Fungal endophyte diversity in soybean



Ann E. Impullitti, Department of Biology, Augsburg College, Minneapolis, MN 55454, USA.




To determine the identity and diversity of endophytes in soybean plants using culture-dependent (CD) and culture-independent (CI) methods.

Methods and Results

Stem samples were collected from three field-grown soybean cultivars grown to a reproductive stage in Minnesota, USA. Samples were surface disinfested, and CD and CI methods were used to assess the endophytes. For the CD method, fungi were isolated and grouped based on colony morphology, and the rDNA ITS region was sequenced to identify the cultures. The most frequently isolated genera were Cladosporium (36%), Alternaria (13%), Diaporthe (9%) and Epicoccum (9%). For the CI method, DNA was extracted from the stems, and the ITS region was amplified, cloned and sequenced for identification. The most prevalent genus detected using CI method was Cladosporium (85%).


Soybean contains a diverse array of endophytic fungi that were identified in this study. The CD method detected greater endophyte diversity (H' = 2·12) than the CI method (H' = 0·66).

Significance and Impact of the Study

The results improve our understanding of the identity and diversity of endophytic fungi that likely have different kinds of interactions with soybean plants. The results suggest that CD and CI methods should be used to study endophytes in soybean and perhaps other annual crop plants.


Endophytic fungi that infect plants are ubiquitous in all environments studied (Carroll 1988; Petrini 1991). Although the diversity and function of clavicipitaceous fungal endophytes that infect grasses are well documented, little is known about the diversity and function of nonclavicipitaceous (NC) endophytes in plants, particularly in economically important species (Clay 1988; Rodriguez et al. 2009). Some NC-endophytes can reportedly reduce plant diseases and enhance plant growth and may be the basis for emerging methods to improve plant growth and production (Sturz et al. 2000; Arnold et al. 2003; Marquez et al. 2007a; Hardoim et al. 2008; Mejía et al. 2008). For example, fungal endophytes in Theobroma cacao and Solanum melongena reduced foliar and root diseases, respectively, and treatment of soybean [Glycine max L. (Merr)] with culture filtrate from the endophyte Cladosporium sphaerospermum increased plant height (Narisawa et al. 2002; Arnold et al. 2003; Mejía et al. 2008; Hamayun et al. 2009).

Although soybean is a major world crop, there is very limited knowledge of its fungal endophyte community. Two studies have reported select aspects of the diversity of fungal endophytes in soybean and both used culture-dependent (CD) methods (Miller and Roy 1982; Pimentel et al. 2006). In Mississippi, USA, 15 genera of fungal endophytes were identified by colony morphology in soybean leaves, pods and seeds (Miller and Roy 1982). Leaves were more frequently colonized by endophytic fungi than pods and seeds, but there were no differences in species diversity among these plant parts. In Brazil, 12 genera of fungal endophytes were identified in soybean leaves and stems (Pimentel et al. 2006). Similar numbers of endophytes were isolated from both plant parts, but the population decreased as plants matured, and endophytes were detected more frequently in lower than upper stem sections. Endophytes in soybean have not been reported elsewhere or using culture-independent (CI) methods to our knowledge.

CD and CI methods are used to identify fungal endophytes in-situ. Each method has advantages and disadvantages. CD methods are based on morphological and molecular identification of fungi that grow from plant tissues on nutrient media. This method is relatively inexpensive and allows for isolation and characterization of culturable fungal endophytes. CI methods are based on extraction of total DNA from plant tissues followed by DNA sequencing and identification of fungi based on reference sequences. The CI methods allow detection and identification of all fungi including those that are cryptic, sterile and unculturable, and CI methods can potentially detect a more diverse endophyte population than CD methods (Arnold et al. 2007). However, identification of fungi at the species level is challenging with CI methods, and PCR bias and chimeric sequences can be problematic (Suzuki and Giovannoni 1996; Acinas et al. 2005). The optimal method to identify fungal endophytes is largely based on the research questions to be answered.

The goal of this study was to identify the composition of fungal endophytes in soybean stems. The objectives were to: (i) determine the identity and diversity of endophytic fungi in different parts of soybean stems; and to (ii) compare the use of CD and CI methods for characterizing endophyte fungal communities.

