Chloroplasts are organelles that play an essential role in photosynthesis, which, due to their origin by endosymbiosis between cyanobacteria and primitive eukaryotic cells, have their own DNA, named cpDNA or chloroplast genome (Margulis and Bermudes, 1985; Rodríguez-Ezpeleta et al., 2005; Sato, 2020). In spermatophytes, a quadripartite circular genome pattern is generally observed with two Inverted Repeat regions (IRa and IRb) separated by a Small Single Copy region (SSC) and a Large Single Copy region (LSC) (Jansen and Ruhlman, 2012; Xiao-Ming et al., 2017). Chloroplast genomes are generally used to investigate the evolutionary history between species, evaluate patterns of genetic diversity and demographic history of populations, develop DNA barcode markers for molecular taxonomic identification, and as targets for genetic transformation and production of organisms genetically modified (OGMs) (Leebens-Mack et al., 2005; CBOL Plant Working Group, 2009; Moore et al., 2010; Kang et al., 2021).

In land plants, the number of genes usually varies from 100 to 140, and in some families, such as Sapindaceae, Poaceae, Leguminosae, and Cactaceae, IR expansion and retraction events significantly influence the total number of genes (Wang et al., 2017; Xiao-Ming et al., 2017; Souza et al., 2019; Köhler et al., 2020; Dong et al., 2021). Although there is some variation in the number, the gene order generally seems to have a phylogenetic pattern of conservation in Angiosperms, showing potential to be used in comparative studies to test evolutionary hypotheses related to collinearity and synteny of plastid genes (Wicke et al., 2011). The chloroplast genome contains protein-coding and functional RNA-coding genes such as tRNAs (trn) and rRNAs (rrn). Most of the protein-coding genes are associated with cell replication mechanisms, such as the genes encoding the large subunit of ribosomal proteins (rpl) and the small subunit of ribosomal proteins (rps), photosystem proteins such as psa and psb genes, and the genes encoding NADH dehydrogenase proteins (ndh) (Allen et al., 2011; Jansen and Ruhlman, 2012).

The Sapindaceae family, also known as the soapberry family, comprises about 1900 species of 144 genera, with approximately 80% of its biodiversity contained in tropical and subtropical regions of the Southern Hemisphere (Acevedo-Rodríguez et al., 2010; Muellner-Riehl et al., 2016; Acevedo-Rodríguez et al., 2017; Buerki et al., 2021). Sapindaceae includes plants of different habits, such as trees, shrubs, lianas (woody climbing plants), and herbaceous climbing plants. It also presents species of high economic importance due to their timber, fruit, and medicinal uses (Ferrucci and Acevedo-Rodríguez, 2005; Buerki et al., 2009; Acevedo-Rodríguez et al., 2010; Muellner-Riehl et al., 2016; Dong et al., 2021). More than 50% of the tribes of the Sapindaceae family do not have a representative with a sequenced chloroplast genome, and only 17 genera contain at least one sequenced chloroplast genome (National Center for Biotechnology Information, July 7th, 2023).

The Sapindaceae family emerged at the beginning of the Upper Cretaceous (105 My) and began diversifying approximately 87 My ago in Eurasia. It later diversified throughout the Southern Hemisphere during the Paleocene. Nowadays, this family is subdivided into four subfamilies: (1) Xanthoceroideae, endemic to China; (2) Hippocastanoideae, predominantly occurring in temperate regions; (3) Dodonaeoideae, distributed in tropical and subtropical regions; and (4) Sapindoideae, with wide distribution in tropical and subtropical regions, and it is the most diverse among the four subfamilies (Buerki et al., 2009). The clades that form tribes in this family are still debated. The most recent botanical revision used proposes 20 tribes, of which 16 belong to the subfamily Sapindoideae, two tribes to the subfamily Dodonaeoideae, two tribes to the subfamily Hippocastanoideae, and in the subfamily Xanthoceratoideae no tribes were classified (Buerki et al., 2021). Although molecular studies of the group have been carried out using classical markers and next-generation sequencing, evolutionary perspectives of chloroplast genomes across tribes have yet to be evaluated.

