A Novel Botrytis Species Is Associated with a Newly Emergent Foliar Disease in Cultivated Hemerocallis

Foliar tissue samples of cultivated daylilies (Hemerocallis hybrids) showing the symptoms of a newly emergent foliar disease known as ‘spring sickness’ were investigated for associated fungi. The cause(s) of this disease remain obscure. We isolated repeatedly a fungal species which proved to be member of the genus Botrytis, based on immunological tests. DNA sequence analysis of these isolates, using several different phyogenetically informative genes, indicated that they represent a new Botrytis species, most closely related to B. elliptica (lily blight, fire blight) which is a major pathogen of cultivated Lilium. The distinction of the isolates was confirmed by morphological analysis of asexual sporulating cultures. Pathogenicity tests on Hemerocallis tissues in vitro demonstrated that this new species was able to induce lesions and rapid tissue necrosis. Based on this data, we infer that this new species, described here as B. deweyae, is likely to be an important contributor to the development of ‘spring sickness’ symptoms. Pathogenesis may be promoted by developmental and environmental factors that favour assault by this necrotrophic pathogen. The emergence of this disease is suggested to have been triggered by breeding-related changes in cultivated hybrids, particularly the erosion of genetic diversity. Our investigation confirms that emergent plant diseases are important and deserve close monitoring, especially in intensively in-bred plants.


Introduction
The emergence of new fungal pathogens in cultivated and wild plants is a major cause for concern [1] and many disparate factors influence such patterns of emergence and establishment. Interspecific hybridisation between different fungal species may be a major contributor to the evolution of novel diversity [2]. In plant pathogens, it has been demonstrated recently that unusually rapid genome evolution and host shifts may occur after hybridisation events. For example, establishment of a new species of Zymoseptoria, Z. pseudotritici, exhibiting an expanded host range on grasses took place within just a few hundred generations of a hybridisation event between Z. tritici and an unidentified species [3]. There is also considerable evidence that the commensal endophytic state and a parasitic, pathogenic state can be highly plastic [2,4]. Indeed, there is confirmation that several fungal diseases of plants can exist as a symptomless, endophytic infection -as, for instance, in the necrotrophic generalist Botrytis cinerea (grey mould) [5,6] and obligate biotrophic Albugo species [7]. Anthropogenic change to natural environments is also likely to be a major factor in promoting the emergence of new pathogens [1]. Important issues such as climate change, degradation of natural environments (such as forest clearance and loss of genetic diversity in natural plant populations), dispersal of plants and their associated fungi to new areas and dependence on agricultural monoculture systems are likely to dramatically enhance the probability of new diseases emerging, establishing and spreading. The initial emergence events are likely to go unreported and uninvestigated not only because of their infrequency, but also because the damage resulting from a new pathogen could be casually attributed to other microbial species or to abiotic stress-induced damage.
The genus Hemerocallis (daylily) of the family Hemerocallidaceae in the order Asparagales has been cultivated for thousands of years [8]. This small genus of petaloid monocotyledons has been valued extensively as a food crop, as a medicinal plant and as an ornamental. The native range extends from eastern Asia, with a centre of diversity in Japan, Korea and China, possibly as far west as central Europe. Widespread cultivation has led to naturalisation of the genus in many parts of the world [8]. Cultivated Hemerocallis in North America, where the numerous hybrids are exceptionally popular ornamental garden plants, were recently challenged by the accidental introduction of a genus-specific biotrophic rust pathogen, Puccinia hemerocallidis (daylily rust) [9,10]. This pathogen caused a rapidly-spreading epidemic. The lack of resistance amongst hybrid cultivars -with the one study showing only a minority of the cultivars tested exhibiting any resistance to even a single P. hemerocallidis isolate -demonstrated the susceptibility of cultivated varieties that have been bred outside of their native range in the absence of any pressure from this specific pathogen [11]. Other emerging threats to cultivated daylily include the first reports of infection by the Armillaria genus [12,13]. Both in the West and in China, anthracnose diseases caused by various Colletotrichum species have been identified and a new species of Colletotrichum, C. hemerocallidis, was recently described from infected leaf and scape material in China [14].
However, the aetiology of one emerging disease of cultivated Hemerocallis in Western gardens has remained elusive. 'Spring sickness' describes a group of disease symptoms that primarily affect the emergent post-winter foliar growth of the plant, causing distortion, stunting, chlorosis, ragged leaf edges and necrotic lesions [15]. These symptoms not only disfigure the foliage but also weaken the plant, resulting in reducing flowering, and -in severe cases -the affected growing point dies. The first reports of this disease come from the U.S.A. in the 1970s [16] and widespread notification of these symptoms has occurred only over the past 20 years. To date, there has not been a report in peer-reviewed published literature of filamentous fungal species being isolated from Hemerocallis material showing such symptoms. Nevertheless, there is considerable speculation that fungal pathogens may be responsible or contribute to disease development [15,16,17].
Here, we report the isolation of a new species of Botrytis from foliar material of Hemerocallis from the United Kingdom that showed a range of symptoms that correspond to 'spring sickness'. We demonstrate that this new species is distinct at the molecular and morphological level from its closest relative, B. elliptica (lily blight, fire blight), and that the two species have diverged in their host range. Our data shows that cryptic fungal species and perplexing fungal diseases can still be identified from common cultivated plants outside of their natural range, suggesting that there may be a significant reservoir of fungal diversity from which new diseases may emerge.

