatae were found clustered at different positions of the phylogram. Moreover, the close proximity to Melilotus reported for these species is also not supported by the present molecular data. Phylogenetic analysis suggests that legume characters like these are likely to be homoplastic in the genus having occurred independently in more than one lineage within Trigonella and Melilotus. Species of clade III-b possess a typical Trigonella pod (linear or oblong and curved, Small 1987a, b) which in part led to their placement in section Falcatulae along with T. balansae and T. anguina. The separate clustering of T. maritima, T. stellata and T. suavissima and T. balansae and T. anguina indicate that the similarity in the shape of the pods in these species is due to parallel evolution. Although morphologically similar, both T. arabica and T. schlumbergeri were placed in separate sections on the basis of the different morphological features of pods (Small 1987a, b). Although legume characters were drastically different, T. arabica, T. schlumbergeri, T. maritima, T. stellata and T. suavissima form a strongly supported clade indicating that they share a common ancestral linkage. Results suggest that although pod characters have been useful to assess infrageneric relationships in Trigonella these characters alone are not sufficient to define the sectional circumscription in Trigonella.

Taxonomic implications

Molecular data has allowed a greater resolution of relationships in Trigonella on the basis of some well supported clades obtained in the combined nrDNA and the trnL intron analysis. A well supported change of taxonomic significance revealed by the present study is the placement of T. arabica and T. schlumbergeri in one section. Sectional delimitation of section Cylindricae is supported. With a moderate support the monophyly of section Vérae and Callicerates is congruent with the sectional delimitation of Širjaev (1928–1932) and Small (1987a, b) based on morphological similarities. However, the phylogeny derived from the combined data sets has also clearly shown incongruence between the classification based on morphological characters and genetic relationships resolved by molecular data indicating that the taxonomic status of some sections needs revision. Section Falcatulae and Foenum-graecum are rendered paraphyletic and thus their current delimitation should be reconsidered. Molecular data indicates that series Anguinae should no longer be recognized. T. suavissima should be included in series Stellatae while T. anguina should be included in subsection Tuberculatae. This placement will highlight the phylogenetic relationship and morphological similarities of T. anguina with T. balansae and that of T. suavissima with T. maritima and T. stellata. T. balansae has the potential to complement the role of annual medics in alkaline soil farming system especially due to the more expensive seeds of annual medics (Howie et al. 2001). This cross pollinated species (Nair et al. 2004) is also compatible with Rhizobium meliloti associated with medic pastures. Close relationship of T. anguina with T. balansae observed in the present study indicates the possibility of the use of the former species in breeding programs in crosses with T. balansae.

A definitive split is indicated for subsection Biebersteinianae and Gladiatae of section Foenum-graecum, which is supported by morphological features related to inflorescence, pods, calyx, stipules and flower color. The phylogram strongly supports the view of Sinskaya (1961) that T. gladiata and T. foenum-graecum share a common ancestral linkage. T. gladiata, also called as sward fenugreek is a forage crop of great interest as suspected ancestor of cultivated fenugreek. This species is attractive for grazing but is characterized by low forage productivity. However, it possesses high drought resistance and grows on poor soil (Sinskaya 1961). The identification of closely related species for the widely cultivated T. foenum-graecum will ensure more efficient use of the wild genetic resources in the improvement of this crop. This is all the more important in the context of susceptibility of fenugreek to various pests and diseases, resulting in low yields (Acharya et al. 2010).

Molecular data indicates that Širjaev’s (1928–1932) taxonomic concept of dividing the genus into three sub genera is not fully consistent among itself. Phylogenetic reconstructions indicated three major lineages supported by a combination of morphological characters related to inflorescence and stipule (Table 3). This study may serve as a basis for future studies in Trigonella, which should include some additional species (especially section Ellipticae) not included in the present study. Moreover, morphological and biochemical similarities among the species in each of these hypothesized groups need to be further investigated.
Table 3

Character state combination of inflorescence, calyx and stipule for clade I–III

Clade

Sections

Clade I: Inflorescence if present capitulum (sessile subumbellate only in T. anguina), stipules entire/dentate, calyx campanulate/tubular

Sect. Samaroideae

Sect. Falcatulae in part (T. balansae and T. anguina)

Sect. Foenum-graecum subsect. Gladiatae

Sect. Cylindricae

Sect. Uncinatae

Sect. Capitatae

Sect. Foenum-graecum subsect. Biebersteinianae

Clade II: Inflorescence umbellate, stipule base dentate with upper part entire, calyx campanulate/subtubular

