Rangewide ploidy variation and evolution in Acacia senegal: a north–south divide?

A recent study of Acacia senegal showed a geographic pattern of genetic variation, which differentiated East and Southern African populations from those in the Sudano-Sahelian region. We build on this previous research to explore variation in DNA content using the flow cytometry method and chromosome number. A geographic north-south DNA content pattern was detected, reflecting the previous results. These results suggest that DNA content may also be important in elucidating the evolutionary history and distribution of the species. Use of external tissues of dried twigs in flow cytometry is new, and presents the opportunity to study numerous other dryland woody species.


Introduction
Polyploidy, also known as whole-genome duplication, is widely recognized as a major driver of evolution, diversification and speciation in angiosperms (Soltis et al. 2009;Wood et al. 2009;Robertson et al. 2010). Cytological, fossil and genomic studies suggest that polyploidy is ancient and widespread, with most angiosperms estimated to have undergone one or more polyploidization events during their evolutionary history (Cui et al. 2006;Soltis et al. 2009;Wood et al. 2009;Jiao et al. 2011). Further analyses of cytogenetic and phylogenetic databases indicate that up to 15 % of angiosperm speciation events are accompanied by ploidy increase (Otto and Whitton 2000;Wood et al. 2009), although this number is likely to grow as more species are studied.
The establishment and maintenance of polyploids will depend on a range of complex ecological and reproductive processes that promote reproductive and/or spatial isolation between the newly formed polyploid plants and their progenitors (Thompson and Lumaret 1992;Petit et al. 1999;Baack 2004). Following genome multiplication, novel genetic combinations arise that may confer polyploids with different reproductive, morphological, ecological, physiological or cytological characteristics from their progenitors, allowing them to exploit new environments (Kolá ř et al. 2009;Manzaneda et al. 2012;Hao et al. 2013).
Variations in 2C DNA content and ploidy levels are not uncommon among closely related species, and can also be useful in delimiting taxonomically complex species (e.g. Kolá ř et al. 2009;Peirson et al. 2012). Furthermore, analysing 1Cx (monoploid) DNA may also be useful in comparing diploid -polyploid complexes across the natural range of the species; several workers have used the constancy in monoploid DNA as evidence of autopolyploidy (Balao et al. 2009;Duchoslav et al. 2013;Krejčíková et al. 2013). In many instances, polyploids and their diploid progenitors may co-occur across geographic regions referred to as contact zones, which can be useful in understanding the processes involved in their origins and evolutionary history (Husband and Schemske 1998;Petit et al. 1999). Diploid -polyploid contact zones may occur in two ways, namely (i) a primary contact zone, arising from the emergence of a new ploidy level within a diploid population, and (ii) a secondary contact zone, formed following contact between diploid and polyploid populations that were geographically isolated and have probably differentiated in allopatry (Petit et al. 1999). Several studies show that contact zones may either be localized or span wide geographic distances, and the dynamics of diploid -polyploid contact zones vary with species or species complexes (e.g. Ranunculus adoneus, Baack 2004; Melampodium spp., Steussy et al. 2004; Dianthus broteri, Balao et al. 2009; Ulmus americana, Whittemore and Olsen 2011; Aster amellus, Castro et al. 2012;Allium oleraceum, Duchoslav et al. 2013; Oxalis obtusa, Krejčíková et al. 2013).
Recent genomic investigations indicate that polyploidy is common in angiosperm lineages including the legume family, and it is postulated that ancient genome-doubling events may be associated with the rapid diversification of this family (Soltis et al. 2009;Cannon et al. 2010). Legumes are reported to have undergone rapid familydiversification since the Tertiary, which corresponded, for instance, with polyploidization within Papilionoideae (Cannon et al. 2010), and emergence of many species as recently as the Plio-Pleistocene (Richardson et al. 2001;Lavin et al. 2005). In Africa, acacias (Acacia Miller s.l., Mimosoideae) diversified during the Pliocene, colonized, expanded and became integral to open, arid-adapted vegetation-the savanna woodlands (Bouchenak-Khelladi et al. 2010;Odee et al. 2012). Globally, the genus Acacia includes .1400 species, and is an important phytogeographic component of the tropics (Lewis et al. 2005). However, genome sizes (2C DNA contents) have been explored in only a few species (Bennett and Leitch 2012).