Isolations and Preliminary Identification
Over a period of 4 years, leaf samples of Hemerocallis cultivars that showed symptoms typical of or closely related to 'spring sickness' such as necrotic patches, necrotic lesions, distortion and chlorosis (Fig. 1a,b) were found to be associated with an unidentified filamentous fungus after the material was rigorously surface-sterilised and plated on malt extract agar (MEA). Occasionally, fungal growth was visible on the symptomatic plant material in the form of dense, short mycelia producing microconidia ( Figure S1a,b) suggesting that a filamentous fungus triggers symptom development. However, no macroconidia could be observed on the plant material. In total, 6 independent isolates were collected (Table 1). Vegetative growth at room temperature (25uC) with ambient light and supplemental near-UV light was rapid, but significant macroconidial sporulation did not occur ( Figure 2). Although microconidia were readily identified, only a very small number of macroconidia could be found on older cultures at the edges of the Petri dish in two isolates (B2 and B4). The overall morphology suggests the isolates are members of the genus Botrytis. Immunological tests of surface washings of plate cultures of the isolates using the Botrytis-specific monoclonal antibody BC-12.CA4 [18,19,20] gave strong positive results (Table 1) indicating that they are all members of the genus Botrytis. These results together with the appearance of Botrytis-like melanised sclerotia in older cultures ( Figure 2) prompted further investigations and DNA analyses.

Sequence Analyses
To determine their identity, all 6 isolates were subjected to DNA sequence analysis. Sequencing of the ITS region confirm they all represent Botrytis species. Interestingly, all 6 isolates shared a unique polymorphism, an indel, in the highly conserved ITS sequence ( Figure 3). BLAST analysis of Genbank data showed that no other known member of the genus possesses this sequence variant.
ITS sequence does not permit sufficient resolution to the species level. To further resolve the relationships of these isolates to other species in the genus Botrytis, analysis of nuclear protein-coding gene sequences that are known to be phylogenetically informative to the species level was undertaken. Five genes previously used to build phylogenies of the genus Botrytis [21,22] were used. Three were single-copy housekeeping genes -encoding glyceraldehyde-3phosphate dehydrogenase (G3PDH), a heat-shock protein (HSP60) and a DNA-dependent RNA polymerase subunit II protein (RPB2) [21] -and two were single-copy genes encoding proteins with roles in phytotoxicity (NEP1 and NEP2) [22].
From the first isolate, B1, combined analysis of 5 sequences (G3PDH, HSP60, RPB2, NEP1 and NEP2) confirmed the distinction ( Figure 4). Analysis showed that except for HSP60 gene, all gene sequences showed sequence differences between B1 and other known members of Botrytis ( Figures S2, S3, S4, S5, S6). The HSP60 sequence was found to be identical to B. elliptica and the other sequences were also close to this species, suggesting that its nearest relation was B. elliptica. Phylogenetic analysis confirmed that this isolate was most closely related to B. elliptica and formed a monophyletic group with B. elliptica and B. squamosa ( Figure 4). Divergence at synonymous sites between the new isolate and B. elliptica and B. squamosa was 1.160.32%, while divergence at synonymous sites between B. elliptica and B. squamosa was 1.060.31%.
For the other 5 isolates, two phylogenetically informative sequences (G3PDH and NEP1) were obtained. The presence of a polymorphism in the ITS1 sequence (Figure 3), as well as multiple sequence differences at other genes, strongly suggests that they represent an undescribed species [23]. G3PDH, NEP1 and ITS sequences from all six isolates were identical confirming genetic similarity of isolates. Genetic similarity of isolates together with significant divergence between the new isolates and B. elliptica and B. squamosa strongly suggests that the new isolates represent an undescribed species.
Further supporting our notion that this species may be involved in development of 'spring sickness' symptoms, PCR amplification of a NEP1 fragment from DNA extracts of symptomatic leaf material generated sequences corresponding to this natural phylogenetic group ( Figure S7). Asymptomatic leaf material did not, however.
Sequences from fungal specimens were submitted to the Genbank database with accession numbers HG799518-HG799538.

