Applied Microbiology and Biotechnology

, Volume 79, Issue 3, pp 499–510

In situ bacterial colonization of compacted bentonite under deep geological high-level radioactive waste repository conditions

Authors

    • Department of Cell and Molecular BiologyGöteborg University Sweden
    • Center for Ecology, Conservation and Evolution, School of Environmental SciencesUniversity of East Anglia
  • R. Athar
    • Department of Cell and Molecular BiologyGöteborg University Sweden
Environmental Biotechnology

DOI: 10.1007/s00253-008-1436-z

Cite this article as:
Chi Fru, E. & Athar, R. Appl Microbiol Biotechnol (2008) 79: 499. doi:10.1007/s00253-008-1436-z
  • 167 Views

Abstract

Subsurface microorganisms are expected to invade, colonize, and influence the safety performance of deep geological spent nuclear fuel (SNF) repositories. An understanding of the interactions of subsurface dwelling microbial communities with the storage is thus essential. For this to be achieved, experiments must be conducted under in situ conditions. We investigated the presence of groundwater microorganisms in repository bentonite saturated with groundwater recovered from tests conducted at the Äspö underground Hard Rock Laboratory in Sweden. A 16S ribosomal RNA and dissimilatory bisulfite reductase gene distribution between the bentonite and groundwater samples suggested that the sulfate-reducing bacteria widespread in the aquifers were not common in the clay. Aerophilic bacteria could be cultured from samples run at ≤55°C but not at ≥67°C. Generally, the largely gram-negative groundwater microorganisms were poorly represented in the bentonite while the gram-positive bacteria commonly found in the clay predominated. Thus, bentonite compacted to a density of approximately 2 g cm−3 together with elevated temperatures might discourage the mass introduction of the predominantly mesophilic granitic aquifer bacteria into future SNF repositories in the long run.

Keywords

Clay bufferSpent nuclear fuelSubsurfaceÄspöGranitic aquifersGroundwater

Introduction

Plans to dispose spent nuclear fuel (SNF) in stable granitic geological formations are advanced in Sweden (www.skb.se). Several sites are currently being explored for their suitability to harbor a long-term SNF storage (Pedersen 2002). At the Äspö underground Hard Rock Laboratory (HRL), trial experiments have been running for over two decades to develop and test the relevant disposal methods (www.skb.se). At the final disposal phase, the waste will be encapsulated in metal canisters and deposited at a depth of 450−500 m in vertical holes dug into granite (Pedersen 2002). Compacted bentonite will be placed around the deposited canisters as one of the multiple barriers. Upon contact with groundwater, the bentonite will expand and develop a mechanical pressure when constrained by the surrounding rock walls leading to the sealing of the repository (Pedersen 2002).

Bentonite deposits are not sterile and mesophilic microbial lineages predominate (Boivin-Jahns et al. 1996; Fukunaga et al. 2005). The bentonite chosen for the Swedish repositories has low microbial counts and a low diversity predominantly associated to the gram-positive bacteria (Pedersen et al. 2000b). Commonly, at a compaction density of 2.0 g cm−3 (the equivalent volume occupied by 1 ml of groundwater), bentonite will expand when saturated with groundwater to occupy a volume of up to 18 times more than its dry weight volume in an open system (http://www.aquatechnologies.com/info_bentonite_clay.htm). This saturation reflects the complete hydration of its interstitial pore sodium and calcium cations with water molecules independent of space (Stewart et al. 2003). Theoretically, when compacted, bentonite saturated with water develops pore sizes about 100 times smaller than that of an average bacterium (Stroes-Gascoyne and West 1997). After compaction, microorganisms introduced into the clay could still be cultured albeit in lower numbers while some were effectively eliminated (Motamedi et al. 1996; Pedersen et al. 2000a, b). On the other hand, granitic aquifers contain a relatively vast, diverse, and largely gram-negative-mesophilic native microbial communities (Pedersen 1997 and 2001). Several of these assemblages include candidates that could negatively or positively influence the safety performance of an SNF repository (Pedersen 2002 and 2005). For example, the sulfate-reducing bacteria (SRB) estimated to be up to 104 ml−1 in granitic aquifers (Haveman and Pedersen 2002) could produce hydrogen sulfide that could induce the corrosion of the copper storage containers. Among the approximately 500 16S ribosomal RNA (rRNA) gene sequences currently available in GenBank from the aquifers at the Äspö Island, about 14% are related to approximately 13 mesophilic SRB genera and approximately 20 putative species. The acetogens in the aquifers provide acetate to the SRB through their metabolic processes thus indirectly fuelling the production of hydrogen sulfide within the aquifers (Pedersen 2001). Within this taxon, up to 104 each of autotrophic and heterotrophic acetogens per milliliter were reported in the groundwater present in the aquifers (Kotelnikova and Pedersen 1998; Haveman and Pedersen 2002).