In this study, we focus on Acacia senegal (see Fig. 1), an insect-pollinated outbreeding species, commonly known as the 'gum arabic' tree, comprising four recognized varieties based on morphological delimitations, namely vars. senegal, kerensis, rostrata and leiorhachis (Fagg and Allison 2004). Together with 20 other rare and closely related species, e.g. A. dudgeoni, and the diploid A. asak as reported by Bukhari (1997b), these form the A. senegal complex (Ross 1975). Acacia senegal is a shrub or tree with considerable economic and ecological importance, producing a natural gum (gum arabic) widely used in the food and beverage industry, pharmaceuticals, other technical applications and provisioning of several ecosystem services in the drylands of tropical Africa (Fagg and Allison 2004;Omondi et al. 2010;Sprent et al. 2010;Odee et al. 2011;Gray et al. 2013). Previous cytological studies of A. senegal have reported only diploids (Atchison 1948;Oballa and Olng'otie 1993;Bukhari 1997b, c), but tetraploids were recently detected in three populations in the Sudano-Sahelian region (Assoumane et al. 2013), which suggested polyploidy might be prevalent in this region and elsewhere across the native range. However, cytogeographic studies to date have not extensively sampled local and regional spatial scales, nor have populations in the southern native range been examined.
A recent phylogeographic and phylogenetic study by Odee et al. (2012) based on ITS and chloroplast DNA (cpDNA) data showed geographic structuring of genetic variation, which separated east and southern African populations (the southern range of the species) from those in west and central Africa within the Sudano-Sahelian region (the northern range). It further inferred that hybridization and introgression were likely to have driven the recent colonization and range expansion of the species in the northern ranges. If the generally held view that polyploidy may confer broader ecological tolerance and colonization ability is true for this species, then we would expect to detect differences in ploidy diversity and distribution within and among populations and regions. Therefore, we employed flow cytometry (FCM) to survey and assess the rangewide distribution patterns of DNA contents and ploidy levels of A. senegal, and specifically addressed the following questions: (i) What is the variation in overall (2C DNA content) and monoploid (1Cx DNA content) genome size? (ii) What is the ploidy variation and distribution pattern across its native range? (iii) Are the ploidy levels phylogeographically and phylogenetically structured? The ploidy data are discussed in the context of origins, evolutionary and recent colonization history of the species. Furthermore, due to difficulties associated with collecting fresh leaf samples in rangewide studies conducted remote from appropriate facilities, we tested a range of tissues, namely seed, fresh leaf and twig, dried leaf and twig and herbarium specimens to evaluate their utility in FCM analysis. Figure 1. The Acacia senegal species: (A) an A. senegal woodland savanna during the rainy season, Ngarendare, Kenya; (B) characteristic alternate bipinnate leaves of A. senegal used in FCM analysis as fresh or desiccated tissue, a pair of secateurs is placed for scale; (C) gum arabic 'nodules' formed on a leafless branch of an A. senegal tree 2 weeks after tapping (debarking) at the beginning of dry season at Dahra experimental field trial, Senegal; (D) shows gum 'nodules' and twigs on protective glove, they were collected from leafless gum-producing trees, the twigs were stored in silica gel and used in FCM analysis; (E) A. senegal tree in Ntumburi, near Isiolo, Kenya; (F) A. senegal trees during rainy and (G) dry season at Dahra, Senegal, also showing a goat feeding on fallen pods and seeds.

