Bacteria associated with arbuscular mycorrhizal fungi within roots of plants growing in a soil highly contaminated with aliphatic and aromatic petroleum hydrocarbons


  • Bachir Iffis,

    1. Institut de recherche en biologie végétale, Département de sciences biologiques, Université de Montréal 4101 Sherbrooke est, Montréal, QC, Canada
    Search for more papers by this author
  • Marc St-Arnaud,

    1. Institut de recherche en biologie végétale, Département de sciences biologiques, Université de Montréal 4101 Sherbrooke est, Montréal, QC, Canada
    Search for more papers by this author
  • Mohamed Hijri

    Corresponding author
    1. Institut de recherche en biologie végétale, Département de sciences biologiques, Université de Montréal 4101 Sherbrooke est, Montréal, QC, Canada
    • Correspondence: Mohamed Hijri, IRBV, 4101 Rue Sherbrooke Est, QC H1X 2B2, Canada. Tel.: +1 514 868 5136; fax: +1 514 872 9406; e-mail:

    Search for more papers by this author


Arbuscular mycorrhizal fungi (AMF) belong to phylum Glomeromycota, an early divergent fungal lineage forming symbiosis with plant roots. Many reports have documented that bacteria are intimately associated with AMF mycelia in the soil. However, the role of these bacteria remains unclear and their diversity within intraradical AMF structures has yet to be explored. We aim to assess the bacterial communities associated within intraradical propagules (vesicles and intraradical spores) harvested from roots of plant growing in the sediments of an extremely petroleum hydrocarbon-polluted basin. Solidago rugosa roots were sampled, surface-sterilized, and microdissected. Eleven propagules were randomly collected and individually subjected to whole-genome amplification, followed by PCRs, cloning, and sequencing targeting fungal and bacterial rDNA. Ribotyping of the 11 propagules showed that at least five different AMF OTUs could be present in S. rugosa roots, while 16S rRNA ribotyping of six of the 11 different propagules showed a surprisingly high bacterial richness associated with the AMF within plant roots. Most dominant bacterial OTUs belonged to Sphingomonas sp., Pseudomonas sp., Massilia sp., and Methylobacterium sp. This study provides the first evidence of the bacterial diversity associated with AMF propagules within the roots of plants growing in extremely petroleum hydrocarbon-polluted conditions.


Arbuscular mycorrhizal fungi (AMF) belong to an ancient group of plant-inhabiting fungi that form symbiotic associations. The arbuscular mycorrhizal association is among the oldest symbioses between plants and fungi on earth, and it has been dated back to Ordovician (c. 450 million years) using fossil records and molecular clocks (Simon et al., 1993; Redecker et al., 2000). AMF belong to phylum Glomeromycota, an early diverging fungal lineage of ecologically and economically relevant microorganisms. Glomeromycota promote plant growth by enhancing mineral uptake, in particular phosphorus, and protect plants against soil-born pathogens (St-Arnaud & Elsen, 2005; Smith & Read, 2008; Ismail et al., 2011, 2013).

AMF are widely distributed and can be found in all ecosystems on earth where plants are able to grow. Therefore, they are usually considered to be generalist plant symbionts, as their diversity is limited to between 200 and 300 described species (Öpik et al., 2010; Redecker et al., 2013). The analysis of a large dataset of 14 961 AMF nucleotide sequences retrieved from 111 studies showed that geographic distance, soil temperature and moisture, and plant community type were each significantly related to AMF community structure, but these factors explain only a small amount of the observed variance in the meta-analysis data (Kivlin et al., 2011). Soil pollutants have been considered as potentially major factors affecting AMF diversity (Vallino et al., 2006; Bedini et al., 2010; Long et al., 2010; Hassan et al., 2011). The role of AMF in polluted soils has been widely studied, and several studies revealed that AMF promote phytoremediation and enhance plant tolerance against trace metals and petroleum hydrocarbon pollutants (Bedini et al., 2010; Aranda et al., 2013; Hassan et al., 2013). For example, Bedini et al., 2010 found nine AMF OTUs associated with plants spontaneously growing in trace metal-polluted ash disposal site containing extreme concentrations of Cu, Pb, and Zn.

