Archaea colonising defined compartments of Scots pine Suillus bovinus or Paxillus involutus mycorrhizospheres developed in forest humus-containing microcosms were investigated by nested polymerase chain reaction (PCR), cloning, restriction fragment length polymorphism (RFLP) and sequencing. Archaea representing six RFLP groups were detected in the system. Sequence analysis of clones representing the different RFLP types confirmed the presence of novel Finnish forest soil Crenarchaeota. Archaeal sequences were identified from mycorrhizas of both P. involutus and S. bovinus, at the margins of the external mycelium and in uncolonised humus but not from non-mycorrhizal short roots. Fungal and compartment-specific crenarchaeal occupation of mycorrhizospheres is discussed in relation to bacterial community distribution in similar systems.
Archaea are distributed over an exceptional variety of habitats ranging from forest soils and fresh water lakes to hot springs and the deep sea trenches [1–5]. Due to the limited success in isolation of Archaea, which is at present restricted to Euryarchaeota and a few thermophilic Crenarchaeota, very little is yet known about general archaeal species richness and physiology [6,7]. However, PCR amplification and phylogenetic analysis of the archaeal 16S rRNA gene allows detection of Archaea in diverse habitats. The abundance of crenarchaeal 16S rRNA in both cultivated and native field soils has been estimated to be between 1 and 2% of the total 16S rRNA .
The first archaeal 16S rRNA genes detected in the humus layer of the soil in a boreal forest ecosystem were confirmed to belong to the division Crenarchaeota . Phylogenetic analyses placed these sequences into a dense novel cluster within a group of unculturable Crenarchaeota, which includes non-thermophilic Crenarchaeota from marine environments, lake sediments and soils. These sequences, termed Finnish forest soil Crenarchaeota (FFS), showed only partial (74–79%) sequence homology to the closest Archaea 16S rDNA sequences available at the time. Crenarchaeal sequences from soil in Wisconsin and the Amazon [1,5] were more closely related to each other than to the FFS sequences. It has also been shown that diversity of boreal forest soil Crenarchaeota changes in response to forest management practices such as clear-cut harvesting or clear-cut and prescribed burning .
Boreal forest soil is typically podzolised comprising defined layers of different physical and chemical attributes . The metabolically most active fraction of the boreal podzol is the uppermost 5–15 cm, the organic humus layer, which is heavily colonised by the mycorrhizal root systems, the mycorrhizospheres, of trees and under storey vegetation . The biomass of ectomycorrhizal external mycelium in forest humus has been estimated to be in the region of 820 kg ha−1 year−1. This extensive fungal component of the mycorrhizosphere is the site of soil nutrient uptake  and supply of plant-derived carbon to the soil [14,15]. Not surprisingly, the presence of mycelium of ectomycorrhizal fungi has been shown to influence the numbers and activity of bacteria in soil [16,17]. Numbers, identity and carbon utilisation patterns of culturable bacteria originating from specific mycorrhizospheric compartments of Scots pine Suillus bovinus and Paxillus involutus mycorrhizospheres (Fig. 1) have been examined in earlier studies using humus collected from the same site as in this study . It has been confirmed that factors such as soil type, fungal symbiont, and location in the mycorrhizosphere have a clear effect on bacterial community structure [18–21].
The detection of Crenarchaeota in the boreal forest humus  and their response to forest management  indicate that Archaea are likely to react to the presence (in intact forest) or absence (clear-cut forest) of tree rhizospheres and mycorrhizospheres. Crenarchaeota have only recently been detected in the rhizospheres of tomato and maize [22,23]. In tomato, Archaea and bacteria were found to prefer different parts of the rhizosphere . The mycorrhizal aspect was not, however, considered in either of these studies.
Thus far, the archaeal communities on the plant root-associated mycorrhizal fungi, which dominate natural sites, have not been studied. Our main aim in this microcosm-based study was to (1) investigate the presence of forest soil Archaea in rhizosphere and mycorrhizospheres of Scots pine and (2) provide preliminary data on potential variations of archaeal colonisation in the different functional compartments [13,14] provided by the mycorrhizospheres of Scots pine and two common ectomycorrhiza-forming fungi, S. bovinus and P. involutus, grown in forest humus.
