Genomic research with NGS technology has developed rapidly, allowing efficient sequencing of complete plastid genomes 54. Molecular differences in the complete chloroplast genome between species and individuals provide a good means of comparison 55. The cp genome offers several advantages over the nuclear genome, such as unique haploid structure, structural conservation, maternal inheritance, and moderate rate of evolution 55,56. In our comparative study of the plastid genomes of N. alba, N. chilensis, and S. strombulifera analyzing gene content, structure, divergence time, and phylogeny we found that the complete chloroplast genomes of N. alba and N. chilensis are conserved in size compared to species of the Strombocarpa genus. The chloroplast of N. alba and N. chilensis showed similar values for genome size and the number of genes compared to Neltuma juliflora and Neltuma glandulosa described by Asaf57, ~ 163.000 bp for both, while the number of genes varied between 131 and 128. The number of genes was similar between the Neltuma and Strombocarpa genera, although S. tamarugo lost the gene psbL (remaining with 127 genes only) 25. The absence of the psbL gene has been observed in some other eudicots, magnoliids, and monocots as well 58. The genome sizes of S. strombulifera (160,569 bp) and S. tamarugo (161,575 bp) were smaller compared to the Neltuma species (~ 163.000 bp) 25. However, the Strombocarpa species presented slightly more GC content (36.0%-36.2%) compared to the Neltuma species (35.9%). These GC values fall within the limit of variation registered in others studies 25,57. Furthermore, a study about several orchid species, showed that the species with a smallest chloroplast size (Pholidota cantonensis, 158,786 bp), had a highest GC content (37.47%) 59, similar to our observations. The chloroplast genome tends to reduce its size during evolution 60, and gene length might be affected by selection during the evolution of spermatophytes 61. The variations in chloroplast genome size among closely related species can be attributed to IRs, LSC, SSC, intergenic regions, and gene numbers 61. In this study, very little variation in IRs and intergenic regions was observed between N. alba and N. chilensis, resulting in very few differences in genome size, while there is a large variation in these regions in the genomes of S. strombulifera and S. tamarugo. Therefore, we assume that Strombocarpa species have been exposed to stronger evolution than Neltuma species.
A total of 70 to 100 chloroplast simple sequence repeats (cpSSRs) were founded in the cp genomes of the species of the Neltuma, Strombocarpa and Prosopis genera. Our results showed high variation values in the number of cpSSRs among Neltuma and Strombocarpa species, being the highest for N. juliflora (100) and the lowest for S. tamarugo (70). The most abundant cpSSR motif types in Neltuma, Strombocarpa and Prosopis were mono-nucleotides, which is the most abundant repeat type in angiosperms cp genomes 62. Only Strombocarpa species did not show mononucleotide C/G motifs, nor dinucleotide motifs and additionally, they had a lower number of trinucleotide AAT/ATT motifs. However, the Strombocarpa species were the only species that presented the pentanucleotide AATAG/ATTCT motifs. It has been shown in Cyatheaceae, that the characteristics of cpSSRs can provide useful phylogenetic information at the genus level, such as phylogenetic relationships, but also about the number, relative abundance, motif type and relative density of cpSSRs 63. In a similar way, our results demonstrate that the cpSSRs among Neltuma and Strombocarpa, both in number and cpSSR motifs, are likely genus specific.
Repeat sequences are considered to play an important role in genome recombination, rearrangements and contain fundamental phylogenetic information 64,65. We found differences in the repeated elements of the cp genome between Neltuma and Strombocarpa species. The highest total number of repeat elements (palindrome, forward, reverse and complement) was found in S. tamarugo (88) and the lowest in S. strombulifera (57). In general, the total number of palindromic repeats was less in Strombocarpa species than in Neltuma species. However, the total number of forward repeats was less in N. alba and N. chilensis than in the Strombocarpa species. On the other hand, the number of complement and reverse (range of 30–39 bp) repeats in S. tamarugo was higher than in the Neltuma species. In the majority of the species in this study, the most abundant repeat elements detected were, in order, forward, palindromic and reverse. This corresponds to other studies about cp genomes of mimosoid species 66,67, although S. tamarugo is an exception in terms of reverse and complements repeats numbers.
