Detection and growth of endophytic entomopathogenic fungi in dicot crop plants

The presence and distribution of fungal endophytes in plants is commonly assessed by re-isolation on agar media or detection by PCR-techniques. Histological studies on the process of colonization of the host plant have only scarcely been performed. In the present study, the development of entomopathogenic fungi on the plant surface and inside the tissue was examined by light and fluorescence microscopy of leaf samples treated with various dyes or, to guarantee the specificity of injected endophytes, with primary polyclonal and secondary FITC-conjugated antibodies; diaminobenzidine-tetrahydrochloride (DAB) was applied as stress test for the detection of hydrogen peroxide. Four species of entomopathogenic fungi were studied and compared with three phytopathogenic fungal species. The host plants were oilseed rape (Brassica napus), faba bean (Vicia faba), and cucumber (Cucumis sativus). When blastospores of selected four fungal species were infiltrated into B. napus leaves they appeared to germinate only on the leaf surface, but not within the mesophyll. Successful re-isolation from B. napus inoculated with B. bassiana, Isaria fumosorosea or Metarhizium anisopliae showed that these entomopathogens were able to persist in the tissue for at least two weeks. Formation of brown precipitates after leaf treatment with DAB in the presence of B. bassiana indicated the production of hydrogen peroxide by B. napus but not by V. faba. Overall, the results indicate a lower endophytic colonization than could have been expected from the literature, suggesting nutrient availability in the plant intercellular space and absence of cell wall and cell membrane degrading fungal enzymes as fungal growth-limiting factors. It is concluded that data on endophytic colonization should generally be supported by histological evidence of the kind and amount of fungal growth in the host tissue.

The presence and distribution of fungal endophytes in plants is commonly assessed by re-isolation on agar media or detection by PCR-techniques. Histological studies on the process of colonization of the host plant have only scarcely been performed. In the present study, the development of entomopathogenic fungi on the plant surface and inside the tissue was examined by light and fluorescence microscopy of leaf samples treated with various dyes or, to guarantee the specificity of injected endophytes, with primary polyclonal and secondary FITC-conjugated antibodies; diaminobenzidine-tetrahydrochloride (DAB) was applied as stress test for the detection of hydrogen peroxide. Four species of entomopathogenic fungi were studied and compared with three phytopathogenic fungal species. The host plants were oilseed rape (Brassica napus), faba bean (Vicia faba), and cucumber (Cucumis sativus).
When blastospores of selected four fungal species were infiltrated into B. napus leaves they appeared to germinate only on the leaf surface, but not within the mesophyll. Successful re-isolation from B. napus inoculated with B. bassiana, Isaria fumosorosea or Metarhizium anisopliae showed that these entomopathogens were able to persist in the tissue for at least two weeks. Formation of brown precipitates after leaf treatment with DAB in the presence of B. bassiana indicated the production of hydrogen peroxide by B. napus but not by V. faba. Overall, the results indicate a lower endophytic colonization than could have been expected from the literature, suggesting nutrient availability in the plant intercellular space and absence of cell wall and cell membrane degrading fungal enzymes as fungal growth-limiting factors. It is concluded that data on endophytic colonization should generally be supported by histological evidence of the kind and amount of fungal growth in the host tissue.

