DNA barcoding and TLC as tools to properly identify natural populations of the Mexican medicinal species Galphimia glauca Cav.

Galphimia glauca is a plant that is endemic to Mexico and has been commonly used since pre-Hispanic times to treat various illnesses, including central nervous system disorders and inflammation. The first studies investigating a natural population of G. glauca in Mexico showed that the plant has anxiolytic and sedative activities in mice and humans. The plant's bioactive compounds were isolated and identified, and they belong to a family of nor-secofriedelanes called galphimines. The integration of DNA barcoding and thin-layer chromatography analysis was performed to clarify whether the botanical classification of the populations in the study, which were collected in different regions of Mexico, as G. glauca was correct or if the populations consist of more than one species of the genus Galphimia. We employed six DNA barcodes (matK, rbcL, rpoC1, psbA-trnH, ITS1 and ITS2) that were analyzed individually and in combination and then compared each other, to indicate differences among the studied populations. In the phylogenetic analysis, ITS1 and ITS2 markers as well as the combination of all DNA regions were the most efficient for discriminating the population studied. The thin-layer chromatography analysis exhibited four principal chemical profiles, one of which corresponded to the populations that produced galphimines. DNA barcoding was consistent and enabled us to differentiate the populations that produce galphimines from those that do not. The results of this investigation suggest that the studied populations belong to at least four different species of the genus Galphimia. The phylogenetic analysis and the thin-layer chromatography chemical profiles were convenient tools for establishing a strong relationship between the genotype and phenotype of the studied populations and could be used for quality control purposes to prepare herbal medicines from plants of the genus Galphimia.

Introduction

Galphimia glauca Cav. (Malpighiaceae), is a plant that is endemic to Mexico and has been traditionally used to treat different ailments, including central nervous system disorders and inflammation [1]. G. glauca is widely distributed in Mexico [2]; however, scientific investigations of the phytochemical and pharmacological properties of this plant have been limited to populations growing in specific localities of the country. The first studies were conducted in natural populations growing in Doctor Mora, Guanajuato, and showed that the plant has anxiolytic and sedative activities in both mice [3] and humans [4]. The bioactive compounds were isolated and identified, and they correspond to a family of nor-secofriedelanes known as galphimines [5,6]. Subsequently, two metabolomic analyses were carried out in seven populations of the plant collected in the states of Chiapas, Jalisco, Morelos and Querétaro. The results of these investigations showed that only two populations produce galphimines [3,7] exhibiting anxiolytic and sedative activities in mice; however, all of them had anti-inflammatory activity in mice [7]. These results indicate that although these plants are morphologically similar, they are different in respect to their biological activity and metabolic profile; these differences may be observed due to divergent environmental conditions or because the plants consist of different botanical species.

An actual procedure to identify species of organisms is DNA barcoding, which includes a short segment of DNA from a standard and agreed-upon position in the genome of the nucleus and organelles [8]. Since the introduction of this method, DNA barcoding has been widely used in ecological, environmental and conservation studies [9-12] or forensic genetics [13,14], as well for the authentication of medicinal plants of different families [15-24]. From the Malpighiaceae family using molecular markers, only a small number of studies of phylogenetic analysis have been published [25-29].

In a previous study conducted by our research group, a DNA barcoding analysis was carried out in seven populations botanically classified as G. glauca by the HUMO Herbarium, CIByC (Centro de Investigación en Biodiversidad y Conservación), UAEM, Mexico. In that work, three genes, matK, rbcL, and rpoC1, were used for the molecular analysis, and the results suggested that the seven populations in the study may have belonged to at least three different species of the genus Galphimia [30].

In the present investigation, individuals from eight natural populations were collected and botanically classified as G. glauca by the HUMO Herbarium, CIByC UAEM, Mexico and the HGOM Herbarium, CIB (Centro de Investigaciones Biológicas), UAEH, Mexico. These specimens (six per population) were collected in different geographical locations in Mexico. To determine if all of the populations were correctly classified as G. glauca or if they belong to different species of the genus Galphimia, they were studied using six DNA barcode genes (matK, rbcL, and rpoC1, used in our first study; and psbA-trnH, ITS1 and ITS2, employed in this study for the first time). In addition, metabolic profiling was performed using thin-layer chromatography.