2Materials and methods
Scots pine (Pinus sylvestris L.) seedlings were aseptically grown in 100-ml test tubes containing expanded clay pellets (Leca™) and inoculated with ectomycorrhizal fungal species, S. bovinus (L. ex Fr.) O. Kuntze (strain SBH1) and P. involutus (Batsch ex Fr.) (strain PIH), as described by Timonen et al. . Flat, 20×20 cm2 microcosms  were prepared containing fully mycorrhizal seedlings, as was earlier described in a related mycorrhiza–bacteria diversity study (Fig. 1) . The growth substrate in the microcosms was sieved (4-mm mesh) humus from a 70-year-old dry Scots pine forest stand in southern Finland (61°06′N, 23°50’E). The microcosms were placed in vertical stacks into cooled (13–16°C) incubation chambers . The seedling shoots were exposed to a 19/5 h (day/night) photoperiod, a photon fluency rate of between 170 and 300 μmol m−2 s−1 and respective day/night temperatures of 19/12°C. The seedlings grew for 3–4 months and the microcosms were watered three times a week with a distilled water spray.
Fully developed Scots pine (PS) S. bovinus (SB) and P. involutus (PI) mycorrhizosphere compartments were individually harvested from five replicate microcosms as follows. Samples of non-mycorrhizal short roots (SR), mycorrhizal short roots (MM) and external mycelium colonising humus (EM, Fig. 1), were removed using fine forceps. The humus-containing samples were: humus under mat of fungal (SB or PI) external mycelium with occasional fine hyphae (hyphosphere humus, HH), uncolonised (non-root or fungal) areas of humus in the microcosms (UH) and sieved bulk humus (BH) from stored material (+4°C) used for this study. Sampled amounts were 0.025 g (fw) for short roots, mycorrhizas and external mycelia and 0.25 g (fw) for the humus samples. Five samples of each sample type were taken from separate microcosms, except PI EM (two replicates, both combined of mycelium from several microcosms) and control BH (only three replicates). In this primary study, the non-colonised Scots pine SR (PSSR) and UH (FSUH) samples were considered equivalent in both mycorrhizal fungal (SB and PI) treatments and were thus sampling from five randomly chosen microcosms. The sample codes were PSPIMM, PSPIEM and PSPIHH for P. involutus mycorrhiza, external mycelium and hyphosphere humus, respectively, and PSSBMM, PSSBEM and PSSBHH for S. bovinus mycorrhiza, external mycelium and hyphal humus, respectively. Bulk humus was coded as FSBH (FS, forest soil).
DNA from the samples was extracted and purified using an UltraClean™ Soil DNA extraction kit (McBio Laboratories, Inc., USA) according to manufacturers instructions with the following modifications. The SR, MM and EM samples were manually pre-homogenised in the kit bead solution buffer in Eppendorf tubes using sterile glass rods and quartz sand. All samples were placed into the bead solution tubes and homogenised by bead beating two times for 1 min, 2000 rpm in a solid CO2 cooled Zell homogeniser (B. Brown Biotech International GmbH, Germany). The yield of total purified DNA was ca. 50 ng (SR samples), 150–250 ng (MM, EM samples) or 1 μg (HH, UH, BH samples).
