Mimulus peregrinus (Phrymaceae): A new British allopolyploid species

Abstract Polyploidization plays an important role in species formation as chromosome doubling results in strong reproductive isolation between derivative and parental taxa. In this note I describe a new species, Mimulus peregrinus (Phrymaceae), which represents the first recorded instance of a new British polyploid species of Mimulus (2n = 6x = 92) that has arisen since the introduction of this genus into the United Kingdom in the 1800’s. Mimulus peregrinus presents floral and vegetative characteristics intermediate between Mimulus guttatus and Mimulus luteus, but can be distinguished from all naturalized British Mimulus species and hybrids based on a combination of reproductive and vegetative traits. Mimulus peregrinus displays high pollen and seed fertility as well as traits usually associated with genome doubling such as increased pollen and stomata size. The intermediate characteristics of Mimulus peregrinus between Mimulus guttatus (2n = 2x = 28)and Mimulus luteus (2n = 4x = 60-62), and its close affinity with the highly sterile, triploid (2n = 3x = 44-45) hybrid taxon Mimulus × robertsii (Mimulus guttatus × Mimulus luteus), suggests that Mimulus peregrinus mayconstitute an example of recent allopolyploid speciation.


Introduction
Th e genus Mimulus (Phrymaceae) comprises more than 120 species, the majority (75%) of which occur in western North America, and the remaining having a world-wide distribution including Eastern North America, South America, Australia, the Himalayas, Japan and Madagascar (Grant 1924, Beardsley and Olmstead 2002, Wu et al. 2007).
Species of Mimulus have been spread outside their native range due to deliberate and accidental introductions. For example, M. guttatus, a native of western North America, is now found in New Zealand and more than 16 European countries (Truscott et al. 2008, van Kleunen and Fischer 2008, Tokarska-Guzik and Dajdok 2010. In some of these areas of introduction, M. guttatus has become naturalized and widely distributed, forming a nontrivial component of the local fl ora (e.g. Wales, Northern England, Scotland, Poland, Germany and New Zealand; Roberts 1964, Stace 2010, Tokarska-Guzik and Dajdok 2010. In the United Kingdom (UK), naturalized populations of Mimulus are widespread (Preston et al. 2002), and the genus is represented here by three currently extant species (M. guttatus, M. luteus and M. moschatus), and a complex array of interspecifi c hybrids, some of which are locally invasive (Stace 2010).
Polyploidization plays a particularly important role in species formation, as chromosome doubling results in immediate and strong reproductive isolation between the derivative and parental species Willis 2007, Köhler et al. 2010). It is therefore not surprising that polyploidization is often thought to be fundamental to angiosperm diversifi cation (Stebbins 1950, Grant 1971, Ramsey and Schemske 2002, Soltis 2005, Wood et al. 2009). In Mimulus, speciation by hybridization and polyploidization may have played an important role during the diversifi cation of this group (Vickery 1995, Beardsley et al. 2004. For instance, allopolyploidization between diploid M. guttatus and M. nasutus has given rise to a widespread North American tetraploid taxon that is strongly reproductively isolated from its progenitors (Sweigart et al. 2008). Despite the importance of hybridization and polyploidization for plants in general, the opportunity to study early events in speciation via this route is limited by the small number of angiosperm species known to have originated via allopolyploidization in the last 150 years (e.g. Spartina anglica (Ayres and Strong 2001), Tragopogon mirus, T. miscellus (Soltis et al. 2004(Soltis et al. , 2012, Senecio cambrensis and S. eboracensis (Abbott and Lowe 2004)). Th e discovery of a recently formed polyploid hybrid species in the wild therefore would provide a window of opportunity to study the evolution and speciation of polyploid taxa.