Materials and methods

Plant material

Plant material (six individuals per population) was collected in the summer of 2015 and 2016 (August and September) from eight different populations growing in five states of Mexico (Table 1): Doctor Mora, Guanajuato (GM); Jalpan de Serra (QJ) and Cadereyta (QC), Querétaro; Zimapán, Hidalgo (HZ); Cuernavaca (MC), Miacatlán (MM) and Santa Catarina (MS), Morelos; and Ciudad Valles, San Luis Potosí (SV). Specimens were deposited at the HUMO Herbarium, CIByC (Centro de Investigación en Biodiversidad y Conservación), UAEM, and at the HGOM Herbarium, CIB (Centro de Investigaciones Biológicas), UAEH, in Mexico. Leaves of G. glauca were kept at −80°C for DNA extraction. Similar populations from GM, QJ, MC and MM were studied in our previous investigation conducted six years prior [30].

Table 1

General data for the studied Galphimia glauca populations.

Population	Voucher (No.)	Locality	Date and time of collection	Position	Altitude (m)	Ecosystem
GM	15189	Doctor Mora, Guanajuato	August 14, 2015, 11:00–12:00 h	N 21.0874W 100.1922	2120	Semidry secondary vegetation
QC	35894	Cadereyta, Querétaro	September 10, 2015, 9:00–10:00 h	N 20.6991W 99.7250	2030	Semidry secondary vegetation
QJ	15018	Jalpan de Serra, Querétaro	September 10, 2015, 16:00–17:00 h	N 21.2850W 99.2857	1548	Pine-oak forest
MC	15011	Cuernavaca, Morelos	August 24, 2015, 10:00–11:00 h	N 18.5891W 99.1348	2204	Pine-oak forest
MM	15426	Miacatlán, Morelos	August 31, 2015, 11:00–12:00 h	N 18.45.57W 99.2175	1004	Seasonal-dry tropical forest secondary vegetation
MS	35896	Santa Catarina, Morelos	August 24, 2015, 12:00–13:00 h	N 18.9756W 99.1709	1600	Seasonal-dry tropical forest secondary vegetation
HZ	1483	Zimapán, Hidalgo	August 20, 2016, 12:00–13:00 h	N 20.9696W 99.3138	1999	Semidry secondary vegetation
SV	35895	Ciudad Valles, San Luis Potosí	September 17, 2016, 10:00–11:00 h	N 22.2225W 98.97.1	310	Seasonal-dry tropical forest secondary vegetation

DNA extraction

Genomic DNA was isolated from frozen leaves (10 mg) using the “Puregene DNA Purification” kit (Qiagen) following the manufacturer’s protocol. The DNA samples were dissolved (50 ng/μL) in T10E1 buffer and kept at −20°C before being used for PCR amplification. The quality and quantity of DNA were determined in a NanoDropTM 8000 spectrophotometer measuring the absorbance at 260 and 280 nm and by visual inspection after being run on an agarose gel.

Oligonucleotides

The oligonucleotides used for PCR amplification of matK, rbcL and rpoC1 genes were previously reported [28,30] (Table 2). The oligonucleotides for the ITS (including ITS1 and ITS2 genes) [18,31-36] and psbA-trnH [31,37-42] regions were selected by an exhaustive search of the previously reported studies investigating angiosperm plants using these markers and by using the BLASTn routine in GenBank database (http://www.ncbi.nlm.nih.gov/genbank/) (Table 2).

Table 2

Oligonucleotide sequences for DNA barcodes.