Morphological Analyses
Non-sporulating colonies are smooth to slightly fluffy and occasionally form aerial mycelia; colonies typically whitish to pale brown in colour on malt extract agar. The distance between septa ranges from 38.5 mm to 127 mm in length. Typically, sclerotia development is seen only as cultures mature; formation varies between isolates and depending on environmental conditions (Table S1 in File S1). Growth in darkness promotes sclerotia formation. On oatmeal agar, sclerotia are formed within 4 weeks at 15uC in darkness. Sclerotia are hemispherical convex in shape and with a concave surface, sometimes hollow in the centre. Sclerotia are black with size ranging from 2 to 6 mm (and on average 3 mm) in diameter.
The initial absence of significant macroconidial development upon isolation indicates that such conditions are sub-optimal for promotion of asexual sporulation. Exclusion of light and exposure to supplemental near-UV light in the presence of daylight individually fails to trigger sporulation. Surface-sterilised host leaf material and sterile, crude extracts of host leaf material also do not stimulate sporulation. Sporulation is more reliably initiated after a minimum of 7 days exposure to near-UV light in the absence of other light sources, though this varies with the isolate and the medium (Table S2 in File S1). One isolate, B2, does not appear to   sporulate under these conditions. The dependence on specific environmental conditions is a contrast to B. elliptica and B. cinerea which efficiently initiate macroconidia formation even when plates are exposed to daylight. The optimum temperature for sporulation of these isolates is 20uC. Conidiophores are erect, septate, opposite branched and slightly swollen at the top. The macroconidia develop in a botryose cluster and are oblong to spherical in shape (but sometimes ovoid), tapering and pointed at one end with their length in a range of 6.5 to 18 mm (mean 12.5 mm, n = 50) and width in a range of 3.5 to 11 mm (mean 7 mm, n = 50). Fungal material from cultures producing macroconidia are illustrated in Figure 5a,b. A scanning electron micrograph of the typical macroconidia is shown in Figure S8. A comparison to other described species can be found in Table S3 in File S1.

Pathogenicity Assays
To determine whether Botrytis isolates induce the symptoms of the plant material from which they were originally isolated, basic assays were performed using excised surface-sterilised leaf material of Hemerocallis on plates. Inoculation of mycelial plugs directly onto the leaf surfaces resulted in a rapid formation of a spreading watersoaked lesion (Table S4 in File S1). Inoculation with B. elliptica did not produce visible lesions after the same period. Testing the novel isolates from Hemerocallis on the leaves of several other petaloid monocots such as Lilium did not produce lesions, indicating that the host range may be restricted. However, B. elliptica did induce similar lesions on Lilium material but not on Hemerocallis or any other species tested (Table S4 in File S1). Inoculation of Hemerocallis leaf material with conidial suspensions did not result in necrosis, though the conidia germinated and grew out into hyphae which after 10 days had covered the leaf surface.
Inoculation of Hemerocallis plantlets from in vitro culture demonstrated that these isolates had the capacity to cause severe necrosis and death of plantlets. Within 10-14 days of inoculation, plantlets showed chlorosis, necrosis, collapse and death of leaf tissue, although the roots and meristematic region were comparatively intact ( Figure 6, Figure 7). Although these experiments on plantlets cannot exactly replicate the 'spring sickness' seen on mature plants, nevertheless the disease phenotypes are sufficient to satisfy basic Koch's postulates. Tests on plant material extracts after inoculation with B1, B2 and B4 isolates using the Botrytis-specific monoclonal antibody BC-12.CA4 [18,19,20] gave strong positive results with signal intensity values ranging between 34 and 46 whilst the control plant tissue without fungal inoculation gave a signal intensity value of 0. B1 isolates appeared to be the most virulent, rapidly causing complete necrosis and collapse of all inoculated plant material within 14 days ( Figure 6, Figure 7). Although microconidia could be found developing on leaf tissue, typical conidiophores producing macroconidia are not produced on the host under these conditions. Plantlets representing a range of different Hemerocallis cultivars were tested to determine if there may be easily recognised resistance to infection with isolate B1. All the tested cultivars developed similar symptoms of tissue necrosis, collapse and death within 14 days of inoculation.

Mating Type
PCR analysis of the mating locus, MAT1, of Botrytis revealed that different isolates had different alleles (Table S5 in File S1). Reciprocal crosses between pairs of strains of opposite mating types, using a method well established for B. cinerea [24,25], in all cases failed to result in the development of apothecia. Possibly this was due to the developmental condition of the sclerotia that had developed on plates and were used in the attempted sexual crosses in this study.

