Plant root carbohydrates affect growth behaviour of endophytic microfungi


  • Franz Hadacek,

    Corresponding author
    1. Department of Comparative and Ecological Phytochemistry, Institute of Botany, University of Vienna, Rennweg 14, A-1030 Vienna, Austria
    Search for more papers by this author
  • Günther F. Kraus

    1. Department of Comparative and Ecological Phytochemistry, Institute of Botany, University of Vienna, Rennweg 14, A-1030 Vienna, Austria
    Search for more papers by this author

*Corresponding author. Fax: +43 (1) 4277 9541.


Peucedanum alsaticum and Peucedanum cervaria represent characteristic umbellifers (Apiaceae) of calcareous grasslands in Central and Eastern Europe. Both accumulate glucose, fructose, mannitol and sucrose as dominant carbohydrates in their roots. The objective of the study was to determine if endophytes utilise host plant carbohydrates differently than rhizosphere and bulk soil microfungi. Inula ensifolia (Asteraceae), Lathyrus latifolius (Fabaceae) and Bromus erectus (Poaceae), all plants that grow along with the two umbellifers, accumulated only sucrose as major sugar in their roots. Germ tube growth of 30 microfungal isolates, recovered from various rhizosphere habitats, was quantified in microdilutions (10–5000 μg ml−1) of a number of substrates, including glucose, sucrose, mannitol and the water-soluble plant root carbohydrate mixtures. Multivariate analysis of variance and subsequent least significant difference analysis of isolate group means revealed that sucrose and mannitol utilisation affected affiliation to a certain fungal lifestyle. Endophytes utilised host plant carbohydrates more efficiently at lower concentrations. Conversely, higher concentrations slowed their growth. Non-host carbohydrates did not cause comparable effects. The results suggest that root carbohydrate diversity may determine fungal diversity in natural rhizosphere environments.


Sloughed-off cells, mucilage and exudates from plant roots represent essential carbon sources for soil microbes such as fungi and bacteria. Root exudates are plant-specific and their composition may also be influenced by physiological conditions or abiotic factors [1–4]. Roots of intact plants harbour an extensive diversity of fungi: arbuscular mycorrhizal fungi of the order Glomales occur in roots of nearly all plants; various asco- and basidiomycetes form an ectomycorrhizal mantle of hyphae around the roots of trees and shrubs; and endophytes comprise asexually reproducing fungi that spend the whole or greater part of their life cycle inside healthy plant tissues but cause no obvious disease symptoms [5–7].

Plant biodiversity is generally considered an important factor that contributes to the functioning and stability of terrestrial ecosystems [8,9]. Microbial communities in the rhizosphere may affect plant communities [10,11] and, conversely, plants may influence microbial communities in their rhizosphere [12,13]. These findings have provided a first step in understanding mechanisms that regulate structure and function of natural plant–microbe communities [14]. Association of fungi with plants dates far back to the arrival of vascular plants [15].

Many studies focus on mycorrhizal fungi as crucial components in the rhizosphere [16]. Endophytes that are the focus of this study are necrotrophic microfungi. The majority of them are common soil fungi that can be found in the rhizosphere as well as in the bulk soil. Such microfungi do not only constitute a big portion of fungal biodiversity in soils, a number of them may develop into perthotrophic pathogens that cause severe damage to crop plants and dramatically reduce yields [17]. In natural systems, pathogen outbreaks do not compare, but still are known to occur [18]. Thus, the diversity of their lifestyles provides us with a unique scenario to study mechanisms involved in adaptation to host plants.

Conventional isolation techniques prevent us from recovering the whole spectrum of fungal diversity from plant tissues [19]. Anamorph microfungi that can be recovered by soil-washing include genera such as Acremonium, Cylindrocarpon or Penicillium. Without too much effort, these strains can also be maintained on artificial media. This provides us with a spectrum of test organisms that allows us to identify possible low molecular signal molecules involved in adaptation towards a lifestyle of an endophyte.

Plant compounds comprise primary metabolites that fungi utilise as nutrients, such as sugars and amino acids [4]. The major portion of chemical biodiversity in plants, however, is created by secondary metabolism. Plant roots also accumulate and may also exude numerous terpenoids, coumarins, flavonoids or polyacetylenes, just to name a few of them. The majority of authors concur that the main function of secondary metabolites is to contribute to the plant's survival by acting as chemical defence against microbial and herbivore predators [20,21]. Independently, plant metabolites may also adopt a further function: as an attractive or behaviour-changing signal to various micro-organisms or animals [22,23].

