Introduction

Acremonium strictum is a widespread soil-inhabiting fungal species formerly known as Cephalosporium acremonium (Gams 1971). Recently, a phylogenetic analysis of the genus Acremonium revealed its place in a smaller Sarocladium strictum clade (Summerbell et al. 2011). A. strictum is distinct from the Neotyphodium group of grass endophytes (Glenn et al. 1996). As a root-associated fungus with a broad host range of monocots and dicots, no pigmentation in mycelia and a low colonization level in healthy plants, it can be assorted into class 2 of non-clavicipitaceous fungal endophytes according to Rodriguez et al. (2009). It was mainly studied as a non-systemic, root-restricted endophyte with no visible effects on the host plant according to the definition of Wilson (1995). But, it has antagonistic characteristics that affect the development and feeding behaviour of herbivorous insects in various host plants (Dugassa-Gobena et al. 1998; Jaber and Vidal 2009). Goswamy et al. (2008) reported A. strictum to be an egg parasite of the tomato root knot nematode in their field experiments. Recently, Lenc et al. (2015) assessed the microbial communities in different wheat production systems. They rated the root-associated pathogen antagonists A. strictum and Acremonium sp. as belonging to the dominant fungal group with >5 % frequency. The capacity of root-colonizing Acremonium species to induce disease resistance against Fusarium wilts in flax and tomato has been described when compared to effects of arbuscular mycorrhizal fungi (AMF) (Grunewaldt-Stöcker and von Alten 2003). Further, an elicitor production by A. strictum has been reported. It released an extracellular elicitor protein which provided a systemic protection against Colletotrichum anthracnose disease and Botrytis grey mould in strawberry when applied as foliar spray. This induced resistance was accompanied by the expression of defence-related genes (i.e. PR1 and Chi2-1). The protein showed characteristics of fungal serine proteinases of the subtilisin family, and its proteolytic activity seemed to be required to induce the defence response (Chalfoun et al. 2013). Few reports on plant pathogenic traits of A. strictum causing wilt disease, e.g. in daisy (Chase and Munnecke 1980), in maize (Tagne et al. 2002) and in tomato plants (Anjum and Akram 2014), have pointed to diverse roles that A. strictum can play within a host plant.

Several isolates of A. strictum were detected in roots of ericaceous salal plants (Gaultheria shallon) in association with mycorrhizal fungi, and thus, it was regarded as a potential ericoid mycorrhizal fungus (ERMF) (Xiao 1994; Xiao and Berch 1996). In molecular genetic analysis, Monreal et al. (1999) separated the root-inhabiting Acremonium group (Hypocreales) from the clusters of the Oidiodendron group (Leotiomycetes, incertae sedis) and Hymenoscyphus ericae aggregate, which include Rhizoscyphus ericae and the related sterile mycelia. They investigated A. strictum isolates from salal that were gathered from a Canadian site and detected microscopically typical mycorrhizal structures in the outer cortical cell layers. However, these A. strictum isolates were described as unusual ERMF due to slow colonization in laboratory tests (Xiao and Berch 1996). Further reports on their physiological growth demands in axenic culture tests revealed essential differences to those of common ERMF isolates of Rhizoscyphus and Oidiodendron species (Xiao and Berch 1999). In fact, the genuine symbiotic relationship of A. strictum with G. shallon was questioned (Berch et al. 2002) and is not confirmed yet. Within a broad molecular ITS analysis of verified ERMF isolates from the same salal stand at SCHIRP site at Vancouver Island, they confirmed that all recovered ERMF belonged to fungal groups which related to ERMF and had been reported from elsewhere in the world (e.g. UK, Australia). However, A. strictum was no more recovered as a mycorrhizal isolate.

