Looking Inside Non-coding Chloroplast Regions of Calophyllum brasiliense (Calophyllaceae) to Understand Its Southernmost Population Distribution
Cecilia Beatriz Percuoco1, Liliana Noelia Talavera Stéfani1, Manuela Edith Rodríguez2, Naiké Lucía González3, Juan Fernando Crivello4, Jorge Víctor Crisci5, Carina Francisca Argüelles1, *
1Laboratorio GIGA, Instituto de Biología Subtropical, nodo Posadas, Universidad Nacional de Misiones, Consejo Nacional de Investigaciones Científicas y Técnicas. Posadas, Argentina
2Cátedra de Biología Vegetal. Facultad de Ciencias Exactas Químicas y Naturales, Universidad Nacional de Misiones. Posadas, Argentina
3Cátedra de Sistemática Teórica. Facultad de Ciencias Exactas Químicas y Naturales, Universidad Nacional de Misiones. Posadas, Argentina
4Cátedra de Introducción a las Ciencias de la Tierra. Facultad de Ciencias Exactas Químicas y Naturales, Universidad Nacional de Misiones, Posadas, Argentina
5Laboratorio de Sistemática y Biología Evolutiva. Facultad de Ciencias Naturales y Museo, Universidad Nacional de La Plata. La Plata, Argentina
To cite this article:
Cecilia Beatriz Percuoco, Liliana Noelia Talavera Stéfani, Manuela Edith Rodríguez, Naiké Lucía González, Juan Fernando Crivello, Jorge Víctor Crisci, Carina Francisca Argüelles. Looking Inside Non-coding Chloroplast Regions of Calophyllum brasiliense (Calophyllaceae) to Understand Its Southernmost Population Distribution. Journal of Plant Sciences. Vol. 3, No. 6, 2015, pp. 310-319. doi: 10.11648/j.jps.20150306.14
Abstract: In recent years the growing interest in the conservation of Paraná River’s riparian forest led to the discovery of botanical novelties for Argentina. Populations of Calophyllum brasiliense Camb. (Calophyllaceae), a typically flooded lowlands species, were identified in the remaining hygrophile forest of northeast Argentina and southeast Paraguay. Deforestation and flooding, due to the construction of dams, have caused these populations to suffer intensive fragmentation. The aim of this work was to infer phylogeographic relationships among five populations of C. brasiliense, three from Argentina and two from Paraguay, which represent the southernmost points of species’ distribution. We also compared them with samples of a C. brasiliense population from Mexico, the northernmost edge of the species distribution. The chloroplast intergenic spacers petG-trnP, psbJ-petA and the trnL-UAA chloroplast intron were amplified from leaves’ DNA. A total of 2234 bp were characterized once the three regions were analyzed. The three chloroplast regions showed nucleotide differences, represented by InDels, inversions and a few SNPs; however, only the trnL intron was selected for further phylogeographic analysis due to the amount of the information obtained for all populations. Based on trnL intron, it was possible to estimate nucleotide and haplotype diversity (π = 0.00237 and Hd = 0.29600, respectively). Three haplotypes were identified, which allowed Argentinean, Paraguayan and Mexican populations to be differentiated. Based on the three haplotypes found, we discuss and propose a model for a C. brasiliense’ geographic dispersion and historical colonization routes, including an alternative new one to the well-known of the Paraná River.
Keywords: C. brasiliense, cpDNA, petG-trnP, psbJ-petA, trnL Intron
The Paranaense and the Amazonic Biogeographic Subregions are separated by the Chaqueña Subregion, all three cover a great portion of the South America and belong to the Neotropical Region  (Figure 1a). The Chaqueña Subregion is known as the major South American disjunction because it is mainly represented by savanna species, with sparse and dry vegetation, which acts as a dispersion barrier between the Paranaense and the Amazonic rich tropical forest [2,3].
Within the Paranaense Subregion, a unique forest formation stands out for growing in almost permanent flooded soils. This forest type shows a differentiated structure and floristic composition, which is known as "floresta higrófila", "mata de brejo"  or "seasonal semideciduous forest" with permanent fluvial influence . There are typical or exclusive species at hygrophile forests, as is the case of Calophyllum brasiliense Camb., which are absent in other forest types .