In this note, I describe a new, fertile, polyploid (2n = 6x = 92) species of Mimulus (Phrymaceae), M. peregrinus, which has currently been found in a single locality in the Lowther Hills, Scotland. A comparison of vegetative and reproductive morphology, DNA content, and chromosome number of this new polyploid species against other British Mimulus, strongly suggests a hybrid origin for M. peregrinus and a close affi nity with the sterile triploid hybrid M. × robertsii. I speculate that M. peregrinus may represent the hexaploid derivative of a hybrid between M. guttatus and M. luteus, although a careful examination of additional populations of both parental and hybrid taxa is required to elucidate the genetic origin, extent and distribution of this new polyploid species. If an allopolyploid origin is demonstrated, M. peregrinus has the potential to serve as a study system to understand the evolutionary processes associated with the origin of species through hybridization and polyploidization following the breakdown of geographic barriers caused by human-assisted dispersal.

Methods
Field surveys in August 2011 uncovered the existence of fertile individuals in a large population of M. × robertsii in South Lanarkshire, Scotland. To further investigate these unusual plants, open-pollinated seeds were collected on 27 August 2011 from multiple seed-bearing fruits in a single patch at Shortcleuch Waters, near Leadhills, South Lanarkshire, Scotland (NS 9029 1578; 55.4237°N, 3.7349°W). Field-collected seeds-accession number 11-LED-seed-were germinated and grown in a controlled environment cabinet (Microclima 1750E; Snijders Scientifi c, Tilburg, the Netherlands) at the University of Stirling under 16 light-hours at 24°C and 8 dark-hours at 16°C, and 70% constant humidity. Individual plants were grown in 0.37 l round pots, fi lled with general purpose peat-sand compost (Sinclair, Lincoln, Lincolnshire, UK), and kept on plastic trays with abundant water. Plants were sporadically treated with SB Plant Invigorator (Fargro Ltd, Littlehampton, West Sussex, UK) to control for fungal infections. Seven plants were brought to fl owering (F 1 generation; 11-LED-seed-1 to 11-LED-seed-7), and each individual plant was used to generate F 2 off spring via manual self-fertilization of emasculated fl owers kept inside the pollinator-free growth cabi-net. A representative individual of this F 2 generation (11-LED-seed-2-14) was chosen as the holotype for the type description presented here (deposited at the Herbarium of the Royal Botanic Garden Edinburgh; E).
Pollen measurements were conducted using fresh pollen fi xed in 1 ml of 70% ethanol and dyed with 50 μl of lactophenol-aniline blue (Kearns and Inouye 1993). Darkly stained grains were considered viable (Sweigart et al. 2006). Pollen diameter was measured at the widest point in expanded pollen grains using image analysis software (analySIS, Olympus Soft Imaging Solutions, Münster, Germany) at 200× magnifi cation in an Olympus BX50 light microscope.
Stomata size was measured in casts obtained from the adaxiall side of healthy leaves. A negative cast was fi rst obtained with polysiloxane precision impression material (Xantoprene VL Plus, Heraeus Kulzer Gmbh, Hanau, Germany), and a positive cast was then generated with quick-drying nail polish. Measurements of stomata length and width were done using a light microscope at 400×. Chromosome counts were conducted by John Bailey (University of Leicester) in mitotic cells from root tips of two F 2 individuals (11-LED-seed-3-21 and 11-LEDseed-5-8). Genome size was measured using DAPI-stained nuclei analysed in a Cy-Flow ML fl owcytometer (Partec GmbH, Münster, Germany) in a commercial facility (Plant Cytometry Services, Schijndel, Th e Netherlands) in six F 1 individuals (11-LEDseed-1 to 11-LED-seed-4, 11-LED-seed-6, 11-LED-seed-7). Vinca major was used as internal standard (2n = 92, 2C = 3.80 pg; Obermayer and Greihulber 2006). Because DAPI preferentially binds to AT-rich regions, the fl ow cytometry results presented here must not be treated as absolute measurements of DNA content.