Plant material
Samples were collected between May 2006 and June 2011 from a total of 157 A. senegal individuals in wild populations, experimental trials and herbarium specimens representing 54 localities across its natural range in sub-Saharan Africa, Pakistan and India (Table 1). Samples were also obtained from Kadiogo, Burkina Faso for A. dudgeoni, a member of the A. senegal complex. Samples of different tissues (seed, leaf or twigs) and ages including herbarium specimens that had been stored for .53 months since collection were used for 2C and 1Cx DNA content analysis [see Supporting Information- Table S1]. During field collections, fresh terminal twig ( 2-4 mm diameter) or leaf (where available) tissue samples were collected (see Fig. 1D), dried on silica gel in sealable bags, transported back to the lab and stored at 4 8C until used.

Chromosome counts
Chromosome numbers were determined from mitotically active root tip meristems of seedlings germinated from seeds from two populations (Kibwezi and Fallatu, Table 1) using standard methods described for acacias (Bukhari 1997a). Chromosome counts for other populations, namely Kidira, Somo and Tourba; and Sodera, representing both diploids and tetraploids have previously been determined (Assoumane et al. 2013). Chromosome counts in other populations from previous studies are presented [see Supporting Information- Table S2].
Flow cytometry, determination of DNA content and ploidy levels All 157 individuals were subjected to DNA content and ploidy-level estimation using FCM. Sample preparation followed the one-step protocol of Doležel et al. (2007), using an internal plant standard and propidium iodide (PI) dye, which is a DNA selective fluorochrome with no base specificity, and Tris . MgCl 2 buffer (Doležel et al. 2007). For each sample, 1 mL of chilled buffer was placed in a small (5 cm diameter) Petri dish, on ice. Twigs were wiped with damp tissue to remove surface debris, and gently scraped with a razor blade, removing tangential slivers of the outer layers of the tissues, up to and including the green cambial layer, and the scrapings (4 mg) were collected into the buffer. For leaf samples, 4 mg of fresh or dried tissue of the test plant were used. Fresh leaf tissue (25 mg) of a DNA primary reference standard of glasshouse-grown tomato (Solanum lycopersicum 'Stupické polní rané', 2C ¼ 1.96 pg DNA), soya bean (Glycine max Merr. 'Polanka', 2C ¼ 2.50 pg DNA) or maize (Zea mays 'CE-777, 2C ¼ 5.43 pg DNA) plants (Doležel et al. 2007) were chopped together with the test samples with a new feather-edge razor blade, cutting pieces measuring 0.5 -2 mm across (taking 60 -90 s). For seeds, radicle tissue was used as described by Śliwiń ska et al. (2005) to avoid endoreduplicated cells in endosperms. The resultant suspension was collected into a 1 mL pipette and filtered through squares of 42-mm nylon mesh (Sefar AG, Switzerland) into standard 5-mL flow cytometer collection tubes, on ice. RNase was added to prevent staining of double-stranded RNA, followed by PI to attain final concentrations of 80 mg mL 21 for both. Samples were then briefly and gently vortexed to mix, and incubated on ice in the dark until they were measured (usually within 10 -50 min).
The use of different tissues was tested in a range of studies, including comparisons of data collected from fresh leaves and twigs from the same plant, and dried leaves and stems from the same plant. Analyses were repeated on several samples to determine the reliability of the data obtained.
Samples were analysed with a Becton Dickinson FACSCalibur flow cytometer (San Jose, CA, USA), equipped with an argon-ion 488-nm wavelength laser, calibrated daily and run on the low flow rate. Counts were collected using the FL2 (orange) detector with a band-pass filter of 585/ 42 nm. The default threshold setting on FL2 was used (channel 52). Instrument settings (voltage and gain) were adjusted to allow for fluorescence intensity of expected ploidy levels, together with the peak of the reference plant, to be located between channels 250 -750 on the x-axis (using a 1024 scale) of plots of counts of nuclei vs FL2A. Forward scatter (FSC) and side scatter (SSC) were set to logarithmic scales, and FL2 was set to linear.
Data were acquired with CellQuest software (v4.0.2; Becton Dickinson) and subsequently analysed using Cyflogic TM software (version 1.2.1; Cyflo Ltd, Finland), determining the DNA content of the test sample against that of the internal reference standard. Gating of the populations of interest was done first on the FSC/SSC dot plot, and then on the FL2A (area) vs FL2W (width) dot plot to remove debris and doublets, and histograms of a number of nuclei vs channel number (relative fluorescence) were then plotted. The peak mean channel number and full peak coefficients of variation (CV) of each peak were determined. DNA contents of the samples were calculated from the ratios of the peak mean channel numbers of the test and reference plant material. The samples were assigned to ploidy levels based on the available chromosome counts and their DNA contents.