In Sapindaceae, the genus Serjania Mill. has approximately 230 species distributed from the United States to Argentina. The phylogenetic relationships between species of the genus need to be better defined, presenting extensive polytomies. Serjania is the unique liana genus in the Sapindaceae family, and other genera of Sapindaceae are trees, shrubs, or herbs. Furthermore, it is considered the most diverse liana genus in the tropics (Acevedo-Rodriguez, 1990; Acevedo-Rodríguez et al., 2017; Buerki et al., 2021). All Serjania species with known karyotypes are diploid (2n=24) (Urdampilleta et al., 2012), and their genomic data are scarce, with no nuclear draft genomes or organellar genomes available in public databases. In our work, Serjania erecta Raldk (subfamily Sapindoideae; tribe Paullinieae) is the first liana species to have its chloroplast genome assembled and the second species of the Sapindaceae family from the Americas (the first one is Dodonaea viscosa) (Saina et al., 2018). Serjania erecta is a plant traditionally used to treat ulcers and hypertension with potential use in the treatment of Alzheimer’s disease, gastric diseases, and anti-inflammatory use (Gomig et al., 2008; Buerki et al., 2009; Hiruma-Lima et al., 2009).

Here, we assembled the chloroplast genome of Serjania erecta, the first one of the Paullinieae tribe (Sapindaceae), and used it in comparative analysis with species of the Sapindaceae family to elucidate evolutionary aspects that occurred in these genomes. The comparative analysis had as its main aims to identify: (1) Variations in the number of genes in plastomes of the family Sapindaceae; (2) Selection pressure on genes in chloroplast genomes of the Sapindaceae family; and (3) Regions with potential use as markers for the genus Serjania.

The chloroplast genomes of the family Sapindaceae have a relatively conserved structure for the twelve species studied ( Table 1 ). They show slight variation in genome size and GC content. Among the regions of the plastid genomes, the IRs showed the highest variation, with variation in the genes that flank the transition between the LSC and IR regions being associated with gene abundance ( Table 1 ; Figure 3 ). These results suggest that IR expansion and retraction events are an essential source of variation in the number of genes in the Sapindaceae family, with a possible phylogenetic relationship. The S. erecta cpDNA has genomic size, structure, number of genes, and GC content within the range expected for chloroplast genomes of the family Sapindaceae ( Table 1 ). The GC content value for this species is most similar to that described for Dodonaea viscosa and Acer buergerianum. Among the twelve species analyzed, the GC content was highly conserved, as described so far for the Sapindaceae family (Saina et al., 2018; Ma et al., 2019; Dong et al., 2021)

A slight variation in the size of the four regions of the chloroplast genomes (LSC, SSC, IRa, and IRb) was observed, considering the coefficient of variation (CV), the sizes of the IRs (CV=0.04) and SSC (CV=0.03) varied more than the LSC (CV=0.01), which is an expected pattern for land plants ( Table 1 ). (Xiao-Ming et al., 2017). This variation is related to expansion and retraction events in the size of IRs, events already described in members of the Sapindaceae family and other angiosperms as an essential evolutionary mechanism in land plants (Dugas et al., 2015; Li et al., 2017; Saina et al., 2018; Wang et al., 2018; Ma et al., 2019; Dong et al., 2021).

The number of genes in the chloroplast genomes of the Sapindaceae family is consistent with previously published studies (Dong et al., 2021), suggesting that this variability in the number of genes is due to three factors: the expansion and retraction in the size of the IRs, the process of pseudogenization of the rps2 gene, and the mobility of the tRNA genes ( Figure 2 ). The evidence of expansion and retraction processes in the IR regions comes from the gene colinearity in the transition of the IR regions with the LSC and SSC regions (Li et al., 2017). Thus, the identified patterns provide evidence for a possible phylogenetic relationship between the IR expansion and contraction events with the number of genes. The genes rpl22, rps19, and rps3, located in the IR in the subfamily Sapindoideae, except for the most basal species Koelreuteria paniculata ( Figure 2 ), are part of a set of genes encoding ribosomal proteins (Christopher et al., 1988). Among these three genes, the rps19 gene is commonly found in the IR in dicotyledons, although in other families, such as Euphorbiaceae, this gene can be found in both the IR and LSC regions (Li et al., 2017). The rpl22 gene, found in the chloroplast genomes of angiosperms except for the Fabaceae family, is located in the IR of the Rutaceae family (order Sapindales), indicating IR expansion events (Gantt et al., 1991; Bausher et al., 2006). Previous studies for the Sapindaceae family also showed expansion and retraction processes in the LSC-IRb region, occurring even within genera of the Sapindaceae family, highlighting that the IR expansion and retraction events are an essential source of variability in the number of genes in the Sapindaceae (Li et al., 2017; Xiao-Ming et al., 2017; Saina et al., 2018; Ma et al., 2019; Köhler et al., 2020).