Peucedanum alsaticum and Peucedanum cervaria represent typical umbellifers of calcareous grasslands in Central and Eastern Europe [24]. Preliminary experiments indicated that, despite the structural diversity of secondary metabolites present in the roots of these two umbellifers, inhibition caused by non-water-soluble secondary metabolites was more uniform than growth stimulation by water-soluble carbohydrates. Thus we focused the role of the water-soluble carbohydrates in the adaptation of microfungi to our model plants, P. alsaticum and P. cervaria. For comparative purposes, Bromus erectus, the dominant grass species, pea Lathyrus latifolius and composite Inula ensifolia were also included. No efficient in situ method for capture, identification and quantification of root exudates exists [25]. In an effort to make our experiments practicable and detect differences in utilisation spectra as efficiently as possible, carbohydrate mixtures isolated directly from the plant roots and selected major components of those extracts were offered in concentration gradients to selected microfungal isolates recovered from bulk soil, rhizosphere soil and root tissues of the two Peucedanum species.

2Materials and methods

2.1Plant material

Eichkogel (367 m asl; long. 16°17′, lat. 48°03′) is a small hill located near the town of Mödling, south of Vienna, Austria, and still carries more or less undisturbed grassland areas. Tap root material of the two umbellifers, P. alsaticum L. and P. cervaria (L.) Lapeyr. was uncovered to about 30 cm below the soil surface and coarsely freed from surrounding soil. Similarly, underground organs of the dominant grass species, B. erectus Huds. (Poaceae) and two other nearby growing herbs, I. ensifolia L. (Asteraceae) and L. latifolius L. (Fabaceae), were also collected. Voucher specimens (P13278BIO-01–P13278BIO-05) were deposited at the herbarium of the Institute of Botany, University of Vienna (WU).

2.2Extraction of water-soluble root carbohydrates

Ground root material (10–20 g) was exhaustively extracted with methanol (MeOH) at room temperature for 3 days. The whole crude extract was evaporated to a watery residue on a rotary evaporator, dissolved in 100 ml deionised water, and twice extracted with approximately 300 ml chloroform to remove lipophilic compounds. The water-soluble phase was then evaporated to about 20 ml and stored at 4°C for further use.

2.3High performance liquid chromatography (HPLC)–UV analyses of the water-soluble root extract fraction

Portions of the extracts were evaporated to dryness, dissolved in methanol and adjusted to a concentration of 10 mg ml−1, of which 10 μl was injected. A Hewlett Packard 1090 Series II HPLC equipped with an UV diode array detector was used. The column was a C18 Hypersil BDS (250×4.6 mm, 5 μm), the column oven was thermostatted at 40°C; binary solvent gradient started at 100% aqueous buffer (o-phosphoric acid (Merck) 0.015 mol, tetrabutylammonium hydroxide (Sigma) 0.0015 mol, pH 3) and rose to 40% acetonitrile (Merck, gradient grade) in 20 min, flow rate was 1 ml min−1; signal wavelength was 230 nm, UV spectra were recorded from 220 to 450 nm.

2.4Gas chromatography coupled with mass spectrometry (GC–MS) analyses of the water-soluble extract fraction

Water-soluble extract fractions were derivatised into their trimethylsilyl (TMS)-oxime ether/esters as described [26]. A Perkin Elmer GC AutoSystem gas chromatograph (GC) coupled to a Perkin Elmer Turbomass quadrupole mass spectrometer (MS) was used for analysis. The column was a Perkin Elmer PE-5ms (20 m×0.18 mm, 0.18 μm film thickness); the GC oven was programmed to stay at 60°C for 2 min, to rise to 100°C at a rate of 20°C min−1, to rise to 250°C at a rate of 5°C min−1, to rise to 330°C at a rate of 8°C min−1, and to hold 330°C for 4 min; helium was used as carrier gas at a constant flow rate of 1 ml min−1; the transfer line to the MS was thermostatted at 280°C and the ion source at 180°C. MS spectra were recorded from 30 to 620 amu per 0.6 s in the electron ionisation mode at an electron energy of 70 eV and a filament emission of 200 μA. One μl of sample was injected in the split mode, the split flow was adjusted to 10 and 30 ml min−1, depending on analyte concentration. Temperature of the injector was programmed to stay at 60°C for 2 min, and then to rise to 320°C at a rate of 180°C min−1. Identification of sugar, sugar alcohol, organic acid and amino acid trimethylsilyl-oxime ether/esters was achieved (1) by comparison with MS spectra and retention times of authentic standards, (2) tentatively, by comparison of mass spectra to the Wiley Registry of Mass Spectral Data, 6th edition or the NIST/EPA/NIH Mass Spectral Library 1.5a, and to a MS library and retention time database of TMS compounds offered for free download from the web site of the Max Planck Institute of Molecular Plant Physiology, Potsdam, Germany for metabolic profiling [27].

