Association of Slugs with the Fungal Pathogen Epichlo ¨E Typhina

Annals of Applied Biology ISSN 0003-4746

RESEARCH ARTICLE Association of slugs with the fungal pathogen Epichloe¨ typhina (Ascomycotina: Clavicipitaceae): potential role in stroma fertilisation and disease spread G.D. Hoffman & S. Rao

Department of Crop and Soil Science, Oregon State University, Corvallis, OR 97331, USA

Keywords Abstract Arion subfuscus; choke disease; Dactylis glomerata; Deroceras reticulatum; fungal Epichloe¨ spp. are endophytes of grasses, and form epiphytic external stromata ecology; mycophagy; Prophysaon andersoni. on flowering tillers. E. typhina was first noticed infecting Dactylis glomerata (= orchardgrass, cocksfoot) stands in the Willamette Valley in 1996, and Correspondence soon became the primary factor limiting the longevity of seed production G.D. Hoffman, Department of Crop and Soil fields. Several species of slugs are present in these fields, and we investigated Science, Oregon State University, Corvallis, OR 97331, USA. Email: their role in E. typhina biology. Pre-dawn surveys of D. glomerata fields in [email protected] 2009 and 2010 found Prophysaon andersoni and Arion subfuscus slugs feeding on the fungal stromata. When unfertilised and fertilised immature stromata Received: 9 August 2012; revised version predominated, approximately 80% of the individuals of these two species accepted: 17 February 2013. that were seen on plants were found on the stromata. As the majority of stromata reached maturity the presence of these species on stromata declined doi:10.1111/aab.12024 to between 20–40%. The common agricultural slug pest, Deroceras reticulatum, was on stromata only 20% of the time early in the season, and declined to <5% at stromata maturity. Observations of frass from slugs determined that the most common constituent was the food sources upon which the slug species was usually found during these surveys. Typically 100% of the frass from P. andersoni and A. subfuscus contained stroma material, compared to 25% for D. reticulatum. Spermatia, and ascospores later in the season, were commonly seen in the frass of slugs that consumed stromata. Some slugs that had no stroma material in their frass appeared to have consumed spermatia and ascospores from the leaf surface. A multiple-choice laboratory test confirmed the different proportional preferences of P. andersoni and D. reticulatum for stroma (0.72 vs 0.20) and leaf (0.07 vs 0.38), respectively. Two laboratory multiple-choice tests, and a field survey, found that P. andersoni preferred unfertilised and immature stroma over mature stroma. D. reticulatum is the most common and abundant slug in Willamette Valley grass seed fields, yet it is the least likely to move spermatia between unfertilised stromata, or ascospores to uninfected plants. P. andersoni and A. subfuscus are mycophagous, frequently transport viable spermatia and ascospores in their frass; yet they are generally confined to field edges. Data and observations suggest the role of slugs in the epidemiology of E. typhina is small compared to other factors.

Introduction in the Willamette Valley of western Oregon (Pfender & Alderman, 2006). Epichloe¨ (Ascomycotina: Clavicipi- Epichloe¨ typhina (Pers.:Fr.) Tul. is the primary factor taceae) endophytes are of considerable interest in agri- limiting the longevity of productive Dactylis glomerata cultural research, as pathogens in cultivated grass seed L. (= orchardgrass, cocksfoot) seed production ﬁelds production ﬁelds (Pfender & Alderman, 1999), as causal

