Sampling was conducted in the Estación Experimental Agropecuaria (EEA) ‘Delta del Paraná’ of Instituto Nacional de Tecnología Agropecuaria (INTA) in Buenos Aires province, Argentina. Soils of the EEAINTADelta (34°10'S, 58°51'W) are in a deltaic plain with clear fluvial influence where nut production with pecan trees began 40 years ago. The climate is temperate humid with an annual average temperature of 17°C and an annual precipitation of 900–1000mm (INTA 1989). The soil texture was classified as silty clay loam, with pH 4.5, total C 4.48%, organic matter 9.96%, total N 0.3%, extractable P 0.29 meq.100 g^-1; K 0.47, Ca 9.8, Mg 5.6 and Na 0.47 meq.100 g^-1. The herbaceous vegetation that appeared under the trees canopy was periodically removed.

In August 2014, twenty pecan trees within the EEA-INTADelta del Paraná were sampled, across an area of approximately 3 ha. Three soil cores of 250 ml (including pecan roots, AM spores and external mycelium) were taken under the canopy of each tree at 20 cm depth, placed in polypropylene bags and stored at 4 °C until processed. All soil cores (sub-samples) from the field were mixed to produce a single complex-sample (T0). This sample was subsequently divided into two parts destined for: i) the molecular and conventional characterisation of AM fungal community; and ii) the establishment of trap cultures in pots with pecan seedlings. Pecan seeds were also collected in the EEA-INTADelta del Paraná during the autumn and stored at 4 °C until use. For stratification, seeds were soaked in tap water, placed in moist perlite and stored at 4 °C for 30 days.

In order to detect the AM fungi colonising pecan roots at sampling time, a microcosm assay was conducted in a greenhouse using the field sample (T0), as AM inoculum. In September, five 10 litre plastic containers were filled with a sterile mix of perlite-vermiculite (1:1) and 400 g of T0. Pecan seeds were surface-disinfected (immersed in 5% sodium hypochlorite solution for 20 min and rinsed with sterile water) and pre-germinated. One seed was placed in each pot. Pecan seedlings were grown under natural light and temperature, watered when necessary and irrigated with 50 ml of Hewitt (1966) nutritive solution without P (to avoid influencing mycorrhization) once a month during 11 months. At the end of the experiment, three sub-samples of the whole root system and rhizospheric soil were removed from each pecan plant and mixed to produce a complex-sample (T1) used for the molecular and morphological characterisation of AM fungal community.

A part of the pecan roots, collected from the field (23 subsamples) and from trap cultures (5 subsamples), were used to evaluate the AM fungal colonisation by observation under a binocular microscope (1000× magnification). Roots were stained by a modified Phillips and Hayman (1970) method: they were bleached with hydrogen peroxide, cleared with KOH (10% w/v, 15 min, 90 °C) and stained with trypan blue in lactic acid (0.02%, 10 min, 90 °C). Intraradical colonisation was quantified by examination of 50 randomly selected root pieces (1 cm length). Frequency (F%) of mycorrhizal colonisation was calculated as the percentage of root segments containing hyphae, arbuscules or vesicles (Declerck et al. 2004). Photos were taken under an Olympus BX51 microscope coupled to an Infinity 1 digital camera.

Spores were recovered by successive wet sieving and decanting of soils and collected with a micropipette under a stereomicroscope. In order to ensure that all AM fungal species were sampled, a species accumulation curve was constructed in T0 and T1 as: number of new AM fungal species observed vs. weight of sampled soil. Five grammes of soil were added at each observation opportunity until no new species were found.

Spores characterisation was performed by mounting them in polyvinyl alcohol-lactic acid-glycerol (PVLG) and a mixture of PVLG-Melzer reagent and examining with a binocular microscope (1000× magnification). AM fungi were identified to species, whenever possible, based on morphological characters and subcellular structure of spores with the descriptions available at Professor Blaszkowski web page (http://www.zor.zut.edu.pl/Glomeromycota/) and at Błaszkowski (2012). Taxonomic assignment was performed according to the MycoBank database (http://www.mycobank.org/).