Sect. Callicerates

Sect. Vérae

Clade III: Inflorescence umbellate, stipules dentate, calyx campanulate

Sect. Pectinatae (+sect. Erosae)

Sect. Falcatulae in part (T. maritima, T. stellata and T. suavissima)

Conclusion

This study provides the most comprehensive reconstruction of phylogenetic relationships to date for the genus Trigonella and will serve as a framework for future taxonomic and evolutionary studies in the genus. The phylogeny derived from the combined data sets provides strong support for the monophyly of the genus Trigonella and indicates that the taxonomic status of some sections like Falcatulae and Foenum-graecum needs revision. Some clades in the phylogram derived from the combined data sets are weakly supported indicating that a number of questions regarding relationships among Trigonella species need to be resolved further. Data from different gene regions along with intensive sampling, mostly in the poorly understood section Ellipticae will be required for taxonomic reassessment and possible taxonomic revision of the genus. However, inflorescence type provides potentially useful synapomorphy for most of the clades in Trigonella and will be one of the useful characters for sub-generic and sectional circumscription in the genus. The phylogeny derived indicates that similarity in legume characteristics has occurred independently in Trigonella and these characters alone are not sufficient for sectional and sub-sectional circumscription in Trigonella.

Notes

Acknowledgments

The authors wish to thank Dr. Steve Hughes, Curator, The Genetic Resource Centre, South Australian Research and Development Institute (SARDI), Waite Research Precinct, Waite Institute, Urrbrae, South Australia for the supply of germplasm used in this study.

Conflict of interest

The authors declare that they have no conflict of interest.

Supplementary material

10722_2015_236_MOESM1_ESM.pdf (222 kb)
Supplementary material 1 (PDF 223 kb)