In all types of soils where AMF are found, their growth is not limited to plant roots; their mycelia extend beyond roots, exploring a larger volume of soil, and producing extraradical hyphae and asexual multinucleate spores. The mycelia of AMF are typically coenocytes that lack septa, allowing cytoplasm, nuclei, and organelles to move freely within hypha (Marleau et al., 2011).

In natural ecosystems, numerous bacterial taxa are closely associated with AMF mycelia where they colonize the surface of extraradical hyphae and spores on which they can form biofilm-like structures (Scheublin et al., 2010; Lecomte et al., 2011; Cruz & Ishii, 2012). Both culture-dependent and culture-independent methods have observed and identified several bacterial taxa belonging to α-, β-, and γ-Proteobacteria and Firmicutes from the surface of mycelia of many AMF species (Roesti et al., 2005; Bonfante & Anca, 2009; Scheublin et al., 2010; Lecomte et al., 2011). In some AMF taxa, bacteria were also shown to live in the cytoplasm as endobacteria [reviewed in (Bonfante & Anca, 2009)]. Using microscopy, Bianciotto et al., 1996 found that an individual spore of the AMF Gigaspora margarita can harbor up to 250 000 bacterial cells in its cytoplasm. However, Jargeat et al., 2004 attempted to cultivate the endobacterium Candidatus Glomeribacter gigasporarum living inside G. margarita using 19 different culture media without notable success. This supports the idea that some endobacteria could be obligate biotrophs that are not able to grow without AMF, which themselves require a host plant to complete their life cycle (Jargeat et al., 2004). Interaction of AMF and bacteria brings another level of complexity to diversity and function of the mycorrhizal symbiosis. Thus, some authors hypothesize that plants, AMF, and bacteria can be considered as tripartite associations resulting in a consortium that promotes plant growth (Bonfante & Anca, 2009). However, the potential roles and infection mechanisms of these bacteria, in particular the endobacteria, are still poorly understood. In addition, the diversity of these associated bacteria has not been explored in polluted soils, neither in extraradical spores nor inside the mycorrhizal roots.

The objective of this study was therefore to describe the bacterial diversity associated with AMF propagules (vesicles and intraradical spores) extracted from the roots of a plant species growing spontaneously in a decantation basin extremely polluted with petroleum hydrocarbons. To do so, we microdissected mycorrhizal Solidago rugosa roots harvested from a polluted site to isolate intraradical AMF propagules. Then, each propagule was subjected to whole-genome amplification (WGA). Bacterial diversity was assessed using cloning and sequencing of the 16S rRNA genes. This approach allowed us to profile the bacteria closely associated with the AMF while reducing the additional diversity of soil bacteria, which can be randomly attached to the surface of AMF extraradical mycelia. 18S rRNA ribotyping was also performed on WGA products to assess AMF taxonomic diversity.

Materials and methods

Site of study, harvesting, and preparation of samples

Sampling occurred on the site of a former petrochemical plant located on the south shore of the St-Lawrence River near Montreal, Quebec, Canada (45°41′55.3″N 73°25′45.0″W). Various plants species were spontaneously growing in an open-air sedimentation basin in which petroleum hydrocarbon wastes were dumped for many decades. Polycyclic aromatic hydrocarbon (PAH) and alkane (C10–C50) concentrations in the basin are shown in Supporting Information, Table S1. PAH and alkane concentrations exceeded by hundreds of time the standards for reuse defined by the government of Quebec for industrial areas. Three individual plants of S. rugosa Mill. growing in the basin sediments were collected in October 2011. Plant roots were cut, were washed several times with tap water to remove rhizospheric soil, and were cut again into pieces c. 1 cm long. A subsample of the roots from each plant was stained for microscopy observations, as described below. The remaining root pieces were washed in sterilized water containing a few drops of Tween 80 to favor removal of the petroleum hydrocarbons attached to the root surface and then in each of the following surface disinfecting treatments: ethanol 90% for 10 s, commercial sodium hypochlorite 5% for 2 min, chlromine-T 4% for 10 min, and streptomycin 0.03% for 30 min. After the last treatment, the roots were washed several times in sterile distilled water and stored in 1.5-mL microtubes prior to microdissection.