Archaeal 16S rRNA gene sequences were amplified using a nested PCR approach as described by Jurgens et al. . Archaea-specific outer primers Ar4F  and Ar958R  were used to target ca. 1000 bp long fragments. As nested primers Ar3F  and a reverse primer, Ar9R , were used. The basic PCR reaction mix contained 1×reaction, 0.8 mM dNTP mix, 1.5/3.0 mM MgCl2 and 10 μM of each primer. 0.625 U/1.25 U of RedHot polymerase (Advanced Biotechnologies, UK) was then added in each reaction. 1 μl of each DNA extract containing 1 ng (SR), 2 ng (MM, EM), 2 or 20 ng (HH, UH, BH) DNA of samples was used in the first PCR and 1 μl of the first PCR was used in the nested PCR. The PCR was optimised based on the programme of Jurgens at al.  with annealing temperatures 50°C, 55°C and 60°C. The PCR was repeated six times for each replicate sample using a Gradient Mastercycler (Eppendorf-Netheler-Hinz GmbH, Germany). Diluted DNA of pure cultured Halobacterium salinarum and water were used as positive and negative controls, respectively. PCR reactions were checked for successful amplification by running ethidium bromide agarose gel electrophoresis .
The ca. 900 bp long archaeal 16S rDNA fragments amplified in the nested PCR were cloned with pGEM®-T Easy vector system II (Promega Corp., USA) according to the manufacturer's instructions. A maximum of 30 clones of each original sample were randomly chosen for screening.
For RFLP analysis, plasmids were isolated by alkaline lysis  and the insert amplified using the DynaZymeII polymerase PCR kit (FinnZymes Oy, Finland). The PCR products (10 μl) were digested overnight at 37°C with 3 U of three different restriction enzymes, HinfI, HspI and RsaI (Promega Corp., USA) in individual reactions. The digested products were electrophoretically separated in a 2.3% Metaphore agarose gel (BioWhittaker Molecular Applications, USA) containing EtBr (0.25 μg ml−1) at 5 V cm−2 for 3 h and imaged using a Bio-Rad Fluor-S MultiImager. The RFLP patterns of the different clones were visually evaluated.
One clone of each RFLP grouping available in each type of sample was chosen for sequence analysis. Since clones displaying digestion pattern A were commonly identified in the P. involutus mycorrhizal samples, two representatives of digestion pattern A were chosen randomly for sequencing. In summary, three clones representing pattern A, two clones of patterns C and D and one clone of patterns B, G and J were sequenced. The specific clones were isolated from 5-ml overnight cultures (Luria broth supplemented with ampicillin, 100 μg ml−1) using the Wizard®Plus SV Minipreps DNA purification system (Promega Corp, USA). Isolated plasmids were precipitated with ethanol, dried and diluted in water to a concentration of 0.1–0.3 μg μl−1. Sequencing was performed on an A.L.F. express DNA Sequencer (Pharmacia Biotech, Sweden), by cyclic sequencing (A.I. Virtanen Institute, University of Kuopio, Finland).
The acquired sequences were manually checked using Vector NTI® Suite programme package, version 6, for PC (InforMax® Inc., USA). Alignment of sequences was carried out as in Jurgens et al. . Briefly, the sequences were compared to all available archaeal 16S rRNA sequences in the EMBL database, using the FASTA programme. Phylogenetic analyses were made using the ARB programme package  (http://www.arb-home.de). New sequences were added to the ARB database and the ARB FastAligner utility was used for automatic sequence alignment, after which the sequences were corrected manually. The accession numbers of the sequences used for the trees are listed in Table 2. Maximum parsimony (PHYLIP package), neighbour joining with Jukes–Cantor distance correction method and maximum likelihood (fastDNAml) analyses, all included in the ARB package, were carried out for the Crenarchaeota sequences. The EMBL accession numbers of the identified Crenarchaeota are AJ419918–AJ419926.
Table 2. Codes for the Crenarchaeota 16S rRNA gene sequences used for the phylogenetic tree, origin, references and accession numbers
PCR fragments of ca. 900 bp were obtained from the P. involutus and S. bovinus mycorrhizal samples as well as from S. bovinus external mycelium, humus colonised with P. involutus or S. bovinus external hyphae, uncolonised microcosm and control humus (Table 1). The frequency of successful amplification was variable. No detectable PCR products were obtained from non-mycorrhizal short root samples under the described assay conditions. The positive control included in each PCR repeat gave a PCR product on every occasion and no products were generated in the negative control. The cloning success of the obtained PCR products was variable (Table 1).