Introduction
Due to problems with resistance development of insect pests against chemical insecticides (BASS et al., 2014;SPARKS and NAUEN, 2015;BUZETTI et al., 2016), the ban of certain groups of pesticides (HILLOCKS, 2012;STOCKSTAD, 2013) and the demand of the public for non-chemical plant protection methods, there is an increasing interest in the use of entomopathogenic fungi in biological plant protection. A number of biocontrol preparations based on entomopathogenic fungi are commercially available. They commonly contain high numbers of fungal spores and are applied onto plant surfaces by spraying (ZIMMER- MANN, 2007;OLIVEIRA et al., 2015). A number of reports in the literature describe the isolation of Beauveria bassiana, Metarhizium anisopliae and other entomopathogenic fungi from plants, suggesting an endophytic life style of these fungi (VEGA et al., 2008;OWNLEY et al., 2010). Infection of insect pests via plant tissues colonized by endophytes, ideally combined with seed transmission of the latter, would be an elegant way of delivery. Therefore, attempts are being made to experimentally introduce entomopathogenic fungi into crop plants. Probably the best studied example in this respect is the entomopathogen B. bassiana. Its establishment as endophyte has been reported for many plant species (JABER, 2015;VIDAL and JABER, 2015;GREENFIELD et al., 2016). It can be expected that to be effective against herbivores, long term establishment of the endophyte in the plant at sufficient densities is important. Re-isolation on agar media and detection by PCR-techniques are at present the preferred methods for monitoring the presence and distribution of endophytes in plants. It has been pointed out that the study of endophytes is very much a method-dependent process (HYDE and SOYTONG, 2008).
Histological observation (e.g. WAGNER and LEWIS, 2000;GÓMEZ-VIDAL et al., 2006;LANDA et al., 2013) appears to be only rarely performed (MCKINNON et al., 2017), despite the fact that it is the best method to evaluate to what extent a fungus actually colonizes the host (SCHULZ and BOYLE, 2005). It not only provides information on the spatial relationship between host and endophyte but may also elucidate physiological reactions associated with host defence.
In order to contribute to a better understanding of the relationship between endophytic entomopathogens and potential hosts, a study was initiated that aimed at documenting the development of different entomopathogenic fungi in the important dicot crop plants Brassica napus, Vicia faba and Cucumis sativus. Following inoculation of plants with blastospores or conidiospores the development of the fungi was studied on the plant surface and inside the tissue by light and fluorescence microscopy. The fungal structures were visualized using common dyes as well as polyclonal antibodies labelled with a fluorescence marker. The work was complemented by re-isolation assays that monitored the survival of the endophytic fungi in the plant tissue.

Entomopathogenic fungal strains and inoculation of plants
All entomopathogenic fungal strains were obtained from the culture collection of the Institute for Biological Control (Darmstadt). Beauveria bassiana strains ATTC 74040 (re-isolate of the product NATURALIS ® ), JKI-BI-1202 and JKI-BI-1133 were grown on MPA. B. bassiana NATU-Originalarbeit RALIS was reported to be a successful endophyte (VIDAL and JABER, 2015). Blastospores of B. bassiana were produced in 300 ml flasks in 50 ml Czapek-Dox broth (Sigma-Aldrich) (CZAPEK, 1902;DOX, 1910). After cultivation on a rotary shaker for 3 days at 25°C the cultures were filtered through gauze (Mullro ® ) and the concentration of blastospores was adjusted to 1 × 10 5 -1 × 10 8 /ml, either after pelleting of the spores by centrifugation or by directly diluting the cultures with sterile tap water. After confirming the spore viability with 0.01% acridine orange under blue fluorescent light (STRUGGER, 1940;DARZYNKIEWICZ et al., 1975;MATECKI et al., 2015), the suspensions of blastospores were used for inoculation of true leaves of B. napus, V. faba or C. sativus by tissue infiltration (KLEMENT, 1990) or by injection into stems (MATECKI et al., 2015). Tissue infiltration was achieved by gently pressing the open end of a 5 ml syringe (without hypodermic needle) on the lower side of the leaf and applying sufficient pressure to introduce the blastospore suspensions through the stomates into the leaf. Additionally, faba bean seeds were surface sterilized by 1% sodium hypochlorite (10 min) and 70% ethanol (2 min). Seeds were washed 3 times with sterile tap water (2 min each). Seeds were coated with spore suspensions in a shaking Intelli-Mixer (Neolab GmbH) for four days at 6 rpm and 25 ± 2°C (16/8 h, RF 58%). Then the seeds were transferred into sterile magenta boxes containing MPA. The entomopathogenic fungal species Isaria fumosorosea JKI-BI-1496, Isaria farinosa JKI-BI-1495, Lecanicillium muscarium JKI-BI-1553 and Metarhizium anisopliae JKI-BI-1339 were cultured in SAMŠINÁKOVÁ (1966) medium (25 g glucose, 20 g corn steep solid, 5 g NaCl per 1000 ml dH 2 O) and similarly applied as B. bassiana.