C. brasiliense (Calophyllaceae) exhibits a broad geographic distribution which contrasts with high habitat specificity, since it exclusively grows in areas saturated with water [5,7]. This ability provides advantages in comparison with other species, which explains its highly concentrated occurrence in damper regions [8-11]. The distribution of C. brasiliense extends from Mexico to northeast Argentina [12,13] and is commonly known as "guanandí" or "cedro de pantano" in Brazil and "arary" in Argentina and Paraguay, among other names. Its economic importance is due to its high quality wood and several bioactive components that are of medical interest . The oil from the seeds has been successfully used to cure some skin diseases  and extracts from different plant parts exhibit antiviral activity by inhibiting HIV-1 [15-17]. In the same way, stem bark extracts have been demonstrated to have anti-leukemic effects [18,19] and a lethal action against the parasitic protozoan Tripanozoma cruzi . Some bioactive compounds isolated from its leaves have proven to have a significant molluscicidal activity against Biomphalaria glabrata  while the antibacterial potential against Mycobacterium tuberculosis has been recently reported .
The Paranaense Subregion has faced threats for decades, which have transformed it into one of the most vulnerable forests in the world with a current rate of change that seems to be accelerated. Most of the original forest has disappeared or diminished, especially in southeastern Brazil, eastern Paraguay and northeastern Argentina [22,23]. While the loss of biodiversity begins with the reduction of ecological interactions and genetic variability, it ends with the local extinction of plant and animal populations .
C. brasiliense populations have been identified in the southern zone of the Paranaense Forest, some of them located in northeastern Argentina and a few more in southeastern Paraguay. The populations are in riverine forest, both sides of the Paraná River upstream Yacyretá dam and in Yacyretá Island (Paraguay). The dynamics and structure of these populations were modified (Cardozo et al., unpublished data) following the last increase in river level associated with dam requirements. The last identified population is nearby the Iberá wetlands natural Reserve, in the Department of Ituzaingó (Corrientes - Argentina). While this small patch has not been affected by increasing water levels, it has been impacted by surrounding forest of Pinus and Eucalyptus.
As a component of biodiversity conservation, the underlying genetic diversity within these populations would be useful in the identification of conservation units as evolutionary significant units (ESUs) or management units (MUs), both fundamental for proper short and long-term management . Currently only a rapid genetic characterization for the first two C. brasiliense populations identified in Argentina exists ; therefore, we would like to expand our understanding to include the origin and evolution of these populations.
Since the first pioneer studies that use chloroplast DNA (cpDNA) variability to solve phylogenetic questions were published [27-30], many research groups have focused in the chloroplast polymorphic regions due to the well-known, advantageous characteristics of haploid genomes. The small size and largely or completely maternal inheritance (except conifers) , as well as the typical absence of recombination and the structural conservative evolution, are properties that make cpDNA an ideal tool for evolutionary studies with species and populations [32-34]. In the last few years as sequencing technology has increased in availability, the taxonomic levels analyzed with cpDNA have decreased from families to species and populations [35-38].
In the present work we hypothesize that the southernmost C. brasiliense populations historically belonged to a continuum of large forest that underwent a profound fragmentation process that resulted in genetic differentiation among the remaining populations. Based on this hypothesis, the aim of this work was to: 1) evaluate three cpDNA regions in the Argentinean and Paraguayan populations of C. brasiliense in order to understand their genetic and phylogeographic characterization, and 2) contrast these populations with individuals from the northernmost point in the species’ distribution.
2. Materials and Methods
2.1. Study Area
In the present work, three Argentinean and two Paraguayan C. brasiliense populations were sampled. Misiones and Corrientes are northeastern Argentinean provinces bordered by the Paraná and Uruguay rivers (Figure 1c, Table 1). Misiones contains the largest remaining tract of Paranaense Forest . Both provinces are characterized by humid, subtropical climate with no distinct dry season . Two of the Argentinean populations are from Corrientes [Puerto Valle (PV) and Rincón Ombú (RO)] and the third one is located in Misiones [Department of San Ignacio (SI)]. In this region, the Paraná River represents the boundary line between Argentina and Paraguay. Inside Paraguay, C. brasiliense populations studied were both located in the Department of Misiones, in Yacyretá Island (IY) and Ayolas city (AY) (Figure 1c, Table 1). In addition to these samples, a set of Mexican (MX) samples representing the species’ northernmost distribution from Jaguaroundi Ecological Reserve, Coatzacoalcos city, Veracruz State was included (Figure 1b, Table 1).
