ORIGINAL RESEARCH
http://dx.doi.org/10.38201/btha.abc.v53.i1.8

 

Evolutionary patterns in South African brambles (Rubus L.) - new insights from molecular markers

 

 

Michal Sochor^I; John C. Manning^{II, III}

^ICentre of the Region Haná for Biotechnological and Agricultural Research, Crop Research Institute, Slechtitelü 29, Olomouc 78371, Czech Republic
^IICompton Herbarium, South African National Biodiversity Institute, Private Bag X7, Claremont 7735, South Africa
^IIIResearch Centre for Plant Growth and Development, School of Life Sciences, University of KwaZulu-Natal, Pietermaritzburg, Private Bag X01, Scottsville 3209, South Africa

Correspondence

 

 

---

ABSTRACT

BACKGROUND: South African brambles (Rubus L., Rosaceae) represent a complex group of six native species and at least 12 introduced taxa with different ploidy levels and varying tendencies to hybridisation. The role of hybridisation, intro-gression and apomixis in the ongoing evolution has been hypothesised based on morphological observations, but it has not been rigorously studied to date, and nor has the phylogeny of the group.
OBJECTIVES AND METHODS: This paper aims to reveal the evolutionary patterns and mechanisms in South African brambles by employing three types of molecular markers: plastid and nuclear ribosomal DNA sequences, and nuclear micro-satellites.
RESULTS: The data confirmed the tetraploid R. thaumasius A.Beek and diploid R. ludwigii Eckl. & Zeyh. as distinct native species, while the other four native species are shown to be closely related and likely derived from three ancestors.
CONCLUSION: Ancient hybridisation and limited gene flow between regions (particularly between winter- and summer-rainfall zones) appear to be the main drivers of current patterns in the tetraploid R. pinnatus Willd. and hexaploid R. rigi-dus Sm. Current hybridisation is also likely, although rare. The mechanism of 'octoploid bridge' is proposed, which overcomes the ploidy reproduction barrier between R. pinnatus (or other tetraploids) and R. rigidus. No gene flow was detected between native and alien taxa, but clonal duplications were discovered in the R. bergii x pinnatus hybrid, which implies the possibility of apomictic spread of homoploid hybrids formed between native and introduced brambles and the potential for a new invasion. On the other hand, heteroploid hybrids (R. bergii x rigidus) are formed recurrently and spread only vegetatively.

Keywords: apomixis, clonal spread, hybridisation, introgression, reticulate evolution

---

 

 

Introduction

Rubus L. (Rosaceae: Rosoideae), commonly known as brambles, blackberries, raspberries, dewberries etc., is a complex genus due to its thousands of species and diverse evolutionary mechanisms, of which hybridisation, polyploidisa-tion and apomixis are among the most important (Sochor et al. 2015), and often exhibiting strong phylogeographic patterns (Sochor & Trávníček 2016; Sochor et al. 2017). The genus has been relatively intensively studied in some parts of the world, e.g., in central and northwestern Europe, where an elaborate morphology-based system of a few sexual species and > 750 recognised apomictic microspecies (i.e., asexual genotypes of certain distribution areas and distinct morphology) organised in series, subsections and sections is in use (Weber 1996; Kurtto et al. 2010). On the other hand, the genus has been relatively neglected in other regions (e.g., North America or the Caucasus: Alice et al. 2015; Sochor &Trávníček 2016). This uneven distribution of knowledge is reflected not only in the taxonomy but also in the phylogenetics, phylogeography and evolutionary biology of the group (biosystematics in a wide sense).

The African continent belongs among the understudied regions of the world. None of the native African Rubus taxa (disregarding North Africa, which is home to a few species of predominantly European distribution) were included in the two worldwide phylogenetic studies (Alice & Campbell 1999; Carter et al. 2019), and just a single DNA sequence from a native sub-Saharan African Rubus species is present in the NCBI GenBank nucleotide database (R. rigidus Sm., accession number U95229). Genetic diversity in African accessions has been analysed only among selected Kenyan specimens (unfortunately mostly undetermined and thus of unknown primary origin) using morphological and mi-crosatellite markers with the aim of characterising potential breeding material (Ochieng et al. 2018, 2019). The evolution and phylogeny of African brambles are therefore unexplored.

Within the African continent, the taxonomy and fundamental biological properties (reproduction mode, poly-ploidy) are best explored in the South African Rubus taxa owing to recent advances in our understanding of the species in the region (Sochor et al. 2018, 2022; Van de Beek 2021). Applying a very narrow morphology-based (micro-)species concept, inspired by the one used for European Rubus apomicts, Van de Beek (2021) distinguished 16 native species in six series in the Cape Flo-ristic Region alone and commented on the existence of a number of other 'species', so far insufficiently studied.

On the other hand, Sochor et al. (2022) incorporated ploidy level and reproductive mode data, and identified six native South African Rubus species in total (see Table 1 for overview), all of them sexual di-, tetra- or hexaploids (2n = 14, 28, 42, respectively), and some of them phenotypically highly variable. In addition, 12 introduced taxa and 12 hybrids were identified, which indicated potential ongoing evolution in South African brambles via hybridisation, introgression and apomixis. However, the real effect of these processes on natural populations could not be evaluated properly based on phenotypic and cytometric data only.