Fungal staining and microscopy
From inoculated leaves cross and paradermal hand-cut sections were prepared and cleared with chloral hydrate solution (1.0 g/ml dH 2 O)/90% lactic acid (2:1) for one to three days. After thoroughly rinsing with dH 2 O, the sections were stained with 0.01% blankophor (Bayer, Leverkusen, Germany) in 0.1 M Tris buffer pH 9.0 for 3 min or with 0.1% solophenyl flavine 7GFE (syn. Direct yellow 96; Shijazhuang Kun Li chemical, China) in 0.1 M Tris/HCl pH 8.5. Prior to staining with the latter dye, the sections were pre-stained with 0.05% safranin for 5 min and rinsed with buffer in order to diminish tissue fluorescence (KNIGHT and SUTHERLAND, 2011). Samples were examined under UV (Leica-filter block A, excitation 340-380 nm, emission ≤ 430 nm) with an Aristoplan epifluorescence microscope (Leica, Wetzlar, Germany). Images were digitally taken with a CCD camera (ColorView II, Olympus) using the software AnalySIS FIVE. The third method used was staining with trypan blue. Leaf segments were boiled in the staining solution (10 ml lactic acid, 10 ml glycerol, 10 ml water, 10 g phenol, 10 mg trypan blue, Merck) for two to four min. After clearing in 100% chloral hydrate solution for at least 30 min, the samples were mounted in glycerol, and microscopically viewed under differential interference contrast optics (KOCH and SLUSARENKO, 1990).

Immunolocalization of fungi
The above described staining techniques are useful to detect fungi in general. To guarantee specific detection of exclusively Beauveria or Isaria and no other fungi, polyclonal antisera against B. bassiana (ATTC 74040) and I. fumosorosea (JKI-BI-1496) were raised in rabbits. For this purpose, fungal biomass from shake cultures was homogenized, centrifuged, and the supernatant used for immunization. IgG fractions were purified as described previously (EIBEL et al., 2005). In Western blotting experiments the main reactive bands for B. bassiana and I. fumosorosea antigen preparations revealed an apparent molecular weight of approximately 27 or 30 kDa, respectively (data not shown). Leaf tissue sections were fixed for 4 h with 4% (w/v) paraformaldehyde in 0.1 × phosphatebuffered saline (0.1 × PBS: 13.7 mM NaCl, 0.15 mM KH 2 PO 4 , 0.65 mM Na 2 HPO 4 , and 0.27 mM KCl, pH 7), washed with microtubuli stabilizing buffer (MTSB: 50 mM HEPES, 5 mM MgSO 4 × 7 H 2 O, 5 mM EGTA, pH 6.9) and 1 × PBS. After dehydration with a graded series of ethanol at 22°C, tissues were embedded overnight in Steedman's wax, a polyester with a low melting point (PEG 400 distearate and cetylalcohol 9:1 [w/w]), according to the method of VITHA et al. (1997). Cross sections of 10 μm thickness were prepared with a microtome at room temperature (Cryocut CM 3050, Leica, Germany) and collected on albumin-glycerol coated slides. Steedman's wax was removed from sections with decreasing ethanol concentrations, rinsed with 0.1 PBS, incubated with buffer for 30 min, and with 100% methanol for 10 min at -20°C. For blocking unspecific binding sites, the sections were incubated with 1% BSA (Aurion) in 0.1x PBS for 1.5 h at 22°C. For immunolabeling, sections were incubated overnight with the anti-B. bassiana (fractions 142/4 or 143/4: 1 mg/ml) or anti I. fumosorosea IgG (fractions 144/4: 1 mg/ml or 144/4-2/2: 2 mg/ml), diluted 1:200 in 1% (w/v) BSA solution. For fungal labelling, green fluorescent FITC conjugate (goat anti-rabbit IgG, H + L, 2 mg/ml, Molecular Probes, Göttingen, Germany) was used as secondary antibody, diluted 1:200 with 0.1 PBS and applied for 2 h at 22°C. After immunolabeling, the sections were mounted on slides in 2.5% 1.4-diazabicyclo[2.2.2]octane (DABCO) in glycerol. Control sections were incubated without primary antibodies either with 1% (w/v) BSA or with 1:200 diluted rabbit preimmune serum (# 170) and did not show secondary antibody-caused fluorescence. Sections were examined under fluorescent blue light (Leica-filter block I3, excitation 450-490 nm, emission ≤ 515 nm) with the same microscope and camera system as stated above. Further details were described by WÄCHTER et al. (2003). Altogether, up to 5000 sections (stained with dyes or immunologically labelled) were microscopically analyzed.