324 Ann Appl Biol 162 (2013) 324–334 © 2013 Association of Applied Biologists G.D. Hoffman & S. Rao Slug consumption of Epichloe¨ stromata agents of livestock toxicosis (Belesky & Bacon, 2009), and of stroma fertilisation have recently been described. as beneficial organisms which reduce insect herbivory as Ascospores released from early maturing stromata can well as increase drought tolerance of host plants (Siegel fertilise late emerging stromata (Alderman & Rao, 2008). et al., 1990; Schardl, 1996; Schardl et al., 2009). During Spermatia dislodged from a stroma by air pressure driven the vegetative growth phase of the host plant, Epichloe¨ water mist, mimicking wind-blown rain, can fertilise is characterised by intercellular hyphal growth with lit- adjacent stromata (Kaser, 2009). tle to no penetration of the host cell wall (Christensen Slugs are serious pests in the grass seed cropping et al., 2002). When the host grass enters the reproductive systems in the Willamette Valley, OR, USA, particularly phase, branched hyphal masses (stromata) form epiphyti- on emerging grass seedlings and in no-till systems (Gavin cally on grass culms, and occasionally on vegetative tillers et al., 2008; Anderson et al., 2010; Mellbye et al., 2011). (Schardl, 1996; Christensen et al., 2008). After stroma fer- The introduced Deroceras reticulatum (Muller),¨ known as tilisation there is a proliferation of white fungal hyphae either the grey garden slug, or grey field slug, is the most leading to a thickening of the stroma and eventual forma- common slug in Willamette Valley agronomic systems tion of perithecia and ascospores. Growth of the stroma (Dreves & Fisher, 2012; personal observation). Other slug mechanically inhibits grass inflorescence development species, particularly those in the genus Arion, are present and production of viable grass seed; this syndrome is in smaller numbers in many grass seed fields. Slugs seen known as choke disease (Kirby, 1961; Bucheli & Leucht- on stromata during choke surveys raised the question of mann, 1996). E. typhina does not produce compounds whether they could be involved in the cycle of stroma toxic to livestock (Leuchtmann et al., 2000), and infected fertilisation, ascospore dispersal and the infection of new plants are not a problem in forage production or grazing. plants. E. typhina, was first recorded in Oregon in 1996, and We investigated the association between the orchard- was likely introduced from Europe where the species is grass pathogen E. typhina and slugs in seed production native (Alderman et al., 1997). By 2000, approximately fields in the Willamette Valley. We documented within 90% of orchardgrass seed production fields in Oregon the plant feeding sites of three slug species: D. reticulatum; were infected by the fungus (Pfender & Alderman, the native slug Prophysaon andersoni Cooper (reticulated 2006). It appears that seed yield loss is proportional to taildropper); and the introduced slug Arion subfuscus Dra- the percentage of flowering tillers choked (Large, 1954; parnaud (dusky Arion). The frass from these slugs was Pfender & Alderman, 2006). While E. typhina spread examined to identify diet components, and confirm that quickly through the Willamette Valley, infecting orchard slug location on the plant during night feeding peri- grass for experimental studies has proven to be difficult ods corresponded to the food consumed. To determine (S.C. Alderman, personal communication). if feeding site locations were due to a true food pref- The sexually reproducing Epichloe¨ are bipolar het- erence, we ran multiple-choice feeding preferences tests erothallic, and are obligate out crossers (White & Bult- in a controlled environment. Corroborative tests deter- man, 1987). Stroma of sexual Epichloe¨ species produce mined the field preference of P. andersoni for stromata of haploid spermatia and receptive hyphae of one of two different developmental stages. This information allowed possible mating types (MAT1-1-1 or MAT1-2-1) (Chung us to make predictions on the role of these slugs in the & Schardl, 1997). It appears that only one mating type infection biology of E. typhina. is found within a host plant (Schardl, 1996). The mei- otically derived haploid ascospores produced within the perithecia are ejected at maturity (Chung & Schardl, 1997; Methods and materials Leyronas & Raynal, 2008). E. typhina is a Type 1 reproduc- Field sampling tive system fungus (White, 1988), requiring ascospores to spread the fungus outside of the host and into susceptible In 2009, we focused our slug surveys on three individuals; it is not transmitted vertically through seed. orchardgrass fields in the mid-Willamette Valley, two For sexual reproduction in Epichloe¨ species, spermatia containing P. andersoni. We began the slug survey shortly produced on the stroma must be transported to a stroma of after the onset of stromata formation (early May). the opposite mating type for gamete transfer and meiosis Sampling was from 5:00 am to 6:30 am on cloudy or to occur. Female flies in the genus Botanophila (Diptera: foggy mornings, at approximately 2-week intervals. The Anthomyiidae) transfer viable spermatia from one stroma exceptions were two 10:00 pm to 12:00 am sampling to another in the process of female feeding, defecation, events (14 May and 3 June), done to compare night and egg laying (Kohlmeyer & Kohlmeyer, 1974; Bultman with the subsequent early morning feeding periods. We et al., 1995, 1998). Spermatia are not carried by wind attempted to record the location of 30–40 individuals (Bultman et al., 1995); however other mechanisms of each species seen on orchardgrass plants during

Ann Appl Biol 162 (2013) 324–334 325 © 2013 Association of Applied Biologists Slug consumption of Epichloe¨ stromata G.D. Hoffman & S. Rao each sampling period. Because D. reticulatum was most The frass generally consisted of a digested, unidentifiable common, we stopped recording information on this granular appearing substance and a portion of undigested species after 40 observations, as we continued to look material whose origin could mostly be identified. The for individuals of the other species. Slugs are extremely identifiable portion was typically >50% of the frass and difficult to find at night among the grass crowns and consisted of several plant cell types (epidermis, vascular detritus on the soil surface so we confined our survey tissue, parenchyma, fibres), fungal material of stroma to slugs on aerial parts. We recorded whether slugs were origin, spermatia, ascospores, anthers, pollen, insect body seen on stromata, leaves, senescing leaves, rust pustules, parts, nematodes, algae, fungal hyphae and spores, yeasts stems or flowers. About one-third of the slugs seen were and soil particles. Different sources of material in the collected for frass collection and laboratory experiments. frass absorbed the stain differently, which facilitated Slugs typically have multiple feeding bouts each identification. night, intersperse with activities such as locomotion We estimated the proportion of the identifiable frass and mating (Hommay et al., 1998). We determined if that was comprised of the major food components. The feeding preferences for these two species changed over identifiable portion of slug frass was not a homogenous this period by comparing the two paired night and early mixture of its components, but rather had semi-discrete morning observation periods. For the first night–morning patches of stroma, leaf material, anthers, etc.; with the pair, feeding sites were condensed to stromata, leaves rarer components disperse throughout. The rare food and other, to fulfil the expected frequency requirements types (insect parts, nematodes, algae, etc.) were combined of the test [Fisher’s Exact Test (SAS 9.1)]. The second into the ‘other’ category. We would observe the entire morning and night comparison for D. reticulatum data sample at 50× to determine its composition, recording needed to be condensed to a two category χ 2-test (leaf and the percent of the identifiable material comprised of other) because only one slug was found on the mature particular food types. When stroma material was present, stromata. we observed five random areas of stroma material at 400× In 2010, we sampled the field with the largest number to count the number of spermatia and ascospores in each of P. andersoni in 2009 (Field 1), starting in April before 0.152 mm2 field of view. In Table 1, we documented stromata were present, and thereafter at approximately the number of frass samples where a particular material 3-week intervals until shortly before harvest. After the was found and the amounts seen in those samples. initiation of stromata formation we began sampling Samples where that material was absent were not used to Field 2, which contained, relative to the 2009 fields, a determine averages. This approach is appropriate when moderate number of D. reticulatum,afewP. andersoni, considering the concentration of spermatia necessary to a large population of A. subfuscus. All sampling was fertilise a stroma, or ascospores needed to infect a host done between 5:30 am and 7:00 am. We recorded the grass. developmental stage of the stroma: unfertilised; fertilised with proliferating thick white fungal mat (immature); Preference tests patchily mature stroma (transitional) and completely mature, orange stroma. We ran three laboratory preference tests to determine if assumed preferences for feeding seen in the field were exhibited in arenas devoid of stimuli not explicitly Frass sampling associated with the plant part, e.g. plant architecture, The presence of a slug on a particular plant part does or humidity, wind and light gradients. Slugs respond to not necessarily mean they have been feeding on this these gradients in the field and laboratory (Cook, 2001). tissue, although in many cases feeding scars were seen We used D. reticulatum (a diet generalist) and P. andersoni on the stromata from which slugs were collected. To (the most mycophagous species) for these tests. document what the slugs were consuming, in June 2009, Experimental slugs came from field collected indi- ◦ and in 2010, slugs were periodically collected during the viduals that had fed on lettuce at 10 C for at least a surveys. They were placed in individual vials without food week to minimise differences in prior feeding experience ◦ and held at 10 C for 24–36 h at which time their frass (Wareing, 1993). Slugs were then isolated individually was collected for content evaluation. For D. reticulatum in Petri-dishes without food for 12 h at room tempera- ◦ and P. andersoni all the retrievable frass was examined, ture (15–18 C) before the experiments. Preference arenas for the larger A. subfuscus only a portion was evaluated. were 12 × 12 cm2 plastic boxes with the bottom covered The frass was placed on a glass slide with a drop of Aniline by a wet natural brown paper towel. The 3-cm high sides blue lactoglycerin stain, squished with a cover slip and were lined on the inside with 2.5 cm wide copper tape to viewed at 50× to 630× with a compound microscope. repel slugs, and thus help confine them within the arena.