Most studies that have evaluated the interactions of bacteria with bentonite buffers under repository conditions have handpicked and introduced microorganisms into the clay (Motamedi et al. 1996; Pedersen et al. 2000a, b). While this approach has enlightened us on bacterial survival in the bentonite at different compaction densities, water activities, and temperatures, the introduced taxa do not necessarily represent the groundwater microbial community. In addition, there is still the need to determine whether the diverse groundwater microorganisms could by themselves interact and colonize the buffer materials. Other studies reported on the bacterial diversity in the clay under certain repository conditions (Stroes-Gascoyne et al. 1997 and 2007). However, little attention has been placed on comparing the populations present in the immediate groundwater to those present in the clay. To bridge the above gaps, the presence of microbial markers specific to granitic aquifers culled from two decades of research literature and from this study was used in an attempt to derive new information on how indigenous groundwater bacteria interact with bentonite buffer materials in the long-term from simulated experiments conducted in the aquifers. We hypothesized that if successful colonization of the clay occurred, groundwater microbial communities present therein will predominate when the bentonite becomes fully saturated with groundwater. Due to an advantage in numbers and diversity and the predominance of gram-negative cells in groundwater as opposed to that in the bentonite, the groundwater microbial communities present in the clay will therefore be reasonably distinguishable.

Materials and methods

Study site

The Äspö underground HRL is located in southeastern Sweden on the Island of Äspö near the town of Oskarshamn. Located in granitic rock, it cuts from the surface down to a depth of 450 m below the Baltic Sea with a tunnel length of 3.6 km (Pedersen 2002). A series of tests spanning over more than 20 years include the investigation of the stability and performance of bentonite buffer under realistic conditions of heat, groundwater, and microbiology as will be expected in a multiple-barrier-engineered SNF repository. The bentonite used in this study was Wyoming MX-80 bentonite (American Colloid Co.), a buffer candidate in many SNF disposal concepts in the world. The characteristics of the Wyoming MX-80 bentonite composed largely of montmorillonite have been previously described (Motamedi et al. 1996).

The long-term test and the canister retrieval test

The experimental design included a geological location of approximately 450 m underground. Briefly, the long-term test (LOT) and canister retrieval test (CRT) consisted of 30-cm-wide blocks of bentonite compacted to a density of about 2.0 t m−3 placed around a copper tube that was supplied with a central electric heater which could provide temperatures of up to 150°C to the blocks (Fig. 1). The aim of the CRT was to demonstrate the retrievability of deposited full-scale SNF copper canisters (1 × 5 m; Sandén et al. 2007). Unlike the CRT, the LOT lacked scale in width and for safety reasons radioactivity was exempted from both experiments. The experiments free of preinoculated microorganisms were left to interact and equilibrate with the deep groundwater for a period of 5 years. When the installations were completely saturated with groundwater measured as described by Sandén et al. (2007), they were retrieved and the blocks containing 100% bentonite sampled onsite. Samples were collected from the LOT and CRT installations that were operated at 19°C and 35−55°C, and 67−110°C, respectively. The blocks were wrapped in sterile aluminum foil. Anaerobic samples were placed in an anaerobic box (COY Laboratory Products, Grass Lake, MI, USA) supplied with H2 (~2.5%), CO2 (~10%), and a ~90% N2 atmosphere immediately after sampling. Pieces of 1−2 g were separately collected for independent analyses by means of a hammer, a sterile chisel, and knife.
https://static-content.springer.com/image/art%3A10.1007%2Fs00253-008-1436-z/MediaObjects/253_2008_1436_Fig1_HTML.gif
Fig. 1

