Introduction

Every species is intricately involved with a myriad of associates—some obligate, some facultative—that profoundly influence their evolution, physiology, and life history [1].

Dutch elm disease (DED) is a highly destructive vascular disease, which caused an extensive loss of mature elms in Europe, Asia, and North America during the twentieth century. The disease is caused by fungi of the genus Ophiostoma (Ascomycota, Ophiostomatales) and, in particular, by Ophiostoma ulmi (Buisman) Nannf and Ophiostoma novo-ulmi Brasier (ONU) [2]. Pathogen spreading and infection of suitable hosts are mainly ensured by elm bark beetles (EBB) (Coleoptera: Curculionidae, Scolytinae) [3]. The synchrony between the life cycles of host tree, fungus, and EBB allows vectors to disseminate ONU when host plants are most prone to infection and temperatures favorable for fungal growth, thus boosting the pathogen’s aggressiveness [4]. Moser et al. [5] showed that phoretic mites carried by EBB in turn transport ONU conidia, ascospores, and in some cases hyphae attached to their body surfaces, in sporothecae and in the digestive system. Mites may therefore contribute to DED transmission by spreading the fungus within the gallery system, enhancing ONU sexual reproduction and promoting an increase in genetic diversity through the fertilization of proto-perithecia. Moreover, mites may contribute to increase the spore load beyond the threshold required for infection [5].

The virulence of ONU might be negatively affected by the presence of a family of naturally occurring viruses, known as “d-factors,” found in the fungus cytoplasm [6] and able to prevent ONU from infecting healthy elms [7]. O. novo-ulmi isolates carrying these viruses exhibit slow, ragged growth, as well as a reduction in sporulation, perithecia production, and viability of conidia [8, 9]. In Europe, the virus was prevented from spreading into the ONU population, probably via the acquirement of the sexual compatibility type (MAT-1) and vegetative compatibility (vic) loci from O. ulmi [10]. Sexual reproduction alone helps to eliminate virus infection in ONU [11] leading to a rapid increase in the diversity of vegetative compatibility phenotypes [12], which reduces virus transmission.

The DED pathosystem is therefore a complex of interactions involving several other organisms in addition to the host plant, the pathogen, and the vector, ideally the entire EBBs’ holobiont [1] as well as the d-factor viruses, whose interplay influences the outcome of the infection.

EBB also transport species of the genus Geosmithia (Ascomycota: Hypocreales) [13], a monophyletic morphogenus of anamorphic ascomycetes that currently includes 32 published and at least 20 unpublished species of mitosporic fungi [14,15,16,17,19]. Geosmithia fungi may live as saprobes on various plant substrates, in soil or foodstuffs, and as true plant endophytes and are in most cases insect-associated [14, 17, 19, 20]. Several Geosmithia species have, at least for some parts of their lives, the same habitat as ONU on dying elms, although they occupy different ecological niches [21].

The existence of a more complicated relationship between ONU and Geosmithias than just occupying the same habitat and having the same vectors has recently been suggested by the discovery of widespread horizontal gene transfer (HGT) of a genomic fragment comprising the cerato-ulmin (cu) gene between the two fungi [22]. Cerato-ulmin, a class II hydrophobin of about 8 kDa produced by the pathogens O. ulmi, O. novo-ulmi, Ophiostoma himal-ulmi, and by the non-pathogen Ophiostoma quercus [23, 24], might play a role in DED by improving the fitness of the fungus [25].

Several hypotheses were made on the ecological role of Geosmithia spp. on host trees, but no conclusive evidence has been provided [17, 20, 26]. The aim of the present study was to examine the occurrence of a relationship between Geosmithia spp. and some Ophiostomatoid fungal species that have the same host plants and vectors and to define the nature of this relationship. The potential consequences of such a relationship on the DED pathosystem are also discussed.

The study focused on the “elm system,” comprising species of Ophiostoma and Geosmithia specific to elms. The elm system was put in comparison with systems comprising Geosmithia and Ophiostomatoid species from other host plants as oak and conifers (“non-elm systems”).

