Hybridization of Oxalis corniculata and O. dillenii in their non-native range

Abstract Many species have been introduced beyond their native ranges and many have become global weeds. Human mediated dispersal has removed the geographic isolation of these species, reversing millions of years of independent evolution. Examples are the Oxalis species in section Corniculatae where several species have become invasive. Here we characterize and formally describe a hybrid between O. dillenii and O. corniculata, which occurs spontaneously in Belgium and Japan. Oxalis corniculata is probably native to Japan, but both species are alien to Belgium and O. dillenii is native to North America. We formally name this hybrid as Oxalis × vanaelstii. Although this hybrid is sterile, it is nevertheless vigorous and perennial. Both parent species grow as weeds in gardens; therefore, it is likely to be more common than currently appreciated in countries where these species co-occur.


Introduction
The human mediated redistribution of plant species, whether by accident or design, facilitates the hybridization of species that were once separated by millennia of independent evolution (Largiadèr 2008). Hybridization is a major force in the evolution of plants and the fingerprint of this breakdown in geographic isolation is likely to be marked on the future evolution of plants (Vallejo-Marín and Hiscock 2016). Hybrids can occur in a geographic range where one parent is native and the other is alien (e.g. Ayres and Strong 2001), but also between two species brought together in a range where they are both alien (e.g. Vallejo-Marín 2012).
The genus Oxalis is one of the weediest genera and many species have become naturalized beyond their native ranges (Pyšek et al. 2017). Therefore, Oxalis species from all over the world are being brought together with sibling species where they may have the opportunity to cross pollinate.
Molecular genetic techniques have shown that hybridization has been an important force in the evolution of Oxalis (Emshwiller and Doyle 2002;Vaio et al. 2013;Vaio et al. 2016). However, few hybrids from the genus have been formally described and, as Salter (1944) and du Preez et al. (2018) point out, field observations of hybrids in the genus are rare.
Studies have indicated that hybridization can occur between species in the section Corniculatae (Mulcahy 1965;Doust et al. 1981). For example, artificial crosses between O. corniculata L. and O. dillenii Jacq. were reported to be fertile irrespective of the direction of the crossing (Mulcahy 1965). However, only recently have sterile wild hybrids of O. dillenii and O. corniculata been characterized from Japan (Fukatsu et al. 2018). Oxalis dillenii is a North American species, which has naturalized in other parts of the world, including Japan and Europe (Mito and Uesugi 2004;. Oxalis corniculata is now a cosmopolitan species, probably native to Japan (Groom et al. 2019 (Turland et al. 2018).

Hybrid origin and habitat
For some years, one of the authors (IH) has maintained a small living collection of Oxalis species in his home garden in Aalter (Belgium, prov. of East Flanders). Among these plants are Oxalis corniculata var. atropurpurea and O. dillenii, both collected as weeds in plant containers offered for sale in garden centers and nurseries (although the former taxon was already present in the garden as a weed years before additional plants were deliberately introduced for cultivation). The precise origin of the cultivated plants is not known. Over the years, plants have escaped from the pots in which they are cultivated, as ripe seeds are catapulted away over distances that frequently exceed 1 m. It is among these escapees that, in 2016, putative hybrids were detected (voucher BR0000025668254V, see the taxonomic description). A few more plants of the possible hybrid were also found in 2017 and 2018. They grew next to plants of the two parents and flowered abundantly, yet produced no fruits. They were observed in a neglected part of the garden where selective manual weeding created opportunities for several different Oxalis to maintain themselves among competitors such as Epilobium sp., Poa annua and other common weeds. The hybrid grew on moderately damp, sandy, humus-rich soil. We note that all populations are homostylous.

Molecular protocols and sequence analysis
We applied a modified CTAB protocol for total genomic DNA isolation (Tel-Zur et al. 1999). Secondary metabolites were removed by washing ground, silica-dried leaf material with 1000 µL extraction buffer (100 mM Tris-HCl pH 8, 5 mM EDTA pH 8, 0.35 M sorbitol). Samples were incubated at 60 °C (1 h) with a CTAB lysis buffer (incl. 1% PVP-40 and 0.3% 2-mercaptoethanol). Extraction was done twice with SEVAG (chloroform-isoamylalcohol (24/1 v/v)) and was followed by an isopropanol precipitation (0.8 volumes). After centrifugation, the pellet was washed in 70% ethanol, air-dried, and dissolved in 50 µL buffer (10 mM Tris-HCl pH 8, 1 mM EDTA pH 8). Amplification and sequencing of ITS and trnL-F was performed using the primers of White et al. (1990) and Taberlet et al. (1991), respectively. PCR reactions of ITS and trnL-F were carried out using a touchdown PCR protocol (25 µL). PCR reactions were initiated with a 3-min heating at 95 °C followed by 20 cycles consisting of a denaturation step of 30 s at 95 °C, a 30 s annealing step starting at 58 °C decreasing each cycle with 0.5 °C, and an extension step of 72 °C for 60 s, and ending with 10 cycles (similar in timing as with the first 20 cycles) at an annealing temperature of 48 °C. PCR products were enzymatically purified using the ExoSap protocol and sequenced by the Macrogen sequencing facilities (Macrogen Europe, Amsterdam, Netherlands). Contiguous sequences were assembled using Geneious v7.0.6 (Biomatters, New Zealand). Automatic alignments were carried out with MAFFT (Katoh et al. 2002) under an E-INS-i algorithm, a 100 PAM/k = 2 scoring matrix, a gap open penalty of 1.3 and an offset value of 0.123. Subsequent manual fine-tuning of the aligned dataset was done in Geneious v7.0.6. Gaps were treated as missing data. All sequences and sample information were uploaded to European Nucleotide Archive and can be found under the project number PRJEB41412.