Gene	Sequence forward	Sequence reverse
matK	GAG GGG TTT GCA GTC ATT GT	CCA ATG ACC CAA TCA AAG GA
rbcL	GAA GGG TCT GTT ACT AAC ATG	TCC CCT TCA AGT TTA CCT AC
rpoC1	GTG GAT ACA CTT CTT GAT AAT GG	CCA TAA GCA TAT CTT GAG TTG G
ITS	GTC CAC TGA ACC TTA TCA TTT AG	TCC TCC GCT TAT TGA TAT GC
psbA-trnH	GTT ATG CAT GAA CGT AA TGC TC	CGC GCA TGG TGG ATT CAC AAT CC

PCR amplification, purification and sequencing

For all six genes, PCR amplification was performed in a Robocycler Gradient 96 (Stratagene) using the following program: a first step of 4 min at 94°C followed by 35 cycles of 1 min at 94°C, 1 min at 53°C, and 1 min at 72°C. A final extension of 10 min at 72°C was performed at the end. PCR analyses were performed in a final volume of 50 μL using 25 μL of Go Taq Green Master Mix, 19 μL of nuclease-free water, 2 μL of each oligonucleotide, and 2 μL of DNA sample. The PCR products were observed in an agarose gel (1%) to verify the presence of unique amplicons. The PCR products were purified using “QIAquick PCR Purification” (Qiagen) following the manufacturer’s instructions. To quantify the PCR product, 2 μL was used in a NanoDropTM 8000 spectrophotometer, which measured the absorbance at 260 and 280 nm. The final concentration of the amplicons was adjusted to 50 ng/μL before the sequencing process.

All PCR products were sequenced in both directions. For sequencing, 1.4 μL of DNA and 1.2 μL of each primer were added to 1 μL of BigDye 3.1 reaction mixture (Applied BioSystems). To complete the reaction volume to 6.4 μL, 1 μL of sequencing buffer was added. The reaction cycle was performed at 96°C for 10 sec, 50°C for 5 sec and 60°C for 10 min. After the reaction cycle, water was added to each tube to complete the volume to ~25 μL. The reactions were purified in Sephadex G50 superfine 96-well filter plates. The collected samples were run on an Applied Biosystems 3730XL capillary instrument (DNA Sequencing Facility, University of Chicago Cancer Research Center). All sequences were visually inspected and edited using Chromas 2.6.5 (Technelysium Pty Ltd). The six forward and six reverse sequences were used to obtain a consensus sequence for each marker in each population. All consensus sequences were deposited in GenBank, and accession numbers were obtained (Table 3).

Table 3

Accession number of the consensus sequences deposited in GenBank.

Population	matK	rbcL	rbcL	ITS	psbA-trnH
GM				MH842122	MK805053
HZ	MK033195	MK033203	MK033199	MH842123	MK805054
MC				MH842124	MK805055
MM				MH842125	MK805056
MS	MK033196	MK033205	MK033201	MH842126	MK805057
QC	MK033197	MK033204	MK033200	MH842127	MK805058
QJ				MH842128	MK805059
SV	MK033198	MK033206	MK033202	MH842129	MK805060

DNA barcoding analysis

Distance-based method

For all six DNA barcodes, the genetic variability using the DnaSP6 program [43] was determined. The alignment and sequence average lengths, the number of sites with gaps, the number of polymorphic and parsimony informative sites, the nucleotide diversity and the number of haplotypes were considered for this analysis. Moreover, the intraspecific and interspecific distances were obtained using Kimura’s 2-parameter (K2P) nucleotide evolution model [44] in the MEGA 7 program [45].

Phylogenetic analysis

For the phylogenetic analysis, we used all the available sequences in the GenBank database (http://www.ncbi.nlm.nih.gov/genbank/) for the genus Galphimia, including all six markers in study, as well as sequences reported previously for G. glauca [30]. Sequences from other phylogenetically closely related species and genera, as well as not closely related genera, were also selected for phylogenetic out groups (Table 4).

Table 4

Accession number of sequences used in the phylogenetic study.