Materials and Methods

Fungal Strains and Media

The fungal species and strains included in this study are reported in Table 1. For the sake of brevity, in this paper, the term Ophiostomatoid fungi will be indistinctly used to refer to fungal species in orders Ophiostomatales and Microascales that share morphological analogies as the result of convergent evolution due to their association with insect vectors [28].

Table 1 Geosmithia and Ophiostomatoid fungi strains used in this work. In particular: experiment a, fungal growth rate in dual culture; experiment b, stereoscopic examination of mycelial interactions; experiment c, observation of hyphal interaction in white light microscopy; experiment d, transformation of Geosmithia strain with the GFP gene; experiment e, observation of hyphal interactions in fluorescence microscopy; experiment f, fertility test; experiment g, pathogenicity test
Full size table

Short-term stock cultures were maintained on malt extract agar (MEA 2%, Oxoid, Basingstoke, UK) at 4 °C and subcultured at 2-week intervals. Long-term stock cultures were maintained on MEA slopes at − 20 °C. Fungal growth rate in dual culture was assessed on MEA and Czapek Dox Agar (CZD, Oxoid, Basingstoke, UK), while mycelial interactions and ONU sporulation were studied in dual cultures growing on elm sapwood agar (ESA) or 2% MEA [29].

Fungal Growth Rate in Dual Culture (Experiment a)

The reciprocal effect of the presence of Geosmithia spp. or Ophiostoma spp. on the growth rate of the other species was assessed in dual culture in several trials designed as follows. For each fungal combination, three Petri dishes (90-mm diameter) filled with 20-ml substrate were inoculated by placing two 6-mm diameter mycelial plugs (one of Geosmithia spp. and one of Ophiostoma spp.), obtained from the edges of actively growing fungal cultures, about 1 cm apart from each other near to the center of dish. Cultures were incubated in the dark at 20 °C and two radii of each colony on the growing edge opposite to the other fungus were measured after 48 h, 3, 5, and 8 days. Three plates per isolate were inoculated with two identical plugs as a control. Daily radial growth rates were compared by one-way ANOVA (Statistica 10, StatSoft Inc.).

Eight Ophiostoma spp. isolates (four ONU and four O. quercus), 11 species of other Ophiostomatoid fungi, and nine Geosmithia spp. isolates (five from elms and four from other trees) were combined in six trials, where Geosmithia from elm and from other trees were grown in dual culture with species from the three Ophiostoma groups (Table 1). Fungal combinations from the non-elm system (oak and conifers) were cultured on both 2% MEA and CZD, while fungal combinations from the elm system were grown on 2% MEA.

Visual Examination of Mycelial Interactions (Experiment b)

In order to determine the existence of a recognition system between Ophiostomatoid fungi and Geosmithia, various species of Ophiostomatoid fungi were grown in dual culture with Geosmithia spp. Inoculations were performed in 90-mm diameter Petri dishes containing 20 ml of substrate (ESA and 2% MEA) as in Experiment a. Three replicates per each fungal combination and medium were prepared and incubated at room temperature in diffuse natural daylight. ESA was used since it had proved very effective for discriminating vegetative compatibility reactions in O. novo-ulmi [29, 32], while MEA is a common medium for growing Geosmithia spp. Colonies were visually examined after 5 and 10 days for the presence of an antagonism zone or a reaction zone in the region of mycelial contact [32, 33]. Ten Ophiostoma spp. isolates (four ONU, four O. quercus, and two O. ulmi), 11 Ophiostomatoid fungi, and ten Geosmithia spp. isolates (six from elms and four from other trees) were combined in dual cultures in eight different trials (Geosmithia strains from elm or from other trees were cultivated with fungi from the four Ophiostoma groups) (Table 1).

Observation of Hyphal Interactions in White Light Microscopy (Experiment c)

The mycelial interactions between several Geosmithia spp. and strains with different O. ulmi and O. novo-ulmi strains were studied by white light microscopic observations.

