Four raised to one equals one: A genetic approach to the Pseudolaelia vellozicola complex does not follow a math rule

Abstract Pseudolaelia is a genus endemic to the eastern Brazilian Atlantic Forest, consisting of 12 accepted species. Some Pseudolaelia species, such as P. vellozicola, P. aguadocensis, P. oliveirana, and P. regentii, referred to here as the PV complex, present extensive intra‐ and interpopulation morphological polymorphism, raising uncertainty regarding their circumscriptions. Although previous morphological analyses were used to solve the generic boundaries in the PV complex, persuasive genetic evidence is lacking. In order to test the hypothesis that the group under investigation contains only one taxon, amplification profiles of five intersimple sequence repeat (ISSR) markers were used to evaluate genetic diversity, genetic structure, and the relationships among the PV complex species. A total of 134 reproductive individuals were sampled in eight insular populations. Intrapopulation genetic analysis indicated low levels of genetic diversity. Analysis of genetic structure revealed that each of the eight sample locations can be considered unique biological populations as they are highly differentiated from each other. The Mantel test showed a high and positive correlation between genetic and geographic distance (r = .841, p < .002), indicating isolation by distance. The results are consistent with that expected for plants with insular geographical distribution. When testing for the null hypothesis, the low levels of genetic variation among species (F CT = 0.155) suggest that the populations constitute only one highly polymorphic species with a wide distribution.

species of the Velloziaceae family), growing in granitic and gneissic outcrops (inselbergs) of the Brazilian Atlantic Forest biome and quartzite outcrops of the Cerrado and Caatinga grasslands (Menini Neto, 2011). Pseudolaelia is a monophyletic genus that is part of a basal clade of the subtribe Laeliinae called Isabelia Alliance (van den Berg et al., 2009). Besides small population sizes and habitat specificity, all Pseudolaelia species are threatened to some degree due to habitat degradation caused by mining and uncontrolled tourism (Menini Neto, 2011).
The orchid genus Pseudolaelia is characterized by homoblastic, fusiform, or piriform pseudobulbs separated by a long rhizome, with cataphylls that disintegrate into fibers. Its indeterminate inflorescence is long, simple, or compound, with pink, yellow, white, or cream maculate or concolorous flowers. Flowers present a simple labellum or are often trilobate, sometimes entire, with entire or erose margins, and the cuniculus present (Menini Neto, 2011).
Although the majority of species morphologies for Pseudolaelia are quite homogeneous, a taxonomic revision based on analyses of herbarium specimens and field studies for the genus revealed remarkable floral polymorphism for Pseudolaelia vellozicola (Hoehne) Porto & Brade populations (Menini Neto et al., 2013). These results, along with recent descriptions of some very similar taxa, viz. Pseudolaelia aguadocensis Campacci (Campacci, 2016), Pseudolaelia regentii V.P.Castro & Marçal (Castro Neto & Marçal, 2007), and Pseudolaelia oliveirana V.P. Castro & Chiron (Castro Neto & Chiron, 2009), have made the taxonomical definition of P. vellozicola species more complex. However, Menini Neto, Berg, and Forzza (2019) carried out a morphological study to better understand if the morphological variability found in P. vellozicola is due to its insular distribution (which would synonymize P. vellozicola, P. aguadocensis, P. oliveirana, and P. regentii), or whether they are different species constituting a complex. Despite the large interpopulation floral variation (that may originate from the relative geographic isolation among insular populations), Menini Neto et al. (2019) reported that there are no strong morphological discontinuities among the populations of P. vellozicola, P. aguadocensis, P. oliveirana, and P. regentii. The species hypothesis resulting from this previous study (Menini Neto et al., 2019) indicated that the morphological differences are insufficient to separate these taxa, suggesting that the PV complex consists of only one species.
Intersimple sequence repeat (ISSR) markers are an extremely useful genetic tool that can be used in taxonomic comparisons and to determine intra-and interpopulation levels of genetic variability.
Furthermore, ISSR patterns are species-specific and, as such, enable the identification of profiles in related plant species (Bozchaloyi, Sheidai, Keshavarzi, & Noormohammadi, 2018;Fajardo, Vieira, & Molina, 2014;Zietkiewicz, Rafalski, & Labuda, 1994). In species with unclear taxonomical definitions, as with some Pseudolaelia species, the use of ISSR markers can help to solve this puzzle. Given the taxonomic challenge of distinguishing P. aguadocensis, P. oliveirana, P. regentii, and P. vellozicola, this study sought to test the hypothesis that the group under investigation contains only one taxon. For this, amplification profiles of five intersimple sequence repeat (ISSR) markers were used to evaluate genetic diversity, genetic structure, and the relationships among the studied populations and species of the PV complex.