2.5Fungal isolates

Recovery of microfungi from soil, rhizosphere and root tissue was carried out several times from April to November in two consecutive years. Root pieces (ca. 5 cm long and 0.5–1.5 cm in diameter) were transferred into glass vials containing 1% (w/v) sodium chloride solution. Soil samples were taken during the digging up of roots; 1–2 g of bulk soil was transferred into glass vials. Processing of the samples on the same day turned out to be mandatory. Soil fungi were isolated by the soil-plating [28] and soil-washing technique [29,30]. The latter method was modified by using sieve sizes of 0.8, 0.4, 0.2 and 0.1 mm. Mineral and organic particles retained on the smallest-sized sieves (0.2 and 0.5 mm) were transferred to 2% malt extract (Merck) and corn meal (Difco) agar-plates. All media used for isolation purposes contained chloramphenicol (Merck) or validamycin (Solacol®, Aventis Crop Science) as antibiotic additives (0.1 g ml−1) [31]. Rhizosphere soil was recovered by vigorously shaking roots in 1% (w/v) sodium chloride solution for 15 min and subjected to the same isolation procedures. Root tissue was thoroughly washed in tap water to remove soil and mucilage as efficiently as possible. Standard surface sterilisation was performed by treatment with 70% ethanol and subsequent flaming [32,33]. Root pieces (5–10×2–3 mm) were transferred to agar plates containing medium identical to that used for isolation of soil microfungi.

Conidia and hyphae of pure isolates were suspended in an aqueous solution of sucrose (140 g l−1) and peptone (10 g l−1) and stored at −80°C. Transfer to freezing temperatures was performed by decreasing temperature not faster than 1°C min−1. For documentation and further use, all isolated strains were deposited in the culture collection of the Institute of Applied Microbiology of the University of Agricultural Sciences (VIAM), Vienna, Austria. Identification was carried out based on morphological characters [34,35].

2.6Conidia suspensions

Petri dishes containing 2% malt extract agar (20 ml) were three-point inoculated with cryoconserved strains and grown for 2–4 weeks at room temperature in the dark. Conidia were harvested by suspending in 1% (w/v) sodium chloride solution containing 5% (v/v) dimethylsulfoxide. Conidia suspension was filtered (pore width 53 or 105 μm, depending on conidia size), concentrated by centrifuging (4000×g) and transferred into 1.8 ml cryotubes and stored at −20°C. Colony forming units (CFU) were determined by plating 20 μl of 10× serially diluted suspensions and counting germinated spores/dilution concentration. Procedures were replicated three times for each suspension.

2.7Germ tube growth assays

The liquid medium was an aqueous solution of sodium nitrate (3.0 g l−1), magnesium sulfate (0.5 g l−1), potassium chloride (0.5 g l−1), iron(III) sulfate (0.01 g l−1) and dipotassium hydrogen phosphate (1.0 g l−1). Sodium nitrate constituted the sole nitrogen source; variable carbon sources were offered by adding pure compounds or root extract fractions. Stock solutions were prepared using this medium containing the test compound or extract at a concentration of 10 mg ml−1 and filtered over sterile filtration units (Nalgene). Assays were carried out in sterile, disposable microtitre plates (Greiner; 96 U-bottomed wells). For result comparability, glucose, sucrose and mannitol were assayed together with the water-soluble root extract fractions within the same plate. Stock solutions were serially diluted. Stock inoculum suspension was adjusted to 105 CFU ml−1. Final concentrations ranged from 10 to 5000 μg ml−1 for each compound/extract. Microtitre plates were incubated at room temperature until growth was visible under the microscope (magnification 140×); plates were regularly checked after 24, 48 and 72 h. To stop germ tube growth and enhance contrast for scoring, 10-μl aliquots of lactophenol blue (Merck) were dispensed into each well. To quantify germ tube growth, a representative section of the well covered with germ tubes was captured to hard disk. Germ tube density was determined by image analysis (Scion Image 3b for Windows [36]) in a representative 150×150-pixel sector of the digitised image. Within the identical microtitre plate, the well showing the highest pixel number was designated as 100%. Pixel numbers obtained from other wells were expressed as percentage of that value, respectively. The 100% value represents also the maximum growth performance of an isolate within a given time period. Replicate experiments were carried out and scored visually to confirm previously observed growth behaviour.