Throughout of the evolution of plastid genomes, structural rearrangements occur, for example in the IRs, which are frequently subject to expansion, contraction or even complete loss 68. An increased length of IR-SSC boundaries plays an important role in mimosoid plastome size variation 69. For example, eight mimosoid plastomes of the tribe Acacia and Inga exhibited an unusual 13 kb IR-SSC boundary shift into the SSC region 67,69, and the size of these plastomes was found significantly affected by a IR-SC boundary shift, as well as by repeat content 67. We observed a slight IR expansion into SSC in S. strombulifera (26.026 bp) and S. tamarugo (25.935 bp) in comparison to the Netuma species. Therefore, the SSC regions of the Strombocarpa species showed contraction, and were the shortest SSC regions compared with those of the Neltuma and Prosopis genera. Asaf et al 57 did not detect IR expansion in Neltuma and Prosopis species, however, they detected a slight expansion in the outgroup species of the genus Adenanthera (with a length of 26,028 bp), similar to what we found the in Strombocarpa species. The study of Asaf Asaf et al 57 did not, however, include Strombocarpa species to compare to the Neltuma and Prosopis species. Similar to Asaf 57, we found a partially duplicated rps19 gene at the beginnings and ends of the IR regions in N. alba, N. chilensis, S. strombulifera and S. tamarugo (including 91 bp in IR). In of most Mimosoideae species, the rps19 is located in the LSC/IRB junction (JLB), with 98–109 bp of the 5′ end of this gene into the IR region 67. The ndhF gene was located closer to the IRB-SSC border (JSB) in Strombocarpa species (up to 67 bp) than in Neltuma and Prosopis species (137 to 156 bp). Likewise, the ndhF gene in the species of the genera Adenanthera, Parkia, Piptadenia, Leucaena and Dichrostachys (Mimosoideae) was found entirely within the SSC region (ranging 11 to 150 away from the JSB junction), however, in species of the tribe Acacia and Inga (Mimosoideae) it was found within the JSB junction, resulting in the duplication of this gene 67. Several models concerning the expansion and contraction of IR regions have been proposed to explain the possible mechanisms that result in shifts in the IR-LSC junctions 70. In our case, we detected that Strombocarpa species had a larger contraction of the LSC region then Neltuma and Prosopis species. The structural differences presented among the plastomes of the Neltuma and Strombocarpa species reinforce the idea and necessity to disintegrate the Prosopis cluster, as proposed by Hughes et al 5. However, for the new genera it would have been recommendable to have kept the names of the sections (Algarobia and Strombocarpa, as proposed by Burkart 4 for the new genera.
The nucleotide diversity (Pi) analysis of Neltuma and Strombocarpa plastomes showed more variations in the LSC and SSC regions than the IR regions. In addition, strong differences of nucleotide diversity value were found between Neltuma and Strombocarpa species. The Pi values between Neltuma species were so low that we found only three variable regions (rps16-trnQUUG, accD-psaI and ycf2- trnICAU), whereas in Strombocarpa species we found ten regions with high Pi values (matK-rps16, trnK-psbI, trnSGCU-trnGGCC, petN-psbM, psaB-psaA, rbcL-accD, psbE-petG, rpoA-rpl36, rps7-ndhB and ycf2). We believe that these ten highly variable regions found in Strombocarpa species, can be of use to resolve uncertainties in phylogenetic analysis of the genus, as well as for DNA barcoding. However, as a very low number of variable regions was found in the species of the Neltuma genus, it will be necessary for further studies to include a sufficient number of samples in order to identify the best regions for identification within the genus Neltuma.
The phylogenetic results (ML and BI) based on 80 protein-coding genes of the plastid genome of nine Mimosoideae species showed that S. strombulifera formed a strongly supported group with S. tamarugo (BP = 100; PP = 1.00), and the Neltuma group appeared as paraphyletic because P. cineraria was part of a well-supported clade (BP = 62; PP = 1.00) with N. juliflora, N. alba and N. chilensis. P. farcta, however appeared as sister group of Neltuma and Strombocarpa clade, as expected. Within the Neltuma clade, N. alba formed a highly supported clade with N. chilensis (BP = 100; PP = 1.00), and so did N. juliflora with P. cineraria (BP = 100; PP = 1.00), whereas N. glandulosa appeared as a strongly supported sister group to both (BP = 100; PP = 1.00). With the exception of P. cineraria (further discussed in the next paragraph), the Neltuma group was monophyletic with Strombocarpa group as its sister clade. Although S. strombulifera and S. tamarugo formed a well-supported group, these two species showed important differences in genome size, number of genes and nucleotide diversity with high degree of variation. These genetic differences in the chloroplast correspond to the findings of Burkart 4 who separated S. tamarugo and S. strombulifera into the Cavernicarpae and Strombocarpae series, respectively. The same was observed by Catalano et al 13 through a three-marker analysis (trnS-psbC, G3pdh, NIA), who found two well supported groups, one of them corresponding to the Cavenicarpae series (including Prosopis ferox and P. tamarugo) and the other formed by North American species of the Strombocarpae series (including Prosopis pubescens and Prosopis palmeri).