Gene	Accession number	Species
matK	JX088054	Galphimia glauca Cav.
	JX088053	Galphimia glauca Cav.
	JX088058	Galphimia glauca Cav.
	JX088055	Galphimia glauca Cav.
	AB233800	Galphimia glauca Cav.
	HQ247275	Galphimia brasiliensis (L.) A.Juss.
	MF349853	Galphimia gracilis Bartl.
	HQ247276	Galphimia mirandae C.E.Anderson
	HQ247277	Galphimia speciosa C.E.Anderson
	JQ588187	Byrsonima crassifolia (L.) Kunth
	KU556672	Silybum marianum (L.) Gaertn.
	KX677063	Vaccinium oxycoccos L.
rbcL	JX125620	Galphimia glauca Cav.
	JX125622	Galphimia glauca Cav.
	JX125624	Galphimia glauca Cav.
	JX125621	Galphimia glauca Cav.
	AB233904	Galphimia glauca Cav.
	KM197480	Galphimia australis Chodat
	HQ247487	Galphimia brasiliensis (L.) A.Juss.
	AF344475	Galphimia gracilis Bartl.
	HQ247488	Galphimia multicaulis A.Juss.
	HQ247489	Galphimia speciosa C.E.Anderson
	JX664036	Byrsonima crassifolia (L.) Kunth
rpoC1	JX125630	Galphimia glauca Cav.
	JX125628	Galphimia glauca Cav.
	JX125625	Galphimia glauca Cav.
	JX125627	Galphimia glauca Cav
	FJ038768	Bunchosia sp. MAG-2009
	GQ429092	Byrsonima crassifolia (L.) Kunth
	KC737454	Mytilaria laosensis Lecomte
	KC481618	Oxalis acetosella L.
ITS1 and ITS2	KR087554	Galphimia australis Chodat
	MF349119	Galphimia gracilis Bartl.
	AY137299	Acridocarpus macrocalyx Engl.
	AF436780	Aspidopterys elliptica A.Juss.
	KR087517	Byrsonima crassifolia (L.) Kunth
	KP675783	Melia azedarach L.
	AJ012364	Nicotiana tabacum L.
psbA-trnH	GQ982369.1	Spachea membranacea Cuatrec.
	KY027037.1	Byrsonima coccolobifolia Kunth
	MH826615.1	Elaeocarpus amplifolius Schltr.
	AB492621.1	Mitella furusei Ohwi

The sequences were aligned using Clustal W [46] with default parameters, and the phylogenetic analysis was performed in MEGA 7.0 [45]. For each marker, we generated separate Neighbor-joining (NJ) as well as Maximum Likelihood (ML), and Maximum Parsimony (MP) phylogenetic trees (S1 Appendix) based on the K2P nucleotide evolution model [44], data-specific model [47] and close-neighbor-interchange algorithm [47], respectively. The sequences for all six markers were combined with a total length of 2964–3022 bp, and phylogenetic analysis was performed using the NJ method. Node support was assessed via 1000 bootstrap replicates. The resulting phylogenetic trees were edited using the iTOL program (http://itol.embl.de/) [48]. We employed the NJ, ML, and MP methods because they have superior computationally intensive analysis in comparison to other methods and to confirm that the choice of phylogenetic method algorithms did not change DNA barcode results.

Thin-layer chromatography

Methanolic extracts were prepared for all eight populations (six individual extracts for each population). The samples (leaves) were dried for 3–4 days in a cool and dry place without direct sunlight. Samples were then pulverized by mortar and pestle. Pulverized dried material (100 mg) for each sample was mixed with 1 mL MeOH. Samples were vortexed for 2 min, sonicated for 15 min and then centrifuged at 10,000 rpm for 15 min. The supernatant was removed, and the material residue containing the pellet was reprocessed four times to achieve exhaustive extraction. All supernatants were combined and applied directly in thin-layer chromatography (TLC) plates, which were eluted using a mobile phase of CHCl3:CH3COOCH2CH3 (2:1 v/v). TLC plates (silica gel 60 UV254 precoated, 10 × 10 cm, 200 μm layer thickness, aluminum-backed, mean particle size of 10–12 μm, particle size distribution of 5–20 μm) were stained with a spray solution containing vanillin/H2SO4 (1 g of vanillin in 100 mL of H2SO4) and heated at 120°C until maximum color formation.