2.8Statistical analysis

Statistical analyses were carried out using SPSS 10.0.5 for Windows [37]. Growth rates obtained from different concentrations in the dilution series represented repeated measurements that may be analysed by one-factor-repeated-measures analysis of variance (ANOVA). However, this type of ANOVA assumes sphericity of variable variances, and present data violate this assumption. Instead, a multivariate analysis of variance (MANOVA) is used that treats the concentration series as multiple independent variables [38]. Assumptions of normal distribution and equality of error variances for variables were fulfiled (one-sample Kolmogorov–Smirnov test for normal distribution and Levene's test of equality of error variances).


3.1Chemical analyses of water-soluble root components

Water-soluble root extract fractions of B. erectus, I. ensifolia, L. latifolius and P. alsaticum amounted to at least two thirds of the whole crude MeOH extract. Water-soluble extract fractions were analysed by HPLC–UV and GC–MS. UV spectra of peaks present in the HPLC chromatogram indicated presence of glycosides of various natural products, e.g. cinnamic acid derivatives. GC–MS chromatograms, however, suggested that their amounts were negligible in comparison to sugars and sugar alcohols. In selected cases, e.g. for caffeic and chlorogenic acid, detectability by GC–MS was additionally confirmed by analyses of derivatised authentic compounds. Consequently, quantification of compounds present in the water-soluble extract fraction was carried out on the basis of results exclusively obtained by GC–MS analysis. Table 1 lists the composition of sugars, sugar alcohols, organic and amino acids that were identified in the extracts offered in the growth assays. Observed accumulation trends were also confirmed by analysis of additional individuals. The two umbellifers, P. alsaticum and P. cervaria, showed different patterns of carbohydrates compared to other investigated plants: mannitol, glucose and fructose were accumulated in similar amounts as sucrose. This disaccharide constituted the main sugar derivative (>70%) in the water-soluble carbohydrate fractions of B. erectus, I. ensifolia and L. latifolius. Besides malic acid, the latter plant also contained amino acids in detectable amounts, of which homoserine and pyroglutamic acid were tentatively identified. MS spectra of all unknowns are listed in the footnotes to Table 1.

Table 1.  GC–MS analysis of the trimethylsilyl/trimethylsilyl-oxime ethers/esters of organic acids, sugars, sugar alcohols and amino acids from the water-soluble fraction of the methanolic root extract
  1. Relative quantities have been calculated from integrated peak areas of the TIC signal; amounts of single compounds represent percentages of the total area of all peaks detected; identified compounds detected in amounts lower than 0.1% are designated as trace (tr) and compounds present in amounts larger than 1% are marked in bold.

  2. aMS [m/z(%)]: 70(29), 74(29), 130(31), 203(100), 204(33), 275(89), 292(56).

  3. bMS [m/z(%)]: 32(100), 45(38), 56(26), 73(99), 75(92), 103(68), 130(22), 146(79), 248(1).

  4. cMS [m/z(%)]: 43(37), 56(100), 73(17), 75(24), 116(5), 130(5).

  5. dMS [m/z(%)]: 73(100), 75(47), 77(26), 103(11), 147(34), 212(17), 254(9), 344(6).

  6. eTentative identification by retention time and MS spectrum comparison with libraries and literature.

  7. fMS [m/z(%)]: 43(76), 76(100), 75(64), 77(19), 117(12), 133(14), 152(7), 212(11), 254(22).

  8. gMS [m/z(%)]: 43(30), 73(63), 75(34), 100(7), 128(100), 147(10), 188(9), 202(6).

  9. hMS [m/z(%)]: 43(47), 73(100), 75(68), 103(23), 147(23), 158(54), 188(21), 229(13), 290(5).

  10. iIdentification by comparison of retention times and mass spectra of authentic compounds.

  11. jMS [m/z(%)]: 73(100), 75(76), 103(21), 147(25), 217(14), 259(3).

  12. kDerivatisation yielded more than one peak in the GC chromatogram.