References

  1. Acharya SN, Thomas JE, Prasad R, Basu SK (2010) Diseases of fenugreek (Trigonella foenum-graecum L.) and their control measures with special emphasis on fungal diseases. In: Arya A, Perello AE (eds) Management of fungal plant pathogens. CAB International, Wallingford, pp 245–261CrossRefGoogle Scholar
  2. Bena G (2001) Molecular phylogeny supports the morphological based taxonomic transfer of the “medicagoid” Trigonella species to the genus Medicago L. Plant Syst Evol 229:217–236CrossRefGoogle Scholar
  3. Bremer K (1994) Branch support and tree stability. Cladistics 10:295–304CrossRefGoogle Scholar
  4. Çeter T, Pinar M, Akan H, Ekici M, Aytaç Z (2012) Comparative seed morphology of Trigonella L. species (Leguminosae) in Turkey. Afr J Agric Res 7:509–522Google Scholar
  5. Chase MW, De AY, Bruijn AV, Cox G, Reeves PJ, Rudall MAT, Johnson L, Eguiarte E (2000) Phylogenetics of Asphodelaceae Asparagales: an analysis of plastid rbcL and trnL-F DNA sequences. Ann Bot 86:935–951CrossRefGoogle Scholar
  6. Cunningham CW (1997a) Can three incongruence tests predict when data should be combined? Mol Biol Evol 14:733–740PubMedCrossRefGoogle Scholar
  7. Cunningham CW (1997b) Is congruence between data partitions a reliable predictor of phylogenetic accuracy? Empirically testing an iterative procedure for choosing among phylogenetic methods. Syst Biol 46:464–478PubMedCrossRefGoogle Scholar
  8. Dangi RS (2013) Diversity in genus Trigonella: molecular and chemical characterization. Ph.D. Dissertation, University of Pune, IndiaGoogle Scholar
  9. Dangi RS, Lagu MD, Choudhary LB, Ranjekar PK, Gupta VS (2004) Assessment of genetic diversity in Trigonella foenum-graecum and Trigonella caerulea using ISSR and RAPD markers. BMC Plant Biol 4:13PubMedPubMedCentralCrossRefGoogle Scholar
  10. Ellison NW, Aaron A, Steiner JJ, Williams WM, Taylor NL (2006) Molecular phylogenetics of the clover genus (Trifolium—Leguminosae). Mol Phylogenet Evol 39:688–705PubMedCrossRefGoogle Scholar
  11. Endress PK, Doyle JA (2009) Reconstructing the ancestral flower and its initial specializations. Am J Bot 96:22–66PubMedCrossRefGoogle Scholar
  12. Farris JS (1989) The retention index and homoplasy excess. Syst Zool 38:406–407CrossRefGoogle Scholar
  13. Farris JS, Kallersjo M, Kluge AG, Bult CJ (1994) Testing significance of incongruence. Cladistics 10:315–319CrossRefGoogle Scholar
  14. Farris JS, Kallersjo M, Kluge AG, Bult CJ (1995) Constructing a significance test for incongruence. Syst Biol 44:570–572CrossRefGoogle Scholar
  15. Fazli FRY (1967) Studies in steroid-yielding plants of the genus Trigonella. Dissertation, University of Nottingham, EnglandGoogle Scholar
  16. Felsenstein J (1985) Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39:783–791CrossRefGoogle Scholar
  17. Feng CM, Xiang QA, Franks RG (2011) Phylogeny-based developmental analyses illuminate evolution of inflorescence architectures in dogwoods (Cornus s. l., Cornaceae). New Phytol 91:850–869CrossRefGoogle Scholar
  18. Gazara M, Kamel W, Haider A (2001) Cladistic analysis of the genera: Trifolium, Trigonella and Melilotus (Fabaceae: Papilionaceae) in Egypt. Egypt J Biol 3:161–170Google Scholar
  19. Gielly L, Taberlet P (1994) The use of chloroplast DNA to resolve phylogenies: non-coding versus rbcL sequences. Mol Biol Evol 11:769–777PubMedGoogle Scholar
  20. Greuter W, Raus T (eds) (1987) Med-Checklist Notulae, 14. Willdenowia 16:439–452Google Scholar
  21. Harris EM (1999) Capitula in the Asteridae: a widespread and varied phenomenon. Bot Rev 65:348–369CrossRefGoogle Scholar
  22. Hector JN (1938) Introduction to the botany of field crops (non cereals). Central News Agency Ltd, Johannesburg, p 1127Google Scholar
  23. Heyn CC (1981) Trifolieae. In: Polhill RM, Raven PH (eds) Advances in legume systematics, Part 1. Royal Botanic Gardens, Kew, pp 383–385Google Scholar
  24. Hillis DM (1995) Approaches for assessing phylogenetic accuracy. Syst Biol 44:3–16CrossRefGoogle Scholar
  25. Howie J, Ballard RA, de Koning C, Sandral G, Charman N (2001) Trigonella balansae—a new pasture legume for the alkaline soils of Southern Australia? In: Proceedings of the 10th Australian agronomy conference, Hobart. Australian Society of Agronomy. http://www.regional.org.au/au/asa/2001/3/d/howie.htm. Accessed 20 Nov 2013
  26. Huber-Morath A (1970) Trigonella L. In: Davis PH (ed) Flora of Turkey and the East Aegean Islands, vol 3. Edinburgh Univ. Press, Edinburgh, pp 452–482Google Scholar
