Background

The Algarrobo tree Neltuma pallida (Humb. & Bonpl. ex Willd.) Hughes & Lewis is a tree of 8 to 20 m in height, with small grayish-green leaves, uninodal axillary spines and yellow pod-shaped fruits [1, 2]. This species belongs to the genus Neltuma (Subfamily: Caesalpinioideae), which includes up to 43 potential species that are arboreal or shrubby, possess uninodal axillary spines, and are distributed in dry tropical and arid regions of America [2]. Neltuma pallida (Fig. 1), specifically, is native to arid regions of Colombia, Ecuador and Peru [1]. In Peru, Algarrobo trees grow in the Equatorial Dry Forest (3.45% of the country’s total area), an ecoregion located in the northern coastal regions such as La Libertad, Lambayeque, Piura and Tumbes [3,4,5,6], being one of the predominant species there.

Fig. 1
figure 1

Pictures of Neltuma pallida trees and its main morphological descriptors. (A) Picture of the healthy tree whose DNA was used in this study. (B) Group of dead trees. (C) Bipinnate leaf. (D) Pubescence over a leaflet. (E) Inflorescence. (F) Single flower. (G) Mature pod. (H) Uninodal axillary spines (pointed by arrows)

Full size image

Algarrobo trees hold significant economic and ecological importance in the South American countries, where they thrive [7, 8]. The species is known as “multipurpose”, offering a range of benefits to local communities [9]. The most common uses of Algarrobo trees are for fuel, medicine, cattle feed, or construction material [10]. Moreover, Algarrobo trees play a crucial role in the maintenance of their environment [11]. They provide a physical barrier to prevent wind erosion, contribute to soil fertility, maintain a microclimate and facilitate bioremediation [12, 13].

Despite the economic and ecological importance of Neltuma spp., their populations in Peru, Argentina, Chile, Venezuela, and the Chaco region (which includes some departments of Bolivia, Brazil, Argentina, and Paraguay) are experiencing a severe decline [14,15,16,17,18]. Regarding Peru, a report by the Peruvian National Forest and Wildlife Inventory [19] shows that as much as 40% of N. pallida trees are standing dead and 9% can be found as stumps. Of the 51% that remain alive, 27% display mild to severe damage. The true cause of the high mortality rate of the Algarrobo tree population is still uncertain and likely complex. Some hypotheses contemplate abiotic factors such as climate change [20] and drought [21] in the northern coast of Peru. Other hypotheses include biotic factors, such as the increasing presence of phytophagous and sucking pest insects [17, 22], now thought to be an effect of the decrease of natural biological controllers due to an ecological imbalance after recent ENSO events, or viruses of the Closteroviridae family [6, 23].

The literature shows few genetic studies on N. pallida, and genomic studies have not yet been carried out. In 2022, the National Institute of Agrarian Innovation (INIA) of Peru, together with the National University of Frontera (Piura - Peru), started a plant breeding project to improve the Algarrobo in Piura, a department in northern Peru [Arbizu, pers. comm]. Also, a project for molecular delimitation of Neltuma species has recently begun in our laboratory [Rivas M. pers. comm].

Some other works that have been published about the species focus on knowing its phenotypic variability, especially its characteristics of agronomic importance [24,25,26]. On the other hand, there are some studies from more than a decade ago that have sought to characterize the species at the genetic level [6]. These use classical molecular markers [27, 28], study the ploidy of the species [29,30,31], or analyze a single nucleotide sequence [32]. The lack of information on such an important and currently endangered species is detrimental to the development of comprehensive conservation and improvement strategies. Thus, the decline of Algarrobo tree populations continues to affect its highly fragile ecosystems, such as the deserts and dry forests of many South American countries. This is expected to lead to a decrease in endemism and biodiversity, as well as harsher living conditions.

A first step in the development of genomic studies in plants is the assembly and annotation of the chloroplast genome. This represents a faster and simpler task than sequencing and analyzing the nuclear genome due to its size and level of complexity [33, 34]. Chloroplast genome sizes range from 120 to 160 Kb, most commonly with 100 to 130 genes [35, 36]. In addition, genomic data is valuable for diversity studies [37, 38], phylogenetic analyses [39, 40], genetic improvement [41, 42], and genetic engineering of the species and closely related species [43, 44].

The objective of this work is to assemble, annotate and analyze the chloroplast genome of Neltuma pallida, and to compare it with other chloroplast genomes of close species. In this study, we discovered the chloroplast genome content of the Algarrobo tree N. pallida (subfamily: Caesalpinioideae) through its assembly, annotation and structural analysis. We also performed a comparative analysis using the sequenced chloroplast genomes of other species Prosopis sensu lato (s.l.), now the genera Neltuma, Strombocarpa and Prosopis sensu stricto (s.s.), and made a phylogenetic reconstruction to identify the relationships of N. pallida within the clade.