Undoubtedly, the biggest inconsistency observed in our phylogenetic analysis was the nesting of P. cineraria within the Neltuma clade. According to the results of Asaf et al 57, P. cineraria forms a group with high support with N. juliflora. It is interesting and unexpected that P. cineraria did not form a group with P. farcta, both of them being Old World species, but nested with N. juliflora, N. glandulosa, N. alba and N. chilensis, which are New World species. However, according to the phylogenetic analysis performed by Catalano et al13, there are more distant relationships among species from the Old World sections and closer relationships among species of the American sections (Strombocarpa, Algarobia, and Monilicarpa sections). Prosopis cineraria is one of the most common trees of the Indian desert, Arabian Peninsula and, in general, is abundant throughout the middle east 57,71, whereas N. juliflora is native to the Caribbean, Central and northern South America 72. However, Neltuma juliflora was introduced to Ethiopia and the Middle East around 1970 and over the years this species has spread outside the plantation areas, adversely affecting natural habitats and rangelands 73. This invasive plant is characterized by vigorous growth which helps it to outcompete indigenous plant species 74. Neltuma juliflora seeds survive in livestock and warthogs’ droppings, which serve as a vehicle for the plant to reach distant areas and to expand their distribution throughout the region 74,75. We hypothesize that N. juliflora might have crossed with some individuals of P. cineraria in a natural way, giving offspring to a hybrid with a phenotype resembling P. cineraria but, when N. juliflora acted as the maternal part, with the cp genome of N. juliflora. This could be a logic explanation for the nesting of P. cineraria within the Neltuma clade, if the samples used by Asaf et al. 57 were obtained from a P. cineraria resembling hybrid.
Estimate of divergence time in plant groups have been important in order to understand their phylogeographic history and evolutionary biology 76. Our molecular dating analysis suggests that Leucaena trichandra as root species diverged in the Middle Eocene (mean = 43.11 Mya; 95% HPD = 37.72–48.07 Mya). Later, P. farcta diverged in the Early Oligocene (mean = 33.52 Mya; 95% HPD = 29.48–37.70 Mya), while P. cineraria diverged together with the Neltuma species in the Pleistocene. Sudalaimuthuasari et al 71, using whole genome sequencing with 76,554 genes, estimated that P. cineraria and P. alba diverged ~ 23Mya. Undoubtedly, this divergence time is closer to what would be expected for species of the genus Prosopis that belongs to the Old World, but not for P. cineraria, whose complete chloroplast genome data show a divergence time of 1.85 Mya (95% HPD: 1.56–2.12 Mya), strengthening our suspicion of this sample being a natural hybrid. Although a previous study indicates that the divergence between Strombocarpa and Neltuma genera occurred in the Oligocene 13, our results show that these genera diverged in the early Miocene (mean = 22.32 Mya; 95% HPD = 19.55–25.08 Mya). The molecular divergence time found in Neltuma and Strombocarpa genera is close to the diversification of the major clades in the subfamily Mimosoideae, which occurred in the Late Miocene 13,50. Our results showed that Strombocarpa diverged in the Late Miocene (mean = 8.70 Mya; 95% HPD = 7.52–9.89 Mya), which is supported by the fossil Prosopisinoxylon anciborae, a Mimosoideae species with a high similarity to genus Prosopis L. (currently re-delimitated), reported to have occurred during the Late Miocene in the Catamarca Province, Argentina 77. Additionally, a similar divergence time, around 9.21 Mya (8.35–10.07), for the genus Strombocarpa was found Catalano et al 13. Our results also showed that the Neltuma genus started diverging in the Pliocene (mean = 2.96 Mya; 95% HPD = 2.62–3.33 Mya) and continued in the Pleistocene. This corresponds to the Mesquite species (e.g. N. alba, N. juliflora, N. glandulosa, N. chilensis, N. alpataco and N. nigra) whose divergence time started in the Pliocene and continued in the Pleistocene, (mean = 3.65 Mya; 95% HPD = 3.31–3.99 Mya) 13.
Tree species such as Neltuma and Strombocarpa are subject to a number of ecological selective pressures due to the hostile conditions of the Atacama Desert. Chloroplast genes are involved in regulatory responses to various abiotic stresses, including heat, chilling, salinity, drought and radiation 78,79. Therefore, the here presented chloroplast genomes of the Neltuma and Strombocarpa species can play an important role in understanding the plants adaptations to these hostile environments.
The chloroplast genome structure of legumes is particularly interesting, because it contains multiple rearrangements, expansions, contractions, and loss of genetic content, which are all very useful for phylogenetic studies 79. Phylogenetic analysis can aid conservation of species through the confirmation of taxonomic status, clarification of evolutionary relationships and consequently the determination fo conservation priorities 80. Additionally, phylogeographic studies offer valuable information for conservation purposes as they describe the geographical distribution of genetic variability among species populations 81. With this study, we discovered differences in chloroplast genomes of Neltuma and Strombocarpa, species improving our understanding of its phylogeny and evolution, in hope to aid the conservation of these valuable species before it is too late and they disappear.