The rps2 gene produces the 2S ribosomal protein and is pseudogenized in Hippocastanoideae and Xanthoceratoideae subfamilies (Zhou et al., 2016; Chen et al., 2019). This gene is often under positive selection in other angiosperm families and is located near a nucleotide diversity hotspot (atpI-rps2). Therefore, the accumulation of nonsynonymous mutations may have led to the pseudogenization of this gene (Wu et al., 2018; Li et al., 2020a; Li et al., 2020b; Sobreiro et al., 2020). The trnH-GUG gene is located exclusively in internal regions of the IRs in X. sorbifolium, so this duplication may not be associated with IR expansion events ( Figure 2 ). Furthermore, other variations in genes encoding tRNAs have been identified, as the absence of trnT-CGU genes in Eurycorymbus cavariei has already been described (Dong et al., 2021). The chloroplasts of D. viscosa show the same absence, perhaps indicating a synapomorphy of the subfamily Dodonaeoideae (Dong et al., 2021). The trnT-CGU gene is not the only gene encoding threonine transporter RNAs, and its absence may be supplied by the other two threonine transporter RNAs in the subfamily Dodonaeoideae.

There is a high level of gene conservation in the cpDNAs of the Sapindaceae family, all of which have the same number of introns. The clpP, rps12, and ycf3 genes have two introns and were observed in all annotated species, unlike some annotations that did not identify the rps12 gene ( Supplementary Table S5 ). The rps12 gene has two 3’ exons in the IR and one 5’ exon in the LSC and is a trans-spliced gene. The trans-spliced rps12 gene has been reported in several Viridiplantae. In some taxonomic groups, such as ferns, one intron of rps12 has been lost (Hildebrand et al., 1988; Ma et al., 2019; Sobreiro et al., 2020). We identified two pseudogenes (infA and one of the ycf1 copies) in all Sapindaceae species. The infA gene is a mobile gene that has undergone several transfer events from the chloroplast to the nucleus, mainly in lineages of the Rosid clade (Millen et al., 2001). Pseudogenization of one of the ycf1 copies is common in angiosperm chloroplast genomes and occurs due to its location in the transition zone between the IR and SSC regions ( Figure 3 ) (Li et al., 2017).

Among the species we identified repetitive regions, K. paniculata and S. erecta had the highest and lowest amounts of wide repeats, respectively. Sapindus mukorossi and S. erecta are the species with the highest and lowest amounts of microsatellite repeats ( Figure 4 ), respectively, confirming a previous study highlighting that S. mukorossi has large numbers of repetitive regions in its chloroplast (Dong et al., 2021). Such repetitive regions are highly informative about genetic diversity within species and are commonly used as molecular markers in studies of biogeography and population genetics (Varshney et al., 2005).

Leucine and cysteine are the most and least abundant amino acids constituting the proteins of the S. erecta chloroplast genome ( Supplementary Table S5 ), as previously described in other species of Sapindaceae and angiosperms (Muellner-Riehl et al., 2016; Ma et al., 2019). The predominant use of some amino acids in chloroplasts may be related to the conservation of genes encoding tRNAs since the composition and number of tRNAs affect the codon composition (Novoa et al., 2012).