Results

Assembly and annotation of the chloroplast genome of Neltuma pallida

We assembled the chloroplast genome and obtained a graphical file of the assembly and the genome sequence. The graphical file was used to examine the structure and sequencing depth of the assembled genome. The assembled N. pallida chloroplast genome (Genbank: OR178743) had a length of 162,381 bp and the classical quadripartite structure (Fig. S1): a long single copy sequence (LSC) of 91,805 bp (~ 56.54% of the genome), a short single copy sequence (SSC) of 18,748 bp (~ 11.55%), and two inverted repeat (IR) regions of 25,914 bp (~ 31.91%) (Table 1).

Table 1 Summary of chloroplast genome features of N. pallida and related species
Full size table

An analysis of the assembled genome structure was performed by examining the nucleotide composition of each region (Table S1). The GC content (GC%) of the whole genome was 35.97%. LSC and SSC had lower GC content: 33.26 and 30.46%, respectively. The IRs presented a higher GC% than the other regions and the whole genome with 42.77%. Among the coding regions, rRNA coding regions (located in the IRs) had the highest GC% with 55.41%, followed by tRNA coding regions with 53.10%. The lowest GC content was found in the protein coding regions with 37.45%.

The annotation of the assembled genome was done using the chloroplast genome of N. juliflora as reference. A total of 132 genes were found in the genome (19 duplicated genes), consisting of 85 protein-coding genes, 39 tRNA-coding genes and 8 rRNA-coding genes. The 85 protein-coding genes correspond to 78,834 bp, the 39 tRNA-coding genes to 2938 bp, and the 8 rRNA-coding genes to 9052 bp (Table 2).

Table 2 Genes annotated in the N. pallida chloroplast genome grouped by category and functional group
Full size table

The annotated genes included the following (Fig. 2): 12 small ribosomal proteins (rps), 8 large ribosomal proteins (rpl), 4 DNA-dependent RNA polymerases (rpo), 4 different rRNA coding genes (rrn), 31 different tRNA coding genes (trn), 5 photosystem I proteins (psa), 15 photosystem II proteins (psb), 11 NADH dehydrogenase proteins (ndh), 6 cytochrome b6/f complex proteins (pet), 6 ATP synthase complex proteins (atp), the major subunit of ribulose-1,5-bisphosphate carboxylase/oxygenase (rbcL), maturase K (matK), the proteolytic subunit of ATP-dependent Clp protease (clpP), membrane envelope protein (cemA), beta subunit of acetyl-CoA carboxylase (accD), cytochrome C biogenesis protein (ccsA), 4 hypothetical proteins of unknown function (ycf), and translation initiation factor 1 (infA). This makes a total of 113 different genes, 18 of which present introns (16 have one intron and 2 have two introns) (Table 3).

Fig. 2
figure 2

Genome map of Neltuma pallida chloroplast genome. It shows its four genomic regions (LSC, SSC, IRA and IRB). Genes located in the inside of the circle are transcribed clockwise, while those located in the outside are transcribed in the opposite direction. Genes are grouped according to their functional group by color codes. The inner circle exhibits de GC content (dark gray) and the AT content (light gray)

Full size image
Table 3 Genes containing introns within the chloroplast genome of N. pallida and the length of their respective exons
Full size table

Functional annotation was conducted on the protein-coding genes sequences, to identify metabolic pathways and processes. Based on information obtained from the KEGG database, these genes belong to four primary classes (metabolism, genetic information processing, cellular processes and organ systems) (Table S2). Metabolism is the class with the highest number of genes, with energy metabolism being particularly prominent. According to the UniProt database, most of the genes are involved in biological processes, photosynthesis, and generation of metabolite precursors and energy (Fig. 3). Additionally, a large number of genes relate to cellular components and the chloroplast.

Fig. 3
figure 3

Histogram of N. pallida chloroplast genes distribution in Gene Ontology (GO) terms

Full size image

Codon usage analysis

Codon frequency (Table S3) and the Relative Synonymous Codon Usage (RSCU) were calculated for the whole exome (protein-coding regions). The most frequently used codon in the genome was AAU, which codes for Isoleucine (n = 1144), followed by AAA, which codes for Lysine (n = 1069). At the other extreme, the least used codons were the three types of STOP codons UGA (n = 17), UAG (n = 18) and UAA (n = 50). The less frequent amino acid-coding codons were UGC, coding for Cysteine (n = 85), and CGC, coding for Arginine (n = 103). Analyzing the RSCU values, it could be seen that, for each amino acid, half of the codons were used with a higher relative frequency than the other half. Also, the preferred codons were A or T/U ending codons (Fig. 4).