  27. Huelsenbeck JP, Larget B, Miller RE, Ronquist F (2002) Potential applications and pitfalls of Bayesian inference of phylogeny. Syst Biol 51:673–688PubMedCrossRefGoogle Scholar
  28. Hutchinson J (1964) The genera of flowering plants. Clarendon Press, Oxford, pp 947–952Google Scholar
  29. Ingham JL (1981) Phytoalexin induction and its chemosystematic significance. Biochem Syst Ecol 9:275–281CrossRefGoogle Scholar
  30. Kakani RK, Singh SK, Pancholy A, Meena RS, Pathak R, Raturi A (2011) Assessment of genetic diversity in Trigonella foenum-graecum based on nuclear ribosomal DNA, internal transcribed spacer and RAPD analysis. Plant Mol Biol Rep 29:315–323CrossRefGoogle Scholar
  31. Kluge AG, Farris JS (1969) Quantitative phyletics and the evolution of anurans. Syst Zool 8:1–32CrossRefGoogle Scholar
  32. Kores PJ, Molvray M, Weston PH, Hopper SD, Brown AP, Cameron KM, Chase MW (2001) A phylogenetic analysis of Diurideae (Orchidaceae) based on plastid DNA sequence data. Am J Bot 88:1903–1914PubMedCrossRefGoogle Scholar
  33. Ladizinsky G (1979) Seed protein electrophoresis in the section Foenum-graecum of Trigonella L. (Fabaceae). Plant Syst Evol 133:87–94CrossRefGoogle Scholar
  34. Lashin GMA (2006) Comparative morphology of pollen grains of some taxa of tribe Trifolieae (Fabaceae: Papilionoideae) from Egypt. Int J Bot 2:270–277CrossRefGoogle Scholar
  35. Lavin M, Doyle JJ, Palmer JD (1990) Evolutionary significance of the loss of the chloroplast-DNA inverted repeat in the Leguminosae subfamily Papilionoideae. Evolution 44:390–402CrossRefGoogle Scholar
  36. Liston A (1995) Use of the polymerase chain reaction to survey for the loss of the inverted repeat in the legume chloroplast genome. In: Crisp MD, Doyle JJ (eds) Advances in legume systematic, part 7: phylogeny. Royal Botanic Gardens, Kew, pp 31–40Google Scholar
  37. Liston A, Wheeler JA (1994) The phylogenetic position of the genus Astragalus (Fabaceae): evidence from the chloroplast genes rpoC1 and rpoC2. Biochem Syst Ecol 22:377–388CrossRefGoogle Scholar
  38. Liston A, Sterner J, Taylor NL, Ellison NW, Williams WM (2001) Phylogeny of New World Trifolium. Identification of a cryptic clade (abstract). 4th International legume conferenceGoogle Scholar
  39. Liu FR, Miyamoto MM (1999) Phylogenetic assessment of molecular and morphological data for Eutherian mammals. Syst Biol 48:54–64PubMedCrossRefGoogle Scholar
  40. Marcilla A, Bargues MD, Ramsey JM, Magallon-Gastelum E, Salazar-Schettino P, Abad-Franch F et al (2001) The ITS-II of the nuclear rDNA as a molecular marker for populations, species, and phylogenetic relationships in Triatominae (Hemiptera: Reduviidae), vectors of Chagas disease. Mol Phylogenet Evol 18:136–142PubMedCrossRefGoogle Scholar
  41. Nair RM, Dundas IS, Wallwork M, Verlin DC, Waterhouse L, Dowling K (2004) Breeding system in a population of Trigonella balansae (Leguminosae). Ann Bot 94:883–888PubMedPubMedCentralCrossRefGoogle Scholar
  42. Niknam V, Sharifizadeh B, Ebrahimzadeh H, Zarre S (2004) Comparative study of proteins in seeds of some species of Trigonella from Iran. Iran Int J Sci 5:1–11Google Scholar
  43. Nixon KC, Carpenter JM (1996) On simultaneous analysis. Cladistics 12:221–241CrossRefGoogle Scholar
  44. Parkin J (1914) The evolution of the inflorescence. Bot J Linn Soc 42:511–563CrossRefGoogle Scholar
  45. Petropoulos GA (2002) Fenugreek the genus Trigonella. Taylor and Francis, London and New York, pp 1–255CrossRefGoogle Scholar
  46. Posoda D, Crandall KA (1998) Modeltest, testing the model of DNA substitution. Bioinfomatics 14:817–818CrossRefGoogle Scholar
  47. Rechinger KH (1984) Trigonella. In: Rechinger KH (ed) Flora Iranica, vol 157. Akad. Druck- Verlagsanstalt, Graz, pp 207–253Google Scholar
  48. Rogers SO, Bendich AJ (1985) Extraction of DNA from milligram amounts of fresh, herbarium and mummified plant tissues. Plant Mol Biol 5:69–76PubMedCrossRefGoogle Scholar
  49. Ronquist F, Huelsenbeck JP (2003) MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19:1572–1574PubMedCrossRefGoogle Scholar
  50. Rouk HF, Mangesha H (1963) Fenugreek (Trigonella foenum-graecum L.) its relationship, geography and economic importance, Exper Stat Bull No. 20, Imper. Ethiopian College of Agric. & Mech. ArtsGoogle Scholar
  51. Sambrook J, Russell DW (2001) Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory Press, New YorkGoogle Scholar
  52. Sanderson MJ, Wojciechowski MF (1996) Diversification rates in a temperate legume clade: are there “so many species” of Astragalus (Fabaceae)? Am J Bot 83:1488–1502CrossRefGoogle Scholar