Statistical analysis
The 2C DNA content was calculated for each sample as the average of the replicates, and monoploid genome size (1Cx) estimated as the amount of nuclear DNA Table 1. Locations of all sampled populations of Acacia senegal, and DNA contents and ploidy levels determined by FCM. 2C DNA values are presented by the mean of all individuals at each population (n, number of individuals) + standard deviation (Mean + s.d.). Populations represented by a single individual are denoted by n/a (not applicable), in the 2C DNA range (min-max) column. Materials originated from field trials, FT; natural fields, NF; procured seeds (PS) and herbarium specimens (H). Chromosome counts determined in populations marked by single (*), this study, and double (**), previous study (Assoumane et al. 2013 divided by ploidy level (Table 1). Either one-way ANOVA followed by Tukey's multiple comparison tests or twotailed unpaired t-test were performed with 2C DNA and 1Cx DNA contents as the dependent variables. Analyses comprised the variation in 2C DNA and 1Cx DNA contents with ploidy ( Fig. 3A and B); the effects of tissue type (seed, leaf and dried leaf and twigs) on diploid 2C DNA content (Fig. 3C); and the effects of storage duration (arbitrarily categorized as 0, between 2 and 41 and .53 months) on 2C DNA content (Fig. 3D). Regional comparisons were carried out among populations from east Africa, west and central Africa and the Indian subcontinent (India and Pakistan) using only diploid individuals for 2C DNA (Fig. 3E) and all ploidy levels and individuals for 1Cx DNA contents (Fig. 3F). Southern Africa was excluded from these regional analyses because possible regional variation could not be separated from variations due to tissue storage duration. All analyses and graphs were performed with GraphPad Prism version 5.04 for Windows, GraphPad Software, La Jolla, CA, USA, www.graphpad. com. The DNA ploidy distribution was mapped using the ESRI software ArcMap 10.1 (ESRI, Redlands, CA, USA) (Fig. 2). Eleven herbarium specimens collected from Mpamalanga province (Republic of South Africa) were grouped and mapped into two geographically proximate quasi populations of four and seven individuals, taxonomically affiliated to vars. leiorhachis and rostrata, respectively (Table 1) [see Supporting Information- Table S1].
The DNA ploidy data of analysed individuals were also located on the phylogenetic tree (maximum parsimony strict consensus tree) of A. senegal based on ITS sequence data (Fig. 4) (Odee et al. 2012). Examples of FCM histograms collected from different tissue types are provided in Supporting Information-Fig. S1.

Chromosome counts
Chromosome numbers for Kibwezi (A. senegal var. kerensis) and Fallatu (A. senegal var. senegal) populations were 2n ¼ 26, corresponding to diploid.    comparable leaf samples [see Supporting Information- Fig. S1B and D]. When fresh leaf and twig were combined in the same extract [see Supporting Information- Fig.  S1E], the peak positions were identical. The dried leaf showed a slightly higher peak position than the dried stem [see Supporting Information- Fig. S1D vs C], as is also indicated in Fig. 3C. Overall, the mean full peak CVs of standards ranged from 1.78 to 2.89 %, while those of samples (including dried leaf and twigs stored 2 -41 months) ranged from 1.76 to 5.09 % [see Supporting Information- Table S3]. The CVs of samples from southern Africa, which had been stored for .53 months ranged from 4.38 to 7.57 %. Fresh material (leaves and seed) generally yielded lower CVs than dried samples. The mean 2C DNA contents of diploids, triploids and tetraploids were significantly different (Fig. 3A). The DNA content of the hexaploid was not included in the statistical analysis but was 3.06-fold higher than the corresponding diploid DNA content. The monoploid DNA content across all the ploidy series ranged from 0.63 to 0.81 pg (Table 1), but the mean monoploid DNA contents among the diploids, triploids and tetraploids were not significantly different (Fig. 3B). Analysis of variance results showed that the 2C DNA value of diploid individuals was dependent on tissue type (seed, fresh leaves, dried leaves and twigs) (F 3,121 ¼ 40.8, P , 0.0001); seed and fresh leaves were not significantly different according to Tukey's post test (Fig. 3C), and were lower than dried leaves and twigs. There was also a significant effect (F 3,56 ¼ 46.7, P , 0.0001) of storage (period between sample collection and FCM analysis) (Fig. 3D). No significant differences between regions were detected in the comparisons of 2C DNA of diploids (Fig. 3E) and 1Cx DNA (Fig. 3F).

DNA content variation and ploidy levels
The 2C DNA contents of all 157 individuals ranged from 1.25 to 4.51 pg (  Fig. 2) [see Supporting Information, Table S1]. Triploid and hexaploid individuals were rare. The triploid individuals occurred in Maroua, Cameroon and Karofane, Niger (Fig. 2, Table 1). Both triploids and the hexaploid detected by FCM are reported for the first time, but have not been confirmed by chromosome counts due to their rarity.