Fig. 4
figure 4

RSCU values of amino acids in 85 protein-coding genes of the Neltuma pallida chloroplast genome

Full size image

Identification of repetitive sequences

We searched for repetitive sequences in the genome. The tandem repeats were identified as microsatellites (SSRs) and tandem repeats in general. A total of 142 SSRs were identified (Fig. 5A), with the most abundant repeats being mononucleotide repeats (n = 78), mostly A/T mononucleotides (Fig. 5B). For the remaining SSR (di- to hexanucleotide repeats), between 4 and 18 repeats were found. Tandem repeats larger than hexanucleotides were also identified, with 164 repeats found (Fig. 5C). Thus, a total of 306 tandem repeats were found.

Fig. 5
figure 5

Repetitive sequences frequency in the Neltuma pallida chloroplast genome. (A) Total SSR frequency. (B) SSR frequency according to its repeat motif. (C) Tandem, palindrome, forward and reverse repeats frequency

Full size image

The rest of the repeats were palindromes, direct repeats and inverted repeats. In the genome, 33 palindromes, 250 direct repeats and 240 inverted repeats were found, most of them having between 15 and 29 nucleotides (Fig. 5C).

Phylogenomic relationships of Neltuma pallida

A phylogeny of the subfamily Caesalpinioideae was constructed using the chloroplast genomes of 30 previously published species to determine the position of N. pallida in the clade. We found that the genus Neltuma was not recovered as a monophyletic group because N. pallida was grouped with the other species of Neltuma spp. and P. cineraria (Fig. 6). Beside this, the genus Neltuma is grouped with the other genera of Prosopis s.l. used in this analysis: Prosopis s.s. and Strombocarpa. Prosopis s.l. also shows closeness to Cylicodiscus gabunensis and the Dichrostachys clade.

Fig. 6
figure 6

Phylogenomic tree of N. pallida within subfamily Caesalpinioideae. The whole chloroplast genomes were used for this reconstruction. The methods employed were Bayesian Inference and Maximum Likelihood, and their statistics values are represented by the numbers on the left of each node. N. pallida position in the tree is highlighted with an asterisk (*)

Full size image

Sequence divergence analysis

Divergence between the sequences of N. juliflora, P. farcta, S. tamarugo and A. ligulata with N. pallida was calculated (Fig. 7). The lowest divergence was observed with N. juliflora, while the highest divergence was observed with A. ligulata. When looking at the genomic regions, it was found that the most conserved regions are the IRs, regardless of the species to which they are compared. The opposite was determined for the single copy regions (LSC and SSC), where the greatest divergence between sequences was found.

Fig. 7
figure 7

Graphical representation of the alignment of the chloroplast genomes of N. pallida and related species. N. juliflora, P. farcta, S. tamarugo and A. ligulata are used for comparison. The graph shows the level of identity (50–100%, Y-axis) along the genomes (X-axis) taking as reference the chloroplast genome of Neltuma pallida

Full size image

As expected, non-coding regions show the highest divergence, highlighting some intergenic regions such as trnK-rps16, trnS-trnG, trnC-petN, psbZ-trnG, trnT-trnL, rbcL-accD, accD-ycf4, rps8-rpl14 and rps3-rps19 in LSC or ccsA-ndhD and rps15-ycf1 in SSC. Also, some intronic regions show a higher degree of divergence such as in trnK, rps16, atpF, rpoC1, rpl2, rpl16 and clpP in LSC or petB, petD and ndhA in SSC.

In the coding regions the divergence was smaller, especially when compared to other Prosopis s.l. species. The genes with the greatest difference in their coding sequences are rpoC2, rpoB, accD, cemA, petA, rpl20, ycf2, ndhF and ycf1. There are also genes that show divergence only in one of the species of this group, this is the case of atpA, psbC, psaA and psbB in P. farcta.

Genetic distance of coding sequences analysis

Genetic distance was determined using the sequences of 74 chloroplast genes taken pairwise with the p-distance algorithm. For this purpose, the genome of N. pallida was compared as a reference with those of N. juliflora, P. farcta, S. tamarugo and A. ligulata (Fig. 8). As expected, the highest average distance was found between N. pallida and A. ligulata (0.0200), while the lowest was found between N. pallida and N. juliflora (0.0007).

