This is the first report of the efficiency of six molecular markers as potential DNA barcode for the evaluation of their genetic diversity and the identification of ethnovarieties of Banisteriopisis spp. used for the preparation of ayahuasca in the Amazon, in the religious context of CEBUDV. Combining the results of the genetic diversity, the ability to differentiate groups of ethnovarieties in the phylogenetic trees, and patterns of genetic divergence. Our results support the designation of ITS as the main barcode DNA locus for the identification of ethnovarieties of Banisteriopisis spp. In addition, the inclusion of at least one of the three analyzed plastid markers (matK, rbcL or ndhF) as an additional locus is recommended since the combination of nuclear and plastid information gives more reliability to the dataset. They did enable the confirmation of the genus Banisteriopsis, although plastid markers do not increase the taxonomic power of identification of species and ethnovarieties.
Genetic diversity and BLAST N
The nuclear markers presented the highest rates of genetic diversity when compared to plastid markers, which is a fact that has been observed for over 54 families of Amazonian angiosperms such as Sapotaceae and Rubiaceae (Gonzalez et al. 2009; Mezzasalma et al. 2017). In forty-nine genera of Amazonian angiosperms, the highest rates of genetic diversity were observed with the ITS (Gonzalez et al. 2009). In eighteen samples of piri-piri (Cyperus spp.), which are also used in traditional Amazonian medicine, eight haplotypes have been observed for the ITS (Mezzasalma et al. 2017). High genetic diversity was observed with 14 haplotypes, for the 15 and 16 samples of Banisteriopsis spp. that were sequenced in the present study for yage5 and ITS, respectively, although we had achieved 80% (16 samples) sequencing efficiency for ITS and 75% (15 samples) for yage5. The difficulty in sequencing for ITS has already been observed in other studies (Gonzalez et al. 2009).
Some regions of the plastid genome have relatively slow rates of nucleotide replacement (Clegg et al. 1994), which may have favored the results of lower indices of genetic diversity found in the analyzed samples (Table 2). Despite the plastid markers matK and rbcL being recommended as DNA barcode for plants (Hollingsworth et al. 2009), they were highly conserved in the 20 samples of Banisteriopsis spp vines. The three plastid markers analyzed (matK, rbcL, and ndhF) presented only three polymorphic sites and four haplotypes (Table 2) and, in this sense, were not variable enough to discriminate the ethnovarieties. In addition, the comparison of plastid markers in the NCBI sequence database with the 19 species of Banisteriopsis identified at the morphological level by Davis and Anderson (2010) presented a percentage of identity ranging from 99–100% for all 19 species (Table S3, S4, and S5). Therefore, for Banisteriopsis spp., the plastid markers on their own are not variable enough to separate and identify species.
On the other hand, for the nuclear marker ITS, the BLAST N analysis resulted in a percentage of identity ranging from 85.33–95.43% after comparison with the only accession of the ITS for B. caapi (Table S2). These molecular identification applications depend on the existence of complete reference libraries of DNA sequences with which new sequences can be compared to perform taxonomic identification (De Lima et al. 2018). If a reference library has poor coverage in terms of species with sequences already available, the implementation of DNA-based identification techniques will remain inefficient. Thus, this work presents an important contribution with new accessions of nuclear and plastid sequences of Banisteriopsis spp. for the NCBI database, especially when we consider Banisteriopsis spp.
Testing the discrimination of ethnovarieties based on phylogenetic trees and distance methods
The three clades (A, B, and C) that were differentiated by the molecular marker ITS are represented by lineages of the three main ethnovarieties of Banisteriopsis spp.; tucunacás, caupuris, and pajezinho, respectively (Fig. 2). These ethnovarieties are recognized by traditional knowledge in the context of CEBUDV, but they can also be recognized by other names when we consider the indigenous context and other religions.
In clade A, it was possible to identify two lineages of Banisteriopsis spp. of the tucunacá ethnovariety, one is a lineage cultivated (I) in the Mestre Angílio nucleus in Manaus and the other a native lineage (II) of Presidente Figueiredo. The only accession of B. caapi available in the NCBI for ITS presented the highest percentage values of identity concerning specimens of lineage II, from the native tucunacá ethnovariety (ranging between 95.24% and 95.43%; Table S2).