Relative abundance (RA%) of each AM species (calculated as: the number of spores of a particular AM species/the total number of identified AM spores) was plotted in rank-abundance diagrams; AM species were ranked from the most to the least abundant.

Six metagenomic DNA isolations were carried out from T0 and six from T1 soil samples with the Mo Bio Power Soil DNA isolation kits (Mo Bio Laboratories, INC., Carlsbad, CA, USA) following manufacturer’s protocol. Given the low DNA yields obtained for AM fungi and in order to increase the total AM fungal DNA isolated, all the AM spores and fine pecan roots from 100 g of T0 and 100 g of T1 soil samples (dry weight) were manually collected under a stereomicroscope for DNA isolation with the same commercial kit. All DNA samples from T0 were pooled together as well as those from T1. It has been previously demonstrated that composite samples could provide more accurate surveys than single samples due to the patchy distribution of AM fungi in soils (Xu et al. 2012).

Glomeromycota sequences of the small subunits (SSU) region of ribosomal DNA were amplified using the AMV4.5F and AMDGR primers. These primers were chosen because of their high AM fungal specificity (Lin et al. 2012; Lumini et al. 2009) with the aim of quantifying the AM fungal community by amplicon sequencing on a 454 Life Sciences Genome Sequencer FLX System (454 GS FLX) and Titanium chemistry (Roche Applied Science). Oligonucleotides were specifically designed for pyrosequencing with the 454 GS FLX Titanium. Amplicon Fusion Primers contain a directional 454 GS FLX Titanium Primer A or B sequence (in bold letters) which includes a four-base library ‘key’ sequence (underlined) at the 5-prime portion of the oligonucleotide, in addition to the template-specific sequence at the 3-primer end. A Multiplex Identifier (MID) sequence or ‘barcoding’ was added to the reverse primer (in brackets) between the B primer and the template-specific sequences in order to sequence multiple samples in a single run. These sequences allowed automated software identification of each sample. Forward (Primer A – Key):

5’-CGTATCGCCTCCCTCGCGCCATCAGAAGCTCGTAGTTGAATTTCG-3’

Reverse (Primer B – Key):

5’-CTATGCGCCTTGCCAGCCCGCTCAG(MID10bp)CCCAACTATCCCTATTAATCAT-3’.

PCR amplification was undertaken on a FastStart High Fidelity PCR system (Roche Applied Science, Mannheim, Germany) following the manufacturer’s instructions. The PCR conditions were 95 °C for 5 min, followed by 30 cycles of 95 °C for 45 s, 57 °C for 45 s and 72 °C for 60 s and a final elongation step at 72 °C for 4 min. The reactions were purified and the pyrosequencing run was carried out on one quarter of the sequencing plate on a 454 GS FLX at the Instituto de Agrobiotecnología de Rosario (INDEAR) following the amplicon sequencing protocol provided by the manufacturer.

Sequence data were quality controlled and de-noised with the ampliconnoise.py script of Quantitative Insights into Microbial Ecology (QIIME) pipeline (Caporaso et al. 2010). This script also eliminated chimeras. Sequences were clustered into molecular operational taxonomic units (MOTUs) using the pick_MOTUs.py script (QIIME) according to the 97% sequence similarity. This value was chosen in accordance with the conventional definition of microbial species (Konstantinidis et al. 2007). Non-AM fungal sequences were subsequently removed from the data sets (Lekberg et al. 2012) and molecular singletons were not considered in the analyses in order to avoid overestimation of species richness (Unterseher et al. 2011) The most abundant sequences from each MOTU were selected as representative and species identification was assigned comparing MOTUs sequences against the MaarjAM database (http://maarjam.botany.ut.ee/) and the GenBank database (http://www.ncbi.nlm.nih.gov). DNA similarity was performed using BLAST servers and only sequences with coverage and similarity values higher than 98% (resulting in e- values close to zero) were considered. Representative MOTUs sequences, analysed in our study, have been deposited in The Sequence Read Archive (SRA) under the accession number: SRA058132.

