The tribe Fabeae (= Vicieae) contains four genera, of which Lathyrus L. and Vicia L. with ca 160 species are the largest (1, 2). Lens Mill. with 4-6 species and Pisum L. with 4 species (Vavilovia formosa (Stev.) Fed. of the monotypic genus Vavilovia Fed. was treated as a part of Pisum as P. formosum (Stev.) Alef. (3)) are small genera. Vicia is paraphyletic and along with Lens, as a monophyletic taxon, formed a sister group relationship with a clade of Lathyrus and Pisum (2-5). Species of Lathrys are distributed throughout the Northern Hemisphere, tropical East Africa and temperate South America. Its main center of diversity is in the Mediterranean and Irano-Turanian regions, with smaller centers in North and South America (2, 6-8). Many species of Lathyrus are used extensively as cover crops, for fodder (L. cicera L., L. hirsutus L. and L. sativus L.), as ornamentals (L. odoratus L. [sweet pea], L. latifolius L. [everlasting pea] and L. sylvestris L.) and as human food (L. sativus [grass pea, Indian pea, chuckling vetch], L. ochrus (L.) DC. and L. montanus Bernh. (with edible root tubers)) and also for erosion control, as green manure and for medicine; toxins are present in some species, causing Lathyrism (2).
More than 60 species of the genus have been distributed in the Southwest Asia (2). In the regional Flora including Flora of Turkey (9), Flora Iranica(10) and Flora of Iran (11), the genus was divided into 9-10 sections. Almost three decades ago, Kupicha (6) based on morphological characters, by assessing the older classifications of the genus, proposed her sectional system encompassing 13 sections, of which three were newly established: Orobus (L.) Gordon, Lathyrostylis (Griseb). Bassler, Pratensis Bassler, Neurolobus Bassler, Orobon Tamamshjan, Orobastrum Boiss, Viciopsis Kupicha, Linearicarpus Kupicha, Lathyrus, Aphaca (Miller) Dumort. Nissolia (Miller) Dumort, Clymenum (Miller) DC. and NotoLathyrus Kupicha. She divided section Orobastrum into three monotypic sections Orobastrum, Linearicarpus and Viciopsis. She included section Cicercula (Medic.) Gren. and Godr. in section Lathyrus and added a new section of South American species, NothoLathyrus. We followed her system of classification in the present study.
Dogan et al. (12) conducted a numerical taxonomic study on 52 species of Turkish Lathyrus using 40 external vegetative and floral morphological characters. They divided the genus into nine sections, of which section Gorgonia Dogan was described as a new taxon. Leht (13) based on cladistic analysis of 210 morphological characters suggest that all species of section Pratensis Bassler should belong to section Aphaca. He placed the monotypic section Orobon (with species L. roseus Steven) within section Orobus (13). Hitherto, several works based on molecular data have been conducted on the phylogeny of Lathyrus. Asmussen and Liston (7) sampled 42 species of Lathyrus based on cpDNA restriction site characters and their study included representatives of all sections but Viciopsis. They generally agreed with the sectional classification of Kupicha (6), yet they suggested that Orobon and Orobastrum should be transferred to the section Lathyrus, and NothoLathyrus to Orobus. Croft et al. (14) used Random Amplified Polymorphic DNA (RAPD) analysis on eight accessions of L. sativus and some allies species. Badr et al. (15) utilized Amplified-Fragment Length Polymorphism (AFLP) data to examine the systematic relationship in section Lathyrus. Ben Brahim et al. (16) studied isozyme variation and phylogenetic relationships between ten species of the genus. Kenicer et al. (8) studied 53 species of the genus based on nrDNA ITS, cpDNA trnL-F and trnS-G sequence data (except members of two Kupicha’s sections, Orobon and Viciopsis). Their results generally supported the monophyly of sections Lathyrus and Lathyrostylis, but questioned monophyly of the larger section Orobus.
In this study, we examined the phylogenetic relationships among 38 species of the genus Lathyrus based on chloroplast gene matK. All species of the genus growing in Iran and some species from Turkey were included. Unlike some previous studies (7, 8) the present study employed several approaches including maximum parsimony, Bayesian and maximum likelihood methods to evaluate the monophyly and relationship within Lathyrus, both at the sectional level and species level.