Fig. 8
figure 8

Pairwise distance of 74 protein-coding genes from N. pallida chloroplast genome with homologs. The gene’s homologs were obtained from N. juliflora, P. farcta, S. tamarugo and A. ligulata chloroplast genomes

Full size image

The 10 genes with the largest genetic distance to N. pallida sequences were clpP (0.0754), ycf1 (0.0231), rpl20 (0.0188), accD (0.0187), psbT (0.0139), petN (0.0139), psbK (0.0134), rps3 (0.0133), rbcL (0.0131) and rps18 (0.0122).

Boundary between regions

The expansion and contraction of the LSC, IRB, SSC, and IRA of the N. pallida chloroplast genome were analyzed by examining the distance between their boundaries with their nearest genes (Fig. 9), and comparing these distances with related species.

Fig. 9
figure 9

Comparison of boundaries between the regions of the N. pallida chloroplast genome with related species. The species used for this comparison were N. juliflora, P. farcta, S. tamarugo and A. ligulata. The genomes are represented as split bars in each region. The boxes above and below the bars are representations of the genes. The arrows indicate the distance in bp between the ends of the genes with the boundaries closest to this one. These representations are not proportional to sequence lengths. rps19* (non-coding) found in IRA is a portion of the complete rps19 gene found in IRB

Full size image

In all Prosopis s.l. species, the genes closest to the boundary between the LSC and IRB regions (called JLB) were rps19 and rpl2. In all cases, the rps19 gene overlaps with both regions, and rpl2 falls within the IRB region. When comparing the distances of rpl2 to JLB between Neltuma spp. and Prosopis s.l. species, a contraction (12 bp) in IRB and an expansion in LSC are found.

Moving to the next boundary, between IRB and SSC (called JSB), in all Prosopis s.l. species, the closest genes are ycf1 and ndhF. When comparing N. pallida with the rest of the Prosopis s.l. species, we found a 17 bp contraction in IRB and a 20–112 bp contraction in SSC.

At the boundary between SSC and IRA (called JSA), the closest genes for Prosopis s.l. species were the other copy of ycf1 and trnN. When comparing the distance between this boundary and the closest genes, it was found that in the N. pallida SSC, there was a contraction of 6 bp with respect to the distance found in N. juliflora and an expansion of 17 bp with respect to the distance found in P. farcta and S. tamarugo. On the other hand, in N. pallida IRA, a contraction of 16 bp was observed with respect to the rest of the Prosopis s.l. species.

Finally, near the boundary between IRA and LSC (called JLA), the rps19 pseudogene and the trnH gene were found in Prosopis s.l. species. In the IRAs of Prosopis s.l., no differences were found between species, starting from the rps19 pseudogene in the IRA and reaching the boundary with LSC. The opposite case was observed in the LSC, when comparing N. pallida with the rest of the Prosopis s.l. species, a contraction of 12 and 182 bp was found with respect to P. farcta and N. juliflora, respectively. With respect to S. tamarugo, an expansion of only 1 bp difference was found.

Different distances were observed between boundaries and their nearest genes, these differences were small among species within Prosopis s.l. We also analyzed the case of A. ligulata and found two differences. First, the boundaries between regions in the chloroplast genome of A. ligulata were at completely different positions than in the Prosopis s.l. species used in this analysis. Second, the genes closest to the boundaries were different.

Discussion

The total genomic DNA of N. pallida was sequenced using Illumina technology, and its chloroplast genome sequences were extracted and assembled with GetOrganelle. This genome was annotated and compared with chloroplast genomes of other species of Prosopis s.l. The assembled chloroplast genome of N. pallida was 162,381 bp (162.4 Kb) long, with a classical circular quadripartite structure: two inverted repeats (IRA and IRB), a short single-copy region (SSC) and a long single-copy region (LSC) (Fig. 2; Table 1). The same quadripartite structure and similar chloroplast genome sizes have been reported in species of the genera Prosopis, Neltuma and Strombocarpa (161.5–163.7 kb) [45,46,47]. The same has been found in other mimosoid species [48, 49] and legumes [45]. In general, chloroplast genomic regions conserve similar sizes among closely related species, as in Prosopis s.l. However, some mimosoid species present regions of different length. This is the case for species of the Inga clade and Albizia spp. [50, 51], which exhibit expansions in their IRs and reductions in their SSCs, generating longer genomes. Our results agree with some studies mentioning that differences in chloroplast genome size are explained by variations in the length of single copy regions, LSC and SSC [46, 52]. Comparing the size differences between IRs and single copy regions in Prosopis s.l., it was found that differences between LSCs and SSCs were always larger than those between IRs.