Browsing the literature of the expanding research on ericoid mycorrhiza distribution in Europe, Australia, USA, Argentina and Canada in the last decades and even including the reports of dark septate endophytes (DSE) in ericaceous plants, we did not find any additional note on A. strictum (e.g. Straker 1996; Chambers et al. 1999; Sharples et al. 2000; Vrålstad et al. 2000; Allen et al. 2003; Addy et al. 2005; Cairney 2006; Bougoure et al. 2007; Chambers et al. 2008; Hazard et al. 2014; Bruzone et al. 2015). It seems that in ericaceous plants, A. strictum is a less abundant, and therefore, a seldom recovered root-inhabiting fungus, or it may have been disregarded and discarded like other unimportant, fast-growing and sporulating fungi of the rhizosphere (Wurzburger et al. 2012).

In our previous microscopic studies on host-fungus interactions with Acremonium root endophytes, including an endophytic A. strictum strain originating from barley roots, we found hyphal structures similar to hyphal coils of ericoid mycorrhizal fungi in the rhizodermal cells of greenhouse-cultured flax plants (Grunewaldt-Stöcker et al. 2007) and later in hair roots of inoculated Rhododendron plantlets in liquid culture. These observations motivated us to pick up the hypothesis of Xiao (1994) on the mycorrhizal nature of A. strictum. The detection of A. strictum with mycorrhiza-like structures in hair roots of inoculated ericaceous plants seemed to be a first necessary step of evidence. But, the relation of structural features to functions of cell invading fungi was already questioned in cases of DSE by Peterson et al. (2008). Besides molecular cytology, they requested live-cell imaging to elucidate early changes in plant cells when challenged with root endophytes mimicking mycorrhiza. Therefore, our objective was to clear the mycorrhizal feature in a valid microscopic test of plant cell vitality during the early colonization process. If A. strictum is a symbiotic fungus, it should exist within a metabolically active host cell. As a mere endophyte, it should not be found in intact root cells but could benefit from nutrients in the apoplast and the disintegrated or dead cells within the root tissue.

Material and methods

Host plant and fungi

Micropropagated Rhododendron plantlets cv. ‘Cunningham’s White’ were raised as reported in Grunewaldt-Stöcker et al. (2013) and served as host plants in all trials.

As root endophytes, we used A. strictum W. Gams from barley roots, identified by W. Gams at the Centraalbureau voor Schimmelcultures, Utrecht, The Netherlands (CBS) (Grunewaldt-Stöcker and von Alten 2003), and deposited as strain DSM 100709 at the Deutsche Sammlung Mikroorganismen und Zellkulturen (DSMZ), Braunschweig, Germany, and A. strictum A.s.T1, a green fluorescent protein (GFP) transformant of A. strictum DSM 100709 (Grunewaldt-Stöcker et al. 2007), deposited at the Phytomedicine collection, Leibniz Universität Hannover.

The following ericoid mycorrhizal fungi (ERMF) were used: Oidiodendron maius Barron, strain DAOM 184108. Rhizoscyphus ericae (D.J. Read) W.Y. Zhuang & Korf 2004, strain Read 100, kindly supplied by Kurt Haselwandter, Innsbruck [=Hymenoscyphus ericae (D.J. Read) Korf & Kernan 1983, = UAMH 6563, = Pezizella ericae D.J. Read 1974]. R. ericae isolate 23II was extracted from a commercial Rhododendron peat substrate ‘Terrasan’ by trap culture with in vitro-raised Rhododendron plantlets cv. ‘Cunningham’s White’. Isolated frequently amongst many other ericoid fungi from mycorrhizal root segments, it was subcultured on potato dextrose agar and malt extract agar for purification and storage. Thereafter, mycorrhiza synthesis was well achieved in the Rhododendron host cultivar and also in several cultivars of Calluna vulgaris. Isolate 23II was identified by CBS as a member of the R. ericae aggregate and is deposited at the Phytomedicine collection, Leibniz Universität Hannover. All fungi were cultured on 1.5 % malt extract agar plates (pH 5.5 ± 0.2) at 22 °C in the dark.

Mycorrhiza test systems

We conducted the experiments in three test systems for mycorrhizal development during different time periods: axenic plant liquid culture for >3 months under growth chamber conditions of 12 h day light and 24 °C as described in Grunewaldt-Stöcker et al. (2013) (A), axenic liquid culture in Petri dish for 4 weeks under growth chamber conditions of 12 h day light and 24 °C (B) and greenhouse culture in γ-irradiated (10 kGy) commercial peat moss substrate (Floragard Rhodohum® Ericaceous Soil, Floragard GmbH Oldenburg, Germany; pH 4.3, salt content 1.3 g/L, premium nutrients NPK and trace elements for 6 weeks) for >3 months at 24 ± 5 °C and additional day light of 12 h (C) (Fig. 1).