Table 1 Identiﬁable contents of the frass from individual slugs collected during pre-dawn surveys in 2010

Spermatia Ascospores Stroma Plant material Pollen Anthers Other

Date Speciesa Total No.b Countc No. Count No. % No. % No. % No. % No. %

Field 1 25 Apr Dr 11 2 25.80 – 0 0.01179.233.60 – 9 24.2 Pa 12 10 159.30 – 936.81251.537.50 – 1120.7 12 May Dr 21 4 69.40 – 514.42188.821.01 0.1217.7 Pa 13 10 42.50 – 1360.31229.454.00 – 1113.0 27 May Dr 11 0 – 0 – 3 29.51081.315.90 – 7 27.5 Pa 7 3 110.00 – 784.0375.00–0–57.6 15 June Dr 10 0 – 4 3.0 0 0.0825.71044.8 8 33.7 7 11.0 Pa 13 0 – 6 4.8 11 55.7812.11110.7 7 48.4 9 14.9 2 July Dr 9 0 – 7 19.8 2 51.4952.7724.4 2 50.0 6 8.6 Pa 9 0 – 8 32.2 9 68.3815.8613.21 5.08 9.4 Field 2 7May As 9 7 81.00 – 713.7965.031.40 – 8 26.8 10 May As 16 15 82.20 – 1529.21649.3510.70 – 1521.2 23 May Dr 4 1 0.00 – 2 4.3381.50–0–314.2 Pa 2 1 1.00 – 232.3246.912.50 – 2 19.6 As 7 2 3.40 – 655.4651.021.00 – 7 8.6 10 June Dr 4 0 – 1 10.0 2 2.0160.0416.7 4 64.3 0 0.0 Pa 3 0 – 2 13.8 3 18.33 0.0316.7 3 58.3 3 6.7 As 19 1 8.0 8 13.7 16 22.51211.71615.5 19 54.5 7 15.7

The plant phenology in the Willamette Valley was 3 week behind normal due to the cold spring. aSpecies code: Dr = D. reticulatum;Pa= P. andersoni;As= A. subfuscus. bNo. is the number of samples that contained that material. cCounts are number of spermatia or ascospores per 0.152 mm2. No count is possible if the material was not found in any of the samples.

Preference trials ran from 10:00 pm to 4:00 am at room For the comparison between P. andersoni and temperature, under faint red light to allow observations D. reticulatum, we examined the proportion of counts and photography. Each slug was introduced into an empty on the stroma, the leaf and arena bottom. After con- arena and allowed to acclimate for 10 min. After introduc- version of the proportions to logits, a Wilcoxson rank tion of the D. glomerata (var. Potomac) plant parts, placed sum test (SAS 9.1, 2002–2003) was used to test whether equidistant from the arena sides and each other, the slug there was a difference between the specific comparisons was placed between the plant parts. The six arenas were of interest. For the P. andersoni – stroma maturity tests photographed simultaneously, at 5-min intervals, with a an initial Exact Friedman’s test examined the overall null mounted camera controlled by computer software. The hypothesis that there was no difference in the prefer- photographs were viewed later and the position of each ence proportions. This was followed by a Wilcoxson rank sum for individual comparisons. In the first P. andersoni slug recorded. – stroma maturity test, we compared their preference for The first test compared the responses of P. andersoni either unfertilised, immature stroma or leaf in the three and D. reticulatum to a fertilised immature stroma and two-way comparisons. In a separate test, we compared orchardgrass leaf tissue in a paired design, n = 12 pairs. the proportion of slugs on the two stroma types, ver- Each pair received half of the same stroma, and a piece of sus the proportion of the leaf and arena bottom. In the the same leaf. The recorded positions were on the stroma, second P. andersoni – stroma maturity test, we compared leaf or arena bottom. their preference for either, an immature stroma, a mature Two subsequent tests compared the response of stroma or leaf in the three two-way comparisons. P. andersoni to three food sources: stromata of two To determine if P. andersoni showed preference for developmental stages and leaf tissue. The first test was stromata of different developmental stages under field the choice among an unfertilised stroma, a fertilised conditions, on 27 May and 15 June 2010 we compared immature stroma, leaf tissue or arena bottom, n = 18. the distribution of slugs on stromata recorded during The second was choosing among a fertilised immature the surveys in Field 1 with the developmental stage of stroma, a mature stroma, leaf tissue or arena bottom, stromata in that portion of the field. After each survey, n = 18. we recorded the developmental stage of 200 stromata