In the chloroplast genome of N. pallida, 132 genes were identified and the 19 genes located in the IRs were duplicated (Fig. 2, Table 2). The genes in the IRs were also duplicated in the chloroplast genomes of P. cineraria and N. juliflora [46]. Duplication of genes in the IRs is common as it has been observed in other mimosoid species, including those with longer IRs. The difference with these species is that their IRs contain a larger number of genes [50, 51]. The same 19 duplicated genes have also been identified as duplicates in other legumes that are not mimosoids [53]. The duplication of the IRs genes in the chloroplast is common regardless of the length of the region and the phylogeny.

Gene prediction annotation identified 15 genes with one intron and 3 with two introns (clpP, rps12 and ycf3) (Table 3). Similarly, to other chloroplast genomes, rps12 was unevenly distributed in LSC (first exon) and IRB (second and third exons) [54, 55]. Additionally, in the chloroplast genome of N. pallida we identified genes that were absent in related species. For instance, we found trnG-GCC, which is not present in P. cineraria or in N. juliflora chloroplast genomes, while it is in N. glandulosa [46]. The gene coding for the transcription initiator factor, infA, was also annotated in N. pallida, as it presents an open reading frame (ORF). This ORF is also present in the chloroplast genomes of Prosopis, Strombocarpa or Neltuma, but has not been annotated in those genomes. This gene is either absent or present only as a pseudogene in many legume species [53, 56, 57]. The ORF of this gene has been identified in various chloroplast genomes of other legumes, such as Albizia julibrissin and Lespedeza spp. [51, 58]. It has been proposed that infA is a gene that has been lost multiple times from chloroplast genomes during the evolutionary history of angiosperms, and has subsequently been transferred to the nuclear genome of plants [59]. Additionally, there is the interesting case of a rps19 segment (103 bp) which is located in IRA as a pseudogene. This rps19 pseudogene has been discovered in all the other species of Prosopis s.l., with incomplete ORFs [46]. The same pseudogene has been found in non-legumes species, such as Cerasus humilis (family: Rosaceae) [60] and Garcinia paucinervis (family: Clusiaceae) [61].

The GC content (GC%) of the whole chloroplast genome of N. pallida was determined to be 35.97% (Table 1). Comparable values have been observed in the chloroplast genomes of P. cineraria and N. juliflora [46]. These values are typical for most sequenced chloroplast genomes, with GC content around 36.2%, except for certain taxonomic groups like Selaginella spp. (family Selaginellaceae), with 54.8% [62, 63]. Lower GC% values were found in single copy regions (33.26 and 30.46% for LSC and SSC, respectively), but GC% was higher in IRs (42.77%). Previous studies have suggested that this increase in GC% value in IRs is due to the presence of rRNA and tRNA genes [46, 64, 65]. Our results support this presumption as the GC% of rRNA and tRNA genes was calculated to be higher than 53%. It is noteworthy that the high GC% found in IRs aids in their preservation, as it has been shown to decelerate the rate of nucleotide substitution [66].

Regarding codon usage in protein-coding genes, a clear preference was observed for those ending in A or T/U, having RSCU values higher than 1. This pattern was observed in Albizia julibrissin [51] and Stryphnodendron adstringens [48], both species from the subfamily Caesalpinioideae, and in more distant species such as Salix floderusii [67] or Chrysosplenium spp. [68]. Also, Duan et al. [69] found a preference for the use of codons ending in A or T/U. They determined that there is a selective pressure favoring the use of these codons. This evolutionary pressure has only been detected in the third position of the codon, but not in the rest of the chloroplast genome. This positive selective pressure facilitates the expression efficiency and conservation of highly important genes [70], which may explain the observed conservation in the third position of codons.

We identified 142 microsatellites (SSRs) in the N. pallida chloroplast genome, with mononucleotide repeats being the most abundant. This is similar to the findings of Asaf et al. [46] in the chloroplast genomes of P. cineraria and N. juliflora, although they found a lower number of SSRs and did not find penta- or hexanucleotides in those genomes. We also identified more repeats of all types compared to other chloroplast genomes of Prosopis, Strombocarpa or Neltuma. This could be attributed to the use of distinct tools to detect repetitive sequences. As noted by Das & Ghosh [71] dedicated software programs for the identification of repetitive sequences commonly produce varying outcomes based on the algorithm utilized. Despite the challenges that arise from algorithmic disparity, it is still feasible to draw general comparisons, such as the higher number of repetitive mononucleotides among SSRs in the chloroplast genomes of distant species [39, 72, 73]. There is a need to establish a standardized methodology for the identification of repetitive sequences in the chloroplast genomes. This will allow for more reliable comparisons. Repetitive sequences analysis within chloroplast genomes is of high importance as they serve to study genetic diversity and phylogeny through molecular markers development, as emphasized in previous work [72, 74]. Likewise, chloroplast genomes exhibit considerable variation in their repetitive sequences’ number of tandem repeats, which may be useful for detecting population-level polymorphisms [75].