Table 1. Sample types, numbers, experiments and RFLP results of the study
aNumber of original samples giving PCR product.
Location of sample
Number of samples
Number of screened clones
Number of RFLP types
RFLP types of clones
S. bovinus mycorrhiza
8, 12, 24
S. bovinus external mycelium
S. bovinus hyphal humus
P. involutus mycorrhiza
10, 21, 20, 8
A, C, G
P. involutus external mycelium
P. involutus hyphal humus
30, 20, 6, 7
A, C, D
sieved bulk humus
H. salinarum control
Pre-screening of randomly sampled clones by digestion with HinfI, MspI and RsaI generated six distinct RFLP groupings (Fig. 2, Table 1). Samples from mycorrhizospheres of P. involutus contained most variation in the RFLP patterns of archaeal sequences. Pattern A was found ubiquitously in the P. involutus samples and in the uncolonised soil. RFLP types G and B were found only in the P. involutus and S. bovinus external mycelium, respectively. Pattern J was only found in bulk humus.
3.3Sequence identification and phylogeny
The phylogenetic analyses (maximum parsimony, neighbour joining and maximum likelihood) of the representative RFLP group sequences generated trees with conserved topology. Of these analyses the maximum parsimony tree with uncultured Crenarchaeota is presented (Fig. 3, sequence codes in Table 2). In all the phylogenetic analyses, RFLP groups A, C, D, G and J were confirmed to belong to the same group as FFS (Finnish forest soil) Crenarchaeota detected earlier by Jurgens et al. , which form a specific cluster within the uncultivated Crenarchaeota (Fig. 3). The sequence analysis shows that the phylogenetic positioning of mycorrhizosphere Crenarchaeota was distant from the recently detected Archaea associated with tomato and maize rhizospheres (Fig. 3).
The mycorrhizospheric (MM, EM, HH) clones (patterns A, C, D, G; clusters M1 and M2) were located in clusters containing mainly FFS sequences detected from humus from clear-cut or clear-cut and burned forests (Fig. 3). The sequence giving restriction pattern J (cluster C1) was only found in control bulk humus and was closely related to sequences found earlier from intact forest humus. The sequence representing pattern B did not match any sequence in the gene bank and was excluded from the analyses. Patterns J and C fell clearly into their own clusters whereas A, D and G formed a single cluster where the sequence similarities were minimally 98.5% (cluster M2). The sequence homology between samples that gave RFLP pattern C was 94% (cluster M1). The lowest sequence homology (89%) in the mycorrhizosphere samples was found between the sample from S. bovinus external mycelium Archaea (PSSBEM10–2, pattern C) and the samples in sequence cluster M2.
This is the first report confirming that Crenarchaeota can be found in mycorrhizospheres of trees. The detection of mycorrhizosphere-associated Crenarchaeota extends the recent discovery of previously undetected members of this archaeal kingdom in the rhizosphere (rhizoplane and endorhizosphere compartments) of two important crop plant species, tomato (Lycopersicon esculentum) and maize (Zea mays) grown in agricultural soils [22,23].
In the analysis of the Scots pine S. bovinus and P. involutus mycorrhizosphere a trend was identified in the occurrence and distribution of Archaea that is linked both to locational ‘niches’ within the mycorrhizosphere and the mycorrhizal fungal species involved in the symbiosis. Under the assay conditions employed archaeal 16S rDNA could not be amplified from non-mycorrhizal short roots. Simon et al.  found that only a small number of archaeal cells were seen on young tomato roots, while archaeal colonisation of older roots was 10 times higher. The authors suggest that the relative abundance observed reflects the slower growth rates of Crenarchaeota. In the case of pine short roots the failure to detect Archaea may be due to the lack of older non-mycorrhizal short roots, as short roots are shown to be rapidly colonised by mycorrhizal fungi in a microcosm study using forest humus . Archaeal signals were detected from most other compartments of the investigated S. bovinus and P. involutus mycorrhizospheres.