(5 stromata at 40 random locations). We used χ 2-analysis A D. reticulatum with Yates correction (OPUS 12 Foundation, Inc., 2010) 100 to compare the distribution of slugs on stromata of four maturities with the distributions of stromata maturities in 80 the field. 60 Stroma Results 40 Leaf Stem The distributions of D. reticulatum and P. andersoni on Rust D. glomerata during night and early morning observation 20 Senescing periods was similar. The distribution of plant parts on which these two species were found on the morning 0 of 13 May and the night of 14 May were the same B P. andersoni (Fisher’s exact test, n = 78, df = 2, PB = 0.071; Fisher’s 100 exact test, n = 67, df = 2, PB = 0.724), respectively. On the morning of 2 June and the night of 3 June the 80 2 = distribution of D. reticulatum was the same (χ , n 207, Location found on Plant (%) 2 = = χ 0.00, P 1.00). The distributions of P. andersoni were 60 different morning versus night (Fisher’s exact test, n = 81, = = df 2, PB 0.002); however the difference was due to the 40 number of slugs on senescent leaves, not the number on stromata. 20 In the 2009 on plant survey D. reticulatum was found more frequently on leaves and other food sources 0 compared to choke stromata. In early May, when 11-May 18-May 25-May 01-Jun 08-Jun stromata were mostly unfertilised or immature, roughly Date (2009) 20% of this species were on stromata. P. andersoni was found approximately 80% of the time on choke stromata Figure 1 The percent of slugs found on orchardgrass plant parts during during this same period; the second most preferred pre-dawn and night observations periods in 2009. Data from the three potential food source was leaves (Fig. 1). For both of these fields are combined. The plant categories were: the choke stroma, leaf species, slug presence on the stromata during feeding surface, tiller stem, rust pustules on leaves, senescing leaf tissue. (A) D. reticulatum, total slug numbers per date from earliest sampling: 12, 36, periods declined as the season progressed, to 2.9% and 45, 25, 79, 49 and 34. (B) P. andersoni, total slug numbers per date from 25.0% for D. reticulatum and P. andersoni, respectively earliest sampling: 21, 43, 24, 24, 41, 40 and 36. (Fig. 1). In Field 1 in 2010, few slugs were found on the initial (pre-stroma emergence) sampling date. By the The 39 frass samples collected in June 2009 were less second sampling date stroma had begun emerging and quantitatively analyzed than those in 2010. Fourteen of P. andersoni was again found on early season stromata the 22 frass samples from P. andersoni contained stroma approximately 80% of the time, while D. reticulatum were material; five of those samples came from slugs collected on stromata 20% or less (Fig. 2). Again prevalence of from leaves or stems. Six of the samples contained the slugs on stromata declined with time, although not ascospores but no identifiable stroma material. Only two greatly as in 2009. Spring 2010 was colder than usual and of the 17 frass samples from D. reticulatum contained the phenology of orchardgrass and choke was delayed stroma material, and four samples contained ascospores compared to 2009. but no stroma. In Field 2, D. reticulatum was even less attracted to The analysis of frass from 180 slugs in 2010 established stromata, spending most of their time on leaves (Fig. 3). that in most cases slugs again had been primarily The seasonal pattern of stromata utilisation by A. subfuscus feeding on the plant parts on which they were found. was similar to P. andersoni, with high stromata use early However they were also feeding on many other food in the season (Fig. 3). For A. subfuscus, the decline in types. For example, on 12 May in Field 1, 13% of preference for stromata as the season progressed may be D. reticulatum were seen on stromata (Fig. 2). Five greater than that seen in P. andersoni. Unlike P. andersoni, of the 21 frass samples contained stroma material. In A. subfuscus was frequently seen feeding on the tiller those five, an average of 14.4% of the identifiable inflorescences. frass material was of stroma origin. Four of those five

A D. reticulatum D. reticulatum 100 100 A

80 80 Stroma Leaf 60 60 Stem Stroma Leaf Rust Stem 40 40 Senescing Rust Senescing 20 Flower 20

0 0 B P. andersoni 100 A. subfuscus 100 B

80 80 Location found on Plant (%) Location found on Plant (%) 60 60

40 40 20 20 0 01-May 01-Jun 01-Jul 0 Date (2010) 01-May 01-Jun 01-Jul Date (2010) Figure 2 The percent of slugs found on plant parts during pre-dawn observations periods in Field 1 in 2010. (A) D. reticulatum, total slug Figure 3 The percent of slugs found on plant parts during pre-dawn numbers per date from second sampling: 23, 38, 48, 73, 46. (B) P. observations periods in Field 2 in 2010. (A) D. reticulatum, total slug andersoni, total slug numbers per date from second sampling: 14, 25, 36, numbers per date from earliest sampling: 7, 7, 22, 21. (B) A. subfuscus, 28, 18. Because of few slugs were found at the first sampling, that data is total slug numbers per date from earliest sampling: 14, 35, 39, 42. not presented in the figure. samples contained spermatia, with an average of 69.4 in the number of frass samples containing stroma material spermatia per 0.152 mm2 (Table 1). Each of the 21 for D. reticulatum. Most frass samples from P. andersoni still samples contained leaf material, averaging 88.8% of contained stroma material, with an average content of the contents. Other biological material was found in approximately 50% (Table 1). smaller amounts (Table 1). Averages do not sum to 100% A similar correspondence between the plant part on because they are based on just the samples containing which the slugs were observed and the contents of the that component, not all the samples. frass was seen in Field 2. For example, on 10 May 71% of On the same day, 76% of P. andersoni were seen the A. subfuscus were on stromata. Fifteen of the sixteen on stromata (Fig. 2). Of the 13 frass samples from frass samples contained boluses of stroma material, with collected slugs, all contained stroma material, which an average content of 29.2%. All of those contained comprised 60.3% of identifiable components (Table 1). spermatia, with an average count of 82.2 spermatia per Ten of the thirteen samples contained spermatia with an 0.152 mm2 (Table 1). On 10 June, when most of the average of 42.5 spermatia per 0.152 mm2. Twelve of the stromata were mature, 35.7% of the A. subfuscus were samples also contained plant material, averaging 29.4%. seen on the stroma. Sixteen of the nineteen frass samples Eleven of the samples contained material in the ‘other’ from this species contained stroma material, with an category, averaging 13.0% of frass contents. Similar average stroma content of 22.5%. Eight of those samples results were found for the other early season collections contained ascospores, with an average count of 13.7 (Table 1). per 0.152 mm2. One of the sixteen samples contained By June 2010, fewer slugs of both species were seen on spermatia from stromata produced late in the season, stromata during feeding bouts. This trend is also present with eight spermatia per 0.152 mm2 (Table 1).