DAB oxidative stress test
For examining possible plant defence reactions towards the endophytic fungi, the production of H 2 O 2 was investigated following the protocol of THORDAL-CHRISTENSEN et al. (1997). Sections of leaves were taken 4 h to 3 d post Originalarbeit inoculation (hpi, dpi), incubated in 10 mM MES-buffer, pH 6.0 for 2 h and then overnight in 0.1% 3.3-diaminobenzidine-tetrahydrochloride (DAB) in darkness. After removing the DAB-solution, leaf sections were incubated in 100% boiling ethanol for 10 min for chlorophyll extraction, followed by overnight incubation in ethanol at room temperature in darkness. The samples were mounted on slides in 70% glycerol and viewed by transmitted-light bright-field microscopy. About 200 sections were analyzed.

Re-isolation experiments
The fungi used were the entomopathogens B. bassiana ATTC 74040, M. anisopliae JKI-BI-1339 and I. fumosorosea JKI-BI-1496, the oil seed rape pathogens Plenodomus lingam (teleomorph Leptosphaeria maculans) strain T12aD34 and Plenodomus biglobosus (teleomorph Leptosphaeria biglobosa) strain NA22 (formerly regarded as aggressive and non-aggressive variants, respectively, of the blackleg pathogen Phoma lingam) (SHOEMAKER and BRUN, 2001) and a strain of Ascochyta fabae. The fungi were cultured on potato dextrose agar (PDA) (B. bassiana, M. anisopliae, I. fumosorosea, A. fabae) or V8 medium (L. maculans, P. biglobosus). Conidial suspensions (10 7 conidia per ml) were prepared by addition of a few ml of sterile 0.0125% Tween 20 to sporulating plates. Cotyledons of 10-day-old B. napus cv. Adriana plants were punctured with a needle (one wound per cotyledon half) and inoculated with 5 μl drops of suspension on each wound, or mock-inoculated with drops of 0.0125% Tween 20. The plants were then returned to the growth room. Cotyledons were harvested 14 dpi and surface sterilized for 2 min each in 0.5% sodium hypochlorite and 70% ethanol, followed by three rinses in sterile dH 2 O. Water of the last rinse was plated on PDA to check the sterilization effect. Discs of leaf were excised using cork borers with 6, 8 or 10 mm diameter, respectively and placed on PDA with 0.002% Rifampicin + 0.005% Streptomycin (PDA + Antibiotics). The plates were incubated at 23 ± 2°C and regularly inspected for mycelium growing from the tissue. Alternatively, cotyledons were infiltrated with conidial suspension as described above.
Similarly, the first and the second pinnate leaves of 10-days-old V. faba plants were wound-inoculated (two wounds per leaflet) or infiltrated with conidial suspensions of the entomopathogenic fungi or A. fabae. Inoculated plants were returned to the growth room and arranged in a randomized design. In order to delay senescence, fresh shoot growth was removed. Fourteen days after inoculation, samples were taken and processed as explained above for oilseed rape cotyledons.