Schema of the overall design of the canister retrieval and long-term test for buffer materials adapted from Claytech (http://www.claytech.se/projects.htm). Diagram is not drawn to scale

Enrichment of sulfate-reducing bacteria and acetogens

SRB and acetogens were enriched from the groundwater and the clay (bentonite) in an anaerobic mineral medium (Widdel and Bak 1992) supplied with 14 mM sodium sulfate (final concentration) in serum flasks. Two types of enrichment media were used; one with 5 ml of 50% lactate solution added (final volume 1 l) as energy source and one without lactate. The pH was adjusted to 7.2 with either 1.0 M HCl or NaOH. Lactate is a preferred electron donor by microorganisms occurring in granitic aquifers including SRB and acetogens over other types of organic carbon sources (Pedersen 2001). The serum flasks and an additional set of sterile rubber stoppers were placed in an anaerobic box (COY Laboratory Products, Grass Lake, MI, USA) supplied with H2 (~2.5%), CO2 (~10%), and a ~90% N2 atmosphere. One gram of bentonite pieces each from the 19°C, 35−55°C and 67−110°C samples were added to separate serum flasks with lactate and without lactate in triplicates. The flasks were resealed with new stoppers and removed from the anaerobic box. To the enrichment cultures without lactate, H2/CO2 (80/20%) at two bars above atmospheric pressure was added as energy and carbon source for autotrophic SRB and acetogens. The enrichment cultures were kept overnight at room temperature for the bentonite to disperse into the medium followed by vigorous shaking until no clumps were visible and incubated at 30°C. Only mesophilic SRB and acetogens are known to occur in groundwater at the Äspö HRL (Pedersen 2001). The formation of black precipitates of sulfide was taken as an indication of the activity of SRB and confirmed by the formation of a brown precipitate upon the addition of 1 ml of the supernatant from each enrichment culture to 2 ml of 5 mM CuSO4. The formation of copper sulfide was detected according to Widdel and Bak (1992). To identify the most abundant SRB in groundwater close to the test site, samples were aseptically collected into sterile 150-ml serum bottles from the aquifers in boreholes KJ0052F01 and KJ0050F01. The samples were serially diluted to extinction in the above specified media and cultured at 30°C.

Culturable aerophilic heterotrophic bacteria isolation

Heterotrophic bacteria (HB) were cultured on agar plates according to a protocol commonly used for isolating HB known to occur in the aquifers at the Äspö HRL (e.g., Pedersen and Ekendahl 1990). Briefly, 1 to 2 g of bentonite pieces from the 19°C, 35−55°C, and 67−110°C blocks were initially suspended in 45 ml of the above SRB medium amended with 10 ml of 50% lactate. An aliquot of 0.1 ml was plated onto the agar plates and incubated at room temperature for a week. Either all the colonies on the culture plates were used for further analysis or if too many a maximum of five were selected dependent on morphology and pigmentation.

Groundwater sampling for genomic DNA preparation

Planktonic (groundwater) and solid-phase samples (precipitates of bacteriogenic iron oxides (BIOS)) were collected from a site located along the tunnel wall at a depth of 297 m denoted BRIC (BIOS reactor, in situ, continuous flow) and approximately 100 m above the engineered bentonite test sites. The BRIC site is unique in that it has a microbial mat dominated by BIOS with microbial communities distributed unevenly between the BIOS and groundwater phases (Anderson et al. 2006). Samples were collected from this site using 60-ml DNA-free syringes. One liter of the groundwater (planktonic) was extracted for further analysis while the BIOS samples were left to settle, the water drained off, and the procedure repeated until about 50 ml of a wet sample was obtained. Samples from the anaerobic aquifers that supplied the BRIC and from two boreholes (KJ0052F01 and KJ0050F01) located at a depth of 450 m close to the engineered experiment sites were collected as follows: 2 l of groundwater was initially drained before sample retrieval to obtain samples from the rock geological formation via PEEK tubing connected to core-drilled packed-off sections intersecting fractured zones containing water running under artesian pressure. All groundwater samples were filtered within minutes of collection through sterile 0.22-μm pore size filters of 47-mm diameter (Millipore, Stockholm, Sweden). The filters were treated with 4% paraformaldehyde and the captured cells were fixed according to the method by Hugenholtz et al. (2002). After fixation, the filters were immediately frozen at −20°C until DNA was extracted.