The chloroplast genome sequence is a powerful tool frequently utilized to elucidate phylogenetic relationships [76, 77]. However, cases of discordance between chloroplast and nuclear phylogenies can also be found [78, 79]. This is why information from both sources should be used concurrently when possible. In this work, N. pallida formed a subclade with N. juliflora, N. glandulosa and P. cineraria. The subclade with these four species, formerly included in the Prosopis genus, was reported by Asaf et al. [46], as a monophyletic group. This differs completely with Hughes et al. [2] whose recent results support that Prosopis s.l. is polyphyletic. Their research included about 1000 nuclear genes sequenced by Ringelberg et al. [80] in combination with morphological characters. The author suggested separating Prosopis s.l. into 4 different genera, placing N. pallida, N. juliflora and N. glandulosa in the genus Neltuma. On the other hand, P. cineraria and P. farcta, remained as species of the genus Prosopis s.s. P. cineraria and P. farcta originated and can be found in western Asia [81,82,83], while N. pallida, N. juliflora and N. glandulosa are native to the Americas [2, 7]. Our outcomes diverged from Hughes et al.’s [2] research since we found that P. cineraria is distanced from P. farcta in our phylogenetic tree (Fig. 6). Literature reports further cases of incongruent phylogeny between nuclear and chloroplast information [84,85,86,87,88,89]. In all these cases, incomplete lineage delimitation or introgression/hybridization was found to have influenced the discordances. Both options are feasible in certain Caesalpinioideae species, such as Prosopis cineraria. There is also evidence of interspecific and intergeneric hybridization in Prosopis s.l. [90, 91]. Nevertheless, due to the lack of reference chloroplast genome sequences for other Caesalpinioideae species, they could not be included in our analysis, which hinders the draw of conclusion.

When analyzing the divergence between chloroplast genome sequences among species of Prosopis s.l., a high level general of conservation was observed. However, it was also found: first, that non-coding regions, including intergenic sequences and introns, display the lowest similarity; second, that LSC and SSC exhibit the lowest identity, as has been seen across different taxonomic groups [92,93,94]; and lastly, that IRs are highly conserved, which aligns with the importance of the biological functions they serve in terms of genomic stabilization and gene conservation [95]. Thus, a single mutation in IRs may cause structural and sequence changes at many other sites in chloroplast genomes [96,97,98]. Another interesting point when comparing chloroplast sequences of Prosopis s.l. is that variations were only detected in 9 genes. A similar finding was made by Asaf et al. [46] in their comparison of N. juliflora and P. cineraria, with N. glandulosa, showing divergence in 5 of these 9 genes. This may indicate that this group’s chloroplast genomes are particularly susceptible to mutations in these genes. Despite the observed variability, the genetic distances were minimal. The clpP gene exhibited the most significant variability (0.0754). This gene, which encodes a subunit of the ATP-dependent chloroplast protease, has been reported to show a high rate of amino acid substitution, which is associated with protein structure variability [99]. The second gene with the highest genetic distance was ycf1 (0.0231). Some studies highlight that ycf1, in combination with the intergenic region between it and the ndhF gene, can be applied as a barcode for land plants [100, 101], indicating its high diversity level.

Boundary shift analysis among chloroplast genome regions indicates that N. pallida differs from other Prosopis s.l. species in the positions of all boundaries. Yet, these differences were minor, with a maximum displacement of a few hundred bases at the beginning of LSC. Asaf et al. [46] reported similar results when evaluating other Prosopis s.l. species. The shifting of boundaries between regions is a primary factor to the variation in chloroplast genome size. This has been observed when comparing closely related species in different groups, and it can be explained by the expansion and contraction of chloroplast regions [102,103,104]. Nonetheless, these variations are typically minor among related species, resulting in similar lengths of their chloroplast genomes and regions.