3. Materials and Methods
3.1. Taxon Sampling
Fifty-two accessions for matK were analyzed. A total of 38 species representing eleven sections of Lathyrus (6) and three species of Pisum (including Vavilovia formosa, see 3) were included in the phylogenetic analyses using each of the separate datasets as well as the combined dataset. The two species of Lens and Vicia were chosen as outgroups according to previous studies (4, 5, 7, 8). List of the analyzed species along with their voucher information and GenBank accession numbers are presented in Table 1.
3.2. DNA Isolation, PCR and Sequencing
Total genomic DNA was isolated from leaf material (stipules of Lathyrus aphaca L.) using the modified Cetyltrimethyl Ammonium Bromide (CTAB) method of Doyle and Doyle (18). The chloroplast gene matK was used. The partial matK gene was amplified using primers matK384F (designed in this study) and trnK2R* (5). Total volume of amplification reactions was 20 µL, which contained 8.2 µL deionized water, 10 µL of the 2 × Taq DNA polymerase master mix Red (Amplicon, Cat. No. 180301, 150 mM Tris-HCL pH 8.5, 40 mM (NH4)2SO4, 3.0 mM MgCl2, 0.4 mM dNTPs, 0.05 units.µL-1 Amplicon Taq DNA polymerase, inert red dye and a stabilizer) 0.5 µL of each primer (5 pmol.µL-1) and 1 µL of template DNA (20 ng.µL-1). The PCR cycles consisted of predenaturation at 95°C for 2.30 minutes followed by 35 cycles: denaturation at 94°C for 45 seconds, annealing at a temperature depending on the region, at 51°C for matK and elongation at 72°C for 1.20 minutes. A final elongation step of seven minutes at 72°C was performed. Each region was sequenced using the ‘Big dye terminator cycle sequencing ready reaction kit’ with appropriate primers in an ABI Prism 3730XL DNA analyzer.
3.3. Sequence Alignment
The dataset was aligned using CLUSTAL (19) and adjusted manually. Positions of insertions and deletions were treated as missing data.
3.4. Phylogenetic Analyses
3.4.1. The Maximum Parsimony Method
Maximum Parsimony (MP) analyses were conducted using the PAUP* version 4.0b10 (20) for phylogenetic analysis. The heuristic search option was employed for the dataset, using tree bisection-reconnection (TBR) branch swapping, with 1000 replications of random addition sequence and an automatic increase in the maximum number of trees. Uninformative characters were excluded from the analysis. Branch support values were calculated using a full heuristic search with 1000 bootstrap (BS) replicates (21) each with simple addition sequence.
3.4.2. Bayesian Method
Model of sequence evolution for the dataset was selected using the program MrModeltest version 2.3 (22) as implemented in MrMTgui (23) based on the Akaike information criterion (AIC) (24). The matK dataset was analyzed using the GTR + I + G model. The program MrBayes (25) was used for the Bayesian phylogenetic analysis. Posteriors on the model parameters were estimated from the data, using the default priors. The analysis was done with four million generations, using Markov chain Monte Carlo search. MrBayes performed two simultaneous analyses starting from different random trees (Nruns = 2) each with four Markov chains and trees sampled at every 100 generations. The first 25% of trees were discarded as the burn-in. The remaining trees were then used to build a 50% majority rule consensus tree accompanied with posterior probability (PP) values. Tree visualization was carried out using Tree View version 1.6.6.
3.4.3. Maximum Likelihood
Maximum likelihood (ML) analyses were performed for the datasets in the program Genetic Algorithm for Rapid Likelihood Inference (GARLI) (26). The model of evolution employed for each data set was the same as that of Bayesian analyses.
Parametric bootstrap values for ML were calculated in GARLI, based on 100 replicates with one search replicate per bootstrap replicate.
3.4.4. Shimodaira-Hasegawa Test
Shimodaira-Hasegawa (SH) test (27) in PAUP*, based on maximum likelihood analyses was used to test the hypothesis that Lathyrus is monophyletic and Pisum is the sister group for Lathyrus. To accept that the hypothesis probability value (p-value) should be more than 0.05.
4. 1. Sequence Dataset
The aligned matK dataset was 1077 nucleotide sites long, of which 186 sites were potentially parsimony informative. The length of matK varies from 623 BP in Lathyrus aphaca, L. ochrus, L. Clymenum L., L. tingitanus L. and L. latifolius (the smaller length of matK gene for these species is due to incomplete sequencing, which was retrieved from GenBank) to 1065 in L. alamutensis Mozaffarian, Ahvazi & Charkhchian. Parsimony analysis of the dataset resulted in 10000 trees of length (L) = 468 steps, Consistency Index (CI) = 0.519 and Retention Index (RI) = 0.798. The 50% majority rule consensus tree resulting from the Bayesian analysis of the matK dataset with posterior probabilities and bootstrap values is shown in Figure 1.