Genomic DNA extraction and amplification

Genomic DNA was extracted using a combination of protocols as shown in Table 1. Briefly, for the direct extraction of genomic DNA from the clay, 10-g samples were initially suspended in 450 ml of a 10× phosphate-buffered saline solution for 24 h at 4°C. Total DNA was extracted and purified from these samples and from the bentonite enrichments by use of the phenol–chloroform method described by Stroes-Gascoyne et al. (1997) with slight modifications. Briefly, 2% β-mercaptoethanol (Sigma Aldrich, Stockholm, Sweden) and 550 mM NaCl (final concentration) was added to the cell lysis buffer with an extended three alternating rounds of thermal shock (70°C and −70°C for 10 min each). A previous test of the protocol with the above modifications on commercial Wyoming MX-80 bentonite was found to improve the efficiency of recoverable bacterial polymerase chain reaction (PCR) amplifiable DNA by over a factor of four. DNA from bacterial colonies on agar plates, a pure Desulfovibrio aespoeensis culture (DSMZ10631), groundwater, and groundwater enrichments, was extracted using the QIAamp mini DNA extraction kit (Qiagen, Stockholm, Sweden) according to the manufacturer’s instructions. DNA from the BIOS material was extracted as previously described (Anderson et al. 2006). Purified DNA was stored at −20°C.
Table 1

Combinations of DNA extraction protocols used in this study on different samples (OPC, optimized chloroform/phenol protocol; Q, Qiagen mini DNA extraction kit; A, according to Anderson et al. (2006). DSMZ (Deutsche Sammlung von Mikroorganismen und Zellkulturen)

Samples

Enrichments

Heterotrophic plates

Groundwater

BIOS

Bentonite (direct)

Desulfovibrio aespoeensis

Lactate

H2 + CO2

colonies

(0.22 μm filters)

Wet weight

Dry weight

DSMZ10631

Bentonite

LOT and CRT

OPC

OPC

Q

OPC

Groundwater

KJ0052F01

Q

Q

KJ0050F01

Q

Q

Planktonic and solid phase

BRIC-planktonic

A

BRIC-solid phase (BIOS)

A

Pure culture

Äspö isolate

Q

PCR reactions were performed in 50-μl volumes containing 1 μl (~1−5 ng μl−1) of template DNA. The 16S rRNA and αβ subunits of dissimilatory bisulfite reductase (dsrAB) genes were PCR amplified with the universal bacterial primers 27f and 1492r and the DSRAB67F and DSRAB698R primers, respectively (Table 2). Cycling conditions were as follows: a 15-min enzyme activation step at 95°C, 35 cycles of 1 min at 94°C, 1 min at 55°C, and 1 min at 72°C with a final one-step extension of 10 min at 72°C at the end of the PCR reaction. For the DsrAB gene analysis, a 2-min extension step was instead used. The PCR products were analyzed on 1% agarose gel containing ethidium bromide (1 μg ml−1). DsrAB gene products (~1.9 kb) were excised from the gel and purified using the QIAquick gel extraction kit (Qiagen, Stockholm, Sweden) while the 16S rRNA gene PCR amplicons were cleaned with the QIAquick PCR purification kit (Qiagen, Stockholm, Sweden).
Table 2

16S rRNA and DsrAB gene primers used in this study

Primer

Sequencea

Gene target

Positionb

Specificity

Reference

DSRAB67Fc

5'-ACSCACTGGAARCACG-3'

DsrAB

63−67

General for SRB

Wagner et al. (1998)

DSRAB698Rc

5'-GTGTAGCAGTTRCCGCA-3'