In this work, we used the Rubus material that was studied in recent biosystematic/taxonomic investigations (Van de Beek 2021; Sochor et al. 2022) and analysed the sampled individuals by employing three types of DNA markers to address South African Rubus evolution from two perspectives. First, plastid and ribosomal nuclear sequence data were used for phylogenetic and phylogeographic reconstructions, and for identifying/ confirming the identity of hybrids and potential intro-gressants. Specifically, we aimed at revealing not only phylogenetic relationships among taxa, but also at detecting any signatures of potential ancient or ongoing gene flow between native species or between native and introduced taxa. Second, simple sequence repeats (SSR, microsatellites) were used for primary evaluation of genotypic diversity and microevolutionary processes in a model group of R. bergii (Cham. & Schltdl.) Eckl. & Zeyh., R. rigidus, R. pinnatus Willd. and their hybrids. In particular, we aimed at detecting clonal duplications and quantifying the degree of apomixis at a regional scale, and thus evaluating the evolutionary and invasive potential of the hybrids. The new DNA data helped us understand the evolutionary history and phenotypic patterns in this relatively young and species-poor (in the context of the Cape flora) but evolutionary complex plant group.

 

Materials and methods

Sampling and DNA extraction

DNA samples (see Supplementary Table S1) were collected during fieldwork for a biosystematic treatment of South African Rubus (see Sochor et al. 2022 for details). The specimens were simultaneously thoroughly studied morphologically and mostly also analysed for ploidy level and reproduction mode. Of the available collections, 224 specimens (particularly from non-apomictic taxa and hybrids) were used for sequencing, whereas only a selection of 45 specimens of R. bergii, R. rigi-dus, R. pinnatus and their hybrids, mostly from Western Cape (see Supplementary Table S1), was used for SSR analysis for primary evaluation of genotypic and allelic diversity, confirmation of phenotypic determinations, as well as assessment of the suitability of the markers for further studies. Six specimens of R. bergii from its native range in Western Europe were included as well. DNA was extracted from silica gel-dried leaves using the CTAB method (Doyle & Doyle 1987). Eight specimens, four of them being the type specimens, were provided by A. van de Beek, which represented his new species or his conception of old species (Supplementary Table S1; Van de Beek 2021; see also Sochor et al. 2022 for further discussion and revised taxonomic concepts); their DNA was extracted from two seeds per specimen using GenElute^TM Plant Genomic DNA Mini-prep kit (Sigma-Aldrich, USA).

Sequencing

Two plastid regions were analysed: the matK intron was amplified and sequenced with primers XFA and AST_R (Dunning & Savolainen 2010) and the trnL-trnF inter-genic spacer with primers c and f (Taberlet et al. 1991). The ribosomal nuclear locus ITS (internal transcribed spacer) was amplified and sequenced with primers ITS1 and ITS4 (White et al. 1990). Polymerase chain reactions (PCRs) were performed using EliZyme FAST Taq mix (Elisabeth Pharmacon, Czechia) according to the manufacturer's protocol in reaction volume of 15 μΙ_. PCR products were checked on agarose gel electrophoresis, purified by precipitation with polyethylene glycol (10% PEG 6000 and 1.25 M NaCl in the precipitation mixture) and sequenced using the Sanger method at Macrogen Europe (the Netherlands). In selected specimens, the ITS amplicon was cloned into a bacterial vector to obtain sequences of different ITS alleles (ribotypes) within one individual. In these cases, PCR was performed using EliZyme HIFI polymerase (Elisabeth Pharmacon) with proofreading activity. The PCR product was purified, its concentration estimated by Nanodrop 2000, and 18 ng of the PCR product was ligated in the total volume of 10 μΙ_ ligation mixture into pJET1.2/blunt cloning vector using CloneJET PCR

Cloning Kit (Thermo Scientific, USA). The plasmid was further used for transformation of Escherichia coli strain DH5a using TransformAid Bacterial Transformation Kit (Thermo Scientific) following the overnight bacterial culture protocol, with a modification that the initial cultivation in C-medium was not longer than six hours and the colony used for its inoculation was not older than one day. Transformed bacterial colonies were used as a template in a colony PCR with primers pJET1.2 forward and reverse (supplied with the cloning kit). PCR products were checked, purified and sequenced with the amplification primers similarly to direct sequencing as described above.

SSR analysis

Ten microsatellite loci (Graham et al. 2004, 2006; Woodhead et al. 2008) were selected based on amplification efficiency and variability in a selection of samples, and amplified using the EliZyme FAST Taq (Elisabeth Pharmacon) in 10μΙ_ reaction volume with 7.5 ng template DNA following the standard manufacturer's protocol (see Supplementary Table S2 for further details). Fluorescent labelling was performed using a nested PCR containing three primers: a template-complementary forward primer with M13 tail at its 5' end (final concentration 0.1 μM), a template-complementary reverse primer (concentration 0.4 μM), and a fluores-cently 5'-modified M13 primer (5'-TGTAAAACGACG-GCCAGT; NED™, PET®, VIC™ or FAM™ modification; concentration 0.4 μM). To facilitate annealing of the universal M13 primer the annealing temperature was lowered to 53°C in the last nine PCR cycles. Such labelled PCR products were separated together with the GeneScan 600LIZ® size standard on an ABI 3730XL capillary sequencer at Macrogen Europe.

Data analysis

DNA sequence editing, alignments and haplotype/ ribotype identification were performed in Geneious 8 (Biomatters, New Zealand). Plastid haplotypes were compared with the sequences of Sochor et al. (2015) and Sochor and Trávníček (2016), and their codes assigned accordingly. A median-joining algorithm was used to create a phylogenetic haplotype network in Network 10.1.0.0 (Bandelt et al. 1999). All sequences were deposited in NCBI GenBank (accession numbers OL899048-OL899299 [ITS], OL954095-OL954503 [matK and trnL-trnF]). ITS data were checked for the presence of contaminations by microorganisms, pseudogenes and PCR recombinants as in Sochor et al. (2015). The filtered alignment was analysed in Network using star contraction (number of mutations set to three) and median-joining algorithms, and in Splits-Tree 4 (Huson & Bryant 2006) using NeighbourNet algorithm with uncorrected P character transformation.