On 10 June two of the four D. reticulatum samples 0 D. reticulatum 100 contained small amounts of stroma material, with an 10 P. andersoni 90 average stroma content of 7.9%. One of these samples 20 contained ascospores, with a count of 10.0 per 0.152 mm2. 80 30 The three P. andersoni contained stroma material, with 70 40 an average content of 22.5%. Two of those samples 60 Stroma 50 contained ascospores, with an average count of 13.8 per 50 Arena 2 60 0.152 mm (Table 1). 40 70 The late season trends in frass content for the three 30 species described here can be seen on the other sampling 80 20 dates documented in Table 1. Generally the portion of 90 10 frass samples containing stroma material was higher than 100 0 the percent of slugs found on stroma. This is the same 0 102030405060708090100 with most other food components, as the frass contains Leaf food eaten in earlier feeding bouts. Slugs appear to be able to consume spermatia and Figure 4 Ternary plot showing the preference of D. reticulatum and ascospores from sources other than feeding directly on the P. andersoni for immature stroma, leaf tissue or the arena bottom. stromata. This is particularly the case with D. reticulatum. It graphically depicts the ratios of the three location preferences as For example, on 25 April in Field 1 none of the 11 frass positions (proportions) along the axes of an equilateral triangle. n = 12 samples contained stroma material, yet two of the samples pairs. contained low numbers of spermatia (Table 1). On 2 July only two of the nine samples contained stroma material, time P. andersoni spent on unfertilised and immature yet seven of the samples contained ascospores. This also fertilised stromata was not different, 0.39 and 0.35, occurred for P. andersoni.On15Julytwoofthefrass respectively (S = 2.5, df = 18, P = 0.92323). A separate samples from this species contained no observable stroma analysis found a greater proportion of slugs on total material but did contain small numbers of ascospores. stroma (unfertilised + immature) 0.62, versus other P. andersoni and D. reticulatum had different preferences (leaf + arena) 0.38 (S = 43.5, df = 17, P = 0.0395). for fertilised immature stroma and leaf tissue in the arena In the third test, P. andersoni preference for either test. P. andersoni individuals clustered in the lower left of immature stroma, mature stroma or a leaf resulted in the ternary plot, indicating a high preference for stroma points in the ternary plot clustered in the lower left and low preference for leaf and arena bottom (Fig. 4). corner (Fig. 5B). This preference for immature stroma D. reticulatum was more dispersed along the leaf and was significant (Friedman’s exact statistic = 10.4, two- arena axes, with only two individuals showing a high sided exact permutation p-value, P = 0.005). P. andersoni preference for stroma (Fig. 4). The proportion of time showed a preference for immature stromata over mature P. andersoni spent on stroma was greater than stromata, 0.66 and 0.16, respectively (S = 77.5, df = 18, D. reticulatum, 0.72 and 0.20, respectively (S = 36, df = 12, P = 0.0002). There was a preference for immature P = 0.0024). D. reticulatum spent more time on the leaf stroma over leaves, 0.66 and 0.18, respectively (S = 71.5, compared to P. andersoni, 0.38 and 0.07, respectively df = 18, P = 0.0008); but not for mature stromata over (S =−38, df = 12, P = 0.0010). There was no difference in leaves, 0.16 and 0.18, respectively (S = 14, df = 18, the proportion of time the two species spent on the arena P = 0.4954). bottom, 0.20 and 0.42 for P. andersoni and D. reticulatum, The field preference test documented that P. andersoni respectively (S =−19, df = 12, P = 0.1514). had a preference for stromata at earlier stages of maturity. The ternary plot of preference of P. andersoni for On 27 May the distribution of slugs on stromata of either unfertilised stroma, immature stroma or a leaf, different developmental stages was different from the suggest significant difference in its response, with no distribution of stromata in the field (Yates χ 2 = 18.08, points (replicates) in the lower left corner of the df = 3, Yates P = 0.0003). A greater proportion of slugs plot (Fig. 5A). The equal distribution of points along were found on younger stromata, and lower proportions both the unfertilised and immature axes suggests that on the two oldest maturities, compared to the distribution there was no preference for these two stroma maturity found in the field (Table 2). On 15 June there were categories. Exact Friedman’s test was not significant no unfertilised stromata in the field (Table 2). There for slug preference along the three food sources was a significant difference in the distribution of slugs (exact statistic = 0.3333, two-sided exact permutation on stromata of the three older developmental stages p-value, P = 0.9003). For example, the proportion of compared to the distribution of stroma maturities in the

AB0 0 100 100 10 10 90 90 20 20 80 80 30 30 70 70 40 40 Unfertilized 60 Immature Immature 60 Mature Stroma 50 Stroma Stroma 50 Stroma 50 50 60 60 40 40 70 70 30 30 80 80 20 20 90 90 10 10 100 100 0 0 0 102030405060708090100 0 102030405060708090100 Leaf Leaf

Figure 5 Ternary plot showing the preference of P. andersoni for either stroma of two maturities, or leaf tissue. The number of times a slug was seen on the arena bottom was dropped for this analysis. It graphically depicts the ratios of the three location preferences as positions (proportions) along the axes of an equilateral triangle. n = 18. (A) Comparisons are unfertilised stroma, immature stroma, and leaf tissue. (B) Comparisons are immature stroma, mature stroma and leaf tissue.