| Study area, population sampling, and DNA extraction
Pseudolaelia occurs in a particular class of residual landforms called inselbergs, which are isolated rock outcrops (height > 100 m) that rise abruptly above the surrounding grasslands (Lima & Corrêa-Gomes, 2015). Inselbergs may be isolated or grouped, separated by only a few or many kilometers across the landscape. Ecologically, they are characterized as having harsh environments, including a high degree of insolation, high rates of evaporation, local and very restricted soil occurrence, and water scarcity, all of which favor the development and occurrence of a large number of endemic species (Porembski, 2007).
Due to its endemism and rarity, eight locations were sampled across the entire geographic distribution of the PV complex ( Figure 1). In order to avoid taxonomic identification errors, the sampling included the same localities that were previously used in a taxonomic revision of the genus (Menini Neto et al., 2013). A total of 134 reproductive individuals of the PV complex were sampled, and the sample size at each locality varied according to the characteristics of the population (Table 1). Specifically, the sampling included one locality (AgD) related to P. aguadocensis, two localities (AgB and Col) related to P. oliveirana, one locality (Mar) related to P. regentii, and four localities (Ata, CamA, CamB, and SLe) related to P. vezicolla  Table 1).
Leaf fragments from each individual were collected and stored in 2.5-ml plastic tubes containing a gel prepared with 2 g of cetyltrimethylammonium bromide (CTAB), 35 g of NaCl, and distilled water to a final volume of 100 ml. Samples were kept refrigerated at 4°C until DNA extraction. Genomic DNA was isolated using the CTAB protocol (Doyle & Doyle, 1987).

| Primer screening and PCR amplification
We tested a set of 20 ISSR primers, previously described by Wolfe (2000). The ISSR primer pairs used in this study were UBC814, The molecular size of the fragments was identified with the aid of a 100-bp DNA ladder. Genotyping was performed manually by comparing banding patterns on the gels.

| Data analyses
After electrophoresis, clear and intense bands were scored and transformed into binary character matrices, indicating the presence (1) or absence (0) of amplified fragments; amplicons of similar molecular size with the same primer were assumed to be homologous.
To evaluate genetic diversity, the presence/absence matrix was analyzed using the POPGENE v. 1.32 software (Yeh, Yang, Boyle, Ye, & Mao, 1997) and the following parameters were estimated: percentage of polymorphic loci (P); allele frequencies; Nei's genetic F I G U R E 1 Sampling localities, geographic distribution, and pictures of the putative species (Pseudolaelia aguadocensis, P. oliveirana, P. regentii, and P. vellozicola) that comprise the PV complex are showed. Colors used for coding the eight sampled localities: yellow to P. aguadocensis, red to P. oliveirana, green to P. regentii, and pink to P. vellozicola TA B L E 1 Locality, abbreviation, coordinates, elevation, sample size, and voucher information for each studied locality of the Pseudolaelia vellozicola complex diversity (h; Nei, 1973); and Shannon diversity index (I). To verify the distribution of private loci across the PV complex, we constructed a Venn diagram using R (R Core Team, 2018). In addition, we estimated genetic diversity indices considering the hypothesis of differentiation between the taxa that comprise the PV complex.

Sample size Voucher
For H E and I, the 95% confidence intervals were calculated to help evaluate differences between estimated means for all species.
In order to infer population structure, a model-based approach was performed using the STRUCTURE 2.2 program (Pritchard, Stephens, & Donnelly, 2000). This method accounts more accurately for the inherent ambiguity of recessive (absent) alleles in ISSR marker data sets. Ten independent runs were performed for K varying from one to nine. Each run consisted of 500,000 Markov chain Monte Carlo (MCMC) iterations, with an initial burn-in of 100,000 iterations. The analysis was performed assuming that the allele frequencies in different populations can be correlated with one another and that alleles carried at a particular locus by a particular individual originated in some unknown population (admixture model). To infer the most probable K to explain the data, we calculated the average of each K likelihood value, log of probability (LnP(D)), through all runs, as suggested by Pritchard et al. (2000), and the ΔK statistic was calculated according to Evanno, Regnaut, and Goudet (2005).
Population genetic structure and differentiation were assessed