SSR chromatograms were analysed and scored manually in Peak Scanner 1.0 (Applied Biosystems). Alleles were coded according to their length in bp and saved as both codominant and binary data. Shannon information index was computed and principal coordinate analysis (PCoA) using the distance method with standardisation was performed in GenAlEx 6.5 (Peakall & Smouse 2012) based on the binary data matrix. Histogram of genetic distances and genotype identification were performed in Genotype 2.0b23 (Meirmans & Van Tienderen 2004).

 

Results

Patterns in plastid DNA variation

Plastid DNA data were obtained for 219 specimens. Among native South African Rubus taxa, 18 plastid haplotypes were distinguished when both single nu-cleotide polymorphisms (SNPs) and indels were considered, and 14 haplotypes when indels were rejected (Figure 1). Two haplotypes characterised R. ludwigii and R. thaumasius, respectively, and grouped separately from other native species. The other haplotypes formed two mutually related groups shared mainly by R. pinnatus and R. rigidus. The haplotype of R. trans-vaalensis (haplotype Rig5) was shared with R. rigidus, and two haplotypes Apel, Ape2) were found only in R. apetalus and its hybrid. Each of the Asian species bore a single unique haplotype. Tetraploid Euro-Caucasian taxa were also characterised by their hap-lotypes, but two haplotypes were detected in the dip-loid R. ulmifolius - one shared with R. bergii and one with R. aff. bergii. Four haplotypes were distinguished among North American taxa, one borne by at least three morphotypes of R. sect. Arguti, one by R. sect. Cuneifolii, one by two undetermined morphotypes (one of them possibly belonging to R. sect. Alleghe-nienses), and one haplotype was shared by R. titanus and R. trichogynus (Ursl; presumably derived from the western North American R. sect. Ursini).

The haplotypes of Rubus rigidus and R. pinnatus exhibited clear patterns in geographic distribution. Rubus rigidus bore only three haplotypes in the westernmost part of the range, all from the A group (Figure 1D), whereas only haplotypes of the C group (with one exception of the Rig6 haplotype detected once near Alexandria, EC) were detected in eastern regions outside of the Cape Floristic Region, and the highest diversity was discovered in KwaZulu-Natal (KZN). A roughly similar pattern (Figure 1C) was detected in R. pinnatus and corresponded to its subspecific classification, in which western R. pinnatus subsp. pinnatus bore only the Pinl haplotype or its derivative Pin4, whereas R. pinnatus subsp. pappei had mostly the Rig6 haplotype shared with western R. rigidus, or one haplotype from the C group differing from the eastern R. rigidus haplotypes only in one indel (Pin3).

All of the studied hybrids of R. rigidus [R. bergii x rigidus - 22 individuals (ind.); R. sect. Arguti x rigidus -11 ind.; R. rigidus x ulmifolius - 3 ind.; R. rigidus x pinnatus - 1 ind.] exhibited haplotypes derived from that species. Similarly, R. thaumasius served as the pistillate parent of all of its studied hybrids (with R. ber-gii - 3 ind.; and with R. pinnatus - 1 ind.), as did R. ludwigii (with R. apetalus - 1 ind.; and R. pinnatus - 1 ind.). Rubus pinnatus served as pistillate parent in all of the studied hybrids with R. bergii (6 ind. representing 4 genotypes) but not in the hybrids with R. apetalus, whose pistillate parent was the latter species (1 ind.). The hybrid R. niveus x transvaalensis shared the haplo-type with the first species.

Variation in ITS

ITS data were generated from 118 individuals in total, of which 90 were sequenced directly (individuals without length variation in the amplicon) and 28 were cloned (Supplementary Table S1). One to eight cloned sequences (185 in total, 6.6 on average) were obtained per individual after the exclusion of contaminants (10 sequences in total) and recombinants (14 sequences). 275 sequences were included in the final analyses. ITS exhibited more variation than plastid DNA, but part of it was not shared among individuals and was thus uninformative. The cloned sequences from hybrids always confirmed their hybrid origin. Similarly, two or more divergent orthologous ITS alleles were detected in alien apomictic polyploids. Except for the (putatively primary) hybrids, no gene flow/introgression was detected between native and introduced taxa.

Among native taxa, R. thaumasius and R. ludwigii formed distinct phylogenetic lineages, while the other species formed three groups (A, B, C, corresponding to the plastid haplotype groups according to their presumed origin), two of which could be subdivided into three subgroups each (Figure 2A). Rubus apetalus formed a separate distinct branch diverging from the base of the A group. Rubus transvaalensis was restricted to the C1 subgroup, which was not shared by any other species, but was placed at the split of C2 and C3 subgroups belonging to R. rigidus. Specimens of R. pin-natus subsp. pinnatus had only B2 ribotypes, but specimens from the transitional zone in the eastern parts of WC, as well as R. pinnatus subsp. pappei from MP bore mostly B1 ribotypes, and the remaining eastern populations had B3 ribotypes (Figure 2B). The B1 subgroup was the only one shared with R. rigidus, although only rarely in KZN and MF! Rubus rigidus was otherwise represented in A and C groups (Figure 2C): A was detected almost throughout the studied area, C3 dominated the lowlands of KZN and C2 was detected in MP always as an ortholog together with B1.