The archaeal diversity appeared to be greater in the mycorrhizosphere compartments compared to the surrounding soil uncolonised by mycorrhizal hyphae. Altogether five different RFLP patterns (A, B, C, D and G) were obtained from the mycorrhizal and hyphal humus samples whereas only two (A and J) were obtained from the uncolonised humus and bulk humus. It has been estimated that between 15 and 28% of the net carbon fixation of a tree is needed for the maintenance of a mycorrhizal system [11,32,33]. Ectomycorrhiza formation alters carbon allocation patterns in the rhizosphere. Ectomycorrhization diminishes the leakage of carbon compounds in the immediate rhizosphere, and through growth and support of the external mycelium redistributes carbon particularly to the tips of the external mycelium. Indeed, it has been shown that in an ectomycorrhizal rhizosphere the rate of decrease in numbers of culturable bacteria is lower with increasing distance from pine roots than is common in non-mycorrhizal systems . Mycorrhizospheres are particularly amino acid- and organic acid-rich niches [13,34]. These acids have been shown to be preferentially utilised by mycorrhizosphere-associated microbial communities [18,21]. Organic acids are also known to be favoured by many culturable Archaea , which may explain enrichment of archaeal sequences in the mycorrhizosphere compartments.
The diversity and frequency of detection of archaeal 16S rDNA differed slightly in the mycorrhizal and hyphal humus samples of the respective Scots pine S. bovinus and P. involutus mycorrhizospheres. This supports the earlier observations by Timonen et al.  and Heinonsalo et al.  where differences in structure and function of microbial communities were related to factors such as soil type, fungal symbiont, and location in the mycorrhizosphere. Uncolonised pine short roots, mycorrhizas and external mycelium have already previously been shown by scanning electron microscopy to harbour distinctly different types of microbial populations . It could be argued that the bacterial monolayers observed in S. bovinus mycorrhizosphere compartments and P. involutus external mycelium  may have excluded Crenarchaeota in these compartments. In the in situ probing of bacterial and crenarchaeal colonisation of tomato roots, Simon et al.  identified a similar trend.
Most crenarchaeal sequences isolated and identified by their different RFLP patterns in this study showed only 89–93% similarity. However, the sequences with RFLP patterns A, D and G displayed 99.5–98.5% similarity and formed a mixed cluster (Figs. 2 and 3). This may reflect multiple species or genetic variation within a species in the samples. In a recent study of Archaea from the maize rhizosphere the Crenarchaeota showing over 98% 16S sequence homology were found to give divergent RFLP patterns when cut with restriction enzymes HaeIII, HhaI . The authors argue, however, that the sequences may represent different species according to an evolutionary ecology definition based on Ward's  studies concerning great adaptive radiation of cyanobacteria with over 97% 16S rRNA homology in a Yellowstone mat. Samples with RFLP pattern C in cluster M1 had only 94% sequence homology, which means that they are likely to represent different species inhabiting pine mycorrhizospheres even though the RFLP patterns did not separate the clones (Figs. 2 and 3).
The analysed sequences all fell into the clade formed mostly by the Crenarchaeota (Fig. 3) previously isolated from Finnish forest humus by Jurgens et al. . On the other hand, the crenarchaeal sequences obtained from the pine mycorrhizospheres differ strongly from the counterparts detected in the tomato and maize rhizospheres by Simon et al.  and Chelius and Triplett . Nevertheless, the differences linked to plant species composition and microbial/mycorrhizal status is also likely to have an effect on the population structure of the Archaea in rhizospheres.
In conclusion, soil crenarchaeal 16S sequences were identified from specific Scots pine mycorrhizosphere compartments developed in boreal forest humus. Fungal and soil habitat specificity was suggested in this study. Further studies using PCR-cloning sequencing approach tied to 16S in situ probing will shed more light on spatiotemporal community and population dynamics of Crenarchaeota in mycorrhizospheres developed in defined podzol horizons.
This study was funded by the Maj and Tor Nessling foundation and the Academy of Finland.