To test the efficiency of the surface sterilization procedure, c. 20 root fragments were transferred to Petri dishes containing TSA or PDA media and incubated for 2 weeks to check for the presence of bacteria or fungi able to grow in these media.

Estimation of mycorrhizal root colonization

The roots of S. rugosa were cleared in a solution of KOH 10% at 80 °C for 1 h, washed several times in deionized water, and stained in a 1% acid fuchsin solution at 60 °C for 1 h. The roots then were washed, cut to small fragments, and mounted on microscope slides using glycerol 60% as a mounting medium. The percentage of AMF root colonization was determined under the microscope using the grid-line intersect method (Giovannetti & Mosse, 1980).

Root microdissection, extraction of AMF propagules, and scanning electron microscopy (SEM)

Disinfected root pieces were soaked in a filtered (0.22 μm) mixture of enzymatic solution of 2% pectolyase (Sigma-Aldrich) and 3% cellulase (Sigma–Aldrich) (w/v) in sterile water at 30 °C for 1 h to digest root cell walls. Root pieces were rinsed and transferred to Petri dishes in which they were microdissected under a stereomicroscope positioned into a horizontal clean bench, using thin sterile forceps and needles. Fourteen fungal propagules were randomly collected and individually put in 0.2-mL microtubes containing 2 μL of sterile water and kept at −20 °C until use. Because clear discrimination between vesicles and intraradical spores that some AMF are able to produce requires destructive examination of the cell wall at high magnification, we used the term ‘AMF intraradical propagules’ to designate both structure types. We also collected an uncolonized root tip sample that was used as a control for assessing bacterial endophytes colonization within S. rugosa roots. Workflow of the experimental approach is summarized in Fig. 1.

Figure 1.

Workflow of the experimental approach used in this study, consisting of the collection of root samples from the field, root sterilization, digestion of cell wall, and microdissection. Isolated propagules were subjected to WGA, followed by PCRs targeting the fungal 18S and bacterial 16S rRNA genes, cloning, and sequencing.

To visually confirm the presence of bacteria on the surface of AMF propagules, root pieces were prepared for SEM following the protocol described in Bozzola & Russell, 1992. A Quanta 200 3D (FEI, Burlington, MA) SEM was used to visualize samples and acquire images.


WGA reactions were performed directly on each individual propagule and on uncolonized root tip using the Illustra GenomiPhi HY DNA Amplification Kit (GE Healthcare Life Sciences, QC, Canada) according to the manufacturer's instructions. All WGA products were stored at −20 °C until use. PCRs were then performed using WGA products as DNA template to amplify 18S rRNA gene fragments using AML1 and AML2 primers (Lee et al., 2008), to identify AMF sequences. Nested PCRs were also performed with the primer pair AM1 and NS31 (Simon et al., 1992; Lee et al., 2008) on the propagules for which no amplification occurs using the AML1/AML2 primers. PCR amplifications were also performed using the 16S rRNA primer pair 27f and 1495r (Bianciotto et al., 1996) to assess bacterial diversity associated with the AMF propagules.

PCRs contained PCR buffer, 0.5 μM of each primer, 0.2 mM of dNTP, 1 μL of WGA product, and 1 U of Taq polymerase (QIAGEN, Toronto, ON, Canada) in a total volume of 40 μL. PCRs were run on a MasterCycler ProS thermocycler (Eppendorf, Mississauga, ON, Canada) under the following program: an initial denaturation at 95 °C for 5 min followed by 38 cycles of 95 °C for 30 s, 54 °C for 30 s, and 72 °C for 90 s, and a final elongation at 72 °C for 10 min. PCR products were separated by agarose gel (1%) electrophoresis, stained with GelRed, and visualized under UV light using a Gel-Doc system (Bio-Rad Laboratories, Mississauga, ON).

Cloning and sequencing

Cloning reactions were performed individually on 16S and 18S rDNA PCR products. The ligation reactions were performed in a volume of 10 μL using pGEM-T Easy Vector System II kit containing chemically competent JM109 Escherichia coli cells (Promega, ThermoFisher, Ottawa, ON, Canada) according to recommendations of the manufacturer. Bacterial colonies were screened by PCR with T7 and SP6 universal primers (Hassan et al., 2011). Positive clones that contained inserts were sent for sequencing at the McGill University and Genome Quebec Innovation Centre (Montreal, QC). Sequences were deposited in GenBank under accession numbers KJ809141KJ809555.