Variation in microsatellites

Ten SSR loci were selected following our previous work (e.g., Király et al. 2017), but Rubus123a was excluded due to poor amplification efficiency. In total, 99 alleles were detected in the studied sample set of 45 individuals across the nine loci (3-21 alleles per locus, mean ± standard deviation 11.0 ± 5.8; Supplementary Table S3). However, only null alleles (no PCR products) were detected at two loci (Rubus26a and ERubLR_SQ01_ G16) in R. bergii and relatives. Distribution of genetic distances among individuals indicated the threshold between within-genotype and among-genotype variation to be set at three mutations (not shown). Applying this threshold, genotype assignment was almost identical to analysis with the threshold of zero (i.e., no mutation within a genotype allowed; Table 2); only R. bergii exhibited three different mutations in three individuals (one per individual; Supplementary Table S3). Despite that, this species was clearly monoclonal in both its native and secondary range (Table 2). Besides R. bergii, clonality was detected in the hybrid R. bergii x pinna-tus. In contrast, R. rigidus, R. pinnatus and R. bergii x rigidus exhibited no clonal duplication. PCoA analysis supported identification of the parents in all of the presumed hybrids (Figure 3).

 

Discussion

Evolutionary history is complex in native species

Both the ITS and cpDNA confirm that R. ludwigii and R. thaumasius are distinct native species that diverged from the common ancestor of all South African Rubus taxa. This finding is contrary to previous interpretations of the origin of R. thaumasius, which was originally presumed to be a hybrid of R. rigidus and some other taxon (Gustafsson 1934) or even of purely European origin (Stirton 1981; Henderson 2011). However, its presumed relationship with tropical African species, such as R. runssorensis Engl. and R. friesiorum Gust. (Van de Beek 2021; Sochor et al. 2022), needs to be confirmed, as no material from tropical Africa was available for this study.

A different pattern was observed in the other four native species. Rubus apetalus is well differentiated for both ITS and cpDNA data and is not participating in the current evolution of the other species. It is undoubtedly closely related to both R. pinnatus and R. rigidus. Rubus transvaalensis is even more closely related to R. rigidus as inferred from phenotype (see Sochor et al. 2022) and DNA sequences (Figures 1, 2). Relationships between R. pinnatus and R. rigidus appear to be complex due to shared haplotypes and ribotypes, but in relation to the geographic distributions, this pattern cannot be explained simply by free recurrent gene flow. Taking into account the phylogenetic relationships among haplotypes and ribotypes and taxonomic and geographic distribution patterns, the following scenario can be hypothesised (Figure 4).

Three ancestral species, possibly already tetraploid or even hexaploid, migrated through the coastal regions from northeast to southwest, occasionally hybridised and further evolved into the species as currently recognised, although the ancestral species themselves disappeared. The first ancestor, 'R. archaeapetalus', is represented in our data as the basal ribotypes and haplotypes of the A group (Figures 1A & 2A). This ancestor evolved directly into R. apetalus but must have contributed to the formation of R. rigidus, as implied from the A ribotypes throughout its range and the A haplotypes in the west (which, however, may have been derived also from the second ancestor despite the fact that the current geographic patterns rather contradict this possibility; Figures 1C & 1D). A second ancestor, 'R. archaepinnatus' (B alleles) gave rise to R. pinnatus with considerable geographic genetic variation between the winter-rainfall and summer-rainfall zones but also contributed to the genome of R. rigidus to some extent, at least in the east (see ITS; Figure 2). A third ancestor, 'R. archaerigidus' (C groups), must have had an identical ribotype (C1) and haplotype (Rig5) as R. transvaalensis and may have therefore also been very similar to this modern species in other respects (e.g., in hexaploidy?). This ancestor probably did not spread to westernmost South Africa as no traces of it have been detected in any modern taxon there. It must, however, have contributed to the formation of R. rigidus (mainly in eastern regions), of R. transvaalensis, and to a lesser degree also R. pinnatus. However, as far as we know, R. pinnatus only bears one haplotype derived from 'R. archaerigidus' (Pin3). This haplotype differs from Rig5 in the absence of one 6-bp repetition, which makes Pin3 the basal-most haplotype within the C group. Therefore, the Pin3 hap-lotype can only be a result of an ancient chloroplast capture, rather than a continuous gene flow from 'R. archaerigidus' to R. pinnatus.

Similar reticulate evolution pathways are often observed in polyploid complexes. For example, Fehrer et al. (2009) revealed very complex evolutionary patterns in both diploid and polyploid accessions of European Hieracium s.str. (Asteraceae). Highly reticulate evolution associated with late Quaternary phylogeog-raphy of sexual ancestors was reconstructed in European blackberries, among which more than 750 species are recognised, but these originate in just around six ancestral diploids, some of them extinct (Sochor et al. 2015, 2017). However, hybridisation has long been recognised as an important process in plant evolution and speciation in general, not only in apomictic genera (Rieseberg 1995; Nolte & Tautz 2010).

Current gene flow among taxa seems to be limited

In our previous paper (Sochor et al. 2022), we reported on the occurrence of 12 hybrid combinations in South African brambles, some of which are locally even more frequent than their parents (e.g. R. bergii x R. rigidus). The hybrid origins of all of these taxa were supported by the molecular data presented here (Figure 3; see also Supplementary Table S1 for plastid haplotypes). The frequent occurrence of hybrids and the successful production of seeds and even the occurrence of facultative apomixis in some of them made us consider the evolutionary potential of hybridisation in South African brambles. Furthermore, two octoploid fertile sexual hybrids derived from R. rigidus (with R. pinnatus or R. sect. Arguti) were also discovered, which implies that such hybrids are not rare (the two specimens represented 4.5% of the 44 hybrid individuals with known ploidy in our dataset). Hypothetically, these octoploids could backcross with the tetraploid parent (2x gamete) due to the formation of regular reduced 4x gametes (see Sochor et al. 2022). The offspring (6x) would then share ploidy level and ± 50% of the genome with the first parent. Therefore, only two generations can be sufficient to overcome the ploidy reproduction barrier between tetra- and hexaploids.