Retention time B. erectusI. ensifoliaL. latifoliusP. alsaticumP. cervaria
5.47unknowna 5.50.5 13.5
6.65unknownb  1.5  
7.31unknownc  0.3  
7.93unknownd  0.4  
8.08homoserinee  1.1  
8.19unknownf  0.2  
8.85malic acide  1.0  
8.96unknowng  0.4  
9.40pyroglutamic acide  0.2  
11.44unknownh  0.5  
14.24arabinoseitr tr  
15.18citric acide0.4 0.2  
15.74unknownj  0.2  
16.37mannosee    1.3
17.21sorbitole0.1   1.7
18.15quebrachitole  2.4  
18.17kgalactosei0.2  1.0 
19.74inositole0.1 0.1  
31.35kmaltosee  0.1  

3.2Selection of fungal isolates from root tissues, rhizosphere and bulk soil

The total number of isolates recovered from the two Peucedanum species were too numerous to be included into this study. Consequently, a number of strains were chosen as representatives of three distinct isolate groups. The three distinguished isolate groups comprised endophytic, rhizosphere soil and bulk soil microfungi (Table 2). Of those identified, C. destructans was the only species occurring ubiquitously, also within root tissues of both Peucedanum species. Penicillium species were recovered as endophytes, rhizosphere soil and bulk soil fungi. Endophytes of this genus, of which we have been able to identify only one isolate so far, were different from those species recovered from rhizosphere and bulk soil. Moreover, morphologically, all four isolates seemed to belong to separate species. Apart from identified isolates, a big number of sterile mycelia also occurred in root tissues. The percentage of such isolates was notably higher among endophytic isolates than rhizosphere soil and bulk soil isolates. However, conidia and/or spores constituted an indispensable prerequisite for germ tube growth assays. Rhizosphere soil contained the greatest diversity of genera and species. Altogether, eight endophytes, 15 rhizosphere soil isolates and seven isolates from bulk soil were chosen for the bioassays.

Table 2.  Microfungal isolates from bulk soil, rhizosphere soil and root tissues of two umbellifers from calcareous grasslands, P. alsaticum L. and P. cervaria (L.) Lapeyr
  1. Strain numbers, assay duration and maximum pixel number determined by image analysis of germ tubes present in a 150×150-pixel section of that well within the microtitre plate that showed maximum growth within the given growth period. All fungal strains have been deposited at the culture collection of the Institute of Applied Microbiology, University of Agricultural Sciences, Vienna (VIAM).

  2. aRecovered by soil-washing.

  3. bRecovered from surface-sterilised root tissue.

  4. cRecovered by soil-plating.

IsolateConserved strainGrowth duration (h)100% growth (pixels)
P. alsaticum
Rhizosphere isolates
Acremonium strictum W. GamsaVIAM-MA23502428 184
Doratomyces stemonitis (Pers. ex Steud.) Morton and G. Sm.aVIAM-MA28084824 986
Fusarium sp.aVIAM-MA10842425 641
Penicillium citrinum ThombVIAM-MA23282417 992
Penicillium manginii Duché and HeimaVIAM-MA20142411 560
P. manginii Duché and HeimaVIAM-MA20162417 262
Penicillium vulpinum (Cooke and Massee) Seifert and SampsonaVIAM-MA28102418 394
Cylindrocarpon destructans (Zinssm.) ScholtenbVIAM-MA34142413 113
C. destructans (Zinssm.) ScholtenbVIAM-MA34152412 151
Penicillium sp.bVIAM-MA33272413 868
Penicillium sp.bVIAM.MA3450243 349
P. cervaria
Rhizosphere isolates
Acremonium murorum (Corda) W. GamsaVIAM-MA28174837 124
Arthrinium arundinis StateaVIAM-MA28142434 335
Beauveria bassiana (Bals.) Vuill.aVIAM-MA28154834 771
C. destructans (Zinssm.) ScholtenaVIAM-MA10594836 011
Fusarium tricinctum (Corda) Sacc.aVIAM-MA27762427 033
Paecilomyces marquandii (Massee) HughesaVIAM-MA19654828 244
Penicillium expansum Link ex GrayaVIAM-MA28112421 412
Trichoderma koningii Oudem.aVIAM-MA10914829 451
C. destructans (Zinssm.) ScholtenbVIAM-MA28194830 833
C. destructans (Zinssm.) ScholtenbVIAM-MA28254830 893
Penicillium sp.bVIAM-MA2824249 055
Talaromyces flavus (Klöcker) Stolk and Samsonb (=Penicillium vermiculatum Dangeard)bVIAM-MA34277242 501
Bulk soil isolates
A. murorum (Corda) W. GamscVIAM-MA19377223 417
Aspergillus ochraceus K. Wilh.cVIAM-MA19454830 500
C. destructans (Zinssm.) ScholtenaVIAM-MA19592415 278
D. stemonitis (Pers. ex Steud.) Morton and G. Sm.aVIAM-MA19874834 455
Penicillium corylophilum DierckxcVIAM-MA19702432 776
P. expansum Link ex GraycVIAM-MA19692422 490
Phoma sp.aVIAM-MA19812417 699