Flow cytometry
Genome sizes were measured with a Partec PAS III flow cytometer equipped with an 20 mW 488 nm solid state laser. Samples were prepared using the commercial kit Cystain PI absolute P (Partec, Germany). Each of the three plants of interest, O. corniculata, O. dillenii and the new hybrid were analyzed separately and chopped with a sharp razor blade at room temperature together with an internal standard (Zea mays, 2C = 5.43 pg, Lysák and Doležel (1998)). The nuclei suspension was filtered through a 50 µm mesh CellTrics disposable filter (Partec GmbH, Münster, Germany), stained with propidium iodide following the specifications of the kit. At least five replicates per sample were incubated for at least 30 min in the dark before measuring with the flow cytometer on two (O. corniculata) to three (O. dillenii, hybrid) different days using the FloMax software (Quantum Analysis, Germany). Average genome size was calculated from the relative fluorescence intensities of the sample of interest and the internal standard with known genome size.

Guard cell length
Epidermal leaf impressions were made from the abaxial side of the leaves in the middle of the leaf, between the midvein and edge. Transparent nail polish (Bourjois Crystal ball) was used to make the impressions which, once dried, were mounted pointing upward with double-sided tape (Scotch) on a microscope slide. Stacked photomicrographs were taken per leaf print (view fields = 0.09 mm 2 ) using a digital microscope (VH-5000 Ver 1.5.1.1, Keyence Corporation) with full coaxial lighting and default factory settings for shutter speed at ×1000 lens magnification (VH-Z250R).

Pollen electron microscopy and viability count
Anthers were collected from mature buds and no distinction was made between the two whorls of anthers. Material for scanning electron microscopy was washed in 70% ethanol for 20 minutes and washed twice with 100% DMM (dimethoxymethane) for 20 minutes while being sonicated each time for a couple of seconds. Then it was washed with 100% acetone. The material was critical point dried using liquid CO 2 with a Leica EMCPD3000 critical point dryer. The dried samples were mounted on aluminum stubs using carbon adhesive tape and coated with a platinum palladium mix with a Cressington JFC-2300/208HR sputter coater. SEM images were obtained with a JEOL JSM7100F field emission scanning electron microscope. Pollen size was measured on 200 grains for each parent and for the hybrid. The diameter of the roughly spherical pollen was measured horizontally on the photograph regardless of orientation of the pollen grain. Pollen viability was measured using the staining protocol of Peterson et al. (2010). Viability was evaluated by counting the proportion of stained versus unstained pollen grains. More than 100 grains were counted per flower for a total of three flowers.

Morphology
Oxalis corniculata var. atropurpurea has purple-brown leaves and O. dillenii bright green leaves (Figs 1, 2). The hybrid is intermediate with pale purple-brown leaves and sometimes green leaves with a purple underside. Guard cells of O. corniculata are about 30% larger than those of O. dillenii (Table 1). However, the guard cells of the hybrid are almost exactly the same size as those of O. dillenii. The hybrid also has an intermediate growth form. Oxalis dillenii is generally erect, though it becomes decumbent with age because the main stem tends to topple over once the crown becomes too heavy. Oxalis corniculata var. atropurpurea is largely prostrate and roots at the nodes. The hybrid also roots at the nodes, but is not as strongly creeping as Oxalis corniculata var. atropurpurea (Fig. 1).
The stipules and stem of the putative hybrid are illustrated in Figure 2. Oxalis corniculata has short adnate stipules with an angular tip. Oxalis dillenii has narrow stipules with a more rounded tip. The hybrid is intermediate, though perhaps closer to O. corniculata. The stipule is distinct, longer than wide, but with more of a rounded tip than O. corniculata.
Stem and petiole hairs are also visible on Figure 2. In Oxalis corniculata var. atropurpurea they are patent and sometimes glandular, whereas in O. dillenii they are antrorse, eglandular and appressed to the stem or petiole. The hybrid is again intermediate in character with eglandular, antrorsely directed hairs, but they are not as strongly appressed as in O. dillenii.   pollen is larger (mean 31 µm, SD = 3.1 µm, n = 200) than O. dillenii pollen (mean 26 µm, SD = 2.6 µm, n = 200) (Fig. 3). On average, both parents have larger pollen grains than the putative hybrid (mean 20 µm, SD = 6.2 µm, n = 200). However, note the large standard deviation and broad spread of hybrid pollen in Figure 3. This is both because some grains are exceedingly small and because about 4% of hybrid pollen grains exceed 40 µm. In pollen viability staining pollen of the parents are strongly stained and all appear to be viable. Viability staining of the pollen of the putative hybrid is 45.6% (SD = 3.0%, n = 3 flowers). Those pollen of the hybrid that were exceptionally large were seen to be stained strongly whereas small grains remained unstained.