Although potentially very effective and explanatory for the extraordinary phenotypic variability of R. rigidus (see Sochor et al. 2022), this 'octoploid bridge' (parallelism of triploid bridge sensu Ramsey & Schemske 1998) does not appear to be a common evolutionary mechanism, because no shared alleles have so far been detected between native and introduced taxa (except for the apparent hybrids), and only a few shared alleles were detected among native species. An example is the Rig6 haplotype in R. rigidus near Alexandria, EC, where this haplotype is shared with R. pinnatus, but R. rigidus bears it in regions much further west (because of shared ancestry) and a transition zone was only documented in the eastern parts of WC (Figure 1D). Another possible example are the B1 ribotypes, which seem to originate from the R. pinnatus/'archaepinnatus' lineage but were found also in R. rigidus in KZN and MP, in all cases together with the C ribotypes in each individual. This last fact could imply that the five R. rigidus individuals (all confirmed hexaploids) can actually be early-generation introgressants, because the ribosomal cistron has not yet been homogenised. However, due to the rather limited sample set, we cannot rule out the possibility of ancient gene flow between the two species and the local preservation of genes of R. pinnatus/'archaepinna-tus' in R. rigidus.

Genetic diversity is geographically structured in R. pinnatus and R. rigidus

Genotypic and allelic diversity and its structuring are crucial information for the management of both introduced and native taxa (especially those of conservation concern), and for understanding their evolutionary behaviour. Our data provide two perspectives. While DNA sequences from the conservative markers provide a wide and superficial overview, the population-genetic data from microsatellites enable much finer and deeper insights, but were restricted in this study to a single model system of R. bergii, R. rigidus, R. pinnatus and their hybrids.

From the wider, phylogeographic perspective, our sequence data imply relatively low genetic diversity in R. thaumasius, R. apetalus, R. transvaalensis, and also R. ludwigii. The latter was, however, included only marginally in this study and its geographic variation may not have been sampled. In contrast, R. pinnatus and particularly R. rigidus exhibit high diversity in plastid DNA and ITS, which is clearly geographically structured. This structuring appears to reflect not only the reticulate evolution discussed above, but also a long-term isolation of populations and limited gene flow among regions. The most conspicuous genetic differences are between the summer-rainfall and winter-rainfall zones (Figures 1 & 2), implying that the haplotypic/ribotypic geographical differentiation may have been accompanied by niche shift, which in turn may be associated with a slight phenotypic shift in R. pinnatus. These geographically linked differences justify its subdivision into two subspecies (Sochor et al. 2022).

In R. rigidus, on the other hand, major phenotypic traits (e.g. structure of leaves, fruit colour, leaf indumentum) do not correspond with haplotypes or ribotypes. Consequently, putatively distinct morphotypes (or species sensu Van de Beek 2021) are widespread across South Africa, but are obviously composed of diverse genotypes of different phylogenetic/genealogical history. In other words, taxonomic treatment of such morphotypes on the species level is contradicted not only by their obligate sexuality (Sochor et al. 2022), but also their diverse polytopic origin. A narrow species concept, such as that used in Europe for apomictic genotypes, is, therefore, clearly inapplicable in South African native taxa.

Clonality implies apomictic spread in R. bergii and R. bergii χpinnatus

Originally, we suspected the R. bergii x R. rigidus hybrids to be partly apomictic and able to persist and spread without recurrent formation of new genotypes via hybridisation (Sochor et al. 2018). This would result in the presence of the same genotype at different locations, and later in the dominance of one or a few successful hybrid genotypes within each region. However, no clonal (i.e., apomictic) duplication was detected among the 11 hybrid individuals in our dataset, despite the fact that the sampling was focused on a small area in westernmost WC (see Supplementary Table S3). This fact supports our later conclusion (Sochor et al. 2022) that these pentaploid hybrids are possibly exclusively sterile and can persist and spread only via vegetative means. On the other hand, the high frequency of occurrence of the hybrid in some regions implies its easy and common recurrent formation.

Surprisingly, clonal duplications were identified in the hybrid R. bergii x pinnatus, although this was not in our primary focus and was therefore represented by only five individuals in our SSR data set. Four of the individuals turned out to belong to a single genotype (Table 2; Supplementary Table S3). The sampled area was very small with distances between the individuals of the clone being 0.33-1.36 km. Such distances, however, seem to be too long to be explained by the spontaneous vegetative spread. As human-mediated propagation can be most likely excluded, the most probable explanation for our finding is asexual dispersal via seeds - apomix-is. We have reported on apomictic seeds in two other homoploid hybrids between native and introduced Ru-bus taxa (R. bergii x thaumasius and R. pinnatus x sect. Arguti; Sochor et al. 2022) but it was not clear whether these seeds were viable and able to secure dispersal.

Similarly, Clark and Jasieniuk (2012) detected (rare) hybridisation among native and introduced Rubus taxa in western United States, as well as apomixis at the level of the embryo. However, seedlings derived from the hybrids exhibited higher allelic variation than would be expected for apomictic offspring, and apomixis, therefore, was not confirmed on the level of seedlings. In contrast, the frequent occurrence of hybrids between Taraxacum officinale (alien) and T. japonicum (native) (Asteraceae) was reported in western Japan despite a very low hybridisation rate (Matsuyama et al. 2018). The number of hybrid genotypes detected in that study in natural populations was surprisingly high but still indicated their apomictic spread. A combination of apo-mixis, high genotypic diversity, and hybrid origin from a native species seemed to promote effective natural selection and propagation of well-adapted genotypes, and thus enhanced invasiveness.