Microscope slides (three per each Ophiostoma/Geosmithia combination) covered by a water-agar film (2% w/v) were inoculated with two mycelial plugs (6 mm in diameter) obtained from the edges of actively growing fungal cultures, placed about 1 cm apart from each other. Microscope slides were observed after 2-day incubation (20 °C in the dark) with a Zeiss Axioscop 50 optical microscope equipped with a Nikon digital camera. Images were processed with the Nikon Digital Sight DS-L1 software.

Transformation of Geosmithia sp. 5 “IVV7” with the Green Fluorescent Protein (GFP) Gene (Experiment d)

A GFP-tagged Geosmithia strain was obtained to gain a clearer and more detailed vision of the interactions between hyphae of the two fungi. Insertion of the GFP gene into the IVV7 isolate of Geosmithia sp. 5 was achieved through Agrobacterium tumefaciens—mediated transformation by using strain AGL-1 (kindly provided by Prof. A. Sesma, Universidad Politécnica de Madrid, Spain) containing the pCAMBgfp vector that includes a modified GFP (SGFP) and the hygromycin resistance gene [34]. Transformation was performed according to [34], while stabilization of transformants was carried out as in [35]. Eight independent IVV7-GFP clones were obtained and GFP expression was observed under fluorescence using a Leica MZ FLIII microscope equipped with a mercury lamp and GFP filters (excitation filter at 480/40 and a barrier filter at 510-nm LP). The number of insertions of the pCAMgfp plasmid was determined by southern hybridization using a digoxigenin-labeled GFP probe [27] (not shown). The growth rates of the IVV7-GFP clones and of their parental isolate were determined by inoculating MEA plates with 7-mm diameter mycelial plugs. Plates (at least three per clone) were incubated in the dark at 20 °C and radial growth was measured daily for 12 days. Differences in growth rate were analyzed with the PAST 3× software [36]. Based on southern blot and growth rate, the Geosmithia-GFP clone 3.2.2, containing one copy of the GFP gene, was chosen for the experiments.

Observation of Hyphal Interactions in Fluorescence Microscopy (Experiment e)

The interactions between the hyphae of O. ulmi and ONU isolates (ONU ssp. novo-ulmi and ONU ssp. americana, Table 1) and the transformed Geosmithia sp. 5 IVV7-GFP were observed in microscope slides (three replicates for each ONU/Geosmithia combination) as described in experiment c. Inoculated slides were incubated in the dark at 20 °C and observed after 2 days under UV light by fluorescence microscopy with a Leica MZ FLIII stereomicroscope (courtesy of Prof. Alessio Mengoni, Department of Biology, University of Florence), equipped with a mercury lamp and GFP filters (excitation filter at 480/40 and barrier filter at 510-nm LP), or white light to verify the autofluorescence of mycoparasite structures. Up to 100 slides per combination were examined.

Fertility Tests (Experiment f)

The effect of the presence of Geosmithia spp. on the production of perithecia in Ophiostoma spp. was assessed both in the elm system and in the non-elm (oak) system. Petri dishes (90-mm diameter, three replicates per species combination) filled with 20 ml of ESA were inoculated as in experiment a with two mycelial plugs, one from Geosmithia spp. and one from Ophiostoma spp. mating type A (mtA). Plates were incubated for 12 days in darkness at 20 °C, followed by 7 days in diffuse light. Spores scraped from the surface of an Ophiostoma spp. mating type B (mtB) colony that served as a donor strain were applied in 2-cm2 patches (five patches per plate) to the plates containing Ophiostoma spp. mtA as a recipient strain in combination with Geosmithia spp. Plates were incubated for 10 days in diffuse daylight at room temperature. The presence and the number of perithecia (no/cm2) were scored under a Nikon SMZ800 stereoscope and data analyzed by means of ANOVA (Statistica 10, StatSoft Inc.). As a control, three plates per species combination were inoculated with only the Ophiostoma mtA strain and fertilized with the respective Ophiostoma mtB strain. In the elm system, 18 Geosmithia isolates were combined with two ONU ssp. novo-ulmi and ssp. americana mtA isolates, respectively, and crossed with the mtB of the corresponding species. In the “oak system,” five isolates of Geosmithia sp. 5 were tested with 5 O. quercus mtA isolates fertilized with a single O. quercus mtB isolate (Table 1). Fertility tests were repeated at least three times for each combination.