Fig. 1
figure 1

Mycorrhiza test systems: axenic plant liquid cultures (A), axenic liquid cultures in Petri dishes (B) and greenhouse cultures in γ-irradiated peat moss substrate (C)

Full size image

The inoculation procedure varied. Long-term test in axenic liquid cultures (A) was conducted with four plants and four inoculum plugs per vessel from freshly grown fungal agar cultures as previously described in detail (Grunewaldt-Stöcker et al. 2013). For short time tests (B), mycelia of the sterile R. ericae isolates and of the sporulating O. maius and A. strictum were collected from the agar surface, transferred into sterilized distilled water and immediately used for a root drench of plantlets from in vitro cultures. For greenhouse tests (C), the inoculation started at the acclimatization step, with a root drench as described above, and was repeated with pre-inoculated substrate when the acclimatized plantlets were transferred to the open box system (C, Fig. 1).

Sample preparation and microscopy

For the sample preparation, plants were gently rinsed in tap water; roots were excised, not fixed and stained in FUN®1 cell stain (F-7030, Molecular Probes™, Thermo Fisher Scientific, Life Technologies, USA; 50-μM working solution in distilled water) for at least 15 min in the dark. After subsequent clearing in sterilized distilled water, root samples were cut into 1-cm segments and mounted on glass slides for microscopy. The function of the fluorochrome was controlled by an additional exclusion test with root samples soaked for 15 min in 50-μM solutions of cycloheximide and actinomycine D.

Root segments were assessed with brightfield DIC microscopy ×400—×1000 and epifluorescence microscopy (Zeiss Axio Imager A2, extinction BP 485/20, beam splitter FT 510, emission LP 515; photo documentation with Axiocam MRC). For detailed studies, confocal laser scanning microscopy (CLSM, Leica TCS SP2; Ar/Kr laser excitation 488 nm, emission 500–530 nm; He/Ne laser excitation 543 nm, emission 560–590 nm) was used with transmission overlay. The A. strictum transformant AsT1 was detected for its GFP protein in live specimens by standard epifluorescence or CLSM technique (Grunewaldt-Stöcker et al. 2007). Staining procedures with acid fuchsine-lacto phenol for CLSM studies of fungal structures were applied as formerly described (Grunewaldt-Stöcker et al. 2013).

Experimental design

The trials for ERM synthesis in Rhododendron plantlets with the fungal strains and a blank control were carried out in three different culture systems (A–C, Fig. 1). Experimental design is given in Table 1 based on the following: the duration of the experiments, the number of repetitions, the number of plants per experiment and the total number of microscopically analyzed plants. The number of analyzed 1-cm root segments per plant differed according to the development of the plants and the test system with ≥20 ≤ 60 segments (laboratory tests A), 15 segments (laboratory tests B) and ≥20 ≤ 60 segments (greenhouse tests C).

Table 1 Experimental array in three test systems (A, B, C) and actually assessed plants to evaluate the mycorrhizal nature of A. strictum
Full size table

Results

Staining for cell vitality assessment

FUN®1 staining conveniently enabled the discrimination of living from dead cells by the formation of moving cylindrical intravacuolar structures (CIVS). This ATP-dependent staining effect could be seen in plant tissues as well as in fungal cells, both in brightfield and in epifluorescence microscopy. Under brightfield microscopy, a vital cell was characterized by an uncompressed nucleus, cytoplasmic movement, visible cytoplasmic threads and vacuoles with red-coloured inclusion bodies which appear at first spherical and later rod shaped. Dead and disintegrated cells with a shriveled cytoplasm did not form CIVS. Metabolically active hair root cells were also discernable by the bright green fluorescence of the nucleus in a light green cytoplasm and the assembly of red fluorescent CIVS in the vacuoles (Fig. 2). In the active fungal mycelia of A. strictum and of the tested ERMF, the nuclei kept the green fluorescence for a longer time, while the small CIVS in the vacuoles were visible only for few minutes under epifluorescence microscopy (Fig. 3). Therefore, in fully colonized root cells, it was important to differentiate carefully whether the CIVS were located in fungal or plant vacuoles. In such cases, the combination of brightfield and epifluorescence microscopy at highest magnification was indispensable for a valid decision.