  53. Schultz OF (1901) Monographie der Gattung Melilotus. Botanische Jahrbücher für Systematik, Pflanzengeschichte und Pflanzengeographie 29:660–735Google Scholar
  54. Seelanan T, Schnabel A, Wendel JF (1997) Congruence and consensus in the cotton tribe (Malvaceae). Syst Biol 22:259–290Google Scholar
  55. Sinskaya E (1961) Flora of the cultivated plants of the USSR: XIII perennial leguminous plants. Part I: Medicago, Sweet clover, Fenugreek. Israel Programme for Scientific Translations, Jerusalem, pp 514–570Google Scholar
  56. Širjaev (1928–1932) Generis Trigonella L. revisio critica Vol 1. Volume 102 of Spisy vydávané Přirodovědecká Fakultou Masarykovy University Přirodověd. Fak., BrnoGoogle Scholar
  57. Small E (1987a) A taxonomic study of the “medicagoid” Trigonella (Leguminosae). Can J Bot 65:1199–1211CrossRefGoogle Scholar
  58. Small E (1987b) Polythetic generic separation in the tribe Trifolieae subtribe Trigonellinae (Leguminosae). Can J Bot 67:1480–1492CrossRefGoogle Scholar
  59. Small E, Jomphe M (1989) A synopsis of the genus Medicago (Leguminosae). Can J Bot 67:3260–3294CrossRefGoogle Scholar
  60. Sorenson MD (1999) TreeRot, version 2. Boston University, BostonGoogle Scholar
  61. Stebbins GL (1974) Flowering plants: evolution above the species level. Belknap Press, CambridgeCrossRefGoogle Scholar
  62. Steele KP, Wojciechowski MF (2003) Phylogenetic analyses of tribes Trifolieae and Vicieae, based on sequences of the plastid gene matK (Papilionoideae: Leguminosae). In: Klitgaard BB, Bruneau A (eds) Advances in legume systematics, part 10, higher level systematics. Royal Botanic Garden, Kew, pp 355–370Google Scholar
  63. Steele KP, Ickert-Bond SM, Zarre S, Wojciechowski MF (2010) Phylogeny and character evolution in Medicago (Leguminosae): evidence from analyses of plastid trnK/matK and nuclear GA3ox1 sequences. Am J Bot 97:1142–1155PubMedCrossRefGoogle Scholar
  64. Swofford DL (2001) PAUP*: phylogenetic analysis using parsimony (*and other methods), version 4. Sinauer, SunderlandGoogle Scholar
  65. Taberlet P, Gielly L, Pautou G, Bouvet J (1991) Universal primers for amplification of three non-coding regions of chloroplast DNA. Plant Mol Biol 17:1105–1109PubMedCrossRefGoogle Scholar
  66. Thompson JD, Higgins DG, Gibson TJ (1994) CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position specific gap penalties and weight matrix choice. Nucl Acids Res 22:4673–4680PubMedPubMedCentralCrossRefGoogle Scholar
  67. Tutin TG, Heywood VH, Burges NA, Moore DM, Valentine DH, Walters SM, Webb DA (1969) Flora Europaea. Vol. 2. Rosaceae to Umbelliferae. Cambridge University Press, CambridgeGoogle Scholar
  68. Vasil’chenko IT (1953) (Bericht über die Arten der Gattung Trigonella) Trudy Bot. Inst. Akad. Nauk. I 10:124–269 (In Russian)Google Scholar
  69. Watson LE, Sayed-Ahmed H, Badr A (2000) Molecular phylogeny of Old World Trifolium (Fabaceae). Plant Syst Evol 224:153–171CrossRefGoogle Scholar
  70. White TJ, Bruns T, Lee S, Taylor J (1990) Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. In: Innis MA, Gelfand DH, Sninsky JJ, White TJ (eds) PCR protocols: a guide to methods and applications. Academic Press, London, pp 315–322Google Scholar
  71. Wojciechowski MF (2003) Reconstructing the phylogeny of legumes (Leguminosae): an early 21st century perspective. In: Klitgaard BB, Bruneau A (eds) Advances in legume systematic, part 10, higher level systematics. Royal Botanic Garden, Kew, pp 5–35Google Scholar
  72. Wojciechowski MF, Sanderson MJ, Steele KP, Liston A (2000) Molecular phylogeny of the “Temperate Herbaceous Tribes” of Papilionoid legumes: a supertree approach. In: Herendeen PS, Bruneau A (eds) Advances in legume systematics, vol 9. Royal Botanic Gardens, Kew, pp 277–298Google Scholar
  73. Wojciechowski MF, Lavin M, Sanderson MJ (2004) A phylogeny of legumes (Leguminosae) based on analysis of the plastid matK gene resolves many well-supported subclades within the family. Am J Bot 91:1846–1862PubMedCrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media Dordrecht 2015

Authors and Affiliations

  • Rakhee Dangi
    • 1
    • 2
  • Shubhada Tamhankar
    • 1
  • Ritesh Kumar Choudhary
    • 1
  • Suryaprakasa Rao
    • 1
    • 3
  1. 1.Plant Sciences DivisionAgharkar Research InstitutePuneIndia
  2. 2.Rajiv Gandhi Institute of Information Technology and BiotechnologyBharati Vidyapeeth Deemed UniversityPuneIndia
  3. 3.Indian Institute of Science Education and Research (IISER)PuneIndia

Personalised recommendations