We successfully sequenced, assembled and annotated the chloroplast genome of N. pallida in this study. Our results allowed us to make comparisons with other species of Prosopis s.l., revealing a high degree of similarity with some differences at the structural and genetic level. We also used the sequence produced along with other published chloroplast sequences to perform a phylogenomic analysis that showed that N. pallida grouped with the other Neltuma species and with P. cineraria. Finally, divergence comparisons with other chloroplast genomes of Neltuma and Prosopis s.l. showed that within the group exists a high level of sequence identity. Nevertheless, certain divergent sequences and genes that could be interesting for the development of molecular markers. The data generated by this research can aid in the development of new lines of research that enhances the understanding of the diversity and preservation of this species in a more effective manner .

Conclusions

The Neltuma pallida chloroplast genome closely resembles those of closely related species. It has a size of 162,381 bp with a classical quadripartite structure and a GC content of 35.97%. The genome contains 132 genes, comprising 85 protein-coding genes, 8 rRNA-coding genes and 39 tRNA-coding genes. The codon usage analysis of the 85 protein-coding genes showed that isoleucine and lysine were the two most prevalent amino acids, and there was a clear preference for codons that have A or T/U in their third position. Also, the repetitive sequence identification enabled us to discover 142 SSR with potential as population-level markers.

Phylogenetic reconstruction revealed that N. pallida grouped together with the other species of the genus Neltuma and with P. cineraria. Additionally, the comparison of the N. pallida chloroplast genome with others from close species exhibited a high degree of similarity, particularly in coding regions. These findings can be useful for further diversity or genetic improvement studies in N. pallida.

Methods

Plant material

Young leaves of N. pallida were collected from an adult Algarrobo tree situated in the Bosque de Pómac Historic Sanctuary, Lambayeque Department, Peru (6°26′39.4″ S 79°48′16.6.6″ W). The collection was carried out under the authorization granted by “Resolución Jefatural de Santuario Histórico Bosque de Pómac N° 003-2020-SERNANP-JEF”. The collected samples were transported in paper envelopes for DNA extraction. The species was identified by PhD (c) Marinoli Rivas from the Laboratorio de Gimnospermas y Monocotiledoneas of the Museo de Historia Natural – UNMSM, using the descriptions made by Burkart [1] and Hughes et al. [2]. The voucher is available at the Museo de Historia Natural - UNMSM herbarium (USM N° 335,439) in Lima, Peru.

DNA extraction and sequencing

For DNA extraction, we used 0.2 g of ground sample in 2% CTAB buffer and followed the protocol published by Doyle [105] with minor modifications. The extracted DNA was purified with a DNAse-free RNAse A treatment at 37 °C for 1 hr. DNA quality and concentration were assessed with Nanodrop™ One C (Thermo Scientific, Massachusetts, USA) and Qubit™ 4 (Invitrogen, Massachusetts, USA), respectively. Additionally, sample integrity was also verified by 1% agarose gel electrophoresis.

DNA sequencing was performed using Illumina methodology by contracting the services of an external laboratory. The TruSeq DNA PCR-Free kit (Illumina Inc., California, USA) was used to construct the sequencing library.

The quality of the generated reads was assessed with FastQC v0.11.9 (https://github.com/s-andrews/FastQC) and then filtered with Trimmomatic v0.39 (https://github.com/usadellab/Trimmomatic) [106].

De novo assembly of Neltuma pallida chloroplast genome

To obtain the sequence of the N. pallida chloroplast genome (Genbank: OR178743), a de novo assembly was performed with GetOrganelle v1.7.6.1 (https://github.com/Kinggerm/GetOrganelle) [107]. This is a toolkit that combines Bowtie2 [108], BLAST [109], SPAdes [110], and Python libraries to identify sequences specific to the chloroplast DNA and assemble the chloroplast genome. GetOrganelle parameters were tuned to use the filtered data in the previous step (forward and reverse reads), and to conduct 1000 rounds of extension iterations. Also, we set up GetOrganelle to search sequences corresponding to plant plastids. The other options were left as default. The obtained GFA file was visualized with Bandage v0.9.0 (https://rrwick.github.io/Bandage/) [111] to explore the chloroplast genome structure. The average coverage of the final genome assembly was 130X.

Genome annotation of Neltuma pallida chloroplast genome

Genome annotation by prediction was performed using GeSeq v2.03 [112] in the Chlorobox web server (https://chlorobox.mpimp-golm.mpg.de/), with N. juliflora (Genbank: NC049132) taken as reference. Chloroplast Inverted Repeats (IR), rps12 interspersed gene, protein-coding sequences, transfer RNAs (tRNAs), and ribosomal RNAs (rRNAs) were all annotated. For proteins and RNAs, 25 and 85% identity were set as thresholds for annotation, respectively. Furthermore, tRNAscan-SE v2.0.7 [113], found on the same server, was used as a secondary tRNA annotator. Additionally, the external annotator Chloë v0.1.0 (https://github.com/ian-small/chloe), which is also found on the web server, was utilized. A manual curation was performed to finish the annotation for each gene, comparing the genes with their homologues found in other chloroplast genomes of species of the genera Neltuma, Prosopis and Strombocarpa.