Fig. 2
figure 2

Metabolically active epidermal hair root cells (vc) of Rhododendron with green nuclei and orange-red CIVS and dead cells (nvc) without distinct organelles and with shrivelled cytoplasm, visualized by vital staining. a Epifluorescence image. b Brightfield image with DIC corresponding to a. c Vital among non-vital cells in compound epidermal tissue layer, epifluorescence image

Full size image
Fig. 3
figure 3

Metabolically active mycelium of R. ericae strain 23II with green-yellow fungal nuclei (FN) and tiny orange CIVS after vital staining. Two-channel confocal laser scanning microscopy, merged with transmission overlay

Full size image

The FUN-staining reaction was tested by comparing root samples treated with cycloheximide and actinomycin D. The cells of these samples were completely inactivated and did not express any fluorescence of nuclei and cytoplasm and did not form CIVS. Furthermore, this staining method also allowed storage of prepared, but non-examined root samples in sterilized distilled water for at least 48 h at 4 °C in the dark without loss of fluorescent signals in vital cells with stained CIVS and nuclei of plant and fungal cells. Thus, the microscopic evaluation could be extended with regard to sample numbers and detection time.

Infection process and colonization potential

Microscopic observations of A. strictum in axenic liquid culture with Rhododendron plantlets revealed infection patterns that gave first hints for a structural similarity with ericoid mycorrhizal infections. By common staining procedures with acid fuchsine and confocal laser scanning microscopy, the penetration and proliferation processes were followed (Fig. 4). Two days after inoculation (days post-inoculation (dpi)), the fungus already entered an epidermal cell with a fine penetration tube from the root surface, from an adjacent cell or from the apoplastic space between epidermal and cortical cell layers. Loop-like hyphae were found in obviously dead cells, and after 2 weeks, infected cells were filled up with hyphal coils.

Fig. 4
figure 4

Observations on A. strictum colonizing micropropagated Rhododendron hair roots in axenic liquid culture; acid fuchsine-lacto phenol staining for confocal laser scanning microscopy and documentation in time series of days after inoculation (dpi). a Surface mycelium and penetration of a dead epidermal cell with a residue of contracted cytoplasm. b Intracellular coiled hyphae. c A loop of vacuolated hyphae in a disordered epidermal cell. d Intercellular spread of hyphae. e A host cell totally filled with hyphae. f–i Intensive intercellular and intracellular colonization

Full size image

With the application of the cell vitality test, we tried to clarify these mycorrhiza-like infection structures in Rhododendron hair roots. In both long-term and short-term laboratory tests (A, B), A. strictum failed to start a symbiotic intracellular interface with vital host cells. However, in liquid cultures, hyphae of A. strictum were tightly associated with the surface and the intercellular spaces around living cells of the epidermal cell layer that were indicated by moving CIVS in distinct vacuoles (Fig. 5) and a green fluorescent nucleus in a light green cytoplasm. The septate hyphae showed vacuoles and nuclei which were easily recognized, and also, metabolically active hyphae harboured tiny CIVS in stained samples. The loose contact of epidermal cells in developing root sections enabled Acremonium hyphae to enter the outer root tissue layers directly without an appressorium formation. This fungal structure built in the first contact with the host was rarely observed and only occurred in liquid culture tests. Preferentially, tissue gaps at the sites of emerging lateral roots were used to colonize the roots. The endophyte was often found in the apoplast surrounding vital epidermal cells, but mainly spread among dead cells (Fig. 6).