Materials and methods

Field experimental design and plant collection

The cultivars AG2107 (Monsanto Co., St. Louis, MO, USA), MN1803 and Parker were planted in June 2008 in double-row plots, 2 m long with 0·7 m row spacing. MN1803 is a single backcross of Parker x (Parker x Resnick BC2F2) from Dr. J. Orf at the University of Minnesota. The plots were located at the University of Minnesota Southern Research Center near Waseca, Minnesota, USA, on a Nicollet–Webster clay loam soil in a field with a history of annual corn and soybean rotation. Four plots each of AG2107 and MN1803 and two plots of Parker were arranged in a randomized block design. Six plants without symptoms of disease were selected arbitrarily from each plot at the R2/R3 growth stage, cut at the soil line, immediately placed on ice and stored up to 72 h at 4°C until processed. Leaves were removed, and stems were washed thoroughly and divided into base (lower 25 cm), middle (26–41 cm above the base) and apex (42–58 cm above the base) sections.

Culture-dependent identification of fungal endophytes

The stem sections were washed a second time in sterile water for 20 s, immersed in 70% ethanol for 5 s, soaked in 0·525% sodium hypochlorite for 2 min, rinsed in sterile water for 10 s and dried for 5 min on a sterile paper towel. Four segments 0·5–0·7 cm long collected at 2·2–3·8 cm intervals from each of the three stem sections were used for isolations. For the stem segments from the base and middle stem sections, the combined dermal and vascular tissues (DV) were separated from pith tissues, and each section cut longitudinally into four quarters to result in eight pieces per segment. Only the pith tissues from the base and middle sections of the stem were analysed. The dermal, vascular and pith tissues from the apex section were inseparable. Stem pieces were placed onto 1/2× acidified potato dextrose agar (APDA, pH 4·8) (BD, Franklin Lakes, NJ, USA) and incubated at 25°C for 5–7 days. As mycelia emerged from the stem pieces, 3 mm2 cubes from the margin of growing cultures were transferred to 1/2× APDA. The plates with little or no growth were maintained at 25°C, checked weekly for 3 months, and any new fungal mycelium was transferred. Fungal cultures were photographed and grouped into morphotaxa based on colony shape, color, hyphal width, hyphal branching and colony growth rate.

The intertranscribed spacer regions (ITS) of representative isolates from each morphotaxon group were sequenced. The number sequenced per morphotaxon was proportional (~1 : 5) to the number of isolates each group contained, or two were selected for analysis if there were <5 isolates in a morphotaxon (total n = 75). Mycelium was removed from each culture plate with a sterile spatula, and DNA was extracted using a FastDNA® kit (MP Biomedicals, Solon, OH, USA). ITS1, 5·8S, and ITS2 of the nuclear rDNA were amplified in 25 μl PCR reactions containing 1× GoTaq Green Master Mix (Promega, Madison, WI, USA), 0·2 μmol/l each of primers ITS1-F and ITS4, 5 μl template DNA solution and sterile water using either an MJ Research Thermal Cycler PTC-100 (Waltham, MA, USA) or Eppendorf Mastercycler (Westbury, NY, USA) using reaction parameters as described previously (White et al. 1990; Gardes and Bruns 1993; Pan et al. 2008). Negative controls containing molecular grade water in place of DNA solutions were included along with all PCR reactions. Reaction products were electrophoresed in 1·5% agarose gels and viewed after staining with SYBR green to ensure the presence of single amplicons of ca. 450–800 bp.

Culture-independent identification of fungal endophytes

Stem tissues from the base, middle and apex sections of three plants from two replications of each cultivar were washed, dried for 48 h at 35°C and ground separately using a Wiley® Mini mill (Thomas Scientific, Swedesboro, NJ, USA) that was cleaned between samples. The dermal, vascular and pith tissues could not be separated after drying. Ground tissue was stored at −20°C until total DNA was extracted (Malvick and Grunden 2005). DNA from the base, middle and apex sections from each plant was pooled, and the fungal ITS region was amplified with primers ITS1-F and ITS4 as described above. The amplicons were cloned with a TOPO TA cloning kit (Invitrogen, Carlsbad, CA, USA). Sixty arbitrarily chosen bacterial clone colonies from each of the five cultivar by treatment combinations were grown at 37°C for 16 h in LB broth amended with 100 μg ml−1 ampicillin. Plasmid DNA was extracted using a PureLink Quick Plasmid MiniPrep kit (Invitrogen), and insertion of the rDNA was confirmed by PCR as described above.