DsrAB

698−702

General for SRB

Wagner et al. (1998)

DSRAB244eFd

5'-ACCTTCAAGGGCCCCATC-3'

DsrAB

239−244

Desulfovibrio aespoeensis

This study

DSRAB585eRd

5'-CCTGGGTGTGCACGATG-3'

DsrAB

585−590

Desulfovibrio aespoeensis

This study

GM5F-GC-Clamp

5'-GC-Clamp-CCTACGGGAGGCAGCAG-3'

16S rRNA

341−357

General for eubacteria

Teske et al. (1996)

907-reverse

5'-CCCCGTCAATTCCTTTGAGTTT-3'

16S rRNA

907−928

General for eubacteria

Teske et al. (1996)

27f

5'-AGAGTTTGATCMTGGCTCAG-3'

16S rRNA

9−27

General for eubacteria

Devereux and Wilkinson 2004

1492r

5'-TACGGYTACCGTTGTTACGACTT-3'

16S rRNA

1513−1492

General for eubacteria

Devereux and Wilkinson 2004

a\({\text{R}} = {\text{A}} + {\text{G, S}} = {\text{C}} + {\text{G, M}} = {\text{A}} + {\text{C, Y}} = {\text{C}} + {\text{T}}\)

bNumbering according to amino acid position (dsrAB) and E. coli number (16S rRNA gene), GC clamp and 5'-CGCCCGCCGCGCGCGGCGGGCGGGGCGGGGGCACGGGGGG-3'

cExternal primers used for PCR and sequencing of DsrAB gene.

dDesigned primers for internal sequencing of the 1.9-kb DsrAB gene from D. aespoeensis.

Cloning and sequencing

Purified PCR products of the 16S rRNA and DsrAB genes were cloned using the ®pCR2.1-TOPO cloning kit (InVitrogen, Stockholm, Sweden) and transformed into chemically competent TOP10 Escherichia coli cells according to the manufacturer’s instructions. Inserts were amplified with M13 vector specific primers supplied with the cloning kit and sequenced by using the BigDye Terminator v3.1 cycle sequencing ready mix kit (Applied Biosystems, Stockholm, Sweden) with primers 27f and DSRAB67F, respectively. The amplified DsrAB gene from D. aespoeensis was directly sequenced. The specific internal primers targeting the D. aespoeensis DsrAB gene listed in Table 2 were designed by comparing the D. aespoeensis DsrAB gene sequence derived from an initial sequencing with the DSRAB67F and DSRAB698R external primers to those of reference SRB in the ARB (arbor, Latin word for tree) software package for sequence data (Ludwig et al. 2004).

Denaturing gradient gel electrophoresis of 16S rRNA gene amplicons

16S rRNA PCR amplicons were obtained from the genomic DNA that was extracted from groundwater, groundwater enrichments, SRB, and acetogenic bacterial enrichments and from bentonite samples by use of the universal bacterial primers GM5-GC (forward) and 907 (reverse) in a pure Taq Ready-to-go PCR beads reaction (Amersham Biosciences, Uppsala, Sweden) according to Teske et al. (1996). The fingerprints were separated by denaturing gradient gel electrophoresis (DGGE; 6.5% acrylamide, 30−60% denaturing gradient (100% denaturant; 7 M urea and 40% formamide)) for 16 h at a constant voltage of 100 V and at 60°C. Following electrophoresis, the gel was stained with SYBR Green 1 for 1 h and viewed under a ultraviolet transilluminator. Prominent bands were excised from the gel and DNA was eluted in 50-μl MilliQ ultra pure water. The eluted DNA was reamplified with the same set of DGGE primers but without the GC clamp, purified with the QIAquick PCR purification kit, and sequenced as described above.