Data analyses

Clustering of bacterial sequences was performed in geneious version 6 (Biomatterts Limited, Auckland, New Zealand), and OTUs were defined at 98% similarity. Singleton sequences were kept and used in a different analysis. Rarefaction analyses were performed in r version 3.0.1 software using the vegan package ( The estimator of sample coverage was calculated using inext online (; (Chao & Jost, 2012).

For phylogenetic analysis, sequence similarities of AMF 18S rRNA genes were obtained from MaarjAM (Öpik et al., 2010) and GenBank databases. Choanoflagellate species Monosiga brevicollis and M. ovata were used as an out-group for this analysis. A phylogenetic tree was generated using a neighbor-joining approach with 1000 bootstrap resamplings using the mega version 5.10 software (Tamura et al., 2011).

Results and discussion

Diversity of AMF in plant roots

Microscopic examination of S. rugosa roots, a plant species spontaneously growing in the sediments of a decantation basin containing very high concentrations of aliphatic and aromatic petroleum hydrocarbons (Table S1), showed typical AMF vesicles and hyphae (Fig. S1). The roots showed mycorrhizal colonization with frequency of 70%. As it is not possible to distinguish AMF species based on microscopic examination of roots, we used WGA, PCR, and cloning of the 18S rRNA gene to assess AMF diversity within roots of S. rugosa. Among the 14 propagules analyzed, 11 led to successful PCR amplification products that were subsequently cloned and provided a total of 41 clones. Clone sequences, which were amplified using AML1/AML2 primers, generated a sequence length of c. 800 bp, except those from propagule 5 which were amplified using AM1/NS31 primers and were c. 550 bp. The number of clones, sequence lengths, and blastn similarity results for each propagule are given in Table 1.

Table 1. Arbuscular mycorrhizal fungal taxa found in propagules isolated from Solidago rugosa roots, based on 18S rRNA gene sequencing
PropagulesNumber of clonesFragment length (pb)Most closely related taxaAccession numberPercentage of identity
  1. a

    Three clones matched with Claroideoglomus sp. VTX00056, while six clones matched with the chitridiomycete Spizellomyces palustris.

13800–801 Diversispora eburnea VTX00060 AM713429 99
21798 Archaeospora schenckii VTX00245 FR773150 96
36795–796Glomus sp. VTX00419 GQ140619 95–97
41805Claroideoglomus sp. VTX00193 HE614988 99
54539–560Claroideoglomus sp. VTX00056 HE615005 98
62798 Glomus irregulare VTX00114 FN600536 99
74805–806Claroideoglomus sp. VTX00056 JN252440 96–97

Claroideoglomus sp. VTX00056

Spizellomyces palustris

Rhizophlyctis rosea






93805–806Claroideoglomus sp. VTX00193 HE614988 99
106799–804 Diversispora eburnea VTX00060 AM713429 98–99
112789–800 Diversispora eburnea VTX00060 AM713429 99

Phylogenetic analysis showed a high AMF species diversity colonizing S. rugosa roots (Fig. 2). The 18S rRNA gene sequences from all propagules clustered within taxa belonging to four different AMF families: Claroideoglomeraceae (five propagules), Diversisporaceae (three propagules), Glomeraceae (one propagule), and Archaeosporaceae (one propagule). Clones from propagule 3 showed 95–97% similarity with an unidentified AMF sequence (accession number: GQ140619.1) closely related to Glomeraceae which was found in trace metal-contaminated soil in China (Long et al., 2010). Interestingly, clones from propagule 6 showed 98–99% similarity with AMF sequences VeGlo8 and VeGlo10, closely related to Rhizophagus irregularis (previously named as Glomus intraradices), which were found in roots of plants growing in an extreme site polluted with Cu, Pb, and Zn in Venice, Italy (Bedini et al., 2010).

Figure 2.