Fig. 5
figure 5

Mycelium of A. strictum in close contact to vital epidermal cells with accumulation of red CIVS in hair roots of Rhododendron plantlets in axenic liquid culture, 26 days after inoculation. Brightfield DIC microscopy after vital cell staining

Full size image
Fig. 6
figure 6

Vital mycelium of A. strictum growing intercellular in the epidermal layer of hair roots of Rhododendron plantlets in axenic liquid culture 26 days after inoculation, verified by vital cell staining. a Epifluorescence image. b Brightfield image with DIC

Full size image

Intracellular colonization was observed with loop structures of hyphae in disrupted, non-vital epidermal cells (Fig. 7a, b). These hyphae later formed intense coil-like structures and totally filled up dead cells (Fig. 8a–f). In comparison, the tested ERMF strains of R. ericae and O. maius started to develop a typical ericoid mycorrhizal association in metabolically active cells visibly 8 dpi. Some impressions of early infection stages are presented from the short-time tests B in Fig. 9.

Fig. 7
figure 7

Intracellular colonization of A. strictum in non-vital cells, 26 days after inoculation, in FUN®1-stained Rhododendron hair roots. a Brightfield image with DIC. b Corresponding epifluorescence image to a

Full size image
Fig. 8
figure 8

Intracellular colonization of A. strictum (arrows) in non-vital hair root cells of Rhododendron in axenic liquid culture, visualized by vital staining. a, b Hyphal complex adjacent to vital cells (vc), as marked with FUN®1 cell stain. c, d Totally filled non-vital epidermal cells. e, f Colonized area of dead epidermal cells at the site of a lateral root primordium, brightfield image with DIC and corresponding epifluorescence image, 26 days after inoculation

Full size image
Fig. 9
figure 9

Early mycorrhizal infection stages in vital epidermal cells (vc) of Rhododendron hair roots, with red CIVS of FUN®1 cell stain. a Mycelium of R. ericae strain 100 at 9 days after inoculation. b Hyphae of O. maius at 16 days after inoculation in a vital cell surrounded by non-vital cells. Plantlets from axenic liquid culture

Full size image

In order to affirm the results from laboratory tests in axenic liquid culture under less artificial conditions, micropropagated rooted Rhododendron plantlets were inoculated at the transfer step of acclimatization in commercial peat moss substrate in greenhouse tests C (Fig. 1). Here, too, early mycorrhizal infections became visible from 8 dpi on, ERMF R. ericae isolate 23II and less frequently O. maius were observed to invade vital cells with hyphae coming from the root surface or from adjacent cells to form the coiled hyphal complex (Fig. 10, 11). Again, the host cell vitality was detected by a fluorescent green nucleus, glimmering cytoplasm and distinct vacuoles containing the luminous red rods of transformed FUN®1 cell stain components which marked the active host cell metabolism. In intensely colonized host cells, however, the CIVS were often absent. Aging ERMF-colonized epidermal cells still showed the marked green fungal nuclei, but no positive FUN reaction of a vital host cell (Fig. 11b). Also, in dead host cells, single non-branched hyphae were observed which appeared to proliferate towards suitable host cells (Fig. 11c).

Fig. 10
figure 10

Hyphae of ERMF R. ericae strain 23II in vital host cells (vc) with red CIVS in FUN®1-stained samples from greenhouse-cultured Rhododendron plants, 28 days after inoculation. a–c Brightfield images with DIC. d Epifluorescence image corresponding to c. pcn plant cell nucleus

Full size image

In the greenhouse test system (C, Fig. 1), the colonization result of A. strictum differed entirely from that of the ERMF. Even though it was inoculated twice with high doses of conidia and mycelium, the infection rate in Rhododendron hair roots was minimal. The root tissue was well developed and metabolically active, but A. strictum was never found in vital cells. Only few mycelia spread out on the root surface, and even dead epidermal cells were rarely colonized. A prolonged culture time exceeding 4 months did not change the observed low colonization potential.