Formal Description
Based on the molecular, morphological and host specificity data presented herein, this fungal isolate from Hemerocallis represents a new species: Description. Botrytis deweyae van Kan, Terhem & Grant-Downton, sp. nov. MycoBank MB804656; Myconame ID 512235.
Culture characteristics: Colonies incubated in the dark on MEA, optimal growth at 25uC, minimum at 4uC, maximum at 37uC. Colony growth rate 0.4-15.8 mm/d; surface dirty white to pale brown, reverse ochreous [26], with moderate to abundant aerial Distribution and host range. Known only from cultivated specimens of Hemerocallis from the UK but likely to be widespread on Hemerocallis in cultivation and possibly in the wild due to the cryptic nature of this species and its broad temperature tolerance.
Etymology. Named after Dr. Molly Dewey in recognition of her outstanding contributions to plant pathology and mycology, in particular relating to the genus Botrytis.

Discussion
We report the discovery of a new species of the ascomycete fungal genus Botrytis from cultivated material of Hemerocallis (commonly known as the daylily) showing a range of symptoms corresponding to those reported for a mysterious disease, 'spring sickness'. Recently, several new species of Botrytis have been reported, such as B. sinoallii from China which is pathogenic on Allium crops [27], B. fabiopsis from central China which is pathogenic on broad bean [28] and B. caroliniana from North America which is pathogenic on blackberry fruits and broad bean leaves [29]. Other as yet undescribed pathogenic Botrytis species are likely to interact with cultivated plants, for example a foliar pathogen of Hosta (also in the order Asparagales) that is closely related to B. tulipae [30]. This new species, B. deweyae, represents the first novel and morphologically distinctive species to be identified in Europe for many years, as compared with recently described B. pseudocinerea that is genetically but not morphologically distinct from B. cinerea [31]. It is likely that B. deweyae has evaded detection as 1) the critical diagnostic morphological feature of macroconidia do not seem to develop on infected plant material nor are they produced from laboratory cultures except under a specific environmental regime, 2) isolation from infected material has not employed sufficiently stringent surface sterilization to prevent overgrowth from less systemic/pathogenic fungi, 3) the disease symptoms are relatively subtle, have a short temporal manifestation and can easily be attributed to another causative agent, and 4) it is probable that this species has a highly restricted host range.
Whilst there have been some reports of investigations of 'spring sickness' disease by specialist growers of Hemerocallis [15,17], the causative agent had not been identified although other fungal species, such as the yeast-like Aureobasidium microstictum (daylily leaf streak) have been implicated [17]. The rigorous surface-sterilisation procedure we employed in isolations is likely to have been an important factor in finding this new species, as other opportunistic microbial species colonising damaged tissue more superficially would have been removed. In our studies, B. deweyae was repeatedly isolated from diseased material over a 4 year period. Whilst our pathogenicity assays using detached leaf material and  plantlets in vitro under laboratory conditions cannot fully replicate mature plants growing in gardens in the colder months of the year, the rapid development of similar leaf tissue necrosis suggests this species has the capacity to be the major contributor to initiating symptom development. Other microbial species, such as A. microstictum, may opportunistically infect as secondary colonisers and aggravate disease development. The rapidity with which plantlet destruction took place was surprising but this is likely to have been enhanced by the absence of any other microbial flora in these cultures. Numerous microbial species have been shown to act antagonistically to Botrytis infections of plants (reviewed in [32]). For instance, infection of Lilium by B. elliptica can be antagonised by specific bacteria [33]. Our hypothesis that B. deweyae is the main agent of 'spring sickness' disease development is supported by a report of its closest relative, B. elliptica, infecting spring growth of Hemerocallis fulva in Korea [34]. In this case, symptoms were similar with necrosis and death of young leaf tissue, but it was accompanied by significant production of macroconidia. Tests in vitro indicated that B. elliptica was only capable of infecting Hemerocalis tissue after physical damage that wounded the plant [34]; this suggests that B. elliptica lacks the capacity to infect uninjured Hemerocallis tissues, as seen in our experimental controls. There have been no other reports of B. elliptica infecting Hemerocallis.
The infection of Hemerocallis with B. elliptica demonstrates that although this species has been most commonly associated with infections of bulbous true lily (Lilium, Liliaceae sensu stricta) -where it is a major disease of cultivated lily bulb crops [35] -it is also capable of infecting other distantly related petaloid monocots such as Hemerocallis (Hemerocallidaceae) and Tricytris (Convallariaceae) [34,36]. There is also evidence that B. elliptica can infect dicotyledonous hosts [37,38], albeit under unnatural conditions. Whether B. deweyae is an emergent pathogen that has evolved from a B. elliptica population that has undergone a host-shift to Hemerocallis is a matter that remains open for further study. The significant morphological and DNA sequence divergences from any described B. elliptica strain suggest that this would not have been a recent event. As 2 mating loci (MAT1) alleles are present in just 5 tested isolates, recent and rapid sympatric speciation from B. elliptica would appear unlikely. To date, only cultivated material from Britain has been examined but it is almost certain that B. deweyae, like its closest relative B. elliptica [35,39], is cosmopolitan and will be detected in cultivated populations of hybrid Hemerocallis elsewhere in the world and perhaps also in native and naturalised Hemerocallis populations in nature.