Rubus bergii x pinnatus, as well as the other two hybrids with apomictic seeds, is only locally common and of rather low importance as an invader at this moment. However, these hybrids may potentially pose an initial phase of new invasion that can take advantage of local adaptations of the native parent (Pfennig et al. 2016), clonal multiplication of a superior genotype (Parepa et al. 2014), potential hybrid vigour (Ayres 2004) or simply of being an evolutionary novelty (Ellstrand & Schieren-beck 2006). Targeted sampling of the tetraploid hybrids with subsequent assessment of genotypic diversity and invasive potential is required to evaluate this hypothesis.

High genotypic and allelic diversity were detected in R. rigidus (Table 2), three or four alleles per locus and individual being no exception, which is consistent with its sexual mode of reproduction and allopolyploid origin. In contrast, R. bergii was confirmed to be monoclonal with no signs of recombination or introgression from other taxa, yet with relatively high allelic diversity (reflecting its allopolyploid origin; Table 2). Mono-clonality in our dataset also confirmed the identity of South African R. bergii and European plants usually treated under the name R. vigorosus PJ.Müll. & Wirtg. (Kurtto et al. 2010; Van de Beek 2014). Such extremely low genotypic diversity is consistent with data from other apomictic Rubus microspecies (Király et al. 2017; Šarhanová et al. 2017). Although the monoclonality is contradictory to the relatively high proportion of sexually derived embryos as detected by flow cytometric seed screen in the demonstrably monoclonal microspe-cies (cf. Šarhanová et al. 2012; Sochor et al. 2022), this paradox appears to be a common phenomenon in Ru-bus, so far without explanation (see also Šarhanová et al. 2017). Similar patterns are therefore presumed to occur in other South African alien apomictic blackberries such as R. armeniacus and R. sect. Cuneifolii (both likely monoclonal in South Africa), and R. sect. Arguti with two widespread clones and several genotypes of local occurrence (Sochor et al., 2022).

 

Conclusion

South Africa is not a hotspot for Rubus diversity, but the genus is taxonomically challenging and has been rather overlooked in this region (Van de Beek 2021; So-chor et al. 2022). A combination of traditional pheno-type-based, molecular, and cytometric methods have improved our understanding of its diversity and evolutionary behaviour.

Contrary to previous concerns and notions that the group (or at least some of the taxa) is a hardly intelligible tangle (Sochor et al. 2018; Van de Beek 2021), the biosystematics of South African Rubus is not intractable. Despite frequent hybridisation, gene flow among modern species appears to be weak, as the hybrids mostly do not contribute to further evolution via hybridogenesis or introgression. However, clonal duplications and asexual-ly derived seeds detected in tetraploid hybrids of native and introduced taxa may indicate incipient new plant invasions, and this process deserves further attention.

High phenotypic variability in some species, which has caused much confusion, can readily be explained by their allopolyploid origin and phylogeographic patterns. For example, the extreme variability in R. rigidus seems to be caused by: 1) its hexaploidy; 2) its origin in (at least) three ancestral species (Figure 4); 3) among-pop-ulation isolation and subsequent differentiation particularly between winter-rainfall and summer-rainfall zones but also within the zones to some extent; and 4) probably weak but possibly continuous gene flow from other species, such as R. pinnatus and R. transvaalensis.

The data presented here and in our previous papers are not exhaustive and should be regarded rather as a foundation for further studies. Besides the invasive potential of the tetraploid hybrids, the most challenging task for the future is to unearth evolutionary links between the South African and tropical African Rubus flora, as well as better characterise the diversity of alien, particularly North American taxa, which seem to be quite rich, yet underexplored in the eastern regions of South Africa. However, our experiences show that new and often surprising discoveries can be expected around every corner of (not only) South African batology.

 

Acknowledgements

We greatly thank Zuzana Sochorová, John Burrows, Grant Martin, Costas Zachariades, Brett Mason and Jan Sochor for their help with fieldwork, Bram van de Beek for sharing his material for analysis and Vojtech Zíla for his specimens of R. bergii from Germany.

Competing interests

The authors declare that they have no financial or personal relationship(s) that may have inappropriately influenced them in writing this article.

Authors' contributions

MS performed the sampling, laboratory work, data analyses and wrote the first draft of the manuscript, JCM contributed to the fieldwork and data interpretations, and edited the manuscript.

Ethical considerations

This article followed all ethical standards for research without direct contact with human or animal subjects.

Funding

MS was supported by the Ministry of Agriculture of the Czech Republic, institutional support MZE-RO0423, and a grant from the National Research Foundation, South Africa. Elizabeth Parker and the Mapula Trust are acknowledged with special thanks for their financial support.

Data availability statement

All sequences were deposited in NCBI GenBank (accession numbers OL899048-OL899299 [ITS], OL954095-OL954503 [matK and trnL-trnF]). SSR data matrix is available in Supplementary table S3. Herbarium vouchers are deposited in public herbaria OL, NBG, PRE, NU and L.