Detection of fungi by staining with blankophor, solophenyl flavine 7 GFE, and trypan blue
All tested dyes effectively stained the examined entomopathogenic fungi, i.e. strains of B. bassiana, I fumosoro-sea, L. muscarium and M. anisopliae (Fig. 1a-g). Blankophor efficiently bound to the hyphae of all fungi on the leaf surface of B. napus (Fig. 1a). On plant surfaces, hyphae were detected up to four weeks after application of spores. Solophenyl flavine 7 GFE effectively stained B. bassiana spores and hyphae. It yielded the best contrast after pre-staining the samples with safranin, which suppresses plant cell wall fluorescence (Fig. 1b, c, f). Comparable patterns were obtained with M. anisopliae on B. napus leaves (Fig. 1e). Trypan blue also provided a good contrast between spores and hyphae on B. napus leaf surfaces (Fig. 1g). Fig. 1g shows that spores and hyphae of B. bassiana (arrows) were considerably smaller than the width of the stomatal opening.
After infiltration of spores into B. napus leaves, blankophor stained hyphae outside the stomates (Fig. 1h, arrow). In cross sections no hyphae were detected inside the tissue. Safranin/Solophenyl 7GFE-staining of the tissue allowed a view into the spongy parenchyma, as shown for uninoculated control tissue of V. faba (Fig. 1i). Different from B. napus, with this method spores and short hyphae could be detected in the mesophyll of V. faba 2 dpi (Fig. 1j, arrowheads).

Detection of fungi by immunolocalization
The fungi examined, B. bassiana, I. fumosorosea, L. muscarium and M. anisopliae, were specifically detected by the anti-Beauveria as well as the anti-Isaria IgGs (Fig. 2a-d).
Infiltration of the spore suspension from the lower leaf surface into the leaves of B. napus was successful and could be followed visibly. Nevertheless, B. bassiana hyphae were detected only on the lower leaf surface at the site of infiltration ( Fig. 2e-g). Spongy and palisade parenchyma were clearly devoid of hyphae in B. napus (Fig. 2f, g, i), as well as the vascular bundles (Fig. 2h).
B. bassiana, I. fumosorosea and L. muscarium were infiltrated also into the leaves of V. faba and C. sativus. In contrast to B. napus, the spores of B. bassiana germinated in the leaf tissue of faba bean and cucumber, and hyphae were detected growing within the intercellular space along the cell walls ( Fig. 2j-l, 3a, b), except within the vascular bundles, where there are no intercellular spaces (Fig. 2h, j). Hyphae also grew in the mesophyll of C. sativus after inoculation with B. bassiana (Fig. 3c, d). Not only B. bassiana but also I. fumosorosea was detected within the mesophyll of V. faba leaves (Fig. 3e, f).
In samples taken over a period of 4 to 8 weeks after inoculation, less and less fungal hyphae remained detectable on and within bean leaves. Most hyphae had a corroded and starved appearance. Hyphae were neither detected in the main stem of the shoot, nor in the shoot apical meristem (SAM).

DAB staining
To detect H 2 O 2 production triggered by fungal hyphae, segments from inoculated leaves were incubated in diaminobenzidine (DAB). By 4 hpi, brownish hyphae within B. napus stomates indicated the plant production of H 2 O 2 upon contact with B. bassiana hyphae, Originalarbeit resulting in the brown precipitation due to dimerization of DAB (Fig. 4a). The reaction was even stronger 1 dpi ( Fig. 4b-d, arrows). In the non-inoculated control leaves, hence in the absence of fungal hyphae, no brown coloured guard cells of the stomates were detected (Fig. 4e). The reaction of B. napus towards I. fumosorosea and L. muscarium was similar as towards B. bassiana. The response was still detectable 7 dpi (Fig. 4f, g). Up to 14 dpi with B. bassiana, I. fumosorosea or L. muscarium, the guard cells of V. faba showed no browning or the latter was much weaker than in B. napus (Fig. 4h, i, k, arrows). As in B. napus, no brown staining was detected in non-inoculated V. faba leaves (Fig. 4j). Only glands reacted with DAB, confirming the method per se to be effective (Fig. 6i,  arrow).