Pathogenicity Tests (Experiment g)

The impact of Geosmithia in the DED pathosystem was investigated in vivo by means of two pathogenicity tests carried out at the IPSP-CNR experimental nursery (Antella (43° 43′ N 11° 22′ E; 170-m elevation, Florence, Italy). Several Geosmithia spp. and ONU strains were inoculated alone and in combination in the elm clone Ulmus “Commelin,” which was chosen for being extremely susceptible to DED [37]. Hundred five-year-old saplings growing in rows (spacing 0.5 m within × 1 m between rows) in a substrate comprising commercial loam to a depth of 2-m drip irrigated were inoculated. The bed was cleared and plowed prior to planting and weeded monthly thereafter. Two pathogenicity tests were performed as follows:

  1. 1)

    In May 2013, Ulmus Commelin (six individuals per fungal strain) was inoculated with each of seven Geosmithia spp. strains with a single wound per plant in the upper third of the main stem. Inoculations were performed following the protocol established by Santini et al. [38] for ONU inoculations, i.e., by cutting through the bark to the younger sapwood with a knife blade bearing two 0.2-ml drops of a 1 × 106/ml fungal spore suspension so that the inoculum was absorbed in the sap flux.

  2. 2)

    In May 2014, 12 Ulmus Commelin individuals were co-inoculated with the same technique as above with a spore suspension containing Geosmithia sp. 5 (IVV7) and ONU ssp. novo-ulmi “H328.” The concentration of each fungus in the inoculum was adjusted to 1 × 106 spores/ml. As a control, 12 trees were inoculated with only Geosmithia sp. 5 (IVV7) and 12 trees with only ONU ssp. novo-ulmi H328. Geosmithia sp. 5 was chosen for the experiment because it is one of the most common species on elm, and IVV7 is our model strain for this species [21, 22, 39]. O. novo-ulmi ssp. novo-ulmi H328 is a well-known and very aggressive strain [38, 40]. Symptoms of disease were observed at 4 weeks (percentage defoliation) and 12 months (percentage of crown dieback) after inoculation by three independent assessors. Pathogenicity data were analyzed by means of ANOVA (Statistica 10, StatSoft Inc.). Arcsine transformation was applied before statistical analyses to correct percentage data for departure from normality assumption.

Results

Fungal Growth Rate in Dual Culture (Experiment a)

Elm System

The growth rate of ONU strains was generally higher in dual culture with Geosmithia spp. isolated from elm than in pure culture, both on MEA and CZD (Fig. 1). In the same trial, the growth rate of Geosmithia did not show such a clear and consistent trend. Within each species, all strains grew at the same rate (non-significant Duncan test, p > 0.05); therefore, different strains were used as replicates in subsequent analyses.

Fig. 1
figure 1

Fungal growth rate in dual culture. Left, growth rate of Geosmithia spp. with Ophiostoma novo-ulmi ssp. novo-ulmi and ssp. americana on MEA (2%) and CZD; right, growth rate of O. novo-ulmi ssp. novo-ulmi and ssp. americana with Geosmithia spp. on MEA (2%) and CZD. Values sharing the same letters are not significantly different based on Duncan’s test (p ≤ 0.05)

Full size image

Non-elm Systems

Both in the oak system (O. quercus in dual culture with Geosmithia from elm, oak or other trees) and in the conifers system (Ophiostomatoid fungi from conifers in dual culture with Geosmithia from elm, conifers, or other trees), the mean radial growth in dual culture was unchanged compared to controls in all tested fungi (results not shown).