Data resources
Th e data underpinning the analysis reported in this paper are deposited at GBIF, the Global Biodiversity Information Facility, http://ipt.pensoft.net/ipt/resource. do?r=mimulus_peregrinus
Ecology. Occurring on the banks of a stream on a substrate of sand and shingle. M. peregrinus is found alongside M. × robertsii, which is locally common. Flowering of Mimulus in this region starts in early June. Seeds of M. peregrinus were collected in August.
Etymology. Th e name is taken from the Latin peregrinus -foreigner, traveller. Preliminary conservation status. Currently known only from a single collection outside of a protected area, M. peregrinus is provisionally assessed as Critically Endangered (CR D; population size estimated to number less than 50 mature individuals) (IUCN 2011).

Discussion
Mimulus peregrinus can be distinguished from closely related Mimulus species and their hybrids in the UK based on a number of morphological and functional characters (Table 1, Fig. 2). Its chromosome number, DNA content, larger stomata and pollen grain size, clearly indicate that M. peregrinus is a polyploid species. Although the parentage of this new polyploid has not been fi rmly established yet, its close affi nity with M. × robertsii suggest that M. peregrinus has been derived from hybridization between M. guttatus and M. luteus and thus it might have arisen through a recent (<140 years) allopolyploidization event. Below I contrast M. peregrinus with related Mimulus taxa in the UK, and end with a brief discussion on its putative origin.  peregrinus. Chromosome number and genome content as measured in fl ow cytometry are also diagnostic characters to distinguish these two species (Table 1, Fig. 3). 2. Mimulus luteus L. (Section Simiolus Green) (blood-drop emlets). M. luteus, is a group of polymorphic perennial herbs comprising several interfertile varieties that are distinguished based on the presence, size and colour of markings on the corolla lobes. Taxa in this group include M. luteus var. rivularis Lindl. 1826, with a single large red spot on the middle lower lip; M. luteus var. variegatus (Lodd.) Hook 1834, with pale yellow corollas tinted with pink at the lobe margins; and M. luteus var. youngana Hook 1834 (= M. smithii Lindl 1835, not Paxton), with deep yellow corollas and lobes with large red spots at the margins (Grant 1924). In the UK, many extant populations of M. luteus likely represent crosses between taxa in this interfertile group (e.g. M. luteus var. rivularis × M. luteus var. variegatus) (MVM pers. obs.), and present highly variable patterns of spots and blotches in the corolla lobes. M. peregrinus can be distinguished from most species and hybrids in the M. luteus aggregate by its more robust habit, elliptical leaves with dentate and slightly irregular margins, and the presence of only a small, faint, elongated central spot in the lower lip. Most importantly, M. peregrinus possesses simple hairs in the calyx, which are always absent from all varieties of M. luteus. Other diagnostic characters of M. peregrinus are pollen grain size, stomata size, DNA content as measured in fl ow cytometry and chromosome number (Table 1, Fig. 3).
3. Mimulus cupreus Dombrain (Section Simiolus Green) (copper monkeyfl ower). M. cupreus with orange to yellow corollas, and which is closely related to M. luteus, has been reported in the UK but most likely in error for the hybrid between M. guttatus and M. cupreus (M. × burnetii S. Arn.) (Stace 2010). In contrast with M. peregrinus, the copper monkeyfl ower usually has orange corollas, more open corolla throat, lacks simple hairs in the calyx, and has a smaller chromosome complement (2n = 62).