Table 2 Proportion of P. andersoni on stromata of different develop- the frass from collected slugs determined that the slugs mental stages compared to the proportion of those stages in the field in were feeding most commonly from the plant part on 2010 which they were seen during night-feeding periods. The Stromata developmental 27 May 15 June range of food types in the frass was large, and the frass from slugs collected from plant parts other than stromata Stage Field Slugs Field Slugs often contained stroma hyphae. Spermatia and ascospores Unfertilised 0.063 0.321 0.000 0.000 were found in the frass of the three slug species. They Immature 0.390 0.393 0.005 0.059 occur more frequently, and in greater abundance, in the Intermediate 0.308 0.286 0.080 0.235 individuals which were found feeding on stromata. In Mature 0.239 0.000 0.915 0.706 na= 200 28 200 18 laboratory preference tests P. andersoni and D. reticulatum showed preference for stroma and leaves, respectively. a For slugs this is number found on stromata during the pre-dawn survey. P. andersoni showed a preference for unfertilised and immature stroma over mature stroma in both laboratory field (Yates χ 2 = 6.428, df = 2, Yates P = 0.0402). There and field experiments. were higher proportions of slugs on the immature and Phytophagous slugs have been described as intermediate stages, and a lower proportion on the mature ‘acceptability-moderated generalists with a hierar- stromata (Table 2). chy of acceptable foods’ (Dirzo, 1980). The preferences of the three slug species for feeding on stromata appears based on the degree to which fungi are a part of their Discussion diet. This could be macrofungi, microfungi, or detritus This is the first study to document slugs serving as vectors colonised by fungus. Slugs in the genus Prophysaon are of spermatia transfer between the two mating types of largely mycophagous (McGraw et al., 2002). A. subfuscus a fungus, and their potential role in the infection of appears to be more omnivorous, and has been reported new host plants. The three slug species investigated all to consume primarily macrofungi (Beyer & Saari, 1978; fed on E. typhina stromata. Early in the phenology of Worthen, 1988), and also decaying vegetation (Jennings stroma production most P. andersoni and A. subfuscus were & Barkham, 1979), and mosses (Davidson et al., 1990). seen on unfertilised and immature stromata during the Several Arion species prefer feeding on leaves infected early morning feeding periods. D. reticulatum is much with rust fungi over healthy leaves (Ramsell & Paul, less commonly found on the stromata early in the 1990). A. subfuscus is also known to have predaceous season. The proportion of each species found on stromata and carrion-feeding behaviours (Quick, 1960; Barker declined when a greater proportion of stromata were & Efford, 2004). In natural habitats D. reticulatum is in the mature stage of development. Examination of primarily a consumer of plant tissue, although a portion

Ann Appl Biol 162 (2013) 324–334 331 © 2013 Association of Applied Biologists Slug consumption of Epichloe¨ stromata G.D. Hoffman & S. Rao of its diet is comprised of leaves colonised by fungi, on D. glomerata are found on unfertilised stromata, and a earthworm, and small insects (Getz, 1959; Pallant, 1969, larger percentage contain stroma material and spermatia. 1972). They tend to prefer healthy over rust infected Because these latter species are generally confined to leaves (Ramsell & Paul, 1990). This species has also been the periphery of fields, near dense field side vegetation found to be an opportunist carnivore (Barker & Efford, (personal observation), they probably have a localised 2004). Typically P. andersoni and A. subfuscus are active influence on stromata fertilisation. The same is probably mycophagists, while fungi are just part of the more true of their potential impact on the infection process. It generalist plant centred diet of D. reticulatum. At Field 2 is the mycophagous habits of P. andersoni and A. subfuscus only P. andersoni and A. subfuscus were seen feeding on that may result in their restricted infield distribution. mushrooms along field borders. They appear to move out of the grass seed fields in the The composition of frass may not correspond exactly early summer and fall to feed on macro- and microfungi to what the slug ate, for it appeared we generally in the adjacent vegetation (personal observation). Arion could identify the less digestible portion of the material circumscriptus Johnson is partially mycophagous (Getz, eaten. For example, leaf epidermis and vascular tissue 1959), and is the Arion species most commonly found were identified more commonly than parenchyma cells. dispersed throughout grass seed fields in the Willamette However, Wolda et al. (1971) found that the food material Valley (personal observation). Their populations were identified in the gut was similar to that found in the higher in the Willamette Valley in years preceding this frass. study, but they could potentially have a greater impact Apparently slugs can consume ascospores, and sper- than either P. andersoni or A. subfuscus.Wesuspectthat matia to a lesser extent, without feeding on stromata. any slug species with mycophagous tendencies would As the frequency of rain declines in May and June, the feed on stromata and could transfer viable spermatia and surface of D. glomerata leaves accumulate a variety of ascospores. biological ‘debris’, such as pollen, anthers, rust spores, Most stromata in infected D. glomerata fields become insects, mites, ascospores and spermatia. Frequently slugs fertilised, even in the absence of populations of are seen feeding on the leaf surface, yet no tissue dam- Botanophila fly (Rao & Baumann, 2004), or slugs (personal age is seen. It appears that the slugs are using the leaf observations). The majority of the fertilised stromata surface as a dinner plate, consuming these easy to ingest develop the post-fertilised hyphal mat uniformly across food items. Immense numbers of ascospores are present their surface, and the perithecia reach maturity (orange) in the air above choke infected D. glomerata fields (Kaser at approximately the same time. When stromata are et al., 2008), and the data in Table 1 suggest that deposi- fertilised by Botanophila flies (Bultman et al., 1998) or tion on leaves is the primary way in which D. reticulatum slugs, initial hyphal growth occurs at the site where frass consumes ascospores. is deposited. In the case of slugs, the development of Laboratory and field tests demonstrated that the mature perithecia only extends a few millimetres beyond preference of slugs for particular fungal or plant parts the frass, and occasionally through the stroma to the seen in field surveys was due to characteristics of the opposite side (personal observation). Thus inoculation part, and not environmental gradients. Thus the pattern with slug frass containing spermatia does not lead to in preferences seen in the field should be relatively the propagation of fertilisation throughout the stromata. consistent regardless of environmental gradients such as This suggests most fertilisation in the field occurs through wind, humidity, and light. other means, such as water splash (Rao et al., 2012) What role do P. andersoni, A. subfuscus,andD. or ascosporic fertilisation (Alderman & Rao, 2008). In reticulatum play in the epidemiology of choke disease newly infected fields with few stromata, slugs and the in Oregon? A portion of the E. typhina spermatia and Botanophila fly could serve an important role in fertilizing ascospores consumed by the slugs are excreted in the the early season stromata. These stromata produce the frass. These remain viable, and the spermatia can fertilise ascospores which may be responsible for much of the stroma of the opposite mating type (Rao et al., 2012; fertilisation during the remainder of the season (Rao unpublished data) and frass containing ascospores can et al., 2012). The feeding activity of slugs on mature initiate infections of D. glomerata when applied to cut stromata, and ascospore consumption from the leaf tillers (unpublished data). surface, can result in new infections of orchardgrass D. reticulatum is the most widespread and abundant slug in a greenhouse setting (unpublished data). However, in grass seed fields in the Willamette Valley, yet it is the the swiftness at which E. typhina infections spread least likely of the three species to feed on stromata and throughout the Willamette Valley (Pfender & Alderman, potentially transfer spermatia and ascospores. Conversely, 2006) suggests other biotic or management factors are approximately 80% of the P. andersoni and A. subfuscus the primary cause.