2.2. Sample Collection
In each population, 2 – 10 individuals were considered (Table 1). Four to five young leaves from adult trees from each population were collected and after cleaning with 70% ethanol were stored dry in individually labeled Ziploc bag containing silica gel. They were maintained at room temperature until DNA extraction. The unique criterion for choosing each individual tree was a minimal distance of 10 m among them.
2.3. DNA Extraction
Completely dried leaves were ground to a fine powder in a ceramic pestle using liquid nitrogen. The total genomic DNA was isolated from leaf tissue (120 mg) based on the CTAB protocol  with modifications which briefly include the incorporation of 2% polyvinylpyrrolidone (PVP), 5 mM ascorbic acid, 4 mM sodium diethyldithiocarbamatetrihydrate (DIECA) and 1.2% β-mercaptoethanol into the digestion buffer in a 2 ml plastic tube to improve the homogenization at 37°C for 3 h . After chloroform: isoamyl alcohol (24:1) organic extractions, the DNA was precipitated with 2/3 isopropanol at -20°C for 1 h. The DNA isolation was verified and semiquantified electrophoretically in 1% agarose gels, stained with 0.5μg/ml ethidium bromide against a mass ladder (BioRad), yielding 10 - 50 ng/μl.
2.4. Primer Selection
According to the conserved nature of chloroplast in higher plants [41,42] primers designed for other angiosperms species were tested [43,35]. Two chloroplastic intergenic spacers, petG-trnP (petG: GGT CTA ATT CCT ATA ACT TTG GC; trnP: GGG ATG TGG CGC AGC TTG G) and psbJ-petA (psbJ: CTC TTT GGT TGA TAG GTA CTG; petA: GGA GAT GCA GAG ATA GTA C), and the trnL-UAA intron were amplified (trnLc(F): CGA AAT CGG TAG ACG CTA CG; trnLd(R): GGG GAT AGA GGG ACT TGA AC). All oligonucleotides were synthesized by IDT® or Macrogen Inc.
2.5. Amplification Reactions
The reactions were optimized with 2 µl of 1/50 DNA dilutions (25 ng/µl), 2 mM MgCl2, 0.2 mM of dNTPs Mix, 5 µg/µl of BSA, 0.1 µM of each primer, and 1U of Taq polymerase (Fermentas) with 1X SO4(NH4)2Taq Buffer, in a 25 µl final volume. For psbJ-petA and petG‑trnP the PCR profile consisted of a 5 minutes denaturation at 80°C, followed by 35 cycles at 95°C for 1 minute, 50°C for 1 minute and 65 °C for 4 minutes . For the trnL intron the PCR profile consisted of an initial 5 minutes step at 94°C, followed by 35 cycles of 93°C for 1 minute, 62°C for 1 minute and 72°C for 2 minutes. A final elongation segment of 4 minutes at 72 °C was necessary. For the trnL intron an internal primer pair was designed using Primer3 software [44,45] and tested with Primer-BLAST (Cbra_F: 5’AAAATGGGCAATCCTGAGC, Cbra_R: ATGGGACTCTATCTTTATTCTCG). PCR profile matched that used for intron trnL but without the initial denaturation step. All amplifications were carried out on a Perkin Elmer GeneAmp System 9600 thermal cycler (Applied Biosystems). To minimize the potential for contamination in all reactions negative controls (no DNA added) were included. Fragment sizes were evaluated electrophoretically in agarose gel (1.5%) using 1X TBE buffer and BrEt (0.5 µg/ml) against 100 bp ladder (Productos Bio-Lógicos®) as molecular marker. Gels were run at 5 V/cm and visualized under UV light .