(4) M. × robertsii Silverside (M. guttatus × M. luteus). A highly pollen-and seedsterile, perennial herb rooting at the nodes, its yellow fl owers are marked with orange to red to brown spots of various sizes in the petal lobes (Roberts 1964, Silverside 1990, Silverside 1998,Stace 2010. Th e corolla is 2.5-4.5cm in length and the throat is partially open (Stace 2010). Th is is a taxon of variable pubescence, but is usually hairy in the upper parts of the plant (Stace 2010) (Stace 2010). In the UK it can be found up to 610 m (Ochil Hills, Scotland), and is suggested to be the commonest taxon of high ground (Preston et al. 2002, Stace 2010. M. peregrinus resembles M. × robertsii rather closely in habit, size and general vegetative and fl oral morphology, suggesting a close affi nity between these two taxa (Table  1). M. × robertsii and M. peregrinus can be distinguished by their diff erences in chromosome number, pollen and seed fertility, pollen grain size, and stomata size (Table  1). M. peregrinus presents consistently high levels of pollen fertility (0.86 ± 0.04) and is capable of producing abundant seed set following artifi cial pollination. In contrast, both natural and artifi cial specimens of M. × robertsii present very low levels of pollen viability (proportion of viable pollen = 0.05 ± 0.01, for both naturalized (N = 7) and synthetic hybrids (N = 15)), and do not set seed following artifi cial pollination (Roberts 1964) (see also Table 1). In addition, the two taxa diff er markedly in chromosome number: M. × robertsii is a triploid (e.g. 2n = 45), while M. peregrinus has twice as many chromosomes (2n = 92), and this diff erence in genome size is clearly seen in fl ow cytometry analysis of DAPI-stained nuclei (Fig. 3). Finally, associated with the diff erent genome size of the two taxa, M. peregrinus has larger pollen grains, larger seeds, and larger stomata than M. robertsii (Table 1) are fertile hybrids with variably-coloured corollas, often copper-coloured or with blotches on the petal lobes. Th ey can both be easily distinguished from M. peregrinus by their corolla colours, lack of abundant simple hairs in the keels of the calyx, and evenly triangular, fl at teeth in the leaf margins. Chromosome numbers for these latter two hybrids are not yet available, but it is to be expected that they are similar to their parental species (2n = 60-62).

Putative origin and distribution of M. peregrinus
Th e intermediate fl oral and vegetative characteristics of M. peregrinus between M. guttatus and M. luteus, as well as its close morphological similarity to M.× robertsii clearly suggest a hybrid origin for this new taxon associated with a polyploidization event. Th e alternative, that M. peregrinus is an autopolyploid derivative of a pure M. guttatus or M. luteus seems highly unlikely based on vegetative and fl oral characteristics of the diff erent taxa (Table 1). Moreover, both chromosome counts and genome size data are inconsistent with the expectations of an early generation autopolyploid of either M. guttatus or M. luteus or a backcross between M. × robertsii and either parent (Fig. 3). Th e fact that M. peregrinus presents approximately twice the number of chromosomes and has double the amount of DAPI-staining DNA than a common cytotype of M. × robertsii (Fig. 3), immediately suggests that the most parsimonious explanation for the origin of M. peregrinus is through hybridization between M. guttatus and M. luteus linked to a polyploidization event. Given that M. peregrinus was indentifi ed amongst a large population of M. × robertsii, a possible origin of this new taxon is via genome doubling of the triploid hybrid.
Th e known distribution of M. peregrinus is currently restricted to a single locality in Scotland. A preliminary examination of herbarium specimens at the Royal Botanic Gardens in Edinburgh did not uncover any hybrid specimens that were obviously fertile. However, the widespread distribution of M. × robertsii in the UK suggests, along with anecdotal records of fertility in hybrids (Silverside 1998), may suggest that M. peregrinus could be signifi cantly under recorded, and future studies are required to determine its actual distribution.
It is well known that polyploidization can act as a mechanism restoring fertility even in highly sterile triploid hybrids (Dobzhansky 1937, Stebbins 1950, Grant 1971, Ramsey and Schemske 1998, Briggs and Walter 2000, and polyploidization has resulted in the evolution of other non-native allohexaploid species from highly sterile triploids in the UK (e.g. Senecio cambrensis, 2n = 6x; Abbott and Lowe 2004). While fi rmly establishing the origin and distribution of M. peregrinus must await further ecological and genetic work, the discovery of this taxon provides an exciting opportunity to study the recent evolution of a new allopolyploid British species.