Visual Examination of Mycelial Interactions (Experiment b)

The reactions observed between the mycelia of Ophiostomatoid fungi and Geosmithia species were here classified into five main types, ranging from fully intermingling colonies to mutual growth inhibition (Table 2, Fig. 2):

  1. Type 1,

    fully intermingling: complete equal bidirectional mycelial penetration. After 10 days, the two colonies were not distinguishable. Neither boundaries nor changes in color were recognizable in the mycelium.

  2. Type 2,

    intermingling: the two colonies were easily recognizable, but no barrage line was visible and hyphae were intermingled along the junction line.

  3. Type 3,

    mutual incompatibility: a diffuse mycelial barrage, 1 to 2 mm large, was clearly visible along the junction line between the two colonies;

    • 3.1: diffuse mycelial barrage developed by Geosmithia spp. No visible barrage was produced by Ophiostomatoid fungi.

    • 3.2: diffuse mycelial barrage developed by Ophiostomatoids. No visible barrage was produced by Geosmithia spp.

  4. Type 4,

    strong growth inhibition and overgrowth: the growth of Ophiostomatoid fungi was inhibited at a distance of about 1–2 mm from Geosmithia hyphae, which later occupied the gap spreading eventually over the mycelium of the co-cultured Ophiostomatoid species.

  5. Type 5,

    mutual inhibition: a 1–5-mm wide demarcation zone, where the aerial mycelium was missing, was visible along the confrontation line.

Table 2 Reaction patterns between paired Geosmithia spp. and Ophiostomatoid fungi on 2% MEA. Key: type 1, fully intermingling: equal bidirectional penetration; type 2, intermingling: colonies intermingling along junction line; type 3, mutual incompatibility: barrages along junction gap; type 3.1, diffuse mycelial barrages developed by Geosmithia spp.; type 3.2, diffuse mycelial barrages developed by Ophiostomatoid fungi; type 4, strong growth inhibition and overgrowth: inhibition of Ophiostomatoid fungi growth by Geosmithia; type 5, mutual inhibition: inhibition zone; nt, not tested
Full size table
Fig. 2
figure 2

Type of mycelial interactions on MEA (2%) (the mycelium at the figure’s top belongs to Geosmithia spp. in all cases): type 1, G. omnicola (CNR8)-Ophiostoma novo-ulmi ssp. novo-ulmi (H327); type 2, G. flava (CNR120)-Ceratocystis minuta (KW/3/4); type 3, G. flava (CNR120)-Rhexographium fimbriasporum (R/4/1/2); type 3.1, G. omnicola (CNR8)-Rhexographium fimbriasporum (R/4/1/2); type 3.2, G. obscura (CCF3422)-O. ulmi (E2); type 3.2, G. putterillii (CCF3342)-Ophiostoma ainoae (KW/Ku/29); type 4, G. flava (CNR120)-Ophiostoma clavatum (AC/1/1/1); type 5, G. ulmacea (CNR23)-Endoconidiophora polonica (KOW/Ku/41). The different types of mycelial interaction are described in the “Results” section (experiment b)

Full size image

A fully intermingling reaction (type 1) was observed every time that two species from the elm system were grown in dual culture. Interactions between species from non-elm systems were generally characterized, with few exceptions, by various signs of mycelial inhibition, from a barrage to a wide gap along the junction line (types 3–5), revealing a recognition system between the two fungi. Compatible reactions of types 1 and 2 were observed only in dual cultures of some Ophiostomatoid fungi with Geosmithia sp. 5 IVV7 (Fig. 2).

Mycoparasitic Interactions Between Geosmithia from Elm and Ophiostoma in White and Fluorescent Light Microscopy (Experiments c, e)

Under the white light microscope, the mycelia of wild type strains of Geosmithia spp. from elm and of O. ulmi or ONU cultured together appeared to grow towards each other, with profuse hyphal growth and production of mycelial tufts (Online Resource 1). Signs of mycoparasitism by Geosmithia on Ophiostoma hyphae, such as the formation of coilings, appressoria-like branches, pseudopod-like structures, or short hooks, were common (Online Resource 1).