Sequence analysis results
ITS polymorphisms between O. corniculata and O. dillenii occurred at nucleotide positions 51,52,102,103,115,122,130,138,152,165,178,187,224,232,235,237,379,423 (Fig. 5). ITS sequencing of the putative hybrid shows a clear ambiguity at each of the aforementioned polymorphic loci consistent with both parental alleles (Fig. 5). We also note a frameshift at position 423 where there are four cytosine base pairs in O. dillenii, versus five cytosine base pairs in O. corniculata. This frameshift blurs the sequencing read in the hybrid beyond this point. The chloroplastic marker trnL-F has a single polymorphism at position 279 (Suppl. material 1: Fig. S1). At this position, Oxalis corniculata has a guanine as character state whereas both O. dillenii and the putative hybrid are characterized by a cytosine.

Discussion
Fukatsu et al. (2018) Kubešová et al. (2010) and Vaio et al. (2013). For tetraploid O. dillenii Kubešová et al. (2010) found a genome size of 1.01 pg, also comparable to the 0.88 pg measured in our study. Fukatsu et al. (2018) also counted chromosomes numbers and concluded that their parental O. corniculata is an octoploid (2n = 48), their O. dillenii is a tetraploid (2n = 24) and the hybrid is a hexaploid. This is also consistent with our hybrid and parental plants. The larger size of the O. corniculata pollen is consistent with a larger genome. The small size of the majority of hybrid pollen is consistent with its sterility as is the poor pollen viability. Similarly, the larger guard cell size agrees with a larger genome size for O. corniculata. Guard cell size is generally positively correlated with genome size in vascular plants and Oxalis (Beaulieu et al. 2008;Jooste 2015).
The DNA sequencing results are consistent with our sterile plant being a hybrid of O. corniculata and O. dillenii. The nuclear ITS sequence has ambiguity at each of the polymorphic loci between the proposed parent species, which suggests the hybrid genome contains alleles of both parents. The chloroplastic marker, trnL-F, differs in only one nucleotide in O. corniculata and O. dillenii. This single polymorphism indicates that O. dillenii is the likely maternal parent. This also accords with the results of Fukatsu et al. (2018) who found that O. dillenii was the maternal parent in 85% of Japanese specimens.
Our results agree with, and add to those of, Fukatsu et al. (2018), giving further support to the parentage of these sterile putative hybrids. The Corniculatae are a group of similar-looking taxa with few distinctive features. The plants described here are distinctive in their sterility, however, fertile hybrids would be considerably harder to recognize, particularly as they are likely to be more closely related and perhaps more similar morphologically. In North America and Japan, where O. dillenii and O. corniculata are respectively native, there is likely to be more genetic variation and perhaps greater chance of fertile hybrids being produced. Much more work is needed to understand the extent of hybridization in this group. Diagnosis. Intermediate in characters between its parents O. corniculata and O. dillenii. Prostrate to ascending, mid-sized stipules, stem hairs antrorse, not tightly appressed to the stem. Flowers with at most weak orange marks in the throat, marks sometimes absent.

Oxalis
Description. A short-lived perennial, prostrate to ascending, herb with a thin taproot. Leaves trifoliolate with three similarly sized, heart-shaped leaflets. A narrow stipule is fused to the base of the petiole and is intermediate in width between O. corniculata and O. dillenii. The stem hair density is moderate with simple, arcuate, antrorse, pointed hairs. Not strongly appressed to the stem as in O. dillenii. The leaves are green or purple-brown, though if purple, not as darkly colored as Oxalis corniculata var. atropurpurea. The flowers are yellow, sometimes with weak orange streaks in the throat. Fruits are unknown.
Etymology. Oxalis × vanaelstii is named to commemorate the Belgian naturalist, conservationist and mycologist Etienne Vanaelst  who, as a volunteer collaborating with mycologists at Ghent University, contributed to a better understanding of the diversity of mushrooms, especially those growing in and around his hometown, Knesselare (prov. of East Flanders).
Habitat. Gardens. Distribution area. Europe and Japan where parental distribution overlaps. The hybrid is also highly likely to occur in North America where the parents also co-occur (Eiten 1963).