The ka/ks ratio is a valuable analysis for understanding the evolutionary process within genes, so values greater than 1 (Ka/Ks > 1) indicate positive selection, values equal to 1 (Ka/Ks = 1) indicate neutrality and values less than 1 (Ka/Ks < 1) indicate negative selection (Nei and Kumar, 2000). We found no evidence of positive selection for any of the cpDNA protein-coding genes (ka/ks > 1) ( Figure 6C ). Only the psaI gene is under neutrality, indicating a high conservation of amino acid sequences produced by chloroplast protein-coding genes in Sapindales. The psaI gene makes up subunit VIII of the photosystem I reaction center and showed evidence of local positive selection in other studies (Plöchinger et al., 2016; Schöttler et al., 2017). Previous tests in Nicotiana tabacum, inducing mutations in this gene, did not cause changes in plant growth and only responded to high light and cold stress during leaf senescence (Schöttler et al., 2017). Although psaI is the only gene under neutral evolution, several of the ka/ks relationships did not show synonymous mutations (ka), and the highest values of the relationships are primarily associated with S. erecta genes ( Figure 7B ). A pairwise comparison of S. erecta with the other species for the psaI gene shows an accumulation of non-synonymous mutations in this species, indicating possible positive selection. This heterogeneous speciation process in the psaI gene may be associated with evolutionary factors of the tribe or species. The S. erecta is the unique Sapindaceae liana with data available; positive selection for psaI may be linked to this habit, as this gene is associated with responses to light and cold. We suggest that comparisons of this gene with orthologous genes from species closer to S. erecta are needed to assess the evolutionary impact of these non-synonymous mutations.

The contrasts between the M8 x M7 models identified 12 genes (atpA, clpP, matK, ndhA, ndhF, petD, psaB, rpl32, rpoB, rpoC2, ycf1, and ycf2) under local positive selection ( Supplementary Table S8 ), among these genes one is an ATP subunit (atpA), one encodes a protease (clpP), one encodes maturase (matK), two encodes NADH dehydrogenase subunits (ndhA and ndhF), one composes the Cytochrome b/f complex (petD), one encodes a Photosystem I subunit (psaB), one encode a ribosomal protein subunit (rpl32), two encode DNA-directed RNA polymerase subunits (rpoB and rpoC2), and two encode protein translocons on the inner chloroplast membrane (ycf1 and ycf2) (Allen et al., 2011). The M2a x M1a and M8 x M8a contrasts, which are more rigorous in identification and identified four genes (ndhF, rpoC2, ycf1, and ycf2) under local positive selection ( Supplementary Table S8 ) (Yang and Swanson, 2002; Swanson et al., 2003). The ndhF gene encodes NADH dehydrogenase F and is present in most vascular plants. Positive selection of this gene is associated with adaptation to greater light incidence at high altitudes in Silene vulgaris (Caryophyllaceae) and salt stress in species of the genus Limonium (Plumbaginaceae), suggesting that these loci under positive selection may be related to adaptation in Sapindaceae (Krüger et al., 2019; Darshetkar et al., 2021). The gene rpoC2 is a crucial gene for the transcription of photosynthesis genes, and mutations in these genes can confer the albino phenotype(Park et al., 2023). The ycf1 and ycf2 are essential plant genes for photosynthetic protein import and plant survival (Kikuchi et al., 2013; Kikuchi et al., 2018). Previous studies also showed local positive selection in the ycf1 and ycf2 genes, both in angiosperm and gimonosperm, related to adaptation to environmental changes (Fan et al., 2018; Zeb et al., 2019). In addition, previous studies without considering the phylogeny indicated positive selection of the ycf2 gene and may, in addition to having loci on positive selection, present clades with a more significant amount of non-synonymous mutations (Saina et al., 2018; Dong et al., 2021).

Previous phylogenies reported for the genus Serjania using data from molecular markers (ITS and trnL intron) indicated the formation of a paraphyletic genus with the inclusion of species from the genera Balsas, Chimborazoa and Houssayanthus (Acevedo-Rodríguez et al., 2017). Thus, using molecular information as support, Acevedo-Rodríguez et al. (2017) suggest that the group has changed to include the genera Balsas, Chimborazoa, and Houssayanthus in Serjania, forming a monophyletic group. Speciation events within the genus have probably occurred during the Neogene, resulting in over 230 species distributed throughout the Neotropics. The high level of diversification may be related to the ability to disperse seeds by wind (Muellner-Riehl et al., 2016; Acevedo-Rodríguez et al., 2017; Buerki et al., 2021) and the emergence of climbing species on lianas. South American Sapindaceae species, especially lianas, show a distribution across Antarctica that occurred during the Middle Eocene (about 44 million years ago) (Buerki et al., 2021) and diversified about 18.5-19 million years ago (Jud et al., 2021).