Fig. 11
figure 11

Hyphae of ERMF R. ericae strain 23II in vital host cells (vc) with red CIVS of FUN®1 cell stain and green-stained nuclei and in non-vital cells (nvc) with green-stained fungal nuclei in FUN®1-stained samples from greenhouse-cultured Rhododendron plants, 23 days after inoculation. a, b Confocal laser scanning microscopy image with transmission overlay. c Epifluorescence image

Full size image

To facilitate the laborious detection of the discrete endophyte and to exclude that fungal infection was overlooked, the GFP-transformed Acremonium strain A.s.T1 was evaluated in non-stained samples by fluorescence microscopy. In correspondence to the wild-type A. strictum, there was nearly no colonization of the clearly identifiable fungus observed over a period of more than 8 weeks. Only in very few cases, fluorescent hyphae were found on the root surface and in obvious dead cells which was confirmed by subsequent vital cell staining (Fig. 12). In summary, there was no fungal coil formation or compact filling within vital cells in the Acremonium treatments.

Fig. 12
figure 12

Intracellular hyphae of the A. strictum GFP-transformant A.s.T1 in dead hair root cells of FUN®1-stained samples from greenhouse-cultured Rhododendron plants. a Brightfield image with DIC, CIVS in vital hyphae of A.s.T1. b Epifluorescence image of hyphae with distinct green fungal nuclei and faint green fungal cytoplasm

Full size image

The results of the different trials in the laboratory and the greenhouse (Table 2) show that there was no evidence for a true mycorrhizal trait of the root endophyte A. strictum. Only by use of the vital cell staining, we could distinguish the ericoid fungal structures from intracellular hyphae of A. strictum.

Table 2 Symbiotic relations of diverse fungal root endophytes in Rhododendron hair roots
Full size table

Discussion

Many diverse root endophytes cannot be clearly assorted into functional categories like saprotrophs, weak pathogens or mycorrhizal mutualists, due to the following: changing behaviour during their lifespan, undiscovered traits or handling difficulties in culture and experiments. Brundrett (2006) postulated main criteria to discriminate arbuscular and ectomycorrhizal fungi from parasites or endophytes. Mycorrhizal fungi should form specialized hyphae in specialized plant organs, and the development of plant and fungal structures should be synchronized. Symbionts should transfer nutrients to the plant, depend on essential host-supplied nutrients and provide strong or weak benefit for the host. Although ERMF differ from obligate biotrophic AMF by a saprophytic phase, they form an interfacial matrix between fungal cell wall and host plasma membrane in epidermal host cells during the symbiotic phase (Peterson et al. 1980; Perotto et al. 1995; Peterson et al. 2004). Such a symbiotic phase can last for some days or weeks, and the mutualistic effects can only be generated as long as both partners are alive. The infection leads to the typical intracellular complex of coiled hyphae which, in the end, can fill the whole cell lumen. But, according to the criteria for a mycorrhizal relation, the ERMF coiled hyphae are not highly specialized (Brundrett 2002) and may resemble other endophytic fungi in morphology.

With ongoing molecular screenings in the fungal microbiome of ericaceous plants, the description of fungal diversity widened the spectrum of identified and non-identified root colonizers (e.g. Sharples et al. 2000; Bougoure et al. 2007; Vohník and Albrechtová 2011; Sun et al. 2012; Hazard et al. 2014). Especially when different fungal species co-exist even in a single root (Perotto et al. 1995), by-catches of non-mycorrhizal endophytic fungi seem to be possible. Therefore, molecular ITS-RFLP-based analysis and genomic sequencing of involved fungi were only valid for the identification of ERMF in combination with positive re-synthesis results from inoculated host plants (Monreal et al. 1999; Allen et al. 2003; Rice and Currah 2006).

In the case of the Acremonium endophytes from ericaceous salal roots, their identity is undoubted. Monreal et al. (1999) found sequences of the unusual ERMF A. strictum clustered in the Hypocreales, and it was similar, but not identical to a GenBank sequence ex type of A. strictum U57495, ATCC 34717 (type culture; source Triticum aestivum). Similarly, our A. strictum isolate was confirmed (by W. Gams at CBS and by molecular comparison at DSMZ) and the strain deposited as DSM 100709. However, Berch et al. (2002) doubted a mycorrhizal nature of A. strictum isolates from salal, even though microscopic evidence of typical coils in hair root epidermal cells had been presented (Xiao 1994; Xiao and Berch 1996, 1999). The mere intracellular occurrence with hyphal structures similar to those of ERMF must not necessarily imply a similar function. Ericaceous plants were shown to be problematic with regard to a wide range of fungi harboured in peripheral hair root cells with typical coiled hyphal complexes, including A. strictum (Rice and Currah 2006). Based on this issue, we began our tests to clarify whether A. strictum is able to infect a living epidermal cell—one primary criterion for the mycorrhizal nature of the fungus.