The Banisteriopsis spp. that were identified in lineage III and IV (clade B) are ethnovarieties of caupuri with nodes and without nodes and presented the highest phenotypic plasticity of the stem (Fig. 2d and Fig. S1), and all were collected in cultivations in Manaus. Of the eight samples (lineage III), four presented the phenotype with nodes and the other four without nodes, corroborating the traditional knowledge that recognizes the caupuri ethnovariety by this characteristic (with nodes and without nodes). The caupuri with nodes was recognized and introduced to the CEBUDV in the 1960s. At this time, marked by the beginning of the CEBUDV in the city of Manaus, Mestre Florêncio Carvalho, who was a rubber tapper in the Amazon Rainforest for many years, carried out some searches for Banisteriopsis spp. in the forests of the banks of the Negro River, where the caupuris (with nodes) were encountered.
Clade C presented the greatest amount of lineages (V, VI, and VII) and these are represented by the ethnovariety identified as pajezinho, which is two natives (V and VI) and one cultivated lineage (VII). Pajezinho is recognized for having one of the strongest effects in ayahuasca after its preparation (Fig. 2). As such, the pajezinho ethnovariety can be confused with the tucunacá ethnovariety because neither possess nodes. However, the lineages of clade C were genetically differentiated from the others, including within the same clade (Table 3), in which it is possible to observe that the cultivated lineage (VII) resulted in a genetic distance of 3.02 and 2.61% for the native lineages V and VI, respectively.
In addition to the phenotype of the vine, which aids in the identification of ethnovarieties, the quality of the visions under the influence of ayahuasca tea is often described as an important characteristic for identification (Schultes 1986). Therefore, we must emphasize that although the phenotype of the vine is very important for the identification of ethnovarieties, these characteristics should be used with caution. When we specifically consider the phenotype of the vine, the ethnovarieties of native lineages (lineages II, V, VI) have phenotypes that are similar to each other (Fig. S1), but are not recognized as being the same ethnovariety by members of the CEBUDV (Table S1), and are genetically different from each other.
Thus, the results presented partially unravel the enigma described by Schultes (1986), who reported the difficulty of recognizing morphological differences as a way of separating ethnovarieties from each other. In our genetic analyses, all known ethnovarieties in the traditional context of the CEBUDV were separated by genetic analyses based on the nuclear molecular marker ITS. However, the comparison of our results with analyses of ethnovarieties known in other religions and traditional peoples is important. Since the revision of the genus Banisteriopsis, there are reports of the difficulty in differentiating species morphologically, due to the morphological floral characters being conserved (Gates 1982). Therefore, part of the botanical riddle described by Schultes (1986) remains.
All the vines analyzed are used in the preparation of ayahuasca tea, although we still do not have morphological evidence that all the samples analyzed in fact belong to the botanical species B. caapi. However, the phenotype of the with nodes stem of two samples (FN2 and MA1) of lineage III (clade B) is similar to the phenotype of the ethnobotanical specimen of B. caapi (formerly Banisteria caapi Spruce ex Griseb.) held in the Economic Botany Collection, at the Royal Botanic Gardens, Kew (EBC 67428), which was collected in the Amazon by Richard Spruce in 1852 (Nesbitt 2014). These observations lead us to believe that representatives of clade B do in fact they can belong to the botanical species B. caapi.
The genetic distance thresholds, resulting from the ITS marker, between the lineages and ethnovarieties identified for the Banisteriopsis spp. vines were larger than we could expect within a single species and presented values of up to 28.73% (Table 3) of genetic distance (mean 14.59%). The assumption for the effectiveness of barcode DNA is that the genetic distances within the species are always smaller than the genetic distances between species, which is called the barcode gap. It is critical to determine these ranges of genetic distances between species of different taxonomic groups (Hebert et al. 2003). However, the genetic distances found in the present study for the vines used in the preparation of ayahuasca tea (Table 3) may indicate that more than one species of Banisteriopsis is used, and not only B. caapi, as most references mention (Gates 1982; Schultes 1984; Schultes et al. 1992). It is evident that integrative taxonomy studies are necessary to test this hypothesis.
For the piri-piri, another Amazonian species, which before was thought to be only a Cyperus articulatus species used in the traditional context, four species were identified, with genetic distance values ranging from 2.7 to 4.7% (Mezzasalma et al. 2017). The genetic distance with the ITS gene for tropical trees of 49 families of angiosperms was 6.23%, which is a value that is lower than the maximum found in the present study (28.73%) for the genus of Amazonian liana (Banisteriopsis).
Although in the results we did not present combinations of two markers (nuclear and plastid) in a single barcode to discriminate the ethnovarieties, we observed that this did not improve the overall performance when compared to single marker (ITS), which was also observed for other Amazonian angiosperms (Gonzalez et al. 2009).