Investigation of nucleotide diversity among families and subfamilies allowed the identification of hotspots of genetic diversity, which can be used as molecular markers to resolve intraspecific relationships and as barcoding markers. In general, intergenic spaces show higher nucleotide diversity, while gene regions are more conserved in species, results like those found by the Percentage of variable characters in embryophytes (Zhang et al., 2011; Fan et al., 2018; Zeb et al., 2019). Our results indicate 10 and 11 nucleotide diversity hotspots for Sapindaceae and Sapindoideae; among these six hotspots (tRNA-Lys — rps16, ndhC – tRNA-Val, petA – psbJ, ndhF, rpl32 – ccsA, and ycf1) are in common for both tribe and family ( Figures 6A, B ). Other studies evaluating nucleotide diversity also indicate the petA-psbJ, rpl32-cssA, ndhF, and ycf1 regions as highly diverse regions, so we strongly suggest using these regions as molecular markers of the group (Zhou et al., 2016; Ma et al., 2019; Dong et al., 2021). Only the ndhF and ycf1 genes showed high nucleotide diversity, genes in which local positive selection was also found ( Figures 6A, B , and Supplementary Table S8 ), which have already been described and used as molecular markers capable of distinguishing angiosperm and gymnosperm species, with great potential for the family of this study (Dong et al., 2015; Fan et al., 2018; Zeb et al., 2019; Amar, 2020; Ramírez-Barahona et al., 2020; Dong et al., 2021).

The availability of data from chloroplast genomes for the Serjania genus can help develop future molecular markers specific to this genus, which can be applied to the definition of interspecific relationships in the genus. However, a more significant number of genomes from phylogenetically close species is required. Although only one chloroplast genome of the genus Serjania has been described, the two gene regions have the potential as markers for phylogenetic studies. They can be used to clarify evolutionary issues of the genus. These include modifying vegetational habits associated with the diversification of the group in lianas and fruit shape (Jud et al., 2021). In addition, they have the potential to be used as markers and DNA barcodes in the molecular taxonomic identification of species, providing very useful tools in the certification of natural medicines, as is the case with species of the genus Serjania.

Our study provides the first assembly and annotation of a chloroplast genome of a species from the tribe Paullinieae (Sapindaceae), Serjania erecta. Three factors are associated with variation in the number of genes in the chloroplast genomes of sapindaceae: (1) expansion events in the IR, (2) pseudogenization of the rps2 gene, and (3) absence and duplication of genes of tRNAs. Protein-coding genes are highly conserved in the Sapindaceae family, and only the psaI gene is evolving under neutrality. We recommend six regions (tRNA-Lys — rps16, ndhC – tRNA-Val, petA – psbJ, ndhF, rpl32 – ccsA, and ycf1) with potential used as a marker for both Sapidaceae and Sapindoideae. The genes ycf1 and ndhF, in addition to showing high nucleotide diversity, also show positive local selection. Our work provides evidence that the ycf1 and ndhF genes may be the most suitable markers for phylogenetic studies of the subfamily Sapindoideae and may contribute to the resolution of taxonomic uncertainties in this group.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: https://www.ncbi.nlm.nih.gov/, SRX18977468, https://www.ncbi.nlm.nih.gov/, NC_072944.1.

LCJC: Data curation, Formal Analysis, Investigation, Methodology, Visualization, Writing – original draft, Writing – review & editing. MS: Formal Analysis, Methodology, Software, Writing – review & editing. LRC: Formal Analysis, Investigation, Software, Writing – review & editing. RD: Formal Analysis, Software, Visualization, Writing – review & editing. RB-F: Formal Analysis, Investigation, Writing – review & editing. CT: Formal Analysis, Writing – review & editing. CS-N: Data curation, Formal Analysis, Investigation, Resources, Writing – review & editing. BB: Investigation, Resources, Writing – review & editing. AS: Resources, Writing – review & editing. JD-f: Conceptualization, Supervision, Writing – review & editing. MT: Conceptualization, Funding acquisition, Project administration, Supervision, Writing – review & editing. RN: Conceptualization, Data curation, Project administration, Supervision, Writing – original draft, Writing – review & editing.