Cytogeography
The overall distribution pattern of ploidy levels is shown in Fig. 2 (Table 1) [see Supporting Information- Table S2 with additional data from previous studies (Bukhari 1997b, c;Assoumane et al. 2013)]. Diploids were the most common and widespread, occurring in the entire natural geographic range of the species. Acacia dudgeoni (analysed in this study) and A. asak (previously reported by Bukhari (1997b)) are also diploids [see Supporting Information- Table S2]. Tetraploids, the rare triploids and hexaploid individuals occur in the Sudano-Sahelian region. Only two tetraploid populations occur in the east African Highlands (Sodera, Ethiopia, and Ntumburi, Kenya), while populations south of the Equator were all diploids. Both diploids and tetraploids were also detected in populations from Pakistan and India in spite of limited sampling of these regions. Interestingly, all detected polyploids belonged to A. senegal var. senegal, although diploids were still the majority in this taxon, and none were detected in the vars. kerensis, leorhachis, rostrata or A. dudgeoni.

Phylogeny and phylogeography of A. senegal diploids and polyploids
The ITS and cpDNA haplotypes were variably shared within and among diploids and polyploids [see Supporting Information- Table S1]. The diploid -polyploid complexes occurred mostly in the terminal clade, which comprised mainly Sudano-Sahelian samples affiliated to var. senegal (Fig. 4). Diploids dominated the basal clades, which largely comprised east and southern Africa populations (southern range), and consisted of individuals affiliated to either var. leiorhachis or vars. senegal, kerensis and rostrata.

Flow cytometry: methodological limitations and innovations
Our main objective to survey and assess the rangewide distribution patterns of DNA contents and ploidy levels of A. senegal was dependent on the utility and reliability of the FCM methodological innovations we developed in order to circumvent the difficulties associated with collecting fresh leaf samples in the species' native range. While other studies have tested the use of herbarium leaf samples, our use of the outer tissues of dried twigs is novel. Flow cytometry comparison of fresh leaf and twig and dried leaf and twig indicated that they yielded similar results [see Supporting Information- Fig. S1].
Both dried leaves and dried twigs showed elevated 2C DNA amounts (Fig. 3C) relative to fresh material. Previous studies by Bukhari (1997c) of six diploid populations across the northern range of the species (Sudan, Senegal, India and Pakistan) [see Supporting Information- Table  S2] yielded estimates of 1.11 + 0.09 -1.17 + 0.04 pg (mean + standard error) with 'White Leghorn' chicken (Gallus domesticus) erythrocytes as internal standard, compared with our fresh leaf mean 2C DNA content of 1.36 + 0.06 s.d. pg, n ¼ 25. We attribute this difference to the internal standards used, especially as a considerable variation in 2C genome size for 'Leghorn' chickens of different strains and sexes has been reported (Mendonça et al. 2010). Our data also showed that triploid, tetraploid and hexaploid individuals had 2C DNA contents commensurate with their ploidy levels. The bulk of plant tissues used in our study were dried twigs. The estimation of DNA content from dried twig and leaf samples mostly yielded results with full peak CVs ,5 %, thus falling within the generally accepted upper limit (Doležel et al. 2007). The CVs of the fresh reference plants, measured simultaneously with the test species were lower, mostly ,3 %. However, CVs of up to 7.57 % were also obtained in some herbarium individuals from the southern African region; these samples were long-stored leaves (up to 58 months) of herbarium specimens. As well as increasing CVs of stored tissues, storage also resulted in increased 2C DNA values; using our fresh leaf mean 2C DNA content for diploids (1.36 + 0.06 pg) as a benchmark, we observed an increase of 15 % for dried leaf and twig tissues stored .53 months (cf. 1.57 + 0.07 pg), highlighting the compounding effects of desiccation and storage duration of tissue material. Nonetheless, these variations (between fresh and dried tissues) are within the ranges reported for various other species that have used either dried tissues and/or storage conditions (e.g. Šmarda and Stančík 2006;Suda and Trávníček 2006;Balao et al. 2009;Cires et al. 2009;Bainard et al. 2011).
In spite of the variation, we could still unequivocally assign samples to the appropriate ploidy. The