3.3Germ tube growth assays to determine utilisation of various carbohydrate sources

At first impression, fungal isolates notably differed in developing variable-sized germ tubes on the various substrates. In general, growth was more vigorous on plant-derived carbohydrates than on pure compounds. A few isolates failed to germinate at lower concentrations. MANOVA was carried out to determine if utilisation of a particular substrate corresponds to affiliation to a particular isolate group (factor lifestyle) or to association with a particular plant species (factor plant). Table 3 lists the tested substrates by increasing Wilks’λ values of the factor lifestyle. Sucrose and mannitol showed the lowest Wilks’λ values that were also significant (P<0.05). The other substrates only showed non-significant Wilks’λ values for the same factor. Among those, the substrate with lowest Wilks’λ was glucose, followed by the water-soluble carbohydrates of P. alsaticum and P. cervaria, the two rhizosphere hosts. The carbohydrate fractions of B. erectus, I. ensifolia and L. latifolius, which represented those plants growing along with the two focussed umbellifers, showed the highest Wilks’λ and P values. For the other factor, plant, no significant Wilks’λ was obtained, and values were generally higher.

Table 3.  MANOVA to test effects of isolate origin (plant: P. alsaticum, P. cervaria, bulk soil) and isolate lifestyle (endophyte, rhizosphere soil, bulk soil) on growth rates on various substrates
  1. Plant-derived carbon sources comprise water-soluble constituents of the methanolic root extract. Significant factors are marked with an asterisk.

Carbon sourceWilks’λFHypoth. dfError dfP
P. alsaticum
P. cervaria
B. erectus
I. ensifolia
L. latifolius

A subsequent pair-wise comparison of least significant differences of group means for the factor lifestyle aimed to explore the nature of existing group differences (Table 4). On the whole, significant pair-wise differences of group means were only detected between endophytes on one hand and rhizosphere and bulk soil isolates on the other hand. Endophytes grew significantly better than rhizosphere soil fungi at 39 and 78 μg ml−1 sucrose, 39 μg ml−1 of glucose, 625 μg ml−1 mannitol and 20 μg ml−1 carbohydrates from B. erectus. Neighbouring concentrations also showed low P values but fell short of the 0.05 significance level due to large standard errors caused by isolate variability. The water-soluble carbohydrates of P. alsaticum reduced the growth of endophytes in the range of 625–5000 μg ml−1. The data of P. cervaria suggested a similar effect. However, differences between group means were not significant. Instead, endophytes grew better at 10 and 20 μg ml−1. The water-soluble carbohydrates of I. ensifolia and L. latifolius did not affect growth of endophytes discernibly different compared to rhizosphere soil and bulk soil isolates.

Table 4.  Pair-wise comparisons of group means of endophyte (E, n=8), rhizosphere soil (R, n=15) and bulk soil (S, n=7) isolates grown in a dilution series of concentrations on eight different carbon sources
  1. Differences significant at the 0.05% level are marked with an asterisk. Data represent the least significant difference post hoc analysis of the factor lifestyle in the MANOVA shown in Table 3.