Neighbor-joining tree of the 18S rRNA gene of the consensus sequences of clones obtained from each of the 11 propagules analyzed in our study except for propagule 8 in which two consensus sequences of different taxonomic origin were found. The tree shows the different AMF families in which each propagule consensus sequence clustered, and it also shows that six clones of propagule 8 clustered within Chytridiomycetes. Bootstrap values lower than 50% of 1000 replicates are not shown on the branches.

No sequence matching with Paraglomeraceae, Gigasporaceae, or Acaulosporaceae was found in this study. However, this was expected as most AMF taxa belonging to these families are not known to form vesicles and spores within plant roots (Oehl et al., 2011).

Arbuscular mycorrhizal fungal diversity in trace metal-polluted soil has been extensively studied. For example, Vallino et al., 2006 identified 13 taxonomic units belonging to Glomeraceae, Diversisporaceae, and Gigasporaceae from roots of S. gigantea naturally growing in a trace metal-contaminated site in northern Italy. Long et al., 2010 studied AMF diversity from roots and rhizospheric soil of five plants species growing in trace metal-contaminated soil and also identified species belonging to Glomeraceae, Claroideoglomeraceae, Acaulosporaceae, and Archaeosporaceae. In another study, Hassan et al., 2011 showed that the community structure of AMF associated with Plantago major plants was determined by trace metal concentrations in the soil and dominated by Funneliformis mosseae in polluted sites. In contrast, although several studies showed that AMF may directly or indirectly influence phytoremediation of petroleum hydrocarbons (Wu et al., 2009; Gao et al., 2010, 2011a,b, Aranda et al., 2013), the effect of these compounds on AMF diversity remains unclear.

In this study, six of nine 18S rRNA gene clones retrieved from propagule 8 matched with Chytridiomycota species, showing 95% similarity to Spizellomyces palustris and Rhizophlyctis rosea (accession numbers: FJ827665 and GQ160454, respectively). This is not surprising because Chytridiomycota are commonly found in wet soils and can be endophytic or pathogens of plant roots (Jobard et al., 2010). However, little is known about the interaction between AMF and Chytridiomycota. Only a few studies from the 1970s and 1980s hypothesized that some chytrids such as Rhizidiomycopsis sp., Phlyctochytrium sp., and Spizellomyces sp. could be hyperparasites of AMF and may negatively affect the mycorrhizal symbiosis (Ross & Ruttencutter, 1977; Schenck & Nicolson, 1977; Sylvia & Schenck, 1983). However, Paulitz & Menge, 1984 also proposed that Chytridiomycota could be saprotrophs of decaying AMF structures and reported that Spizellomyces punctatum infected mainly nongerminated or dead spores of Gigaspora margarita. Tzean et al., 1983 suggested that Phlyctochytrium kniepii may be vector for transmitting bacteria such as Spiroplasma-like organisms to the cell wall and cytoplasm of AMF spores. It has also been reported that some AMF spores could be infected by other fungi belonging to Ascomycota (Hijri et al., 2002).