|San Ignacio (SI)||San Ignacio (Misiones - ARG)||27º16'34.4"S - 55º34'11.9"W||11|
|Rincón Ombú (RO)||Ituzaingó (Corrientes - ARG)||27º24'54.42"S - 56º29'43"W||11|
|Puerto Valle (PV)||Ituzaingó (Corrientes - ARG)||27°35'23.84"S - 56°31'34.17"W||10|
|Yacyretá Island (IY)||Misiones (PY)||27°24'53.88"S - 56°45'21.25"W||3|
|Ayolas (AY)||Misiones (Misiones - PY)||27°24'6.48"S - 56°51'17.75"W||2|
|Jaguaroundi Reserve (MX)||Coatzacoalcos (Veracruz Llave- MX)||18°7'20.27"N – 94°21'45.15"W||6|
2.6. Sequencing Reactions and Analysis
Amplicons of individuals randomly selected from each population were sequenced by Macrogen Inc. using forward and reverse primers. Products were previously purified using GFX PCR DNA #28-9034-70 (GE). The three amplified regions were sequenced for 2-11 individuals of each population. For each analyzed individual, forward and reverse reads were edited and a consensus sequence was created. Three separated alignments were done to identify polymorphic sites and to describe the observed haplotypes for each analyzed region. Additional alignments were carried out among haplotypes found in C. brasiliense and the most identical GenBank sequences using the Basic Local Alignment Search Tool BLASTn 2.2.25 : Hevea brasiliensis (psbJ-petA and petG-trnP) and Calophyllum inophyllum (trnL intron). All consensus sequences were aligned for SNPs and InDels identification. All sequences edition and alignment were carried out using Bioedit v.22.214.171.124  with ClustalW setting default parameters.
2.7. Phylogeographic Analysis
Based on the quantity and quality of sequences obtained per population, trnL intron was selected for the phylogeographic analysis. The alignment analysis allowed estimation of haplotype number (h), gaps, polymorphic sites (S), allele number per polymorphic site, parsimonious or informative sites, and singletons (non-informative sites). Haplotype diversity (Hd) and nucleotide diversity (π) were computed . The inter-population differentiation was assessed through FST statistic. Tajima’s D  and Fu’s FS  were estimated to provide inference on sequence neutrality and to evaluate past departures in population size with a confidence of 99% (p<0.01). All parameters estimations were performed using DnaSP v.5.10.1 .
Phylogeographic relationships among C. brasiliense populations were determined using haplotype networks constructed with the median-joining algorithm in Network v.4.611 . The evolutionary divergences among haplotypes were illustrated by means of a phylogenetic tree reconstructed under maximum parsimony criteria using MEGA v.5.05 . A trnL intron sequence from C. inophyllum (AB817676.1) was included as outgroup to set the tree root and give directionality to the observed changes.
We successfully amplified the three chloroplastic regions in a variety of individuals. Furthermore for trnL intron the specific primers designed (Cbra_F and Cbra_R) allowed us to obtain good quality reads for all 43 individuals analyzed. All populations were represented by two or more sequences of the three chloroplastic regions, although psbJ-petA and petG-trnP intergenic flanking primers failed to amplify for some individuals. As the total number of sequences was different for each region, they were analyzed independently. The three regions added up 2234 bp.
psbJ-petA: A total of 30 sequences of 1200 bp in length were aligned, which enabled us to recognize two haplotypes (H1_JA, KJ093668 and H2_JA, KJ093669) that were differentiated by two InDels, a substitution and a Tn mononucleotide chloroplastic microsatellite (cpSSR) (Table 2a). H1_JA was exclusively observed in individuals belonging to Argentina and Paraguay populations whereas H2_JA was shared among the six Mexican samples analyzed.
N: any nucleotide; - : gap; · : identity
petG-trnP: A total of 19 sequences of 449 bp in length were aligned, which again enabled us to recognize two haplotypes (H1_GP, KJ093665 and H2_GP, KJ093666) that were differentiated by a unique mutational step consistent in a 8 bp inversion flanked by two 10 bp inverted repeats in both directions (Table 2b). As in the previous region, H1_GP was unique for Argentinean and Paraguayan C. brasiliense populations whereas H2_GP was observed in all six samples from Mexico.
trnL intron: A total of 43 sequences of 450 bp in length were aligned, which enabled us to recognized three haplotypes (H1_Lint, KF854307; H2_Lint, KF854308; H3_Lint, KF854309) that were differentiated by a 7 bp InDel and 4 informative SNPs (293, 316, 317, 416) with two variants, but no singletons were observed (Table 2c). Two of these positions were adjacent (316, 317) which seems to be the result of an inversion that could involve a unique mutational step. On the other hand, the InDel consisted of a 7 bp duplication from 335 to 341 nt. H1_Lint was present in all Argentinean and Paraguayan populations, H2_Lint was found in the Ayolas population whereas H3_Lint was common to all Mexican individuals.