Acknowledgements Cook A. (2001) Behavioural ecology: on doing the right thing, in the right place and the right time. In The This research was supported by funding from the biology of terrestrial molluscs, pp. 447–487. Ed G.M. Barker. USDA/CSREES Grass Seed Cropping Systems for a Wallingford, Oxon, UK: CABI Publishing. Sustainable Agriculture, Oregon Seed Council and the Davidson A.J., Harborne J.B., Longton R.E. (1990) The Oregon Orchardgrass Seed Producers Commission. We acceptability of mosses as food for generalist herbivores, thank the orchardgrass growers who allowed us to slugs in the Arionidae. Botanical Journal of the Linnean sample their fields. Cliff Pereira assisted with the statistical Society, 104, 99–113. analysis of the preference tests. Dirzo R. (1980) Experimental studies on slug-plant interac- tions: I. The acceptability of thirty plant species to the slug References Agriolimax caruaneae. Journal of Ecology, 68, 981–998. Dreves A.J., Fisher G. (2012) Slug Control. In Pacific North- Alderman S.C., Rao S. (2008) Ascosporic fertilization of west Insect Management Handbook. Ed C.S. Hollingsworth. Epichloe¨ typhina in Dactylis glomerata seed production fields Corvallis, OR: Oregon State University Extension in Oregon and implications for choke management. Plant http://uspest.org/pnw/insects. Chapter: IPM, Symphs, Bio- Health Progress. DOI: 10.1094/PHP-2008-0421-01-BR. control, Slugs. Section: Slug Control. Alderman S.C., Pfender W.F., Welty R.E. (1997) First report Gavin W.E., Hoffman G.D., Banowetz G.M. (2008) Control of of choke, caused by Epichloe¨ typhina, on orchardgrass in the gray field slug during annual ryegrass establishment. In Oregon. Plant Disease, 81, 1335. Seed Production Research at Oregon State University, pp. 71–76. Anderson N.P., Hoffman G.D., Dreves A.J. (2010) Evaluation Ed W.C. Young, III. Corvallis, OR: Oregon State University of newly formulated molluscides for control of slugs in the Extension Service. western Oregon grass seed fields. In Seed Production Research Getz L.L. (1959) Notes on the ecology of slugs: Arion at Oregon State University, pp. 15–18. Ed W.C. Young, III. circumscriptus, Deroceras reticulatum,andD. laeve. The Corvallis, OR: Oregon State University Extension Service. American Midland Naturalist, 61, 485–498. Barker G., Efford M. (2004) Predatory gastropods as natural Hommay G., Lorvelec O., Jacky F. (1998) Daily activity enemies of terrestrial gastropods and other invertebrates. rhythm and use of shelter in the slugs Deroceras reticulatum In Natural Enemies of Terrestrial Molluscs, pp. 279–403. Ed G.M. Barker. Wallingford, Oxon, UK: CABI Publishing. and Arion distinctus under laboratory conditions. Annals of Belesky D.P., Bacon C.W. (2009) Tall fescue and associated Applied Biology, 132, 167–185. mutualistic toxic fungal endophytes in agroecosystems. Jennings T.J., Barkham J.P. (1979) Litter decomposition by Toxin Reviews, 28, 102–117. slugs in mixed deciduous woodland. Ecography, 2, 21–29. Beyer W.N., Saari D.M. (1978) Activity and ecological Kaser J.M. (2009) Epichloe¨ typhina (fungus) – Botanophila distribution of the slug, Arion subfuscus (Draparnaud) lobata (fly) interaction: An invasive ‘‘pollinator’’ system in its (Stylommatophora, Arionidae). The American Midland introduced range in western Oregon. MS Thesis, Oregon State Naturalist, 100, 359–367. University, Oregon. Bucheli E., Leuchtmann A. (1996) Evidence for genetic Kaser J.M., Rao S., Alderman S.C. (2008) Seasonal differentiation between choke-inducing and asymptomatic production of infective ascospores of the choke pathogen, strains of the Epichloe¨ grass endophyte from Brachypodium Epichloe¨ typhina, in orchardgrass in the Willamette Valley. sylvaticum. Evolution, 50, 1879–1887. In Seed Production Research at Oregon State University, pp. Bultman T.L., White J.F. Jr., Bowdish T.I., Welch A.M., 11–15. Ed W.C. Young, III. Corvallis, OR: Oregon State Johnston J. (1995) Mutualistic transfer of Epichloe¨ University Extension Service. spermatia by Phorbia flies. Mycologia, 87, 182–189. Kirby E.J.M. (1961) Host–parasite relations in the choke Bultman T.L., White J.J.F., Bowdish T.I., Welch A.M. (1998) disease of grasses. Transactions of the British Mycological A new kind of mutualism between fungi and insects. Society, 44, 493–503. Mycological Research, 102, 235–238. Kohlmeyer J., Kohlmeyer E. (1974) Distribution of Epichloe¨ Christensen M.J., Bennett R.J., Schmid J. (2002) Growth typhina (Ascomycetes) and its parasitic fly. Mycologia, 66, of Epichloe¨/Neotyphodium and p-endophytes in leaves of 77–86. Lolium and Festuca grasses. Mycological Research, 106, Large E.C. (1954) Surveys for choke (Epichloe¨ typhina)in 93–106. cocksfoot seed crops, 1951–53. Plant Pathology, 3, 6–11. Christensen M.J., Zhang X., Scott B. (2008) Regulation Leuchtmann A., Schmidt D., Bush L.P. (2000) Different levels switching of Epichloe¨ typhina within elongating perennial of protective alkaloids in grasses with stroma-forming ryegrass leaves. Mycological Research, 112, 1056–1062. and seed-transmitted Epichloe¨/Neotyphodium endophytes. Chung K.R., Schardl C.L. (1997) Sexual cycle and horizontal Journal of Chemical Ecology, 26, 1025–1036. transmission of the grass symbiont, Epichloe¨ typhina. Leyronas C., Raynal G. (2008) Role of fungal ascospores in Mycological Research, 101, 295–301. the infection of orchardgrass (Dactylis glomerata)byEpichloe¨