In the elm system, the formation by Geosmithia on Ophiostoma hyphae of structures that are typically observed during mycoparasitic attack was confirmed with increased evidence when the Geosmithia sp. 5 IVV7-GFP clone 3.2.2 was observed in dual culture with both O. ulmi (not shown) and ONU (Fig. 3a–c) under UV light.

Fig. 3
figure 3

a–c Interacting hyphae of Ophiostoma novo-ulmi (ONU) and Geosmithia sp. 5 “IVV7”-GFP (IVV7T). On the left both species are observed under white light, while on the right only Geosmithia hyphae are visible in UV light (under UV light a green specific signal is due to the GFP transformation). Arrows indicate possible parasitic structures formed by Geosmithia

Full size image

Fertility Test (Experiment f)

In the elm system, ONU ssp. novo-ulmi mtA (H327) fertilized by a mtB strain (H328) produced a significantly higher number of perithecia (Duncan test, p < 0.05) in dual cultures with Geosmithia spp. isolates than in control crosses where Geosmithia was absent (Fig. 4). On the contrary, ONU ssp. americana cultivated with Geosmithia spp. did not produce perithecia after fertilization with the opposite mating type.

Fig. 4
figure 4

Fertility test between species from the “elm system.” The number of perithecia formed by Ophiostoma spp. in dual-culture with Geosmithia spp. is shown in red, while gray bars correspond to the control. Duncan’s test was applied to test for differences in means. Values sharing the same letters are not significantly different (p ≤ 0.05)

Full size image

In the oak system instead, all strains of O. quercus mtA fertilized with opposite mtB isolates produced an equal number of perithecia whether or not they were grown in dual culture with Geosmithia sp. 5 (results not shown).

Pathogenicity Tests (Experiment g)

The addition of spores of Geosmithia species to the suspension of ONU spores used for artificial inoculations reduced DED symptoms, both defoliation (4 weeks after inoculation, Duncan test p ≤ 0.05) and dieback (12 months after inoculation, Duncan test p ≤ 0.05), in the elm clone Ulmus Commelin compared to controls inoculated with ONU alone (Table 3). While inoculation with ONU produced severe DED symptoms, elms remained substantially asymptomatic after inoculation with only Geosmithia (Table 3), regardless of the Geosmithia species applied (results not shown).

Table 3 Pathogenicity test. Mean disease scores not sharing a common letter differ significantly by Duncan’s test (p < 005). N number of tested isolates, mean % mean percentage of symptoms, SE standard error
Full size table

In particular, 12 months after inoculation, when the plant reaction is stable and can be considered as conclusive, dieback was much more severe in the plants inoculated with ONU alone than in the plants co-inoculated with ONU and Geosmithia spp. (55.5 vs. 17.6%, respectively). O. novo-ulmi was always successfully re-isolated from xylem of inoculated trees, while none of the Geosmithia species used was re-isolated.

Discussion

Millions of elms vanished from Europe and North America over the last 100 years because the alien fungi responsible for DED established in the areas of introduction a new association with native EBBs that became extremely efficient vectors of the disease. The same beetles also have a high-fidelity association with fungi of the genus Geosmithia [16].

Geosmithias are generally considered as saprotrophs or endophytes [17]. However, in elms, they have never been isolated from dead wood or from healthy trees, but they were only found in beetles’ galleries [21]. High frequency HGT of the cerato-ulmin (cu) gene between ONU and Geosmithia spp. [22] suggests that between the two species exists a relationship that goes beyond simple sharing of habitat and vectors and is much closer.

The present study provides direct and indirect evidence of mycoparasitism on ONU by many Geosmithia isolates specific to the elm system. Should this be the case, it might be concluded that the transfer of the cu gene observed between ONU and Geosmithia may be described as prey-derived HGT. The integration into the genome of sequences derived from organisms consumed as food has frequently been reported, leading to the “you are what you eat” hypothesis [41], both in phagotrophic eukaryotes harboring genes from food sources [42, 43] and in prokaryotes such as the bacteriolytic Bdellovibrio bacteriovorus HD100 [44, 45].