Relative abundance (RA%) of each AM MOTUs (calculated as: number of sequences of an AMMOTU/the total number of identified MOTUs) was plotted in rank-abundance diagrams; AM MOTUs were ranked from most to least commonly collected according to their abundance in the samples.

A rarefaction curve was constructed to assess sampling efficiency: the number of new AM fungal MOTUs observed vs. the number of sequences read. To calculate AM molecular richness within T0 and T1 samples, observed species (OS) and Chao 1 richness estimator were performed. Curve analyses were constructed with QIIME by randomly selecting a series of subsets from libraries in different sizes; this procedure was replicated by the programme ten times for each subset sample.

After sampling 75 g of both pooled soils (T0 and T1), new AM fungal species were not observed (Fig. 1a) indicating that all AM fungal species were already detected. Based on morphological taxonomic determinations, the AM fungal richness was 6 species in T0 and 9 in T1. Several differences in the AM community structure were found amongst field soil samples and one year old trap pot culture samples: rank-abundance species diagrams (Figs 2a,b) showed that two thirds of the identified spores at T0 belonged to Claroideoglomus lamellosum (Dalpé, Koske & Tews) C. Walker & A. Schüssler (37.5% of spores) and Rhizoglomus microaggregatum (Koske, Gemma & P.D. Olexia) Sieverd., G.A. Silva & Oehl (31%), followed by Funneliformis coronatum (Giovann.) C. Walker & A. Schüssler (12.5%). Spores of Entrophospora infrequens (I.R. Hall) R.N. Ames & R.W. Schneid., Funneliformis mosseae (T.H. Nicolson & Gerd.) C. Walker & A. Schüssler and Gigaspora margarita W.N. Becker & I.R. Hall were also detected (6.25% for each species) in pecan rhizospheric soils. The number of isolated spores at T1 was nearly nine times higher than at T0. In these samples, C. lamellosum was also the dominant AM fungal species (49% of spores), followed by Claroideoglomus etunicatum (W.N. Becker & Gerd.) C. Walker & A. Schüssler (34%), which was not detected in T0 samples and F. mosseae (6%), E. infrequens, Cetraspora pellucida (T.H. Nicolson & N.C. Schenck) Oehl, F.A. Souza & Sieverd., R. microaggregatum, Diversispora eburnea (L.J. Kenn., J.C. Stutz & J.B. Morton) C. Walker & A. Schüssler, Fuscutata rubra (Stürmer & J.B. Morton) Oehl, F.A. Souza & Sieverd. and Septoglomus constrictum (Trappe) Sieverd., G.A. Silva & Oehl were also identified with frequencies under 5%.

Figure 1. 

Morphological AM fungal species accumulation curves (a) and rarefaction curves (± errors, ten replicates for each subset) of observed AM MOTUs (b) detected on C. illinoinensis rhizosphere in T0 (field) and T1 (containers) samples.

Figure 2. 

Rank-abundance diagrams of morphological AM fungal species (a–b) and rank-abundance diagrams of AM MOTUs (c–d) detected on C. illinoinensis rhizosphere in T0 (field) and T1 (containers) samples. RA: Relative abundance. Cl: Claroideoglomus lamellosum, Rm: Rhizoglomus microaggregatum, Fc: Funneliformis coronatum, Ei: Entrophospora infrequens, Fm: Funneliformis mosseae, Gi Gigaspora margarita, Ce: Claroideoglomus etunicatum, Cp: Cetraspora pellucida, De: Diversispora eburnea, Fr: Fuscutata rubra, Sc: Septoglomus constrictum, Re: Rhizoglomus irregulare, Ri: Rhizoglomus intraradices, Di: Dominikia iranica, Da: Dominikia indica, Gsp: Glomus sp.