Staining of B. bassiana in plants after seed coating
In faba bean seeds inoculated with blastospores of B. bassiana and placed on MPA, hyphae were detected up to 25 days after planting the inoculated seeds. Hyphae were seen in thin hand-cut sections in the intercellular space of the young stem by staining with solophenyl flavine 7GFE (Fig. 5b, c, arrows). When these plants were transferred into sterile soil, hyphae of B. bassiana could be observed up to 35 days after transplanting. They proliferated vigorously along the roots and also the shoot at a distance of Histochemical staining of hyphae of Beauveria bassiana and Metarhizium anisopliae, a Beauveria bassiana on Brassica napus leaves, 5 dpi, blankophor and b, c safranin/solophenyl 7GFE, 7 dpi, d Beauveria bassiana, still 28 dpi detectable by blankophor staining, e Metarhizium anisopliae, 3 dpi, blankophor staining and f safranin/solophenyl 7GFE staining, 7 dpi, g Germinating conidia of Beauveria bassiana stained with trypan blue (arrow), 6 dpi, h Cross section of Brassica napus leaf, blankophor staining, 9 dpi, no hyphae within the spongy parenchyma, but outside attached to a guard cell (arrow), i Control leaf of Vicia faba without fungal hyphae, view into the spongy parenchyma, safranin/solophenyl GFE staining after tissue clearing, j Vicia faba leaf with Beauveria bassiana, 2 dpi, spores and small hyphae (arrowheads) are visible in the spongy parenchyma after staining as in i.  Originalarbeit up to about 3 cm from the cotyledons. At or within the distal stem and shoot tissue no hyphae were detected. They were predominantly found on the surface of root ( Fig. 5e-g) and stem, and also within the stem epidermis and root rhizodermis (Fig. 5a) and hypodermis/protective exodermis (Fig. 5d). In particular, good growth was observed around and inside dead root hairs (Fig. 5e, f). The root cortex zone, vascular bundles and pith were free of hyphae (Fig. 5e, g, asterisks).

Re-isolation of fungi from inoculated leaves
After wound inoculation of oilseed rape cotyledons with conidia of the pathogen P. lingam, grey-green, non-sporulating lesions about 5 mm in diameter became visible around 7 dpi, whereas restricted dark brown lesions (1-2 mm diameter) appeared after inoculation with P. biglobosus (corresponding to interaction phenotypes 7 and 2, respectively, according to MENGISTU et al., 1991). The only visible reaction on mock-inoculated cotyledons and on cotyledons inoculated with the entomopathogens was the formation of callus at the wounding sites, accompanied by very limited browning of the tissue in some cases (not shown). In the two experiments performed, the pathogens P. lingam and P. biglobosus could be re-isolated from all inoculation sites. The entomopathogens could also be re-isolated, but at a lower frequency. The time period between placement of the leaves on the agar until appearance of the fungal mycelium was shorter for the phytopathogens than for the entomopathogens (Table 1). Whereas P. lingam could be isolated also from the surrounding tissue, mycelium of P. biglobosus and the entomopathogens developed only at the point of inoculation (Fig. 6).
On V. faba wounding of leaves followed by inoculation with the pathogen, A. fabae resulted in the formation of a necrotic area about 5 mm in diameter. Pycnidia were not formed. Inoculation with the entomopathogens B. bassiana, M. anisopliae and I. fumosorosea led to tissue necrosis at the point of inoculation. Neither callus nor necrosis developed at mock-inoculated wounding sites (Fig. 7).

Originalarbeit
Attempts to re-isolate the fungi 14 days after wound-inoculation were successful in the case of A. fabae as well as for the entomopathogens, although at a lower frequency than from the cotyledons of oilseed rape (data not shown). Two weeks after infiltration of conidial suspensions, both phytopathogens and the three entomopathogens could be successfully re-isolated from all inoculated oilseed rape cotyledons and faba bean leaves (data not shown).

Discussion
The dyes and histological methods used in the present study proved to be reliable for detection of the entomopathogenic fungi B. bassiana, I. fumosorosea, L. muscarium and M. anisopliae. Trypan blue, blankophor and solophenyl 7GFE had been used before for staining fungi belonging to different taxonomic groups. While trypan Root about 1 cm proximal to the seed, 36 dpi, abundant growth of hyphae (arrow) in the rhizosphere, b, c Culture of coated seeds on MPA with hyphae, 25 dpi, in the intercellular space of the seedling shoot (arrows), d-g Culture of coated seeds in sterile soil, 36 dpi, d Stem tissue with hyphae in the epidermis, e Hyphae at and in root hairs (arrow) and rhizodermis (arrowhead), f, g Hyphae only on and in the rhizodermis (arrowhead), but not within the root cortex tissue (asterisks).