Concentration (μg ml−1)IJ Mean difference I−JStandard errorP Mean difference I−JStandard errorP Mean difference I−JStandard errorP Mean difference I−JStandard errorP
5000ERglucose4.838.560.58mannitol−7.4110.010.47B. erectus−5.267.580.49I. ensifolia0.77210.80.94
  S 7.3410.110.47 −6.3511.820.60 −0.378.950.97 −1.8712.780.88
2500ER 2.197.830.78 11.149.520.25 −4.397.910.58
  S 0.939.250.92 8.6711.250.45 11.699.340.22 10.2010.870.36
1250ER −0.577.480.94 12.738.630.15 7.159.440.46 −1.488.290.86
  S −0.908.830.92 14.3710.190.17 15.8411.150.17 1.749.790.86
625ER 7.306.900.30 20.00*9.310.04 −1.319.490.89 15.579.350.11
  S 14.5011.000.20 12.9211.220.26 17.4011.050.13
312ER 2.656.560.69 10.127.750.20 10.369.830.30 10.2810.840.35
  S 1.667.750.83 12.449.150.19 14.8011.610.21 7.7912.810.55
156ER 1.155.870.85 8.387.140.25 9.539.950.35 6.008.970.51
  S 0.676.930.92 7.968.430.35 10.3811.760.39 5.3110.600.62
78ER 6.926.120.27 4.797.300.52 11.519.300.23 5.699.070.54
  S 3.747.230.61 5.838.620.51 11.2910.980.31 −1.0510.720.92
39ER 12.89*5.600.03 7.299.360.44 4.077.940.61
  S 6.286.610.35 8.258.530.34 13.3011.050.24 0.729.380.94
20ER 11.235.650.06 15.67*6.450.02 4.856.870.49
  S 8.496.680.22 1.657.300.82 14.61*7.620.07 2.928.110.35
10ER 6.195.650.28 9.915.620.09 −0.086.470.99
  S 4.967.230.50 7.016.670.30 5.786.640.39 3.977.640.61
5000ERsucrose4.778.400.58P. alsaticum−22.84*5.680.00P. cervaria−16.818.770.07L. latifolius−1.888.330.82
  S 10.219.920.31 −17.50*6.710.02 −16.3610.360.13 −4.139.840.68
2500ER 3.467.710.66 −20.56*8.640.03 −8.0110.460.45 −9.8010.250.35
  S 5.909.110.52 −18.1110.210.09 −6.7012.360.59 −5.0412.110.68
1250ER −22.01*8.070.01 −3.3511.830.78 −3.648.700.68
  S 4.359.470.65 −19.559.530.05 −5.6313.970.69 −1.4710.270.89
625ER −20.39*9.160.04 −6.7110.710.54 −2.159.300.82
  S 14.178.580.11 −13.0610.820.24 −4.1512.650.75 5.6410.980.61
312ER 7.387.910.36 −15.3210.370.15 1.899.370.84 4.73−7.790.72
  S 10.189.350.29 −12.8712.250.30 −1.0111.060.93 6.854.730.81
156ER −12.079.360.21 −1.877.830.81 3.539.540.71
  S 10.088.470.25 −11.3311.050.32 − 8.0311.270.48
78ER 14.59*6.460.03 −4.799.600.62 1.627.920.84 7.988.170.34
  S 12.967.630.10 3.2611.340.78 −4.859.350.61 7.839.650.42
39ER 13.47*5.730.03 −4.008.450.64 3.966.440.54 3.157.740.69
  S 7.796.770.26 2.239.980.82 0.917.600.91
20ER 7.795.570.17 10.18*4.790.04 6.256.480.34
  S 3.066.580.65 2.527.380.74 8.265.660.16 8.007.660.31
10ER 5.774.880.25 −3.775.220.48 13.11*4.980.01 2.865.850.63
  S 1.005.770.86 −2.876.160.64 5.786.640.39 1.746.910.80


Multivariate analyses of variance and subsequent least significant difference analyses of group means of fungi recovered from different rhizosphere habitats suggested that plant sugars or sugar alcohols may constitute signals that facilitate adaptation of certain fungi to a specific host plant. Various results support this interpretation: (1) affiliation to a particular isolate group was most notably reflected by growth patterns on those substrates that contained a single sugar or sugar alcohol (MANOVA); (2) carbohydrates of the two host plants, P. alsaticum and P. cervaria, yielded the lower Wilks’λ values with lower P values than non-host plants; and (3) at various concentrations, extracts of host plants changed growth behaviour of endophytes.

The two umbellifers, P. alsaticum and P. cervaria, differed from other plants growing around them by accumulating large amounts of fructose, glucose and mannitol in addition to sucrose in their roots (Table 1). However, sucrose was also detected as the major sugar derivative in the roots of B. erectus, I. ensifolia and L. latifolius, despite the fact that these three plant species are classified in quite heterogeneous plant families. Recent investigations demonstrated that external sugar concentrations may affect gene expression in ectomycorrhizal basidiomycetes [39]. The present results also suggest that sugars or sugar alcohols may constitute important signals to soil fungi. Higher plants are also capable of sensing changing levels of sugars, especially of glucose and sucrose [23,40]. The occurrence of sucrose in the roots of diverse plants might turn efficient utilisation of this monosaccharide into a general constraint for the endophytic lifestyle of a fungus.