Sequencing and data analysis

The ITS samples amplified from the CD and CI methods were purified using a PureLink PCR kit (Invitrogen) and sequenced to identify fungal endophytes. ITS samples obtained with the CD method were sequenced using either ITS1-F or ITS4. Five isolates and 10 clones were sequenced in both directions to determine whether one and two directional sequencing resulted in the same taxonomic identification. The ITS region in the clones from the CI method was sequenced using universal M13 primers (Messing 1983). Sequencing was completed at the BioMedical Genomics Center at the University of Minnesota – Twin Cities using an ABI Prism 3730xl capillary DNA Analyzer (Applied Biosystems, Foster City, CA, USA). The 450–650 bp sequences were edited, and consensus contigs were determined by the ‘assemble automatically’ function in Sequencher v4.8 (Gene Codes, Ann Arbor, MI, USA) using 40 nucleotides for the minimum overlap and a sequence similarity value of 99%.

Comparison of sequence similarities at the 90, 95, 98 and 99% levels yielded similar consensus groupings, and therefore, only the groupings from the 99% level were used for complete analyses. Each consensus sequence was compared with other DNA sequences in the ‘nr’ database with Blastn (Altschul et al. 1990). Sequences were considered a match to a genus if the E-value was <0·001, identity was 100%, and the alignment score was >200. Endophytes could not be identified to the species level because multiple species matches were obtained with sequence similarities >98%. The Blastn feature in GenBank is widely accepted and is an important tool for studies such as these, but it is also important to note GenBank has limited quality control. Sequences are deposited in GenBank (Accessions: KC662222 - KC66248). Comparison of the BLASTn results from one to two directional sequencing produced the same identification results (data not shown); hence; data from one directional sequencing only were used for multiple samples.

Differences in the infection frequencies for the number of isolates within individual cultivars were tested for significance using a chi-square test. The sufficiency of sample size for the CD and CI methods and species richness were compared by constructing species accumulation curves with EstimateS, version 8.2 (Colwell 2006). Curves were generated by sampling without replacement. Diversity was assessed using the Shannon Index (Magurran 2004). Differences in the number of isolates and frequency of colonization for the cultivars and tissue types were tested for significance using chi-square analysis (GraphPad v6.01).


The fungal endophyte colonization in the plants was influenced by the tissue type, section of the stem and soybean cultivar. More endophytes were isolated from the dermal and vascular stem tissues (DV) than the pith (Table 1). A greater number of isolates were detected in the base of the plants compared with the middle and apex of the DV of the stems (Fig. 1). DV tissue was colonized more than the pith tissue. Greater than 60% of the DV tissues were colonized, while <30% of pith tissues were colonized (Table 1).

Table 1. Colonization percentage, number of endophytic fungal morphotaxa and the number of isolates detected in dermal and vascular tissues (DV) and pith tissues of three soybean cultivars using a culture-dependent (CD) method
CultivarColonization (%)MorphotaxaaNo. of Isolates
  1. a

    Isolates were grouped into 20 morphotaxa based on colony shape, colour, hyphae and growth rate. Numbers indicate how many morphotaxa were represented.

  2. b

    Isolates collected using the CD method from DV and the pith.

  3. c

    Significant differences between DV and pith tissues for each cultivar were determined using a chi-square (χ) analysis.

MN 1803637na716830
χ2 P-valuec<0·005na<0·005
Figure 1.

The number of fungal endophyte isolates identified from the base, middle or apex of stems using the culture-dependent method from three different soybean cultivars grown in field plots. (a) The number of isolates collected from the pith tissues of the stem. Differences in the number of isolates collected from the various positions on the stem were not significantly different based on a chi-square analysis (P > 0·1) (b) The number of isolates collected from the dermal and vascular tissues of three different cultivars. The number of endophytes detected at each position on the stem differed based on a chi-square analysis (P < 0·05). The apex is not included in panel (a) because pith tissues could not be separated from vascular tissues in the apex of stems. (■) AG2107; (□) MN 1803 and (image_n/jam12164-gra-0001.png) Parker.