 

References

Alice, L. & Campbell, C., 1999, 'Phylogeny of Rubus (Rosa-ceae) based on nuclear ribosomal DNA internal transcribed spacer region sequences', American Journal of Botany 86(1), 81-97, https://doi.org/10.2307/2656957.         [ Links ]

Alice, L.A., Goldman, D.H., Macklin, J.A. & Moore, G., 2015, 'Rubus', in Flora of North America, Editorial Committee (ed.) Flora of North America North of Mexico, vol. 9, New York and Oxford, http://www.efloras.org/florataxon.aspx?-flora_id=1&taxon_id=128837.         [ Links ]

Ayres, D.R., Zaremba, K. & Strong, D.R., 2004, 'Extinction of a common native species by hybridization with an invasive congener', Weed Technology 18, 1288-1291, http://www.jstor.org/stable/3989636.         [ Links ]

Bandelt, H.J., Forster, P & Rohl, A., 1999, 'Median-joining networks for inferring intraspecific phylogenies', Molecular Biology and Evolution 16(1), 37-48, https://doi.org/10.1093/oxfordjournals.molbev.a026036.         [ Links ]

Carter, K.A., Liston, A., Bassil, N.V., Alice, L.A., Bushakra, J.M., Sutherland, B.L., Mockler, T.C., Bryant, D.W. & Hummer, K.E., 2019, 'Target capture sequencing unravels Ru-bus evolution', Frontiers in Plant Science 10, 1-18, https://doi.org/10.3389/fpls.2019.01615.         [ Links ]

Clark, L. V. & Jasieniuk, M., 2012, 'Spontaneous hybrids between native and exotic Rubus in the Western United States produce offspring both by apomixis and by sexual recombination', Heredity 109, 320-328, https://doi.org/10.1038/hdy.2012.45.         [ Links ]

Doyle, J.J. & Doyle, J.L., 1987, 'A rapid DNA isolation procedure for small quantities of fresh leaf tissue', Phytochemical Bulletin 19:11-15.         [ Links ]

Dunning, L.T. & Savolainen, V., 2010, 'Broad-scale amplification of matK for DNA barcoding plants, a technical note', Botanical Journal of the Linnean Society 164(1), 1-9, https://doi.org/10.1111/j.1095-8339.2010.01071.x.         [ Links ]

Ellstrand, N.C. & Schierenbeck, K.A., 2006, 'Hybridization as a stimulus for the evolution of invasiveness in plants?', Euphytica 148, 35-46, https://doi.org/10.1007/s10681-006-5939-3.         [ Links ]

Fehrer, J., Krak, K. & Chrtek, J., 2009, 'Intra-individual polymorphism in diploid and apomictic polyploid hawkweeds (Hieracium, Lactuceae, Asteraceae): disentangling phylo-genetic signal, reticulation, and noise', BMC Evolutionary Biology 9, 239, https://doi.org/10.1186/1471-2148-9-239.         [ Links ]

Graham, J., Smith, K., MacKenzie, K., Jorgenson, L., Hack-ett, C. & Powell, W., 2004, 'The construction of a genetic linkage map of red raspberry (Rubus idaeus subsp. idae-us) based on AFLPs, genomic-SSR and EST-SSR markers', Theoretical and Applied Genetics 109(4), 740-749, https://doi.org/10.1007/s00122-004-1687-8.         [ Links ]

Graham, J., Smith, K., Tierney, I., MacKenzie, K. & Hackett, C.A., 2006, 'Mapping gene H controlling cane pubescence in raspberry and its association with resistance to cane botrytis and spur blight, rust and cane spot', Theoretical and Applied Genetics 112(5), 818-831, https://doi.org/10.1007/s00122-005-0184-z.         [ Links ]

Gustafsson, C.E., 1934, 'Rubi africani', Arkiv för Botanik 26A, 1-68.         [ Links ]

Henderson, L., 2011, 'Rubus species - brambles, blackberries and others', SAPIA News 19, 1-9.         [ Links ]

Huson, D.H. & Bryant, D., 2006, 'Application of phylogenetic networks in evolutionary studies', Molecular biology and evolution 23(2), 254-67, https://doi.org/10.1093/molbev/msj030.         [ Links ]

Király, G., Sochor, M. & Trávníček, B., 2017, 'Reopening an old chapter: a revised taxonomic and evolutionary concept of the Rubus montanus group', Preslia 89, 309-331, https://doi.org/10.23855/preslia.2017.309.         [ Links ]

Kurtto, A., Weber, H.E., Lampinen, R. & Sennikov, A.N. (eds), 2010, 'Atlas florae Europaeae. Distribution of vascular plants in Europe, vol. 15: Rosaceae (Rubus)', Committee for Mapping the Flora of Europe and Societas Biologica Fennica Vanamo, Helsinki.         [ Links ]

Matsuyama, S., Morimoto, M., Harata, T., Nanami, S. & Itoh, A., 2018, 'Hybridization rate and genotypic diversity of apomictic hybrids between native (Taraxacum japonicum) and introduced (T. officinale) dandelions in western Japan', Conservation Genetics 19, 181-191, https://doi.org/10.1007/s10592-017-1014-y.         [ Links ]

Meirmans, PG. & Van Tienderen, PH., 2004, 'Genotype and genodive: two programs for the analysis of genetic diversity of asexual organisms', Molecular Ecology Notes 4(4), 792794, https://doi.org/10.1111/j.1471-8286.2004.00770.x.         [ Links ]

Nolte, A.W. & Tautz, D., 2010, 'Understanding the onset of hybrid speciation', Trends in Genetics 26, 54-58.         [ Links ]

Ochieng, J.A., Oyoo, M.E., Gesimba, R.M., Korir, PC., Okwiri Ojwang', PP & Owuoche, J.O., 2018, 'Genetic diversity of blackberry (Rubus subgenus Rubus Watson) in selected counties in Kenya using simple sequence repeats (SSRs) markers', African Journal of Biotechnology 17(39), 12471264, https://doi.org/10.5897/ajb2018.16613.         [ Links ]

Ochieng, J.A., Gesimba, R.M., Oyoo, M.E., Korir, P.C., Ow-uoche, J.O. & Miheso, M., 2019, 'Morphological characterization of blackberry (Rubus subgenus Rubus Watson) genetic resources in Kenya', African Journal of Plant Science. 13(11), 297-308, https://doi.org/10.5897/ajps2018.1703.         [ Links ]