| RE SULTS
Five of the 20 tested primers showed high polymorphism and resulted in 94 fragments varying from 100 to 2,000 bp. Populationlevel genetic parameters are summarized in Table 2. Each primer amplified between 14 and 22 fragments, and the ISSR population analyses revealed low levels of within-population genetic diversity, with the percentage of polymorphic loci (P) ranging from 11.8% to 43.4% (Table 2). The values for Nei's genetic diversity (H E ) and Shannon index (I) were low ( compared to P. aguadocensis (n = 1), P. oliveirana (n = 3), and P. regentii (n = 5; Figure 2).
Using Bayesian inference in the STRUCTURE program, the analysis of genetic structure produced an optimal value of K = 8. This result indicates that each sampling locality can be considered a unique population (Figure 3, Figure S1a,b). The significant variation in genetic structure across sampling localities was also evident in the hierarchical multilocus evaluation of genetic differentiation performed using AMOVA, which indicates that a greater proportion of the overall genetic variation exists among populations (Table 3). At the species level, only 15.5% of the overall genetic variation was observed among Pseudolaelia species (F CT = 0.155; Table 3), with the variation among localities within species explaining the greatest proportion of overall genetic variation (52%; Table 3). and geographic distance (r = .841, p < .002), indicating IBD.

| D ISCUSS I ON
Our study presents no evidence that P. aguadocensis, P. oliveirana, P. regentii, and P. vellozicola are different species. From the AMOVA analysis, the small proportion of total variance that can be attributed to variation among species strengthens our conclusion that Pseudolaelia should be recognized as a single and widely distributed species. In addition, the strong relationship between genetic and geographic distance matrices indicates that historical gene flow has occurred infrequently across the populations.
Although the maximum genetic diversity (i.e., H E ) observable using dominant markers such as ISSRs is 0.5, low levels of within-population genetic diversity were observed for the PV complex. These values are lower than those reported by Nybom  Millar, Coates, & Byrne, 2013;Pinheiro et al., 2011Pinheiro et al., , 2014Tapper et al., 2014aTapper et al., , 2014b. Inselbergs are naturally isolated and fragmented environments not only due to the surrounding vegetation, but also as F I G U R E 2 Venn diagram representing the distribution of private loci across the putative species (Pseudolaelia aguadocensis, P. oliveirana, P. regentii, and P. vellozicola) that comprise the PV complex F I G U R E 3 Bar plot representation of the eight genetic populations of the Pseudolaelia vellozicola complex inferred using Bayesian analysis (STRUCTURE) of 134 individuals from eight localities a result of urban growth and the expansion of farmland (Porembski, 2007), which further undermine the survival of pollinators of species living in outcrops (Barbará et al., 2007). This geographical context could explain the high genetic differentiation found among the studied populations. Furthermore, aspects of reproductive biology (e.g., morphological flower traits and pollination system) are important factors that influence the observed genetic structure.
Although no information about the reproductive biology is available for the PV complex, morphological similarities with Pseudolaelia geraensis Pabst, studied by Borba and Braga (2003), can help explain the high interpopulation genetic variation observed herein. Pseudolaelia geraensis presents a pollinator deception mechanism that is not a mimetic pair, but rather mimics the reduced population size (Blambert, Mallet, Humeau, & Pailler, 2016;Hmeljevski et al., 2017;Tapper et al., 2014aTapper et al., , 2014b. These conditions, which are characteristic of inselberg environments, support random genetic drift and limit gene flow and have been suggested as key factors in promoting plant diversification and evolution on terrestrial islands (Hmeljevski et al., 2017;Millar et al., 2013;Pinheiro et al., 2011Pinheiro et al., , 2014Tapper et al., 2014aTapper et al., , 2014b. Additionally, small effective population sizes seem to be characteristic of epiphytes, which adds to the reduced levels of gene flow and promotes population differentiation and speciation events (Phillips, Dixon, & Peakall, 2012).
Regarding the definition of the species, our results do not support the species delimitation inferred by Castro Neto and Marçal (2007) and Castro Neto and Chiron (2009

CO N FLI C T O F I NTE R E S T S
None declared.

S TATE M E NT O F H U M A N A N D A N I M A L R I G HT S
This article does not contain any studies with human participants or vertebrate animals performed by any of the authors.

O PE N R E S E A RCH BA D G E S
This article has earned an Open Data Badge for making publicly available the digitally-shareable data necessary to reproduce the reported results. The data is available at https://doi.org/10.5061/ dryad.4mw6m 9070.