The higher growth rate observed in ONU in dual culture with Geosmithia might be regarded as a sort of “escape in space” (sensu Janzen) [46] of ONU from Geosmithia towards an area free from the “enemy.” The absence of mycelial interaction between Geosmithia and ONU is consistent with the hypothesis that the two organisms represent a newly formed host-pathogen system. In the oak system and in the conifers system, recognition between Geosmithia spp. and Ophiostomatoid fungi was the norm, with very few exceptions. A weak intermingling reaction with no mycelial barrages along the junction line was observed in dual culture (1) between Geosmithia sp. 5 IVV7 and many Ophiostomatales, (2) in all the combinations of Geosmithia spp. with Ophiostomatoid fungi in non-elm systems, and (3) when Geosmithia sp. 20 was co-cultured with Leptographium piceaperdum, the most common Ophiostomatales species associated with spruce beetles [47]. In no instance, however, structures typical of parasitic behavior were formed in these combinations.

Therefore, Geosmithia parasitic behavior seems to be specific to the elm system. In fact, in most of the non-elm systems challenged here, similar interactions were not observed. If this hypothesis proves true, then IVV7 is the isolate displaying the most evident mycoparasitic behavior. This behavior explains its ability to overtake the host defense mechanisms and to grow over it. In this system, ONU represents a widely available carbon source exploitable by Geosmithia fungi.

Comparing the present results with what is known for Trichoderma, a fungal genus well known as a mycoparasite and biocontrol agent [48,46,47,51], several similarities can be identified. Trichoderma attraction to and growth towards its host seems to be stimulated at a distance by the recognition of diffusible signals, such as oligochitins [52]. Mycoparasitism in Trichoderma spp. involves hydrophobins and hydrophobin-like proteins, such as cerato-platanins. Class II hydrophobins HYTLO1 and TvHydII1, isolated respectively from Trichoderma longibrachiatum MK1 [53] and T. viride [54], are required for mycoparasitic activity against phytopathogenic fungi to grow over their hosts. Trichoderma harzianum cerato-platanin Epl-1 [50, 55] also has key functions in the mycoparasitic process, as a self-recognition factor or by modulating hyphal coiling and mycoparasitism-related gene expression, and in the interaction with the host plant [55].

Similarly, in the Geosmithia-ONU system, the attraction signal seems to act at a distance without physical contact. Upon contact, Geosmithia hyphae coil around or grow along ONU hyphae, forming appressoria-like structures that may be used for penetrating ONU (Fig. 4). Geosmithia fungi produce a class II hydrophobin, GEO1, which could be involved in the attachment to other hydrophobic structures, e.g., insect exoskeleton and hyphae of other fungi [25, 56]. The mode of action and the mechanisms involved in the Geosmithia-ONU-elm interaction are still unknown, but GEO1 might play a similar role as Trichoderma hydrophobins and Epl-1, promoting mycoparasitic activity and inducing local and systemic defenses in plants [53,51,55].

Brasier [57,55,59] showed that Trichoderma could trigger sexual reproduction in many isolates of the Phytophthora A2 compatibility group by producing volatile antibiotics, an effect which is more likely a defense mechanism specifically evolved in Phythopthora than an incidental phenomenon. In the present study, Geosmithia spp. tested in fertility trials showed on ONU a similar effect as Trichoderma spp. on Phytophthora. Within the elm system, Geosmithia was shown to induce a significantly higher production of proto-perithecia in all isolates of ONU mtA and of perithecia when fertilized by the opposite mtB. A possible interpretation is that Geosmithia (predator) stimulates in ONU (prey) the “escape from the predator in time” [46] reaction, possibly increasing the evolutionary potential of ONU populations by boosting sexual reproduction and recombination. Such an effect was not observed in the oak system.