Results

Eight populations botanically classified as G. glauca (six individuals per population) were analyzed using a molecular approach by DNA barcodes and phytochemical analysis by TLC. Individuals from four of these populations (GM, MC, MM and QJ) were studied previously by our group [30]. In this new research, we considered these four populations because three new DNA barcodes (psbA-trnH, ITS1 and ITS2) were introduced in the molecular analysis and to compare the actual chemical profiles by TLC analysis of individuals collected in the four abovementioned localities six years previously.

DNA barcode features and distance-based methods

In this study, we used six different molecular markers (matK, rbcL, rpoC1, psbA-trnH, ITS1 and ITS2) that have been previously reported with high levels of species identification in angiosperm [34,49-51]. For each marker, all six individuals in each population showed 100% identical DNA sequences, with the exception of psbA-trnH, which showed between 2 and 4 nonspecific nucleotides in the sequences. The sequences for matK, rbcL and rpoC1 that were obtained in our previous work [30] were included in this work to make an accurate analysis and comparison of the genetic variability of the G. glauca populations in study. The lengths of the DNA fragments that were generated were 793–829, 669, 599, 433–449, 256–258 and 214–218 bp for matK, rbcL, rpoC1, psbA-trnH, ITS1 and ITS2, respectively. The sequences were similar among members of each population, but single nucleotide polymorphisms were observed among all six markers.

The largest sequences of all DNA barcodes were obtained for matK, with median and alignment lengths of 797.5 and 829 bp, respectively. The shortest sequences were recorded for the ITS2 marker, with median and alignment lengths of 216.8 and 221 bp, respectively. For the rbcL barcode, smaller values of variability were obtained without any parsimony informative site and a nucleotide diversity of 0.003. In the same way, matK and rpoC1 showed low genetic variability with only 2 parsimony informative sites and a nucleotide diversity of 0.004 and 0.003, respectively (Table 5). The intergenic regions psbA-trnH, ITS1 and ITS2 were the most variable DNA barcodes in this study. In particular, the psbA-trnH intergenic spacer showed the highest values of polymorphic (equal to 48) and parsimony-informative (equal to 18) sites. The spacer regions ITS2 exhibited the highest value of nucleotide diversity (equal to 0.055). Conversely, psbA-trnH, ITS1 and ITS2 markers presented the largest number of haplotypes (equal to 4) (Table 5).

Table 5

Molecular features of Galphimia populations DNA barcodes.

Barcode	Average sequence length(bp)	Sequence alignment length(bp)	No. of sites with gaps	No. of polymorphic sites	No. of parsimony informative sites	Nucleotide diversity	No. of haplotypes
matK	797.5	829	36	11	2	0.004	3
rbcL	669	669	0	9	0	0.003	3
rpoC1	599	599	0	5	2	0.003	3
ITS1	257.8	261	8	34	10	0.044	4
ITS2	216.8	221	10	40	7	0.055	4
psbA-trnH	443	475	58	48	18	0.042	4

The intraspecific distances were evaluated considering the botanical classification as G. glauca of all populations in study. For the interspecific distances, the comparison was performed between the populations of G. glauca in the study and the species selected as out groups for the analysis (Fig 1). In the case of the intraspecific distances, matK, rbcL and rpoC1 showed low values, that is, below 0.0127. However, psbA-trnH, ITS1 and ITS2 exhibited high values of intraspecific distances, ranging from 0.0616–0.172, 0.0399–0.111 and 0.0427–0.1826, respectively (Fig 1).

Fig 1

Frequency histograms of the pairwise intra- and interspecific distances for barcodes matK, rbcL, rpoC1, psbA-trnH, ITS1 and ITS2 and for all combined barcodes obtained for Galphimia populations.

The alignments for all six markers were developed, and three bootstrap consensus trees were constructed for each marker by means of NJ, ML, and MP methods. All phylogenetic trees showed strongly supported clades with high bootstrap values. Additionally, the bootstrap consensus trees obtained by the three methods were highly similar for each of the six DNA barcodes studied. The bootstrap values observed by ML and MP (S1 Appendix) were also highly similar with the trees of NJ analysis (Figs 2–8). The ML and MP methods support the results obtained by the NJ method.