typhina agent of choke disease. Journal of Plant Pathology, Rao S., Alderman S.C., Kaser J.M., Hoffman G.D. (2012) 90, 15–21. Fertilization of Epichloe¨ typhina in cultivated Dactylis McGraw R., Duncan N., Cazares E. (2002) Fungi and glomerata by factors besides Botanophila flies.InEpichloe,¨ other items consumed by the blue-gray taildropper slug endophytes of cool season grasses: implications utilization, (Prophysaon coeruleum) and the papillose taildropper slug and biology, pp. 122–126. Eds C.A. Young, G.E. Aiken, (Prophysaon dubium). Veliger, 45, 261–264. R.L. McCulley, J.R. Strickland, C.L. Schardl. Ardmore, Mellbye M.E., Young W.C. III, Garbacik C.J. (2011) Long- Oklahoma: The Samuel Roberts Noble Foundation. term evaluation of annual ryegrass cropping systems for Schardl C.L. (1996) Epichloe¨ species: fungal symbionts of seed production – year 6. In Seed Production Research at grasses. Annual Review of Phytopathology, 34, 109–130. Oregon State University, pp. 1–4. Ed W.C. Young, III. Schardl C.L., Scott B., Florea S., Zhang D. (2009) Epichloe¨ Corvallis, OR: Oregon State University Extension Service. endophytes: Clavicipitaceous symbionts of grasses. In The OPUS 12 Foundation, Inc. (2010) http://www.opus12. Mycota V, pp. 276–306. Ed H.B. Deising. Berlin Heidelberg: org/Chi-Square_Calculator.html. [accessed on 16 July Springer. 2010]. Siegel M.R., Latch G.C.M., Bush L.P., Fannin F.F., Rowan Pallant D. (1969) The food of the grey field slug, (Agriolimax D.D., Tapper B.A., Bacon C.W., Johnson M.C. (1990) reticulatus (Muller),¨ in woodland. Journal of Animal Ecology, Fungal endophyte-infected grasses: Alkaloid accumulation 38, 391–397. and aphid response. Journal of Chemical Ecology, 16, Pallant D. (1972) The food of the grey field slug, Agriolimax 3301–3315. reticulatus (Muller),¨ on grassland. Journal of Animal Ecology, Wareing D.R. (1993) Feeding history – a factor in determin- 41, 761–769. ing food preference in slugs. Journal of Molluscan Studies, Pfender, W.F., Alderman, S.C. (1999) Geographical distri- 59, 368–370. bution and incidence of orchardgrass choke, caused by White J.F. (1988) Endophyte – host associations in forage Epichloe¨ typhina, in Oregon. Plant Disease, 83, 654–758. grasses. XI. A proposal concerning origin and evolution. Pfender W.F., Alderman S.C. (2006) Regional development Mycologia, 80, 442–446. of orchardgrass choke and estimation of seed yield loss. White J.F. Jr., Bultman T.L. (1987) Endophyte – host Plant Disease, 90, 240–244. associations in forage grasses. VIII. Heterothallism in Quick H.E. (1960) British slugs (Pulmonata; Testacellidae, Epichloe¨ typhina. American Journal of Botany, 74, 1716–1721. Arionidae, Limacidae). Bulletin of the British Museum Wolda H., Zweep A., Schuitema K.A. (1971) The role of food (Natural History). Zoology, 6, 103–226. in the dynamics of populations of the landsnail Cepaea Ramsell J., Paul N. (1990) Preferential grazing by molluscs nemoralis. Oecologia, 7, 361–381. of plants infected by rust fungi. Oikos, 58, 145–150. Worthen W.B. (1988) Slugs (Arion spp.) facilitate Rao S., Baumann D. (2004) The interaction of a Botanophila mycophagous drosophilids in laboratory and field experi- fly species with an exotic Epichloe¨ fungus in a cultivated ments. Oikos, 53, 161–166. grass: fungivore or mutualist? Entomologia Experimentalis et Applicata, 112, 99–105.