Artificial inoculation with ONU resulted in typical symptoms of DED in elms, while no sign of disease was observed when Geosmithia alone was inoculated. In the case of co-infection, the presence of Geosmithia reduces DED symptoms. This could be attributed either to its mycoparasitic activity or to the enhancement of defense mechanisms in elm. A similar effect is well known in Trichoderma fungi that not only protect plants directly by killing other fungi and nematodes but also induce resistance against plant pathogens [51]. Based on these results, Geosmithia is not a pathogen on elm, in contrast with the observation by Hänzi et al. [60]. In no case, we were able to re-isolate Geosmithia from artificially infected elms, nor was it reported among the endophytic cohort of saprotrophs of elm trees [61]. The amount of the fungus in elm tissues could be too low to be detected with standard techniques and require a more sensitive method such as a specific qPCR assay. It could as well be moved to a district of the tree different from the xylem.

If mycotrophy towards many plant pathogenic fungi has long been the original lifestyle of Trichoderma, in Geosmithia, it appears to be a recent event. The DED epidemics that occurred in Europe during the past century created the conditions for Geosmithia development, reproduction, and dissemination by increasing the number of suitable habitats for both ONU and Geosmithia spp. These conditions may have favored the discovery and systematic study of the genus Geosmithia by the scientific community [13]. This hypothesis is supported by the finding that the cu gene was transferred to Geosmithia from ONU, but not from O. ulmi [22]. As the appearance of ONU in Europe can be dated at around the 1960s [2], HGT between the two fungi should be a very recent and currently ongoing event in Europe. The lack of recognition between Geosmithia and ONU in the elm system confirms that they were geographically isolated and interacted only recently. The cu gene was not found in any of the Geosmithia isolates obtained from the non-elm system.

A DED epidemic outbreak is governed by the population dynamics of the host, the pathogen and its vector, and also by the rate of sexual reproduction of the pathogen, which can influence the risk of fungus viral disease outcome and, lastly, by the presence of mycoparasitic fungi as Geosmithia species [22].

The system can be described as a classical Lotka-Volterra model in which the predator, ONU, supported by beetles as vectors, consumes the prey, leading to depletion of elm population and, consequently, of both the predator and the vector populations. When the predator population is low, the prey is able to thrive, thereby putting the ecosystem through cycles of “boom-and-bust.” In the long run, the intervention of new factors may lead to stabilization of the population dynamics. Many polyphagous organisms are able to switch to different carbon sources over time in response to variation in the local ecosystem. Therefore, as ONU became more and more abundant in the community (getting in contact more frequently with organisms sharing the same habitat and vectors), we expect that another organism, even mildly pathogenic as Geosmithia, might have adapted to attack this new host species and reproduce on it, which would lead to an increased degree of parasitism [62].

In the early 1980s, many researchers focused on possible agents of biological control of DED as bacteria [47, 63,61,62,63,64,65,66,70]. Unfortunately, none of these authors could provide evidence that any of these microorganisms might become a successful and widespread competitor or parasite of DED fungi. The main reasons for these drawbacks are that these antagonistic species either are limited by environmental factors [69] or have no vectors able to spread them. On the contrary, Geosmithia species benefit from a widespread distribution and a strict association with effective insect vectors.

Here, it was shown that Geosmithia is an important element in the DED network, making it even more complex, yet probably less detrimental for elms, and more stable over time. There is increasing evidence that the health or disease status of a given organism is not just the result of the interaction between host and pathogen but depends on a complex interaction between each partner and its microbial community (holobiont), which in the end determines the outcome of the infection. Therefore, the fate of the infected elm is not determined only by ONU, but it rather depends on the DED network which may be defined as a holobiont, i.e., the totality of all beings involved comprising ONU, d-factor viruses, EBBs, mites, and also Geosmithia.

Moreover, as Geosmithias living in the elm system are able to mycoparasitize ONU and to reduce DED symptoms in artificially inoculated plants, these fungi might be used as biocontrol agents against ONU. Further research is certainly needed to assess the mechanisms that allow Geosmithias, when co-inoculated with ONU, to attenuate DED symptoms, and to define both how to exploit this effect and how to artificially spread “elm Geosmithias.” However, such a holistic approach would reinforce the conviction that a different management of diseases in natural ecosystems is possible.