Bacterial diversity associated with AMF propagules

Among the 11 AMF propagules successfully identified, six propagules belonging to different AMF species (propagules 1–6) were used to assess the associated bacterial diversity. A total of 428 clones were sequenced from the six propagules and from a control noncolonized root tip. Of these clones, 53 sequences matched with plant chloroplastic genes and were removed from the analysis. The remaining 375 sequences matched with bacterial 16S rRNA genes and were clustered at 98% sequence identity resulting in 27 OTUs and 23 additional singleton sequences. The bacterial genera obtained from the different OTUs groups and singletons are shown in Table 2. Rarefaction analyses were performed on each propagule sequence dataset (Fig. 3). The highest coverage was obtained from the propagules 2, 5, and 6 with recovery of 91.7%, 92.6%, and 90.4%, respectively, while the lowest coverage was obtained from the propagule 4 with 66.2%. Propagules 1 and 3 showed intermediate values with 80.4% and 87%, respectively. The control root tip showed a saturated rarefaction curve with 100% of sample coverage and was represented only by two OTUs (Fig. 3). The most dominant bacterial OTUs associated with AMF propagules belonged to Sphingomonas sp. (28.2%), Pseudomonas sp. (15.7%), Massilia sp. (14.4%), Methylobacterium sp. (11.7%), and unidentified bacterium (9.8%). Other OTUs belonging to Bradyrhizobium sp., Bacillus sp., Bosea sp., and Paenibacillus sp. were found at lower frequencies (Fig. S2). The highly abundant bacterial OTUs were observed in almost all propagules with variable frequencies between the propagules. Those that were less abundant were only found in specific propagules. This may be due to the sampling effort, which did not cover all the bacterial diversity associated with propagules, or perhaps there is a specific affinity of bacteria for AMF species. For example, Pseudomonas sp. was the dominant OTU in propagules 1 and 5 with proportions of 33% and 42%, respectively, but it was not detected in propagule 2 and was only found at low frequencies in other propagules. The same is true for Sphingomonas sp. which was detected as the dominant OTU for propagules 2, 4, 5, and 6, while only one clone belonging to this taxon was found in propagule 1 (Fig. S3). Interestingly, the bacterial OTU richness found in the root sample was extremely low in comparison with that found in AMF propagules as it was represented by only two OTUs, which were also found in AMF propagules, and these belonged to Pseudomonas sp. (14 clones) and Streptococcus sp. (15 clones). Pseudomonas sp. was largely dominant (45 clones from 5 of the 6 AMF propagules), while Streptococcus sp. was found in propagules 1 and 2 only (with seven clones in total). Pseudomonas spp. were reported as an endophytic as well as AMF-associated bacterium, able to promote both plant growth and the symbiotic association between AMF and plants (Strobel & Daisy, 2003; St-Arnaud & Elsen, 2005; Scheublin et al., 2010; Gaiero et al., 2013). The low bacterial richness found in the root may be due to the effect of the surface disinfection of root fragments or to the fact that meristematic cells forming a significant part of the root tip are mainly exempt of bacterial colonization.

Table 2. Detection frequencies and identity of bacterial genera associated with AMF propagules isolated form Solidago rugosa based on 16S rRNA gene sequencing
Bacterial taxaP-1P-2P-3P-4P-5P-6RFragment lengthPercentage of identityAccession number
  1. OTUs were clustered at 98% of sequence similarity; then, OTUs showing similarities with the same bacterial genera were grouped together. P means propagule; R means roots.

Sphingomonas sp.13926141115664–146391–98.5%EF061133, JQ659520, JX566637, AY749436, HF558376, AF181571, AB109749, JX879745
Pseudomonas sp.2211111014325–153898–99%AM421016, FJ889609, KF011692
Massilia sp.52616511827–150594–95% JX566602
Methylobacterium sp.43325604–148792–99%AB698713, FJ025133, CP001029, JF905617
Uncultured bacterium 114175791–150696–99%HE576045, AY672523, JF429334, DQ129631, JX647723, HE798198, JN023771, JX271950, FJ984639, HM845051, JQ769980, GU563747
Streptococcus sp.2515876–151792–99%FR875178, FQ312041
Bosea sp.7411846–145397% DQ440827
Afipia sp.1114842–145399% DQ123622
Brevundimonas sp.221912–143098% KC494321
Bradyrhizobium sp.41461–148698% FJ390912
Paenibacillus sp.31034–151797% JX566644
Lysobacter sp.12946–100297–99% JQ746036
Acinetobacter sp.21868–150499%HE651920, AB099655
Stenotrophomonas sp.2890–151499% KF150351
Propionibacterium sp.111495–149999% CP003877
Agrobacterium sp.2816–89899% AY174112
Legionella sp.11883–93895–97%AM747393, JF779686
Pseudacidovorax sp.2951–100097–98% HQ834240
Azospirillum sp.186995% AP010946
Bacillus sp.1152197.50% JQ435679
Lactobacillus sp.1154099% EU855223
Leptothrix sp.1123897.80% AF385534
Pseudoxanthomonas sp.192998.70% DQ337597
Figure 3.

Rarefaction curves of the bacterial OTUs associated with AMF propagules and roots. OTUs were assigned at 98% of sequence similarity. The estimation of sample coverage for propagules 1–6 and the uncolonized roots were respectively 80.4%, 91.7%, 87%, 66.2%, 92.2%, 90.4%, and 100%.