The number of endophytic fungal genera identified in soybean stems with the two methods used to detect and identify them were not significantly different (Fig. 2). Using the CD method, 12 genera were identified from the pith of soybean. The most frequently isolated genera and their isolation frequencies were Cladosporium (36%), Alternaria (13%), Diaporthe (9%) and Epicoccum (7%) (Table 2). The least commonly detected genera were Plectosphaerella and Phialophora. Using the CI method, six genera were identified from soybean stems. Cladosporium was the most frequently detected genus, representing 85% of all isolates (Table 2). The remaining six genera were detected at frequencies of 1–7%. The CD method detected greater diversity when the data from all cultivars were combined, and when each cultivar was analysed individually when compared to the CI method (Table 3). Species accumulation curves based on CI and CD data did not plateau, and the curve for the CI method was shallower than that for the CD method (Fig. 3).

Table 2. Frequency of detection of fungal endophyte genera in soybean stems of three field-grown soybean cultivars combined. Endophytes were detected using either culture-dependent (CD) or culture-independent (CI) methods and identified based on ITS sequencing
Fungal generaaFrequency of detection (%)
  1. a

    Genera identified from BLAST based on 99% sequence similarity of the ITS region.

  2. b

    Fungal endophyte genotypes that could not be identified to genus with BLAST.

Alternaria 131
Cladosporium 3685
Davidella 02
Diaporthe 90
Epicoccum 72
Fusarium sp. 220
Fusarium sp. 320
Phialophora 10
Phoma sp. 120
Phoma sp. 220
Phomopsis 50
Plectosphaerella 17
Verticillium 00
Trichoderma 42
Unidentified fungal endophytesb101
Table 3. The diversity of endophytes as measured with Shannon's diversity index (H′) from three different soybean cultivars using culture-dependent (CD) and culture-independent (CI) methods
CultivarShannon's diversity index (H')
All cultivars2·120·66
Figure 2.

The total number of endophytic fungal genera detected from three different soybean cultivars using culture-dependent (CD) and -independent methods. For the CD method, fungi were isolated from disinfested soybean stems at the R3 growth stage, and for the culture-independent (CI) method, DNA was extracted from soybean stems at the R3 growth stage and the fungal ITS region was amplified and sequenced. There were no significant differences in the number of genera collected from the cultivars based on a chi-square analysis (P > 0·05). (□) CD and (image_n/jam12164-gra-0002.png) CI.

Figure 3.

Species accumulation curves for endophytic fungi identified using either a (a) culture-independent or (b) culture-dependent method. Dotted lines indicate 95% confidence intervals.


Studies of fungal endophytes in many environments are an active area for research, but the endophytes in soybean have never been systematically characterized, and methods for characterizing the endophyte populations have not been compared previously. This work expands our understanding of the endophytic fungi in soybean plants and reveals advantages and disadvantages of CI and CD methods to study endophytes.

This is the first report of fungal endophyte diversity in soybean stems in North America to our knowledge. Stems were the focus because many soybean pathogens commonly colonize stems, and stems have not previously been investigated in North America. Many of the fungal endophytes identified in this study are not known to be soybean pathogens, and the functional associations between these fungi and soybean are unknown. Some of the fungi identified are known soybean pathogens, although no symptoms of disease were observed during this study.

The most prevalent endophytic fungal genus detected in the vascular and pith tissues of soybean was Cladosporium. The high frequency of isolation of Cladosporium suggests it was the dominant member of the soybean endophytic community. Our data and the GenBank database were insufficient to identify the species of this genus that were detected in this study. Cladosporium was also one of the most frequently identified endophytes from the leaves and stems of soybeans grown in Brazil and from corn grown in Minnesota, USA (Pimentel et al. 2006; Pan et al. 2008). Cladosporium can be seed borne in soybean, but is not reported to be pathogenic on soybean (Hartmann et al. 1999). Cladosporium has been investigated for its potential to improve soybean growth, whereby culture filtrates of C. sphaerospermum produced gibberellin-like compounds that increased soybean biomass and height (Hamayun et al. 2009).