DNA ploidy variation, diversity and distribution
To the best of our knowledge, this is the highest polyploid series reported for A. senegal. Acacia tortilis, which also has a diploid chromosome count of 2n ¼ 2x ¼ 26, is reported to be 2x, 3x, 4x and 8x (e.g. Oballa and Olng'otie 1993;Bukhari 1997b, c;El Ferchichi Ouarda et al. 2009), and the Australian acacia, A. dealbata (2x, 3x and 4x, Blakesley et al. 2002). However, unlike A. senegal, other studied African acacias, namely A. tortilis and A. nilotica, appear to be predominantly polyploids (Bennett and Leitch 2012). Our results report populations of mixed ploidy in the species for the first time. Rangewide determination of DNA ploidy in A. senegal found diploids, triploids, tetraploids and a hexaploid including some populations with multiple ploidy levels (2x + 4x, 2x + 3x and 2x + 6x). Diploids were the most common and widespread, occurring throughout the range (Fig. 2). Tetraploids constituted 17.8 % of the samples analysed, and together with the rare triploids and hexaploid, they were only found in the northern range and only with var. senegal, mostly in the Sudano-Sahelian regions of central and west Africa. The east African region, represented by the three common varieties (vars. senegal, kerensis and leiorhachis), had relatively higher sampling intensity than the other regions. In this region, sampled populations immediately south of the Equator were all diploids. Further south (South Africa and Namibia), localities were represented by several single herbarium specimens of vars. rostrata and leiorhachis, and were also all diploids. Thus, our results suggest that if polyploids are present in the southern range, they are likely to be in low frequencies.
Origins and evolution of polyploidy in A. senegal, and range expansion The wide spatial distribution of polyploids, coupled with their occurrence at different levels in the phylogenetic tree (Figs 2 and 4) suggests that polyploids have arisen on multiple independent occasions, including contemporary ecological times to the present. The diploid -polyploid complexes shared the major haplotypes indicating gene flow among the different ploidy levels (Fig. 4) [see Supporting Information- Table S1]. In particular, the distribution of diploid -polyploid complexes in the terminal clade, comprising mainly the west and central African (Sudano-Sahelian region) populations, is also consistent with the hypothesis of a recent range expansion of the species in that region based on ITS and cpDNA data (Odee et al. 2012). Young polyploid complexes are characterized by many diploids and a few tetraploids (Stebbins 1971, cited by Meyers and Levin 2006). In the same Sudano-Sahelian region, Assoumane et al. (2013) detected three tetraploid populations not sharing the same chlorotypes, and suggested multiple origins including allopolyploidy. However, the constancy in monoploid DNA content among our diploid -polyploid populations in the Sudano-Sahelian and Indian subcontinent regions also suggests the possibility of autopolyploid origin. Lack of current data on species composition, distribution and ecology, especially of A. senegal and potential progenitors of the ploidy series observed in this expansive and ecologically heterogeneous geographical region, does not allow for precise inferences on the origins. The documented localities of early herbarium collections (e.g. Ross 1975Ross , 1979Fagg and Allison 2004) show extensive distribution of the A. senegal complex and other congeneric populations, with similar base chromosome numbers and therefore potential sources for allopolyploidy. Clear evidence of allopolyploidy within acacia is demonstrated by the triploid Acacia laeta (2n ¼ 39), which is a hybrid between A. senegal var. senegal and A. mellifera subsp. mellifera found in the Sudano-Sahelian region, as well as eastern Africa and the Arabian peninsula (Khan 1951, cited by Fagg and Allison 2004;Ross 1979). Our study did not sample any known A. laeta population, but we detected triploids only in three individuals found in two A. senegal var. senegal populations, Maroua and Karofane, with the latter occurring as a minority cytotype within a diploid population (2x + 3x). The two samples from Maroua were both from seed, which might not have had long-term viability, while the sample from Karofane was taken from a mature tree. These populations with triploid individuals suggest two possible scenarios: the triploids could have arisen either from a diploid × tetraploid cross, or from combination of reduced and unreduced gametes of a diploid (Petit et al. 1999). However, there are no known diploid or tetraploid populations in proximity of Maroua population, Cameroun (Fig. 2).
In this study, the Sudano-Sahelian and the Indian subcontinent regions broadly represent the contact zones for diploid -polyploid complexes in A. senegal.
Notwithstanding the relatively low representation of DNA content and ploidy data from the southern region, it is clearly evident from this study that there is a north -south divide in diploid -polyploid distribution of the species. The tetraploid Ntumburi population, from Kenya (00811 ′ 54.7 ′′ N, 37831 ′ 00.2 ′′ E), was the southernmost polyploid population; all of the populations south of the equator were diploids. The Acacia senegal species complex is young in evolutionary terms predating the Pleistocene (Bouchenak-Khelladi et al. 2010;Odee et al. 2012), and the terminal clade of the phylogenetic tree (Fig. 4) depicts a relatively recent and rapid expansion into the northern range from eastern Africa (Odee et al. 2012). This expansion may have been facilitated by the polyploidization events in the distinctive var. senegal, which is drought-tolerant and high gum-producing compared with the other varieties (Fagg and Allison 2004). The association of var. senegal with polyploids may also suggest that polyploidization is driving its differentiation from the other varieties. Furthermore, the diploid -polyploid contact zones also appear to coincide with cpDNA haplotype disjunctions located in westernmost part of west Africa (western Senegal) and central Africa, proposed to be refugial zones for A. senegal during the climate oscillations of the Plio-Pleistocene (Odee et al. 2012).