In the present study, for all trees derived from MP, Bayesian and ML methods, Pisum was nested among Lathyrus species. Two members of the section Clymenum (Lathyrus ochrus and L. Clymenum) with Pisum subclade formed a weakly supported clade as sister to larger polytomy comprising the remainder Lathyrus species.
Here, we described the relationship between the taxa based upon the tree from Bayesian analysis (Figure 1). All multi-specific sections of Lathyrus (Lathyrostylis, Lathyrus and Clymenum), with the exception of Linearicarpus, Orobus and Pratensis were well-supported monophyletic groups. Pratensis and the monotypic section Aphaca were the closest taxa. Results of the SH test based upon the dataset showed that Lathyrus is a monophyletic taxon and Pisum is sister to this taxon (P > 0.05).
5. 1. Phylogenetic Status of the Genus Lathyrus
Kupicha (6) and Lock & Maxted (2) stated that Lathyrus is a monophyletic genus. In studies of Steele et al. (4) and Kenicer et al. (8), Pisum sativum L. along with species of Lathyrus formed a clade with bootstrap value of 100%. In the latter work, Lathyrus species, with the inclusion of sections Nissolia, Clymenum and Neurolobus, formed a weakly supported clade (bootstrap value of < 50%). In the present study, at the base of the tree, the Pisum clade was weakly allied with L. ochrus and L. clymenum (sect. Clymenum) and thus, Lathyrus was not monophyletic (Figure 1). SH test (p > 0.05) of the dataset showed that all analyzed Lathyrus species formed their own clade and Pisum was sister to them. Furthermore, when we removed these two Lathyrus species, the analysis retrieved Pisum as a well-supported clade being sister to Lathyrus calde (tree not shown).
Lathyrus and Pisum are similar in some morphological characters, including presence of pollen brush on the adaxial side of the style; non-brochiododromous leaflet venation (i.e. veins reaching the margins) and accumulate pisatin, a phytoalexin that is not found in other members of Fabeae. In addition, both genera lack canavanine (4, 8, 28-31). The two genera, however, differ in some diagnostic characters as conduplicate or folded leaflet ptyxis and folded longitudinal style in Pisum (except P. formosum for the former character), but super volute or rolled ptyxis and non folded longitudinal style in Lathyrus (1, 4). Moreover, the wider hybridization experiments between Lathyrus and Pisum species have shown cross-incompatibility (32).
5. 2. Relationships within Lathyrus
The tree resulting from the dataset composed of several clades. Some of the clades had also appeared in previous molecular studies (7, 8); albeit the relationships were well resolved and supported in our tree. Below we examine some of these lineages.
5. 3. Section Orobus
Kupicha (6) accepted the heterogeneous section Orobus with 54 species having a perennial habit, multijugate leaves with or without tendrils. As she noted, “the section contains a diversity of vegetative and floral forms and many of its members appear to be rather distantly related”.
Our matK analysis did not retrieve section Orobus as a monophyletic group. This is in agreement with the conclusion made by Kenicer et al. (8). Lathyrus vernus (L.) Bernh, L. incurvus (Roth) Willd and L. aureus (Steven) Brandza, are successive sisters to a subclade of three species L. quinquenervus (Miq.) Litv. L. japonicus Willd. and L. humilis (Ser.) Sprengel (all of sect. Orobus) (BS=81%, 89% and pp = 1.00) and sect. Linearicarpus (L. woronowii Bornm, L. inconspicus L. and L. sphaericus Retz.). Kupicha (6) claimed that a few species of section Orobus (e.g. L. quinquenervus, L. alpestris (Waldst. & Kit.)) are closely related to members of section Lathyrostylis, whereas our analysis did not show a direct relationship between these two sections.
5. 4. Sections Aphaca and Pratensis
The present study is concordant with previous reports (6-8) regarding the close relationship of sections Aphaca and Pratensis. They do share large hastate stipules, a unique pattern of nodal anatomy and distinctive wing petal architecture. Aphaca is a ditypic section and Pratensis is a small section comprising six species (6). The former section is represented by a single species (L. aphaca) with three accessions, and the second one by three species with three accessions (L. Pratensis L., L. laxiflorus (Desf.) Kuntze and L. czeczottianus Bassler). In our analysis, L. Pratensis formed a sister group relationship with the Aphaca clade not its own section.