Phomopsis and Diaporthe were commonly isolated genera and relatively well understood to be opportunistic pathogens that cause disease as plants mature or are stressed. These two genera are reported to be common in stems during reproductive growth stages and cause the diseases stem canker and pod and stem blight (Sinclair 1991; Hartmann et al. 1999; Harrington et al. 2000). Other endophytic fungi that were identified in this study and have poorly understood associations with soybean include Epicoccum, Verticillium and Phoma (Hartmann et al. 1999). These genera, however, can have beneficial or negative associations with other crops. Epicoccum purpurascens reduced Sclerotinia stem rot in dry beans and sunflower (Huang et al. 2000; Pieckenstain et al. 2001). Several species of Verticillium and Phoma are pathogens on a wide range of plant species, but are not known to cause soybean disease. The significance of endophytism, if any, for Epicoccum, Verticillium, Phoma, Phomopis and Diaporthe in soybean is unknown and merits future investigation.

The endophytic fungal species detected in plants may be influenced by many factors, including the type of tissue sampled, when plants were assayed, and perhaps the climate and location in which they were grown. We studied the tissues of the stem collected from the base, middle and apex of soybean plants. The greatest number of isolates was collected from the base, and the number declined in samples collected at progressively higher portions of the stems. Greater numbers of endophytes near the soil line in stems were also observed for fungi in soybean in Brazil and for bacteria in maize in the USA (Fisher et al. 1992; Pimentel et al. 2006). These studies also suggested that endophytes may exclusively colonize certain tissues (Fisher et al. 1992; Pimentel et al. 2006). For example, Colletotrichum was only isolated from leaves of soybean and not stems in Brazil (Pimentel et al. 2006).

The diversity of endophytes detected and identified from soybean stems using CD and CI methods was less than previously reported from tropical plants and grasses, but similar to the endophytic diversity found in corn in Minnesota (Arnold et al. 2000; Marquez et al. 2007b; Pan et al. 2008). Diversity of fungal endophytes within a plant could be influenced by whether the plant is grown in a monoculture or polyculture, plant age or the cropping history of the field. Soybeans used in this study were grown in a monoculture and within a field that had a history of corn and soybean rotations. Furthermore, the diversity of the endophytic community in soybean grown in temperate vs tropical climates may differ. Research on fungal endophytes in leaves of multiple plant species in locations from Canada to the tropics indicated that diversity of endophytes can be greater in tropical than temperate regions (Arnold et al. 2000; Arnold and Lutzoni 2007).

A trend for greater diversity of fungal endophytes was detected in soybean stems using a culture-based method compared with a CI method, but differences were not statistically significant. The trend for greater diversity using the CD method was unexpected. The recovery and isolation of fungi using CD methods are limited by culture media, incubation environment and other specialized conditions that some fungi require for growth. The relatively low number of genera identified using the CI method is not well understood, but could be attributed to PCR bias during cloning (Suzuki and Giovannoni 1996; Yang et al. 2001; Acinas et al. 2005; Arnold et al. 2007). However when we increased the number of clones sequenced, the number of genera identified did not increase. For example, when 50 clones from AG2107 were sequenced, one was Alternaria and 49 were Cladosporium. The species accumulation curve for the CD method was steeper than the CI method, suggesting more endophytes could be identified with additional samples with the CD, but not the CI method.

Some of the endophytes identified in this study have the potential to promote soybean growth, and some could be latent pathogens. Advantages and disadvantages to using CD and CI methods were demonstrated, and the results suggest using a combination of both methods is best to study the diversity of endophytes. Endophytes may be important organisms to improve sustainable production of crops, although their identities and functions in a range of plants are just beginning to be revealed.


This project was supported by the Doctoral Dissertation Fellowship Program and a Thesis Research Grant awarded from the University of Minnesota Graduate School, and the Minnesota Soybean Research and Promotion Council. We would like to thank Anna Testen and Nora Powers for assistance with culture maintenance and DNA extractions.