The sequences of the annotated protein-coding genes were blasted in the KEGG (https://www.kegg.jp/) and UniProt (https://www.uniprot.org/) databases. In this way we sought to perform functional annotation.

Codon usage of protein-coding sequences

This analysis was solely performed on protein-coding sequences. Codon usage was analyzed by calculating codon frequency and the Relative Synonym Codon Usage (RSCU) values. If the RSCU was greater than 1, this codon was considered to be used more frequently, whereas if the RSCU was less than 1, the opposite was considered to be true. DAMBE5 v7.3.2 (http://dambe.bio.uottawa.ca/index.aspx) [114] was used to perform this analysis.

Identification of repetitive sequences in the chloroplast genome

Several programs were employed for the identification of repetitive sequences. The identification of tandem repeats, both microsatellite repeats (SSR, Short Sequence Repeat) and long sequences was performed with Phobos v3.3.12 (https://www.ruhr-uni-bochum.de/ecoevo/cm/cm_phobos.htm) [115]. The threshold for determining repeats was a minimum of 10 repeats for mononucleotides, 8 repeats for dinucleotides, 4 repeats for trinucleotides and tetranucleotides, and 3 repeats for pentanucleotides and hexanucleotides.

For the other repetitive sequences: palindromes (P), forward (F), and reverse (R), the RepEX web server (http://bioserver2.physics.iisc.ac.in/RepEx/index.html) [116] was used. Additionally, IUPACpal was used (https://sourceforge.net/projects/iupacpal/) [117] for palindromic sequences, and Vmatch v2.3.1 (http://www.vmatch.de/) [118] for forward repeats. For all of these repeats, a minimum length of 15 bp was considered, and 90% identity with respect to their template as threshold.

Phylogenomic relationships of Neltuma pallida

The N. pallida chloroplast genome was aligned with previously published chloroplast genomes of 30 species of the Caesalpinioideae subfamily (Table S7). Sequences were retrieved from the NCBI database (https://www.ncbi.nlm.nih.gov/) and aligned using the MAFFT web server (https://mafft.cbrc.jp/alignment/software/) [119] with default options.

Phylogenetic inference was done with two types of algorithms: Bayesian Inference (BI), in Beast2 v2.7.3 (https://www.beast2.org/) [120], and Maximum Likelihood (ML), in RAxML-HPC2 v8.2.12 [121]. For BI, the GTR + I + G substitution model (Yang 1994) with a Markov Chain Monte Carlo of 1,500,000 steps was used, removing the initial 10% of these. For ML, the GTR + I + G model with 1000 Bootstrap replicates was used. We used jModelTest 2.1.10 v20160303 (https://github.com/ddarriba/jmodeltest2) [122] to determine the substitution models.

Sequence divergence of chloroplast genomes

The Vista web server (https://genome.lbl.gov/vista/index.shtml) [123], with the mVista function in Shuffle-LAGAN mode was utilized for sequence divergence analysis. This mode enables the detection of rearrangements within genomes. The chloroplast genome sequence of N. pallida was used as a reference along with its annotation. The comparison was made with the chloroplast genomes of N. juliflora, P. farcta (Genbank: NC061227), S. tamarugo (Genbank: NC060845) and A. ligulata (Genbank: NC026134).

Genetic distance analysis of coding sequences

Sequences of 74 chloroplast genes from N. pallida, N. juliflora, P. farcta, S. tamarugo and A. ligulata (Genbank: NC026134) were aligned using ClustalW [124] in Bioedit v7.2.6 [125]. The generated alignments were used to calculate the genetic distance of the sequences of all species from N. pallida. The genetic distance was calculated using the p-distance algorithm with Mega X v10.1.8 (https://www.megasoftware.net/) [126].

Boundary shift analysis between chloroplast genome regions

A comparison was made of the position of boundaries between single copy regions (LSC and SSC) and inverted regions (IRA and IRB) in 5 species of the subfamily Caesalpinioideae: N. pallida, N. juliflora, P. farcta, S. tamarugo and A. ligulata. These boundaries were called JLB (LSC-IRB boundary), JSB (IRB-SSC), JSA (SSC-IRA) and JLA (IRA-LSC). For the positions of the boundaries, the distance between the boundaries and the genes closest to or over the boundaries was calculated.