Fig 2

Bootstrap consensus tree generated by the Neighbor-joining method for matK sequences obtained for Galphimia populations.

Numbers below the branches are bootstrap values expressed as a percentage of 1000 replicates.

Fig 8

Comparison of Neighbor-joining tree for Galphimia populations analyzed in this study based on matK, rbcL, rpoC1, psbA-trnH, ITS1, and ITS2 combined sequences and TLC profiles of methanolic extracts of individuals from these populations.

The TLC mobile phase was CH3COOCH2CH3-CHCl3 (2:1). The individuals of each population are numbered (1–6). The control (C) is a fraction of galphimines that was obtained from individuals of GM. The presence of galphimines is indicated with a black frame.

Three and two major clades were formed for matK, rbcL, and rpoC1 (Figs 2–4), and for psbA-trnH, ITS1 and ITS2 (Figs 5–7) genes, respectively. In the case of matK, four well-supported sister clades enabled us to separate the populations of GM, HZ, QC, and QJ from the populations of MC, MM, and MS. Furthermore, for the matK gene, the population of SV separated independently from the rest of the populations analyzed in this study (Fig 2). Similarly, for rpoC1, four well-supported sister clades showed discrimination between the populations collected in GM, HZ, QC, and QJ from the populations of MC, MM, and MS, and once again, the population of SV was kept apart in a different group from the rest of the studied populations (Fig 4). In accordance with the previous results, psbA-trnH allowed the segregation of the populations of GM, HZ, QC, and QJ from the populations of MC, MM, and MS, as well as the independent separation of the SV population (Fig 5). The ITS1 and ITS2 markers presented high rates of polymorphism, and once more, it was possible to discriminate among the populations of MC, MM, MS, and SV from the populations that came from GM, HZ, QC, and QJ (Figs 6 and 7). The phylogenetic analyses of ITS1 and ITS2 also showed that SV discriminated in an independent group (Figs 6 and 7). The rbcL gene did not show sufficient variation to differentiate among most of the studied populations but enabled us to discriminate one of them (SV) (Fig 3).

Fig 4

Bootstrap consensus tree generated by the Neighbor-joining method for rpoC1 sequences obtained for Galphimia populations.

Numbers below the branches are bootstrap values expressed as a percentage of 1000 replicates.

Fig 5

Bootstrap consensus tree generated by the Neighbor-joining method for psbA-trnH sequences obtained for Galphimia populations.

Numbers below the branches are bootstrap values expressed as a percentage of 1000 replicates.

Fig 7

Bootstrap consensus tree generated by the Neighbor-joining method for ITS2 sequences obtained for Galphimia populations.

Numbers below the branches are bootstrap values expressed as a percentage of 1000 replicates.

Fig 6

Bootstrap consensus tree generated by the Neighbor-joining method for ITS1 sequences obtained for Galphimia populations.

Numbers below the branches are bootstrap values expressed as a percentage of 1000 replicates.

Fig 3

Bootstrap consensus tree generated by the Neighbor-joining method for rbcL sequences obtained for Galphimia populations.

Numbers below the branches are bootstrap values expressed as a percentage of 1000 replicates.

Thin-layer chromatography analysis

The TLC profiles of Galphimia populations showed differences between them (Fig 8). Using this separation technique, it was possible to distinguish four principal groups: the galphimine-producing (marked with a black frame) populations (GM, HZ, QC and QJ) (group 1) and the nongalphimine-producing populations, which were separated in MM and MS (group 2), MC (group 3) and SV (group 4) (Fig 8). The bootstrap consensus tree based on matK, rbcL, rpoC1, psbA-trnH, ITS1 and ITS2 combined sequences presented four sister clades that corresponded with the four different groups observed in the TLC chemical profiles of the populations investigated in this study (Fig 8).

Discussion

The analysis of the genetic variability revealed that rbcL was the less polymorphic barcode, exhibiting no sites with gaps in the alignment and parsimony informative sites and exhibiting the lowest number of mutation and haplotypes. These results were similar to those previously obtained by our group [30] and to the findings of other studies developed in woody plant species from Brazil [37].