Bacteria found to be associated with AMF propagules and within roots may either colonize the AMF cytoplasm or may be attached to their external surface, although both situations were reported to occur for some bacteria (Bonfante & Anca, 2009). To test whether bacteria were attached to the surface of the intraradical AMF propagules, microdissected mycorrhizal roots previously disinfected, and processed under sterile conditions were examined using SEM. Coccoid bacteria and biofilm-like structures attaching to the surface of AMF propagules were clearly visible inside the cortex of S. rugosa roots (Fig. 4). The size of these bacterial cells ranged from 0.5 to 1.5 μm.

Figure 4.

Scanning electron micrographs of dissected roots of Solidago rugosa showing AMF hyphae (H) and propagules (P) on which bacterial cells and biofilm-like structures (orange arrow) can be seen attached to their surface (panels a to e). Panels b and c are magnifications of the sections selected in the panels a and d, respectively. Panel f shows an AMF spore (S) isolated from the rhizospheric soil of S. rugosa roots sampled from the contaminated soil. Although the spore surface was washed several times with sterilized water, many microorganisms are visibly still attached to its cell wall surface.

Based on 16S rRNA genes, most bacterial OTUs retrieved from AMF propagules matched with bacterial taxa already reported to attach on the surface of spores and extraradical mycelia of AMF in soil (Roesti et al., 2005; Scheublin et al., 2010; Lecomte et al., 2011). Our data support that these bacteria are not only able to interact with the external AMF mycelia in soil, but they also colonize the AMF propagules inside roots. However, the effect of these bacteria on AMF remains unclear. Dominant bacterial taxa found in this study, belonging to genera Sphingomonas, Pseudomonas, Massilia, and Methylobacterium, were previously identified in hydrocarbon-polluted soils and were shown to be involved in biodegradation of PAHs (FDennis & Zylstra, 2004; Van Aken et al., 2004; Zhou et al., 2006; Ní Chadhain & Zylstra, 2010; Zhang et al., 2010). Interestingly, the nondominant taxa found here, which belong to genera Bosea, Brevundimonas, Bradyrhizobium, and Paenibacillus, have been reported elsewhere to improve mycorrhizal colonization of roots and plant nutrient uptake (Frey-Klett et al., 2007; Tarkka & Frey-Klett, 2008; Bonfante & Anca, 2009). The nondominant taxa we have found have also been considered to be members of a group called mycorrhiza helper bacteria (MHB), which includes phosphate-solubilizing bacteria (PSB) and nitrogen-fixing bacteria (Garbaye, 1994; Marschner & Timonen, 2006). The relationship between MHB and AMF has not yet been investigated in detail, but it has been suggested that MHB obtain their carbon resources from AMF hyphae (Bonfante & Anca, 2009) and that in return, these bacteria produce signaling compounds that can enhance the AMF stimulation of root exudates (Barea et al., 2005). It has also been reported that Paenibacillus sp. can support the growth and sporulation of the AMF Rhizophagus irregularis (formerly named G. intraradices), independently from the presence of the plant (Hildebrandt et al., 2006), although this report was controversial. Our results also support that intraradical AMF propagules can also harbor obligate endobacteria living inside AMF spores. For example, the clone G15GN sequenced from the propagule 1 (accession number KJ809239) showed a sequence similarity of 97% with an obligate endobacterium (accession number FJ984641) found in the cytoplasm of an AMF species (Naumann et al., 2010).


In this report, we described AMF diversity and associated bacteria from individual intraradical propagules isolated from roots of plants spontaneously growing in sediments of an extreme petroleum hydrocarbon-polluted basin. Based on WGA, cloning, and sequencing on individual AMF propagules, our results showed that intraradical propagules of AMF harbor a highly diversified bacterial community. However, further investigations will be needed to determine the functions and putative roles of these bacteria in mycorrhizal symbiosis and in phytoremediation.


This project was supported by the Genome Canada and Genome Quebec funded Genorem project which are greatly acknowledged. We thank Petromont Inc. (ConocoPhillips Canada) for allowing us to access to the Varennes field site. We also thank Franck Stefani for support in sequence analyses and Karen Fisher-Favret for comments on the manuscript.