Fig. 6.
Re-isolation of fungi from inoculated oilseed rape cotyledons. The cotyledons were punctured with a needle and inoculated with conidial suspensions. After 14 days they were harvested and surface sterilized. Leaf discs were excised with cork borers with different diameters and placed on PDA supplemented with antibiotics, a mock-inoculated, b Plenodomus lingam, c Plenodomus biglobosus, d Beauveria bassiana, e Metarhizium anisopliae, f Isaria fumosorosea.
blue is a classical unspecific dye for light microscopy, the fluorophores blankophor and solophenyl flavine 7GFE are known to bind to polysaccharides (RÜCHEL and SCHAFFRINSKI, 1999;WALLACE and ANDERSON, 2012). The method of combining solophenyl flavine 7GFE with pre-staining with safranin to suppress plant cell wall fluorescence (KNIGHT and SUTHERLAND, 2011) proved efficient also in the present experiments. To detect specific endophytes, specific antibodies were raised. Immunofluorescence microscopy with primary polyclonal and secondary FITC-conjugated antibodies also provided excellent detection of hyphae and spores of all four entomopathogens. Due to cross-reactivity, both antibodies (i.e. anti-B. bassiana and anti-I. fumosorosea) detected all four entomopathogens which belong to the order Hypocreales. The cross-reactivity of both purified rabbit IgGs was confirmed by plate trapped antigen ELISA (PTA-ELISA) (ROHDE and RABENSTEIN, 2005) and western blotting experiments (WB). No reaction was observed with bovine serum albumin as negative control in PTA-ELISA. The observed immunological cross-reactivity of antibodies seem to be common in fungus serology (KAUFMAN and STANDARD, 1987;NOTERMANS et al., 1998) including the plant invading fungi (DEWEY, 2002) and suggests that several fungi share common antigenic determinants. Cross-reactivity may occur with related fungi but also, to a lesser extent, with non-related fungi (SCHMECHEL et al., 2006;THORNTON and WILLS, 2015). There was no labelling of e.g. powdery mildew (data not shown).
Using light and electron microscopy, WAGNER and LEWIS (2000) observed germination of the conidia of B. bassiana on corn leaves. In some cases, hyphae developing from the spores penetrated the cuticle directly or, at a lower rate, grew through stomates into the leaves, where they branched and followed the leaf apoplast. The present studies indicated that stomates are appropriate openings for entry of spores and hyphae of all tested entomopathogenic fungi also in case of B. napus, V. faba, and C. sativus as the size of conidia and blastospores of the entomopathogens is significantly lower compared to the size of the stomatal openings of the tested plants.
When spores of B. bassiana were infiltrated into B. napus leaves, they appeared to germinate only on the leaf surface but not within the leaf. Nevertheless, no systemic fungal growth was observed in leaves, stems, vascular bundles or shoot apical meristems of B. napus. This was not only valid for B. bassiana, but as well for I. fumosorosea, L. muscarium and M. anisopliae and for all B. napus cultivars tested.
The experiments described here were performed with conidiospores of M. anisopliae, I. fumosorosea, L. muscarium and with blastospores or conidia in the case of Originalarbeit B. bassiana. Blastospores of entomopathogenic fungi are formed in liquid media as described in detail e.g. by KLEESPIES (1993), KLEESPIES and ZIMMERMANN (1994), KASSA et al. (2004). They generally germinate fast and at a high rate and were therefore used in the experiments dealing with the histology of the infection process. However, in a separate study the development of B. bassiana on plants was identical irrespective whether blastospores or conidia were employed as inocula (unpublished results). Based on this, conidiospores were used in the present re-isolation experiments, and it is assumed that the results of the microscopical examination can be related to those of the re-isolation experiment although different spore types were used for inoculation.
The lack of systemic hyphal growth observed in the microscope studies was in agreement with the re-isolation experiments. These showed that the entomopathogens did not grow away from the point of inoculation. In this respect, they behaved differently from the pathogen P. lingam (teleomorph: L. maculans), which spread out into the adjacent tissue, and similar to P. biglobosus which was previously regarded as an avirulent variant of Phoma lingam. The re-isolation experiment further illustrated that the entomopathogens were able to persist in the tissue for at least 2 weeks, irrespective of whether they had been inoculated into wounds or infiltrated into the tissue. Since spreading fungal hyphae were not detected, the fungi must have persisted in the tissue as ungerminated spores.