Phylogenetic analysis

DNA sequence homology searches were conducted on the Basic Local Alignment Search Tool (BLAST) server (www.ncbi.nlm.nih.gov) using the nucleotide–nucleotide BLAST algorithm. For both the DsrAB and 16S rRNA gene phylogenies, all regions of ambiguous homology, insertions, and deletions were excluded. Two thousand and 100 random bootstrap replications were generated for the 16S rRNA and DsrAB gene trees, respectively, to statistically test the reliability of the tree topology. For reconstructing the 16S rRNA gene phylogeny, the sequences were imported alongside their closest relatives into ARB and aligned using the fast aligner tool, manually checked and formatted against the secondary structure of the E. coli 16S rRNA. Approximately 400 aligned nucleotide positions complementary to all the sequences was bracketed and used for phylogenetic inference. A distance tree was constructed by neighbor joining in ARB and corrected with the Felsenstein correction model (Felsenstein 1988). For the DsrAB gene phylogeny, after importation into ARB, the sequences were automatically aligned in ClustalW alongside reference sequences. The alignments were manually checked and translated into amino acid sequences. A maximum parsimony tree was constructed using Protein Sequence Parsimony Method from approximately 300 concatenated amino acid positions. The sequences reported in this study can be accessed from GenBank under the accession numbers DQ839141–DQ839180, DQ903911–DQ903929, and AF492838. The ~500 16S rRNA gene sequences from the Äspö HRL together with the large number of environmental sequences currently available in GenBank provided a useful resource for the verification of the presence of groundwater bacteria in the clay samples. Reference papers reporting on microbial lineages that dominate in the aquifers at the Äspö HRL can be obtained from the review by Pedersen (2001).

Results

16S rRNA and DsrAB gene distribution in unenriched and enriched samples

Overall, the generated 16S rRNA gene clone library revealed that a substantial proportion (43.5%) of the retrieved sequences belonged to the bacilli, the only bacteria that were discovered in the DNA that was amplified directly from the clay (Table 3). Bacillus benzoevorans and Bacillus koguryoae were the only bacteria identified in the ≥67°C clay samples. Other identified clones were putatively related to the Paenibacillus with up to 99% identities. Bacillus megaterium a common granitic aquifer bacterium was found in the 35−55°C clay samples. When the samples from the 19°C blocks were enriched for acetogens, the clostridia and Sedimentibacter were detected (Fig. 2) with two clones closely related to the acetogenic Clostridium hydroxybenzoicum which was reclassified to Sedimentibacter hydroxybenzoicus (Breitenstein et al. 2002).
https://static-content.springer.com/image/art%3A10.1007%2Fs00253-008-1436-z/MediaObjects/253_2008_1436_Fig2_HTML.gif
Fig. 2

A 16S rRNA gene phylogenetic tree (distance) for representative clones and isolates from engineered bentonite. Numbers on branching nodes (50 and above) signify 2,000 bootstrap replications reported in percentages. The scale bar stands for the number of nucleotide substitutions per position. OTU denotes operational taxonomic units

Table 3

GenBank analysis of representative 16S rRNA gene sequences identified in compacted bentonite

Detection method

Identified relative

Sequence identity (%)

Classification

Temperature (°C)

Clones—direct DNA extraction

LOTb

Bacillus firmus (1)

98

Firmicutes

19

LOTd

Paenibacillus wynnii (1)

97

Firmicutes

19

LOTa, c, e_f

Bacillus benzoevorans (6)

100

Firmicutes

19

LOTg-h

Bacillus benzoev orans (2)

100

Firmicutes

67–110

LOTi

Bacillus litoralis (1)

98

Firmicutes

19

LOTj–k

Bacillus koguryoae (2)

100

Firmicutes

19 and 67–110

LOTl

Bacillus pichinotyi (1)

94

Firmicutes

19

CRT1–2

Bacillus pichinotyi (2)

93

Firmicutes

35–55

CRT3

Bacillus koguryoae (1)

99

Firmicutes

35–55

CRT25

Paenibacillus polymyxa strain GBR_1 (1)

99

Firmicutes

35–55

CRT18

Bacillus megaterium (1)

99

Firmicutes

35–55

CRT4r

Bacillus borophilicus (1)

99

Firmicutes

35–55

Heterotrophic isolates

LOT1, 1a_d

Pseudomonas Stutzeri (5)

99

γ-Proteobacteria 19

 

LOT3

Devosia riboflavina (1)