Given the lack of any report of this species from indigenous Hemerocallis populations, its morphology and the nature of disease development, it is plausible that B. deweyae may be an endophyte that is undergoing the transition to a more aggressive pathogenic state. There was apparent variation in pathogenicity amongst three isolates tested on plantlets, with the isolate B1 being the most  pathogenic, whilst isolates B2 and B4 were visibly less aggressive. Endophytic Botrytis have been identified using sequencing methods from naturalised populations of Centaurea stoebe [40], clearly demonstrating that outside of their native range plant species can retain or acquire endophytic Botrytis species. Five putative new endophytic species were discovered in this study alone. Intriguingly, there is also evidence that pathogenic B. cinerea (grey mould) can exist as a systemic endophyte without causing pathogenesis in plants [5,6]. By some mechanism, B. cinerea must be capable of substantially down-regulating gene expression responsible for aggressive, necrotrophic pathogenesis thereby permitting it to co-exist and grow within plant tissues without damaging them or even triggering defence responses. In the case of B. deweyae the absence of identifiable macroconidia in nature, after growth on the host in vitro and even the absence of sporulation on plates (except under specific conditions) suggests that this development is usually highly suppressed, as might be expected of an endophytic lifestyle [4]. In common with other specialised Botrytis species that infect petaloid monocotyledons [35], additional studies are required to identify the precise details of the life cycle. Whether, as in B. elliptica [39,41], sexual reproduction is important in natural populations of B. deweyae requires further investigations.
As this disease has become prevalent only in the past 20 years, one factor that may be precipitating the evolution of pathogenesis is the increasing loss of genetic diversity in the host plant through in-breeding of the cultivated hybrid lines, as demonstrated by an AFLP study [42]. It is known from studies of Colletotrichum that the ability for a fungal isolate to exhibit pathogenic or mutualistic behaviour can depend on the host genotype [43]. The emergence of in-breeding depression and loss of vigour amongst Hemerocallis hybrids may favour the emergence of strains that are able to effectively take advantage of weakened resistance systems in the plant and hence exploit the host in a more parasitic manner. In hybrid daylily, it is conceivable that the effects of combining divergent genomes followed by multiple generations of in-breeding and selection for novel traits have compromised the effectiveness of mechanisms that regulate pathogen resistance and endophyte activity in planta. Such a biotic environment may facilitate the emergence of new diseases, including transitions from endophytes to parasites and host shifts.
Other factors may have assisted the emergence of this disease. The clonal propagation of cultivars by division, which creates large wounds, and the tendency to plant Hemerocallis as monocultures or near-monocultures by enthusiasts may assist spread of strains with emerging pathogenic tendencies. A role for bulb mites (for example members of the genus Rhizoglyphus) in spreading fungal material such as mycelium fragments to new infection sites appears plausible [17,44,45]. Anecdotal reports suggest that both treatment by both acaricides and fungicides may reduce the incidence of 'spring sickness', supporting this view [17]. Young, emerging leaf tissues in late winter and spring may be particularly vulnerable to pathogenesis not only due to injury by fluctuating environments (such as tissue damage by freezing weather) but also due to rapid mobilisation of stored compounds from the tuberous roots to sustain new growth. Both factors may tip the balance in favour of fungal development.
Our discovery suggests that other potentially important pathogenic fungi could be easily overlooked and many are likely to remain unknown to science. The large reservoirs of endophytic fungi residing in plants [46], for example those found in wild grasses (Poaceae) [47], may be a source of important biological diversity for the evolution of new plant pathogens, especially with continued erosion of genetic diversity in cultivated plants and in fragmented natural habitats [1]. As threats from emerging pathogens to crop and wild plant resources continue to grow, investment in surveillance and detection systems for new plant diseases should be made a priority. This case in particular demonstrates that diagnostic tools using immunological and DNA sequencing methods, in combination with more conventional morphological and pathology assays, will need to work together in the future if other new plant diseases are to be identified effectively. Widespread 'cryptic' diseases of this kind with a small detrimental impact on the host may be important indicators of host and pathogen groups in which more virulent diseases are likely to be emergent.

Materials and Methods
Permission to use plant material for the study was provided by the owner of the private collections. The studies did not involve endangered or protected species.