Based on 97% sequence similarity, a total of 49 MOTUs were obtained, 41 of them belonging to Glomeromycota phylum, proving the Glomeromycota specificity of the selected primers, since the majority of detected sequences corresponded to this phylum (98.5% and 94.8% in T0 and T1, respectively), followed by the Basidiomycota fungi Piriformospora indica (0.73% and 4.85% of sequences in T0 and T1, respectively) and Cyphellopsis anomala (0.034% of sequences in T1). Sequences of some Chytridiomycetes and other Eukaryota were also detected (Table 1).

A total of 2196 and 5876 sequences were obtained for T0 and T1, respectively (there were no chimera sequences). The number of Glomeromycota sequences was 7736 (average read length of 268±16bp). Number of sequences ranged from 2164 reads (TO), to 5572 reads (T1), defining 30 and 29 MOTUs, respectively (Table 1). The OS (Fig. 1b) and Chao1 (Data not shown) rarefaction curves reached the plateau phase over 2000 (T0) and 5000 (T1) sampled sequences. The numbers of obtained and estimated MOTUs were close; AM richness appeared to be similar for T0 and T1 with both estimators: 30.8 and 28.9 (OS of T0 and T1) and 37.1 and 35.2 (Chao1 of T0 and T1).

Considering all the Glomeromycota genera identified, the proportion of Rhizoglomus sequences was always dominant in pecan rhizosphere (83% and 91% in T0 and T1, respectively). Funneliformis (7.2% and 0.16% in T0 and T1, respectively) and Glomus sequences (8.2% and 4% in T0 and T1, respectively) were more abundant in field samples but clearly diminished in pots. Particularly, Claroideoglomus had the opposite behaviour, as their sequences’ abundances were lower at field (1.4%) than at pot (4.9%) level. A few sequences of Paraglomus laccatum (Błaszk.) Renker, Błaszk. & Buscot and E. infrequens (less than 0.05%) were also detected in both samples.

The rank-abundance species diagrams (Figs 2c,d) showed that, in T0, almost 60% of the total sequences belong to Rhizoglomus irregulare (Błaszk., Wubet, Renker & Buscot) Sieverd., G.A. Silva & Oehl and into the 95% of accumulated sequences, Rhizoglomus intraradices (N.C. Schenck & G.S. Sm.) Sieverd., G.A. Silva & Oehl, F. mosseae, Dominikia iranica (Błaszk., Kovács & Balázs) Błaszk., Chwat & Kovács, Dominikia indica (Błaszk., Wubet & Harikumar) Błaszk., G.S. Silva & Oehl and C. lamellosum were also detected. Meanwhile in T1, 90% of total sequences were of R.irregulare and C. lamellosum was only detected in 95% of the accumulated sequences. The rest of sequences represented only 5%.

Most identified species were common to both soils, some were found only in field samples as D. iranica, while others only appeared (E. infrequens) or were much more abundant (Sclerocystis sinuosa Gerd. & B.K. Bakshi) in containers after one year of pecan culture.

At sampling time, the mycorrhization percentage of T0 natural pecan root was 17% (±6.8), while T1 roots showed 6.33 (±3.53), 28.33 (±15.9) and 29 (±3.6)% of mycorrhization, when observed at 60, 100 and 360 days of seedling growth, respectively.

Appressoria at entry points and intraradical longitudinal hyphae in the outer layers of the pecan root cortex were observed (Fig. 3a). Typical arbuscules were very frequent; however, the presence of vesicles was less commonly detected (Fig. 3b).

Figure 3. 

Arbuscular mycorrhizal intraradical colonisation in pecan roots (a–b). A: arbuscules, AP: appressoria, ILH: intraradical longitudinal hyphae. Spore of Claroideoglomus lamellosum (c), Entrophospora infrequens (d), Cetraspora pellucida (e), Rhizoglomus microaggregatum (f) inside another, dead AMF spore, resembling E. infrequens.