Low concentrations of sucrose stimulated the growth of endophytes. Raw data suggested a similar effect for glucose. However, significance closely missed the 0.05% level. Interestingly, the sugar alcohol mannitol also stimulated growth of the endophytic isolates, but only at higher concentrations. High amounts of mannitol were only detected in the two umbellifers, the host plants of the tested endophytic isolates. These findings suggest that species-specific components of the root carbohydrates, that are accumulated in larger amounts, may also affect endophytes. Mixtures of sugars cause effects that cannot be explained merely by adding the effects of the major single constituents. This becomes evident in the comparison of the variation in the responses of the endophytic isolates to the carbohydrates of the two Peucedanum species that show variable quantities of the same major sugars. Carbohydrates of the dominant grass species, B. erectus, stimulated the growth of the majority of the tested isolates most efficiently of all tested substrates. This concurs with expectations that plant species present in large numbers of individuals may shape the utilisation behaviour of soil microfungi more than those present in lower numbers. The results point to a mechanism of adaptation that may be facilitated by fine-tuning the combination of variable substrate utilisation and/or signal perception capabilities of a single fungal strain. These factors may also participate in the postulated reciprocal determination of plant and fungal diversity.

Endophytes constitute an essential proportion of fungal diversity. Among the fungi recovered from the root tissues of the two Peucedanum species, a large number were sterile isolates. A big proportion of these were dark-pigmented mycelia, many more compared to those recovered from rhizosphere and bulk soil. This suggests that endophytes comprise more highly specialised strains than rhizosphere and bulk soil isolates. Our results also corroborate the general impression that rhizosphere soil harbours a higher proportion of fungal diversity than bulk soil [2,3]. However, isolates from both soil habitats did not markedly differ in their bandwidth of substrate utilisation.

Among those endophytic isolates that we could identify on the basis of morphological characteristics, we found strains belonging to the genera Cylindrocarpon and Penicillium. Cylindrocarpon isolates have frequently been reported as colonisers of plant root surfaces. C. destructans and its teleomorph, Nectria radicicola Gerlach and L. Nilsson, are known to cause root rot, together with various Fusarium species [34]. Other identified endophytes comprised four Penicillium species. Penicillium represented the most species-rich genus among the isolates, especially in the rhizosphere of P. alsaticum and, to a lesser extent, in the bulk soil. The four endophytes were not affiliated to any of the soil-derived Penicillium isolates and we have failed to identify three of them so far. Penicillum species have already been reported to occur as endophytes in various trees and shrubs [41–43] and may be more widespread than we currently assume.

The biology of endophytic fungi within plants is still not understood completely. Endophytes are considered to be latent pathogens on one hand [44]; on the other hand, they are also regarded to be mutualists that, by producing toxic secondary metabolites, may even protect their host plants against herbivores or nematodes [45,46]. The majority of anamorphic microfungi in the soil are necrotrophs [34]. The endophytes in our model umbelllifers obviously adapt to their host plants by utilising root carbohydrates more efficiently at lower concentrations. This might be helpful during the infection process. Concurrently, once within the root's tissue, the growth rate of the endophyte is slowed by higher concentrations of the same compounds. This reduced growth might help to avoid the induction of plant defence mechanisms that are usually triggered by more vigorously growing pathogenic fungi [47]. Later, often many years after infection, the endophytes might profit from the fact that they are able to exploit the dying plant from within its tissues.

Offering substrates in a range of concentrations offers additional parameters and thus improves the detection of differences in the responses of fungal isolates to a range of differently composed substrates. Usually, growth of a specific fungus or fungal community on a single concentration of defined sugars, sugar alcohols, organic and amino acids after a given period or in a defined time course is compared [13,48]. Although the presented study is of pilot character, conclusions drawn from the obtained results offer an interesting working hypothesis for similarly aimed investigations in the future and may also benefit research in discovering and understanding gene regulation mechanisms governing plant–fungus interactions. Secondary metabolites have not been the focus of this study, but they certainly are not to be neglected because they may play an essential role in the fine tuning of the species composition within a fungal community. Flavonoids have already been identified as serving as important stimulant signals to arbuscular mycorrhizal fungi [49].


This research was supported by Grant P13278-BIO from the Austrian Science Fund. We are grateful to Dr W. Gams (CBS, Utrecht, The Netherlands) and Dr. Katja Sterflinger (Institute of Applied Microbiology, University of Agricultural Sciences, Vienna) for advice on isolate identification, Prof. K.-H. Prillinger (Institute of Applied Microbiology, University of Agricultural Sciences, Vienna) for general support, and Dr Elvira Hörandl (Institute of Botany, University of Vienna) for critical comments.