The total haplotype diversity (Hd) and the total nucleotide diversity was 0.28600 and 0.00237, respectively; however, values distinct to zero were obtained for the Yacyretá Island population (Table 3). The estimated FST indexes did not differentiate the three Argentinean populations (SI, RO, PV) or the two Paraguayan ones (IY, AY); however, high estimates were obtained between the Paraguayan and the Mexican populations (FST= 0.750, p<0.01) and between the Argentinean and the Mexican populations (FST= 1.000, p<0.01) (Table 4).
The neutrality tests were only carried out on Paraguayan populations because they exhibited haplotype diversity greater than zero. Both Tajima’s and Fu’s test were not significant at p<0.01 (Table 5), indicating that the variants found within them are in accordance with the neutral evolution model and do not reflect past demographic events. The evolutionary relationships among the three identified haplotypes based on trnL intron are represented in the phylogenetic tree shown in Figure 2, where the C. inophyllum trnL intron sequence (#AB817676.1) was included. The haplotype frequencies and distribution along all populations are summarized in the haplotype network (Figure 3) where the Paraguayan populations were grouped together.
The use of non-coding chloroplast regions for phylogeographic inferences has proven to be useful in several population studies; however, there is agreement in recognizing that each region has its own characteristics differing across taxonomic groups [36,55,56]. Therefore when choosing a region suitable for intraspecific analyses it is important to take into account that chloroplast genome is extremely conserved within species, at least in the most widely used intergenic spacers. However, there are other more variable unexplored regions that are rarely employed .
The three chloroplastic regions analyzed in the present work were able to highlight haplotype variants among C. brasiliense populations. The intergenic regions psbJ-petA and petG-trnP have not been frequently used in phylogeographic analyses; however we could identify InDels in both regions for the analyzed populations adding a nucleotide substitution and a mononucleotide microsatellite in the case of psbJ-petA. These two spacers allowed us to distinguish the southernmost Argentinean and Paraguayan populations from samples that represented the northernmost populations in C. brasiliense distribution range. Among the studies that have used petG-trnP, a 350 bp intergenic spacer was sequenced in the Asian tropical tree Trochodendron aralioides and reported four substitutions analyzing four individuals from each of the 24 populations in Taiwan and Japan . Similarly, two polymorphic sites and two InDels were detected in the 441 bp spacer while inferring phylogeographic relationships among 17 populations of Cunninghamia konishii, an Asian conifer .
|Total||4||0.00237 (0.00068)||0.28600 (0.081)|
|Neutrality test||Value||99% CI|
|Tajima’s D||0.31826||-1.85076 – 2.83370|
|Fu’s FS||2.099||-4.84068 – 7.10335|
It is interesting to note that the type of mutations selected to reconstruct the phylogeographic history of a taxon indicate the importance of considering its particular characteristics and the possibility of them being homoplasies. For example, chloroplastic SSRs (cpSSRs) often arise independently in different lineages; therefore, several authors have excluded the frequent poliA/T cpSSRs considering them inappropriate for phylogeographic analysis at population level [35,58,60]. Another type of variant extensively distributed along the majority of chloroplast genomes are shorts InDels (1-10 bp) corresponding to the 25-50% of mutations in non-coding regions [39,61]. This type of polymorphism was observed in the psbJ-petA spacer of C. brasiliense populations. On the other hand an 8 bp inversion flanked by inverted repeats was observed in petG-trnP, which probably made possible the formation of a stem-loop structure favoring intra-molecular recombination events. This kind of inversions has been reported for other intergenic and intronic regions [62-64] in many cases suggesting that they are common enough to occur independently provoking evolution parallelism . However in the present study the difficulties to obtain amplicons from some individuals and the fact that no variants were found among the southernmost populations brought us to exclude petG-trnP and psbJ-petA from the following analysis.