Starting from the observation of mycorrhiza-like structures in Rhododendron hair roots, we used a staining technique with the vital cell stain FUN®1 which is able to permeate the plasma membrane and to mark metabolically active cells in the presence of ATP with CIVS in plant and fungal cells (Millard et al. 1997; Pina-Vaz et al. 2001; Hua et al. 2007). The staining procedure, already applied in tissue of Asparagus officinalis (Rempe-Verspermann et al. 2014), was adapted for the hair root structures of ericaceous plants, with a reduced reaction time at room temperature. So, the method used here is a valuable, easy-to-handle tool for vitality assessment of both ericoid fungal mycelia and ericaceous hair roots cells. Previously described methods of fluorescence microscopy for living root tissues (Vierheilig et al. 2005) appeared to be much more laborious and less suited for our objective. Therefore, we recommend application of the FUN vitality cell test for precise differentiation of mycorrhizal from endophytic root fungi. Moreover, it might be useful for the study of the assumed temporary symbiotic phase of DSE (Jumpponen 2001) to elucidate the function of these root endophytes. The superficial and internal root colonization pattern of A. strictum is a reminder of early phases described for DSE in general and for Phialocephala fortinii in detail (reviewed in Jumpponen and Trappe 1998), although A. strictum lacks dark pigmented cell walls and the development of microsclerotia.

The many cited reports in the reviews of Yuan et al. (2010) and Rodriguez et al. (2009) on the roles of class 2 non-systemic endophytes within their host plants show that there is still a large lack of knowledge about their function. This is also true for the Acremonium root endophytes. But, there were several hints that A. strictum is a non-mycorrhizal root colonizer. Bargmann (1993) observed Acremonium kiliense (later re-defined by W. Gams as A. strictum) in electron microscopy studies only in dead cortex cells of axenically raised tomato roots. Xiao and Berch (1996) recorded nutritional demands of A. strictum in axenic culture that differed significantly from those of ERMF like R. ericae and O. maius.

As we failed to give evidence for at least one infected living epidermal root cell in our studies of hundreds of Rhododendron root samples derived from different culture systems, the hypothesis of a mycorrhizal nature of A. strictum is almost invalid. The penetration of disintegrated epidermal and cortex cells and the development of coil-like structures filling dead host cells rather point to a saprophytic phase in the lifespan of this root endophyte. It might co-exist with ERMF in mycorrhizal roots, thus explaining the early findings of Xiao and Berch (1994).

The low infection level of A. strictum in Rhododendron, under greenhouse culture growth conditions which were favourable for the tested ERMF, confirms the poor colonization potential observed by Xiao and Berch (1996) in their laboratory tests with Gaultheria plants and further corroborates its difference from ERMF.

The tight association of A. strictum hyphae to the intercellular space around living cells enables this endophyte to interact in wide latitude with the host plant, even as a beneficial inducer of disease resistance (Grunewaldt-Stöcker and von Alten 2003; Grunewaldt-Stöcker et al. 2007). So far, this discrete root endophyte A. strictum is fitting into the range of multifunctional endophytes looked at in depth by Porras-Alfaro and Bayman (2011). It remains open whether the capability to form a mycorrhizal association is a strain-specific trait and whether A. strictum strains from different hosts differ in their qualification for a symbiosis. It might also be that in this genus, more host-fungus specificity occurs as compared to known ERMF. We should not extrapolate the lack of specificity in Rhizoscyphus and assume that all ERMs behave this way. In our tests, an ericaceous isolate of A. strictum was not at our disposal, but further comparative studies with isolates from Ericaceae could answer the question of specificity.