Parepa, M., Fischer, M., Krebs, C. & Bossdorf, O., 2014, 'Hybridization increases invasive knotweed success', Evolutionary Applications 7(3), 413-420, https://doi.org/10.1111/eva.12139.         [ Links ]

Peakall, R. & Smouse, P.E., 2012, 'GenAlEx 6.5: genetic analysis in Excel. Population genetic software for teaching and research - an update', Bioinformatics 28(19), 2537-2539, https://doi.org/10.1093/bioinformatics/bts460.         [ Links ]

Pfennig, K.S., Kelly, A.L. & Pierce, A.A., 2016, 'Hybridization as a facilitator of species range expansion', Proceedings of the Royal Society B: Biological Sciences 283, 20161329, http://dx.doi.org/10.1098/rspb.2016.1329.         [ Links ]

Ramsey, J. & Schemske, D.W., 1998, 'Pathways, mechanisms, and rates of polyploid formation in flowering plants', Annual Review of Ecology and Systematics 29, 467-501, https://doi.org/10.1146/annurev.ecolsys.29.1.467.         [ Links ]

Rieseberg, L., 1995, 'The role of hybridization in evolution: old wine in new skins', American Journal of Botany 82, 944-953.         [ Links ]

Šarhanová, P, Sharbel, T.F., Sochor, M., Vašut, R.J., Dančák, M. & Trávníček, B., 2017, 'Hybridization drives evolution of apomicts in Rubus subgenus Rubus: evidence from microsatellite markers', Annals of Botany 120, 317-328, https://doi.org/10.1093/aob/mcx033.         [ Links ]

Šarhanová, P, Vašut, R.J., Dančák, M., Bures, P & Trávníček, B., 2012. 'New insights into the variability of reproduction modes in European populations of Rubus subgen. Rubus: how sexual are polyploid brambles?', Sexual Plant Reproduction 25, 319-335, https://doi.org/DOI10.1007/s00497-012-0200-9.         [ Links ]

Sochor, M., Hrones, M. & Manning, J.C., 2022, 'Guide to the genus Rubus L. (Rosaceae) in South Africa - disentangling a taxonomic Gordian knot with the help of ploidy and reproductive data', South African Journal of Botany 147, 511-567, https://doi.org/10.1016/j.sajb.2022.01.044.         [ Links ]

Sochor, M., Šarhanová, P., Pfanzelt, S. & Trávníček, B., 2017, 'Is evolution of apomicts driven by the phylogeography of the sexual ancestor? Insights from European and Caucasian brambles (Rubus, Rosaceae)', Journal of Biogeography 4, 2717-2728, https://doi.org/10.1111/jbi.13084.         [ Links ]

Sochor, M. & Trávníček, B., 2016, 'Melting pot of biodiversity: first insights into the evolutionary patterns of the Colchic bramble flora (Rubus subgenus Rubus, Rosaceae)', Botanical Journal of the Linnean Society 181, 610-620, https://doi.org/10.1111/boj.12436.         [ Links ]

Sochor, M., Trávníček, B. & Manning, J.C., 2018, 'Biosyste-matic revision of the native and naturalised species of Ru-bus L. (Rosaceae) in the Cape Floristic Region, South Africa', South African Journal of Botany 118, 241-259, https://doi.org/10.1016/j.sajb.2018.07.015.         [ Links ]

Sochor, M., Vašut, R.J., Sharbel, T.F. & Trávníček, B., 2015, 'How just a few makes a lot: Speciation via reticulation and apomixis on example of European brambles (Ru-bus subgen. Rubus, Rosaceae)', Molecular Phylogenet-ics and Evolution 89, 13-27, https://doi.org/10.1016/j.ympev.2015.04.007.         [ Links ]

Stirton, C.H., 1981, 'Notes on the taxonomy of Rubus in southern Africa', Bothalia 13(3), 331-332.         [ Links ]

Taberlet, P, Gielly, L., Pautou, G. & Bouvet, J., 1991, 'Universal primers for amplification of 3 noncoding regions of chloroplast DNA', Plant Molecular Biology 17, 1105-1109, https://doi.org/10.1007/Bf00037152.         [ Links ]

Van de Beek, A., 2014, 'Rubus costifolius and R. bergii in the National Herbarium of Victoria', Muelleria 32, 52-57, https://doi.org/10.5962/p.295687.         [ Links ]

Van de Beek, A., 2021, 'Rubi Capenses: a further contribution to the knowledge of the genus Rubus (Rosaceae) in South Africa', Phytotaxa 515(1), 1-71, https://doi.org/10.11646/phytotaxa.515.1.1.         [ Links ]

Weber, H. E., 1996, 'Former and modern taxonomic treatment of the apomictic Rubus complex', Folia Geobotanica Phytotaxonomica 31, 373-380.         [ Links ]

White, T.J., Bruns, T., Lee, S. & Taylor, J.W., 1990, 'Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics', in M.A. Innis, D.H. Gelfand, J.J. Sninsky, & T.J. White (eds), PCR Protocols: A Guide to Methods and Applications, pp. 315-322, Academic Press, New York.         [ Links ]

Woodhead, M., McCallum, S., Smith, K., Cardle, L., Mazzitelli, L. & Graham, J., 2008, 'Identification, characterisation and mapping of simple sequence repeat (SSR) markers from raspberry root and bud ESTs', Molecular Breeding 22(4), 555-563, https://doi.org/10.1007/s11032-008-9198-y.         [ Links ]

 

 

Correspondence:
Michal Sochor
e-mail: michal. sochor@volny.cz

Submitted: 27 January 2022
Accepted: 10 November 2022
Published: 14 April 2023

 

 

Supplementary Data

The supplementary data is available in pdf: [Supplementary data]