Isolation from Plant Material
Samples were collected from collections of hybrid Hemerocallis growing in gardens in England, from Oxfordshire, Wiltshire and Somerset. Pieces of asymptomatic and symptomatic leaf tissue, each approximately 6 mm 2 , were surface-sterilized by immersing in 30% (v/v) sodium hypochlorite-based domestic bleach containing detergent (Parazone; Jeyes, Cambridge, UK) for 30 min followed by 465 min washes in sterile distilled water. Surfacesterilized leaf pieces were plated out on 2% (w/v) malt extract agar (MEA) (Oxoid, Basingstoke, UK) and grown under natural ambient light. The 6 isolates, labelled B1, B2, B4, B5, B6 and P1, were routinely grown on 2% MEA and stock cultures were maintained at 4uC in the dark. Stocks were also preserved as excised mycelial material from plate cultures in Eppendorfs in sterile 50% (v/v) glycerol at 280uC.

Immunological Tests of Fungal Isolates and Plant Extracts for Botrytis Antigens
Plate cultures of fungal isolates were washed with 5 ml phosphate buffered saline, pH 7.4 (Sigma) plus 0.05% Tween 20 (Sigma) (PBST). 1 ml was removed by suction, centrifuged briefly at 13,0006g in a microcentrifuge and 400 ml of the supernatant tested with EnviroLogix Botrytis QuickStixs (Portland, Maine, USA) which employs the anti-Botrytis-monoclonal antibody, BC-12.CA4, raised and employed in previous studies [18,19,20]. Tests were performed by incubating a QuickStix in the supernatant for 10 minutes, the lower pad was then removed and the intensity of the test line (signal intensity, SI) was determined using an EnviroLogix Quickstix reader [19,20]. Extracts of leaf tissues were made by crushing leaf material in an extraction bag (Noegen, UK) with PBST, 1:5 (w/v). 400 ml of the resulting non-particulate extract was tested with EnviroLogix Botrytis QuickStix system as above.

DNA Extraction, PCR Amplification and Sequencing
Genomic DNA was extracted from plugs of cultured mycelial material that were frozen and ground in liquid nitrogen. DNA was purified from tissue powder using the DNEasy kit (Qiagen, Manchester, UK) and quantified using a Nanodrop spectrophotometer. PCRs were performed using approximately 4-12 ng genomic DNA at in a 50 ml reaction using the proof-reading Phusion polymerase kit (Finnzymes, Thermo Scientific, UK) according to the manufacturer's instructions. Cycling conditions for the amplifications were as follows: 98uC for 1 minute; 10 cycles of 98uC for 10 seconds, primerspecific annealing temperature 1 for 30 seconds, extension at 72uC for 45 seconds; 15 cycles of 98uC for 10 seconds, primer-specific annealing temperature 2 for 30 seconds, extension at 72uC for 45 seconds; final extension at 72uC for 5 minutes.
The 18S rRNA sequence was amplified using ITS1 primers previously described [21] (Table S6 in File S1). The primerspecific annealing temperatures 1 and 2 were 60uC and 56uC respectively. The amplification of the G3PDH, HSP60 and RPB2 sequences was achieved using primers for the genus Botrytis described in [19] (Table S6 in File S1). Primer-specific annealing temperatures 1 and 2 were 62uC and 60uC respectively for G3PDH and HSP60, and 60uC and 56uC respectively for RPB2. NEP1 was amplified using primers NEP1for and NEP1revB [22] (Table S6 in File S1) with primer-specific annealing temperatures 1 and 2 of 62uC and 60uC respectively. NEP2 was amplified using primers NEP2forE and NEP2revE (Table S6 in File S1) [22] with primerspecific annealing temperatures 1 and 2 of 62uC and 60uC respectively. PCRs were run on 1.5% (w/v) agarose gels, the bands were cut out with a clean razor and extracted using Qiaex II kit (Qiagen, Manchester, UK). Purified PCR fragments were then cloned into pJET sequencing vector (CloneJET PCR cloning kit; Thermo Scientific, UK) and transformed into E. coli DH5a competent cells. Transformants were selected on LB-ampicillin plates (with 100 mg ampicillin per ml of medium) and colonies with inserts identified by PCR screening. Plasmid DNA was extracted from selected colonies using GeneJET Plasmid Miniprep Kit (Thermo Scientific, UK) and sequenced by Source Biosciences (Nottingham, UK).
Genomic DNA was extracted from leaf material showing symptoms and from leaf material of the same cultivar without symptoms, collected on the same day. Leaf material was derived from the cultivars 'Gerda Brooker' and 'Free Bird'. Small sections of leaf material approximately 262 cm were ground to a fine powder in liquid nitrogen. DNA was purified from tissue powder using the DNEasy kit (Qiagen, Manchester, UK) and quantified using Nanodrop spectrophotometer. PCRs were performed using approximately 100 ng genomic DNA as template at in a 50 ml reaction using the proof-reading Phusion polymerase kit (Finnzymes, Thermo Scientific, UK) according to the manufacturer's instructions. NEP1 primers that amplify the promoter to 39 end -NEP1(2207for) and NEP1(+1124rev) - [22] (Table S6 in File S1) were used in a primary PCR with cycling as follows. 98uC for 1 minute; 10 cycles of 98uC for 20 seconds, 64uC for 30 seconds, extension at 72uC for 45 seconds; 10 cycles of 98uC for 20 seconds, 60uC for 30 seconds, extension at 72uC for 45 seconds; 15 cycles of 98uC for 20 seconds, 58uC for 30 seconds, extension at 72uC for 45 seconds; final extension at 72uC for 5 minutes.
A 2 ml aliquot of primary PCR was then used as a template in a 50 ml reaction with NEP1for and NEP1revB primers [22] (Table  S6 in File S1) as follows: 98uC for 1 minute; 15 cycles of 98uC for 10 seconds, 64uC for 30 seconds, extension at 72uC for 20 seconds; 15 cycles of 98uC for 10 seconds, 60uC for 30 seconds, extension at 72uC for 20 seconds; final extension at 72uC for 5 minutes.
PCRs were run on 1.5% (w/v) agarose gels, the PCR fragments were then removed, purified, cloned and sequenced as described above.