The values for the intraspecific distances for matK and rpoC1 were low but enabled us to visualize a small differentiation between the populations that produce galphimines (GM, HZ, QC, and QJ) from those that do not produce these compounds (MC, MM, MS, and SV). These results are also consistent with the three haplotypes presented for matK and rpoC1. In addition, the psbA-trnH region showed three different haplotypes, which corresponded with the populations that produced galphimines (GM, HZ, QC, and QJ) and nonproducers that separated into two groups: one was integrated by the populations from Morelos (MC; MM and MS), and the other was composed only of the SV population.

Conversely, the psbA-trnH region presented high values of intraspecific distances that strongly suggest that the populations in the study belong to more than one species of the genus Galphimia. For the ITS1 and ITS2 intergenic regions, the results of the genetic variability analysis showed four haplotypes, and the high interspecific distances among the sequences of these markers suggest the presence of four different species of the genus Galphimia separated into four groups: the galphimine-producer populations [GM, HZ, QC, and QJ] and the nonproducers, which can be separated into three groups: [MC], [MM and MS] and [SV]. The results for psbA-trnH, ITS1 and ITS2 were expected because these intergenic regions have been reported as highly polymorphic barcodes in angiosperm [31,37-39], including medicinal species [18,32,52-54].

The lower values for the intraspecific distances displayed by the combination of all six DNA barcodes compared with the three analyzed intergenic regions are the result of small intraspecific distances of matK, rbcL and rpoC1 markers. However, the results obtained for the combination of all six markers also support the presence of four different species of the genus.

Numerous DNA sequences of matK and rbcL deposited in GenBank for the genus Galphimia were incorporated in this analysis with the objective of investigating the relationship with the samples described in this study. In the case of both rpoC1 and psbA-trnH, we could not find any sequence for the genus Galphimia. For ITS1 and ITS2, we found one and two sequences of the genus Galphimia deposited in the GenBank database, respectively. Moreover, other sequences that belong to species related and unrelated to Galphimia were included in the phylogenetic analysis to construct out groups for all six markers.

Most of the sequences of matK and rbcL for the genus Galphimia were reported previously [29], but some plants studied in that investigation did not receive an accession number and were not available in GenBank. The accession numbers in GenBank for G. glauca were reported by Tokuoka and Tobe [28] before the botanical classification for the genus Galphimia was published by Anderson [2]. The taxonomic classification as G. glauca related to these two sequences was made in the Kyoto Botanical Garden [28]. It is possible that this analysis was misinterpreted with other species of the genus because the population growing in Dr. Mora, Guanajuato (GM), was classified as G. glauca [2], but in this work and in a previous study developed by our group [30], it was demonstrated that the GM population is genetically different from G. glauca sequences reported formerly [28] for both matK and rbcL genes (Figs 2 and 3). The samples from SV grouped individually from the rest of the populations in study for matK and rbcL sequences (Figs 2 and 3). These samples were more closely related to G. gracilis for these two markers. In addition, SV was the population that presented the highest rate of polymorphisms among all of the populations in study. The results of the analysis of the matK, rbcL and rpoC1 genes of this investigation corroborate our previous study in which matK and rpoC1 were the markers that enabled us to discriminate among the galphimine producer populations and suggest the presence of three different species of the genus Galphimia.

In this work, we report the phylogenetic analyses of psbA-trnH, ITS1 and ITS2 sequences in G. glauca for the first time. These barcode sequences show a high level of DNA variation among the eight natural populations of G. glauca that we analyzed. The high level of variability observed for these three markers was expected because spacer DNA sequences, as noncoding DNA regions, show low selection pressure and exhibit a high level of mutation rates [55,56]. The phylogenetic trees that were obtained by NJ, ML, and MP based on psbA-trnH, ITS1 and ITS2 also allowed the discrimination of the galphimines producing populations (GM, HZ, QC, and QJ) from the rest of the populations, which do not produce galphimines (MC, MM, MS, and SV). With the use of ITS1 and ITS2 sequences, for the first time, it was possible to discriminate two groups from the populations growing in Morelos: group 1 is formed by the MC population, and group 2 comprises MM and MS populations (Figs 6 and 7). This result suggests that both MS and MM belong to the same species of the genus Galphimia. Although the psbA-trnH barcode was highly polymorphic, it did not enable the differentiation of the populations growing in the state of Morelos.