5. 5. Section Lathyrus
In our results, all members of the section Lathyrus with inclusion of L. roseus Steven(of the monotypic section Orobon) formed a well-supported clade (BS = 71%, 77% and pp = 0.97). Within this calde, L. tingitanus is well allied with the remainder species. The remainder species made up two subclades. One subclade included only annual species, L. cicera, L. pseudocicera Pampan. and L. sativus (BS = 99%, 98% and pp = 1.00), which were previously classified under either section Cicercula (Medic.) Godr. (9, 10, 33) or section Clymenum (11). These species are characterized by a canaliculate style. The second subclade comprised of both annual and perennial species (BS = 97%, 95% and pp = 1.00). Members of this subclade were already placed in sections Cicercula (annuals), Lathyrus (perennials) and the monotypic Orobon (L. roseus) (9, 10, 33). Kupicha (6) questioned the monophyly of section Cicercula and synonymized it with sect Lathyrus. This treatment was confirmed and followed by Kenicer et al. (8) and the present study. However, Kupicha maintained section Orobon (L. roseus) by suggesting that a bushy habit and isodiametric and strongly wavy-walled epidermal cells distinguishes this group from the species of section Lathyrus. Yet, the flower and fruit characteristics agree entirely with those of section Lathyrus (6). She claimed that this species is related to the delicate perennials (e.g. L. tuberosus L.) of the section Lathyrus. Indeed, in the present study and Asmussen and Liston’s (7) analyses, this species was allied with L. tuberosus. Hence, we are in agreement with Asmussen and Liston (7) in that the section Orobon should be combined with section Lathyrus.
5. 6. Section Lathyrostylis
Lathyrostylis (= Platystylis) is a relatively uniform section comprising 21 species (6, 12), of which 10 species were included. Our matK as well as the combined (nrDNA ITS, trnL-F and trnS-G) sequence data of Kenicer et al. (8) indicated that this section formed a well-supported clade (BS = 94%, 65%, PP = 1.00). Czefranova (33) placed species of this section within section Orobus. However, Bassler (34) and later Kupicha (6) suggested that the two sections are distinct. The current study and previous molecular studies clearly showed that members of section Lathyrostylis are not related to section Orobus, supporting the recent morphology-based treatments (6, 34). It is important to note that the section is represented by three species L. cyaneus (Stev.) C. Koch, L. boissieri Sirj. and L. alamutensis in Iran only, of which, the latter was recently established as a new species from western Alborz mountain ranges. This new species has very specific characteristics with two flap-like appendages above the claw of standard (35).
Linearicarpus is a small section with seven annual species of which four were analyzed here (6). Analysis of matK sequence data revealed that the section at the current status is not monophyletic. Its three species L. inconspicus, L. sphaericus and L. woronowii are closely related taxa and allied with some members of section Orobus. Yet, L. vinealis Boiss.& Noe (fourth species), falls distantly from them in a trichotomy. Previous morphology (12) and molecular-based (7, 8, 15) analyses questioned the monophyly of this section. It should be mentioned that in Flora of Turkey (9) and Flora Iranica(10), L. inconspicus, L. sphaericus, L. vinealis and L. woronowii plus some others were placed in section Orobastrum, whereas in Flora of Iran (11), they along with L. nissolia were classified under sect. Nissolia. Kupicha (6) combined section. Orobastrum with sect. Linearicarpus and recognized Nissolia as a monotypic section.
5. 8. Section Clymenum
The section Clymenum is composed of only three annual species, L. gloeospermus Warb. & Eig., L. Clymenum and L. ochrus. Previous researchers (7, 8) by analyzing all these three species questioned the monophyly of this section. They all reached the same conclusion as ours, claiming that L. Clymenum and L. ochrus are closely related taxa. Although, L. gloeospermus was well separated from the remainder of the section. In contrast to the latter species, L. Clymenum and L. ochrus do share several floral features including hollow, finger-like pouches on the standard, an unusual type of style (spathulate with apex ended to sterile fleshy mucro or arista which divides the stigma into two halves) and pods with winged sutures (6).
The authors thank the staff of the herbaria of West Azerbaijan Research Center of Natural Resources and Agriculture Herbarium, Ferdowsi University of Mashhad Herbarium, Gazi University Herbarium, Tehran University Herbarium, Herbarium of the Research Institute of Forests and Rangelands, for providing leaf materials.
All authors participated equally in the study
There was no conflict of interest.
Funding was provided by the Islamic Azad University, Ayatollah Amoli Branch.