Phylogenetic Analysis
We obtained sequences of Botrytis genes from GenBank (www. ncbi.nlm.nih.gov/genbank) and aligned them with newly sequenced genes from the six isolates using MUSCLE [48]. Phylogenies for single gene and combined datasets were reconstructed using a maximum-likelihood inference conducted with RAxML version 7.2.6 [49] via the raxmlGUI interface [50]. We conducted five independent runs from different starting points to assess convergence within two likelihood units of the best tree, which was consistently selected. The parameters of partition were allowed to vary independently under the GTRGAMMA model of evolution as implemented in RAxML. Maximum-likelihood nodal support was calculated by analysing 1000 bootstrap replicates.

Induction of Sporulation and Sclerotia Formation
To determine whether presence of host material stimulated macroconidia production, 1% water agar on which surfacesterilized pieces of greenhouse-grown Hemerocallis 'Jurassic Spider' young leaf tissue had been placed. Plates were also made with 1% water agar with 0.5 ml plant extract added. The plant extract was made by thoroughly crushing young leaf tissue in an extraction bag (Neogen, Ayr, UK) with PBST at 1:5 (w/v). The extract was passed through a 0.25 micron filter (Millipore) attached to a syringe for sterilisation before addition to molten agar. To attempt to stimulate production of macroconidia, MEA plates were incubated without light and under a mixture of day light and near UV light at room temperature (,20uC) for 7 to 10 days. Plates were also incubated in the dark except for a continuous near-UV light source. Isolates were also grown on oatmeal agar (OMA) (Difco, Becton Dickinson BV, Breda, The Netherlands), potato dextrose agar (PDA) (Oxoid, Basingstoke, UK), Czapeks Dox medium (Oxoid, Basingstoke, UK), V8 agar (prepared with 200 ml V8 vegetable juice (Campbell Soup Company, Camden, NJ, USA), 20 g agar, 800 ml water, pH adjusted to 6.0 with NaOH).

Light Microscopy
Sporulating structures were mounted on slides with filtersterilised MilliQ water. Observations were made with a Nikon Eclipse 90i (Nikon Instruments, Badhoevedorp, The Netherlands) compound microscope with a Nikon DS-5MC camera attached. Measurements were performed using N.I.S. Elements AR 2.30 software (Nikon Instruments, Badhoevedorp, The Netherlands).

Electron Microscopy
Spores were removed from the edges of mature cultures of B2 and B4 using a sticky pad mounted in a SEM stub. Samples were directly coated with gold/palladium in a Polaron SC7640 sputter coating unit (Quorum Technologies, Ashford, UK). Spores were also trapped using poly-lysine coated slides and subsequently treated in osmium tetroxide vapour for 3 hours, followed by 4% paraformaldehyde in phosphate buffer for 3 hours, dehydrated then subsequently sputter coating as described. Images were taken using a JEOL JSM-5510 scanning electron microscope unit (JEOL, Welwyn Garden City, UK) operating at 15 kV.

Nomenclature
The electronic version of this article in Portable Document Format (PDF) in a work with an ISSN or ISBN will represent a published work according to the International Code of Nomenclature for algae, fungi, and plants, and hence the new names contained in the electronic publication of a PLOS ONE article are effectively published under that Code from the electronic edition alone, so there is no longer any need to provide printed copies.
In addition, new names contained in this work have been submitted to MycoBank from where they will be made available to the Global Names Index. The unique MycoBank number can be resolved and the associated information viewed through any standard web browser by appending the MycoBank number contained in this publication to the prefix http://www.mycobank. org/MB/. The online version of this work is archived and available from the following digital repositories: PubMed Central and LOCKSS etc.