The observed lack of fungal growth in B. napus leaves is somewhat conflicting with other reports describing Brassicaceae and B. napus as hosts of various fungal endophytes (ZHANG et al., 2014;CARD et al., 2015). From re-isolation experiments, VIDAL and JABER (2015) concluded that 12 isolates of B. bassiana (including isolate ATCC 74040 used in the present study) were able to colonize B. napus and V. faba leaves; two isolates could not be re-isolated in their study. Different from B. napus, intercellular hyphal growth of B. bassiana, I. fumosorosea and L. muscarium was observed in V. faba and cucumber leaves in the present study, best documented by detection with antibodies and enhancement by secondary FITCconjugated antibodies. However, growth appeared to be transitional and restricted to the inoculated leaf area. When seeds of V. faba were coated with B. bassiana and placed on MPA, the parts of the shoot and root adjacent to the cotyledons became colonized with hyphae, and the latter were still present up to 35 days after transplanting into sterile soil. However, the observed hyphal growth may have been supported by the availability of nutrients from the cotyledons and may thus not have been truly endophytic. This would indicate that nutrient availability is indeed an important factor for endophytic establishment. Unequivocal detection of fungal hyphae by PCR methods may be problematic due to DNA from inocula remaining on the plant surface even after sterilization (MCKINNON et al., 2017). This is supported by own observations that in spite of repeated (more than 35) washing steps with paraformaldehyde, PBS, MTSB, various ethanol concen-trations and methanol, fungal spores and hyphae still remained adhered on the leaf surface (Fig. 2 e, f). Similar results have been reported for bacterial adherence on plant surfaces after disinfection or washing treatments (FORNEFELD et al., 2015).
A rapid defence mechanism of plant cells is the production of reactive oxygen species (ROS), such as H 2 O 2 . Hydrogen peroxide is assumed to have various roles in plant defence reactions against pathogens (LAMB and DIXON, 1997) but has also been demonstrated in host-endophyte interactions (SCHULZ and BOYLE, 2005). In the present study, browning of plant cells in the presence of fungal hyphae after treatment with DAB was much more pronounced in B. napus than in V. faba and cucumber. The response occurred already within four hours after inoculation. This indicates a direct antimicrobial activity of H 2 O 2 produced by plant cells (PENG and KUC, 1992).
To summarize, it was not possible to observe endophytic establishment of the tested entomopathogens in oilseed rape leaves, and in faba bean and cucumber colonization was restricted to the area of inoculation. We suggest this scarce colonization to be due to defence reactions like production of H 2 O 2 or formation of glucosinolates (BEDNAREK et al., 2009), inability of the fungi to acquire nutrients in the apoplast, or a combination of both. Our results are in agreement with the general statement that endophytic colonisation by non-balansiaceous endophytes of above-ground organs in many interactions remains limited (SCHULZ and BOYLE, 2005). However, in the case of B. bassiana a number of reports not only suggest movement within the plant (OWNLEY et al., 2008) and from soil and seeds into stems and leaves (GURULINGAPPA et al., 2010) but also vertical transmission via seeds (QUESADA-MORAGA et al., 2014). Endophytic colonization by B. bassiana is known to depend on the inoculation method, fungal isolate and plant species (RUSSO et al., 2015), however, the detection method used may also be important as was recently emphasized by MCKINNON et al. (2017). From the present results it is concluded that wherever possible data on endophytic colonization should be supported by histological evidence of the kind and amount of fungal growth in the host tissue.
Originalarbeit and theoretical struggle of advanced students and young scientists at the JKI Institute for Biological Control in Darmstadt.