The plastid trnL intron is one of the most used in plant systematics and phylogeographic studies since the 1990’s [65-67]. Some authors have suggested it as an appropriate region for DNA barcoding . In contrast, others have found that resolution within this region is low , in part due to a highly conserved stem-loop structure, P6-loop, mixed with less constrained variable regions used by ecologist for studying highly degraded DNA . Intraspecific variation was found in the phylogenetic study of Clusia genera where three mononucleotide and one dinucleotide microsatellites, all four within a 671 bp sequence of trnL intron, were reported . In this study, we observed only a mononucleotide repeat, whereas the three others were interrupted when comparing C. brasiliense with Clusia sp. sequences (Percuoco, 2014, data not shown). However, the polymorphisms found among the C. brasiliense populations analyzed in the current work support the use of trnL intron for population approaches in C. brasiliense. As reported, three haplotypes were found among the C. brasiliense populations studied and nucleotide diversity was due to four transversions and a 7 bp InDel. As was mentioned above, two of the four transversions were adjacent positions that could probably correspond to an inversion that was shared with the Mexican samples and one individual from Paraguayan Yacyretá Island. It is necessary to analyze a greater number of individuals in order to determine the phylogeographic significance of this inversion because we could not exclude that this mutation has arisen independently in these populations. A similar situation was reported in the analysis of Trochodendron aralioides populations, where an 11 bp inversion was observed in three individuals from two distinct populations . These authors sampled four individuals per population and also suggested that analyzing more individuals was needed in order to discuss the significance of this variant.
Our results for the trnL intron region allowed us to differentiate the Argentinean and the Paraguayan populations, where two haplotypes were identified (H1_Lint and H2_Lint). A third haplotype (H3_Lint) was observed exclusively in the Mexican samples. With respect to the proposed hypothesis, these results seem to be in accordance with a common origin of the Argentinean and Paraguayan populations all sharing H1_Lint, which could indicate that the five analyzed populations conformed a forest continuum in the past. The H2_Lint found in one of the Paraguayan Yacyretá Island (IY) could be indicating an early divergence from Argentinean populations, assuming their common origin. H1_Lint found within all five populations evaluated in the present work demonstrated that the use of cpDNA genetic data reveal stronger genetic structure when compared with other molecular markers. Therefore this characteristic should be highly valued in the case of C. brasiliense phylogeography studies from other geographical regions. An average of 46 π estimates from phylogeographic studies that used cpDNA showed a mean of 0.00070 . Later this value was doubled with the extended use of sequencing technology, and ranged from 0.00065 to 0.15000 .
In plant phylogeography studies it seems to be common to sample a few individuals from many populations, emphasizing inter- over intra-population differences [35,61]. It has been demonstrated that in order to have more precise estimates is more convenient to sample few individuals from many populations than many individuals from a few populations . Eight individuals are enough to obtain reliable estimates, providing a 2000 bp alignment [74,75]. In this first approach for the trnL intron, we analyzed a 450 bp alignment of 43 sequences from individuals belonging to six C. brasiliense populations. According to the above considerations, the number of populations is limited and the alignment extension should be increased. Both issues will be address in future analyses that will include unexplored C. brasiliense populations from Perú, Bolivia and northern Paraguay. In addition, other cytoplasmic and nuclear regions are being optimized for C. brasiliense, which will be useful in extending the alignments.