DNA barcoding is a tool that allowed us to distinguish among the specimens of the genus Galphimia, as in the previous work of our group [30]. The use of the DNA barcodes ITS1 and ITS2 was useful to show a genetic variability that had not been identified previously.

TLC is a simple, reproducible and easy technique that represents a useful strategy to analyze chemical profiles. We used TLC to correlate the molecular results with the chemical profiles that were obtained in this study. In previous studies, it has been demonstrated that metabolic profiling is a reliable approach to study the chemical profiles of Galphimia populations [30], suggesting that galphimines could be a chemotaxonomic marker for G. glauca species. The variation in environmental factors is tightly related to the differential production of metabolites in plants. However, the genetic information variability is the primary cause of any phenotypic variation in organisms; consequently, the genetic variation that was demonstrated in this study could be the answer to the production of galphimines in four (GM, HZ, QC, and QJ) of the eight populations studied and correlates perfectly with the chemical profiles that were obtained by TLC (Fig 8). Potentially, these differences are related to the different genetic information among the populations described in this study. Similar results were found in our previous research [30].

The TLC profiles enabled us to differentiate among the populations of Morelos, MM and MS in contrast with MC. The first two populations presented a group of blue-colored compounds, revealed by vanillin/H2SO4, that were absent in the samples from the MC population. The differences in the production of metabolites of the populations that we studied could be related to the diverse ecological regions from which they originated, but the molecular analysis also demonstrated that these differences could be due to different genetic information between these populations. Genetic and environmental factors are closely related, and the interaction between them modulates the response as a particular phenotype of all living organisms.

Integration of phylogenetic and thin-layer chromatography analysis

The bootstrap consensus tree based on matK, rbcL, rpoC1, psbA-trnH, ITS1 and ITS2 combined sequences was compared with the TLC chemical profiles of the populations described in this study studied (Fig 8). The chemical profiles obtained by TLC were uniform, stable and highly reproducible. DNA barcode analysis was well supported by TLC profiles of these plant samples (Fig 8). Thus, TLC was effective in discriminating among populations that produce galphimines, populations with blue-colored compounds, and those that do not produce these two groups of metabolites. The different concentrations of the same metabolites among individuals of the same population are a consequence of each individual’s ability to respond differently to environmental changes. The phylogenetic analysis and the TLC chemical profiles evidently matched. The genotype analysis in the phylogenetic tree presented four sister clades that corresponded with four different phenotypes in the TLC chemical profiles. These results were similar to those obtained in the previous investigation performed by our group [30].

Phylogenetic analysis and TLC chemical profiles were convenient tools to establish a strong relationship between the genotype and phenotype of the populations in study. The phylogenetic analysis of all six DNA barcodes analyzed in this study suggests that the populations in this study belong to at least three different species of the genus Galphimia. This analysis would be complemented with an ongoing morphological investigation, as well as cytogenetic studies, to establish a powerful and integrative methodology for quality control purposes that allow the preparation of effective herbal medicines from plants of this genus.

Phylogenetic trees generated by maximum likelihood and maximum parsimony methods.

(PDF)

Click here for additional data file.

Herbarium specimens of the natural population in study botanically classified as Galphimia glauca.

(PDF)

Click here for additional data file.

matK sequence alignment.

(XLS)

Click here for additional data file.

rbcL sequence alignment.

(XLS)

Click here for additional data file.

rpoC1 sequence alignment.

(XLS)

Click here for additional data file.

psbA-trnH sequence alignment.

(XLS)

Click here for additional data file.

ITS1 sequence alignment.

(XLS)

Click here for additional data file.

ITS2 sequence alignment.

(XLS)

Click here for additional data file.