Pathogenicity Assays
Pathogenicity tests were carried out using detached leaves of the following plants: Alstroemeria hybrid; Tricyrtis formosana; Lilium Oriental hybrid; Hemerocallis fulva and Hemerocallis 'Jurassic Spider'. Leaves were placed on 0.3% water agar in large Petri dishes Four replicates were made with the leaf adaxial and abaxial sides upwards. The exposed surfaces were inoculated with a 0.5 cm mycelial plug from MEA cultures of isolates B1, B2 and B4. The Petri dishes were sealed and incubated at 25uC under 12 hours light and monitored for up to 14 days. As a control, plugs of B. elliptica (isolate 9601), grown on MEA agar under identical conditions, were used. A sterile MEA agar control was also performed.
Macroconidia of B4 and B5 from sporulating plates were collected in sterile potato dextrose broth (1.2 g/l) and adjusted to concentrations of 1610 5 and 1610 6 ml 21 . Excised Hemerocallis leaves on water agar were inoculated with 5 ml droplets of spore suspensions on their upper sides and three replicates performed for each. The material was sealed in a plastic box with a .90% humidity at 25uC and monitored for up to 10 days after inoculation.
Pathogenicity tests were also carried out on axenic plantlets of Hemerocallis 'Jurassic Spider' as follows. Previously established cultures of Hemerocallis 'Jurassic Spider' were maintained at 4uC on a 0.6% agarose growth medium composed of Murashige and Skoog macroelements [51] with Heller's microelements [52] and 5 ml/litre of 1% ferric ammonium citrate solution, adjusted to pH 5.5 before autoclaving. The clumps of shoots were divided, with the leaves and roots trimmed. The plantlets derived from division were grown in Magenta boxes containing sterilised 100 ml vermiculite plus 80 ml liquid growth medium (as above, but lacking the 0.6% agarose). In each box, 5 divisions were planted and chilled at 4uC in the dark for 2 weeks, then removed in a welllit growth room at 20uC with 16 hour photoperiod for 2 weeks to allow wound healing and establishment prior to inoculation. Individual plantlets were inoculated with a 0.5 mm mycelial plug from a plate culture of one of the fungal isolates (B1, B2, B4) and incubated at 25uC under 16 hours direct light in a growth room for up to 10 days. For controls, B. elliptica (isolate 9601) and sterile MEA plugs were used. For each fungal isolate and for each control, three replicate Magenta boxes were inoculated.
To confirm Botrytis infection of Hemerocallis tissue, extracts of plant tissues were made by crushing sections of aerial material in an extraction bag (Neogen, UK) with PBST, 1:5 (w/v). 400 ml of the resulting non-particulate extract was tested with EnviroLogix Botrytis QuickStix system as above.
To determine if pathogenicity varied depending on the host cultivar that was challenged, plantlets of a range of cultivars were tested using the same method except that one plantlet per Magenta box was used. Two replicates were performed for each plant genotype. These pathogenicity assays were all performed with the B1 isolate.

Identification of Mating Types and Sexual Crosses
Five isolates were analysed (B1, B2, B4, B5 and P1) to identify the mating type alleles. Gentra Puregene DNA purification kit (Qiagen, Venlo, The Netherlands) was used for DNA extraction from freeze-dried mycelia following the manufacturer's instructions. 10-50 ng genomic DNA was used in 25 ml reaction volume. PCRs were performed with GoTaq polymerase (Promega) according to manufacturer's instructions. Primers used were MAT1-1 forward/reverse and MAT1-2 forward/reverse (Table  S6 in File S1). Amplification conditions were as follows: 95uC 5 minutes, then 35 cycles of 94uC for 30 seconds, 52uC for 30 seconds and 72uC for 2 minutes, followed with a final extension of 72uC for 5 minutes. PCR products were visualised on gel to determine the mating type of each isolate. Crosses were set up between isolates carrying different mating types [24]. Isolates B1 and B5 (each of the MAT1-1 mating type) were mated with isolates B2 and B4 (each of the MAT1-2 mating type). To develop sclerotia, strains were plated on oatmeal agar and incubated in darkness for 1 month at 15uC, followed by incubation at 0uC in darkness for 1 month. For mating, sclerotia were sampled from the plates, rinsed in water with a soft toothbrush and place in a 6-well microtitre plate. Mycelial cultures were flooded with sterile water and a suspension of mycelial fragments and microconidia was obtained by gently rubbing the surface with a spatula. The sclerotia were fertilised by addition of this suspension at 3 ml per well. Reciprocal crosses were set up in this manner, with each partner as a female (sclerotial) or male (microconidial) parent. As a control, sclerotia that were not exposed to microconidia and kept in sterile water were used. The microtitre plates were sealed and incubated at 12uC in normal artificial light with a 12 hour photoperiod. File S1 Supporting information file containing Tables S1-S6. (DOCX)