It is interesting that the Mexican and Paraguayan populations shared an inversion in the trnL intron. Although this must be confirmed through further analysis, the finding of this shared sequence variant led us to question about past dispersion routes of C. brasiliense from Mexico to Paraguay. At first, the Paraguayan populations appear to be remnants of Paranaense Forest Province that cover eastern Paraguay. In this sense, there are three records of C. brasiliense populations on the left side of the Paraguay River but no specimens have been registered along the banks of the Paraná River further North of San Ignacio, Argentina. One possible explanation for the absence of C. brasiliense along High Paraná River tract could be the result of the basaltic riverbed. The basaltic layer results in the river running fast and turbulent between steep banks making seed deposition difficult to occur. Sandstone outcrops in San Ignacio becomes the river broader for a short stretch letting sedimentation to occur. Once the river changes directions in the southwest plain, the current definitely turns calm and slow, allowing light seeds to be transported by flotation . Thus, the species geographic discontinuity could be the consequence of a) the geomorphology characteristics and b) the temperature and rainfall patterns (Cardozo et al., unpublished data). All three could be limited factors for C. brasiliense colonization success; however, we can not exclude alternative colonization routes for the southernmost C. brasiliense populations analyzed in the present study. Another possible route to consider is the Paraguay River, which flows into the Paraná River, confluence that used to be further south and which has migrated towards the present Paraná River channel in its northward displacement. Could the Paraguay River have been a past southern communication bridge between the Paranaense and the Amazonic Biogeographic Subregions? This model seems to be supported by a phylogeographic study carried out in twenty-four C. brasiliense populations (23 from Brazil and one from Costa Rica) . In the mentioned study the authors analyzed the chloroplastic intergenic region trnH-psbA and found a common haplotype within 14 populations from the Paranaense Subregion. Another nine haplotypes were identified in two and seven populations from the Amazonian and the Chaqueña Subregions, respectively, all of them restricted to a single population. These authors proposed that the Caatinga and Cerrado Provinces, that belongs to the Chaqueña Subregion, acts as a corridor joining the Amazonic and the Paranaense Forest Subregions. In the same way, the Chaco Province (also within the Chaqueña Subregion) located in the South of the Cerrado could also be consider a similar corridor. This proposition is in accordance with the fact that C. brasiliense is distributed all along the Neotropical Region. Interestingly, when we compare the updated distribution map of C. brasiliense including Paraguayan and Argentinean location references with individual traces published , the major coincidence seems to be with the typical neotropical trace of the Hemipthera Rhinacloa (Figure 35, in Morrone, 2001). The analysis of samples derived from northern Paraguay, the western Brazilian States, Bolivia, Perú and northern and central populations that occupy the Paranaense Subregion would be crucial to elucidate the processes that originated the observed distribution patterns and to test the dispersion model we have proposed. If this was true and Paraguayan populations would have dispersed by the Paraguay River, our first hypothesis should be rejected at least in regard to the common origin.
The second assessment set out in our hypothesis affirms that the fragmentation process has provoked the differentiation of the analyzed C. brasiliense populations. This hypothesis has not been proved by the results obtained in the present work given that the Argentinean and Paraguayan populations all shared a largely distributed haplotype and represent a unique management unit (MU) in terms of conservation plans. However, the time passed since fragmentation events began could not be enough to allow mutations to accumulate and fix in the chloroplast genome of C. brasiliense, but would explain the high genetic differentiation observed between San Ignacio and Rincón Ombú populations carried out using arbitrary primers for nuclear genome, where the authors identified two MU . Currently SSRs for nuclear C. brasiliense genome are being developed in order to analyze all known southern populations and as a whole, these data will increase the molecular information about these and other C. brasiliense populations. A deep population analysis using nuclear and cytoplasmic markers along with the ecological data that we are relieving about these hygrofile areas will largely contribute to our understanding of these austral forest patches and to drawing up management plans for these particular habitats conservation.
Three haplotypes were identified through which Argentinean, Paraguayan and Mexican populations could be differentiated. The haplotypic variants that have been found in the present work led us to reconsider and propose a model on C. brasiliense’ geographic dispersion and colonization used in the past, both through the Paraná River and alternative routes.
The present work was partially supported by Consejo de Investigación y Desarrollo Tecnológico (CIDET Project 16Q409 to C.F.A.) and Fondo para la Investigación Científica y Tecnológica Project (PICTO UNaM 2011 N°122 to C.F.A). We thank Ministerio de Ecología y Recursos Naturales Renovables of Misiones (MEyRNR) for providing the corresponding permits to access populations and Entidad Binacional Yacyretá (EBY) for providing all the logistic to access to populations located in Paraná River. We also thank Dr. Leonardo Amarilla for the collection and shipping of C. brasiliense samples from Mexico and we are grateful to Dr. Karen De Matteo for the valuable comments and English revision. The authors would also like to thank Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET) for providing posgraduate fellowships to C.B.P and L.N.T.S. and Comité Ejecutivo de Desarrollo e InnovaciónTecnológica (CEDIT) for the research fellowship to L.N.T.S.