Conclusions
Through rangewide analysis of ploidy levels, we have shown in this study that A. senegal is predominantly diploid, but has new ploidy levels which we report for the first time. The occurrence of diploid -polyploid complexes in the northern range is congruent with existing phylogenetic and phylogeographic data and supports the hypothesis that polyploidization may have been crucial to the recent colonization and range expansion in these regions. However, further studies that employ greater sampling intensity at various spatial scales, focussing on the contact zones of the Sudano-Sahelian region, as well increasing resolution of southern African populations, will be necessary to detect whether polyploidy exists in localities and regions where it has not been reported. This study demonstrated the use of dried twigs as a good alternative source of leaf tissue for determinations of reliable DNA content and ploidy levels, which provides opportunities for rangewide screening of numerous other understudied dryland species using FCM. 032233]. D.W.O. was also supported by a Marie Curie Fellowship [Contract MIIF-CT-2006-39216].

Contributions by the Authors
D.W.O. and J.W. designed the experiment and performed FCM analysis; all authors participated in sample collection, data analysis and writing of the manuscript. They all have seen and agreed to the submitted manuscript.

Conflicts of Interest Statement
None declared. providing herbarium material from southern Africa, and our anonymous reviewers for their constructive comments. Seeds of FCM reference plant material were kindly provided by Prof. Dr Jaroslav Doležel, Institute of Experimental Botany, Olomouc, Czech Republic. Plant material was imported to Scotland under licence from the Scottish Agricultural Science Agency, number PH/34/2009, and IMP/SOIL/27/2010, in accordance with the Plant Health (Scotland) Order 2005.

Supporting Information
The following additional information is available in the online version of this article - Table S1. Locations of sampled Acacia senegal populations, DNA content and ploidy levels analysed by FCM, and corresponding nuclear ITS and chloroplast PCR-RFLP trnH-psbA haplotype data derived from a previous study (Odee et al. 2012). Table S2. Species, locations, 2C DNA content and ploidy levels (chromosome counts) of Acacia senegal from previous studies. Table S3. 2C DNA content and ploidy of Acacia senegal determined from dried twigs, dried leaves and fresh leaves, with Solanum lycopersicum 'Stupické polní rané ' as the internal standard and Tris.MgCl 2 buffer. Dried twigs had been stored for 41 months and dried leaves for 2 months. Samples measured were each taken from individual plants, not repeat measurements on the same plant or same extract. Figure S1. Number of nuclei vs relative fluorescence intensity (FL2-A 1024 linear channel scale) obtained after simultaneous extraction and analysis of nuclei isolated from (A) fresh twigs and (B) fresh leaves of DNA diploid Acacia senegal (Kirbou) (G1 peak) and fresh internal reference standard, Solanum lycopersicum 'Stupické polní rané ' (2C ¼ 1.96 pg DNA), showing G0/G1 and G2/M peaks of the solanum cell cycle), (C) dried twigs and (D) dried leaves of DNA diploid A. senegal (Bissiga) and fresh internal solanum reference standard and (E) a combined extract of fresh twig and leaf of DNA diploid A. senegal (Fallatu Forest) and fresh internal solanum reference standard. Horizontal bars represent the full peak widths used for determination of full peak coefficients of variation. Solanum and acacia samples A, B and E were from fresh glasshouse grown material, acacia samples C and D had been stored in silica gel for 41 months before extraction and analysis.