95

α-Proteobacteria

19

LOT2

Ornithinimicrobium sp. (1)

99

Actinobacteria

19

LOT4

Dietzia daqingensis (1)

99

Actinobacteria

19

CRT11

Microbacterium phyllosphaerae (1)

98

Actinobacteria

35–55

CRT11a

Streptomyces albidoflavus (1)

100

Actinobacteria

35–55

CRT12

Arthrobacter sp. KT1115 (1)

99

Actinobacteria

35–55

CRT13, 17

Pseudomonas stutzeri (2)

100

γ-Proteobacteria

35–55

SRB and acetogens enrichment

Lace1_2 (LOT)

Clostridium sp. EBR.02E_0599 (2)

92

Firmicutes

19

Lace3_5 (LOT)

Clostridium sp. (BN II) (3)

96

Firmicutes

19

Lace6 (LOT)

Clostridium hydrobenzoicum (1)

97

Firmicutes

19

Lace7 (LOT)

Bacillus megaterium strain GP S10 (1)

99

Firmicutes

19

Lace8 (LOT)

Sedimentibacter sp. JN18_A14_H (1)

97

Firmicutes

19

2Ben5 (LOT)

Clostridium hydroxybenzoicum (1)

98

Firmicutes

19

3Ben5 (LOT)

Clostridium sp. EBR-02E-0599 (1)

94

Firmicutes

19

4Ben5, 5Ben5 (LOT)

Desulfospo rosinus sp. (2)

99

Firmicutes

19

7Ben5 (CRT)

Desulfospo rosinus sp. (1)

99

Firmicutes

35–55

Numbers in parenthesis represent the number of similar clones or isolates

The 16S rRNA gene analysis could not detect SRB-related sequences in the clay (enriched and unenriched). Sulfate reduction was however indicated in the enrichments by the detection of hydrogen sulfide supported by darkened solutions and a rotten-egg smell. The presence of sulfide was consistent in all the enrichments including the samples from the ≥67°C blocks even when the SRB present were not identified. When the DsrAB gene common to all known SRB was targeted, surprisingly, DsrAB gene markers could not be detected in the DNA that was extracted directly from the clay nor its enrichments, confirming the 16S rRNA gene results that were negative for SRB. The DsrAB gene PCR reactions worked for positive controls (D. aespoeensis) providing proof that the PCR conditions were optimal. In addition, DsrAB gene markers were identified in all the groundwaters using the same PCR protocol that was applied on the clay samples. The detection of non-SRB 16S rRNA genes in the DNA isolated from the clay by PCR eliminated possibilities of poor DNA quality. The optimized protocol for extracting DNA from the clay ensured the acquisition of reasonable amounts of PCR-amplifiable DNA for our analysis.

The DsrAB gene clones from the groundwater and the BIOS samples were related to different SRB taxa previously known to occur in the aquifers while some were newly identified (Table 4 and Fig. 3). This analysis confirmed the presence of SRB in groundwater above and around the test site. DsrAB gene sequences related to D. aespoeensis were found in all the groundwater samples analyzed and represented the most predominant DsrAB genes detected in the groundwater (21.4%). DsrAB genes were not detected in the clay samples.
https://static-content.springer.com/image/art%3A10.1007%2Fs00253-008-1436-z/MediaObjects/253_2008_1436_Fig3_HTML.gif
Fig. 3

A phylogenetic tree (maximum parsimony) of the α-subunit of the Dsr gene (DsrA). Numbers (%) on branch points represent 100 bootstrap analyses. Only bootstrap values ≥50 are shown. Bold entries represent clones and isolates. BIOS and R represent clones from the solid phase and planktonic samples collected at the BRIC site, respectively, while BIN represents clones obtained from the anaerobic groundwater in the aquifer supplying the BRIC. Green and Red represent clones from the KJ0052F01 and KJ0050F01 boreholes, respectively. Scale bar stands for the number of amino acid substitutions per position

Table 4

BLASTN (GenBank) closest relatives of DsrAB gene nucleotide sequences

Known relative

Site

Depth (m)

Identity (%)

Desulfovibrio aespoeensis

KJ0052F01–KJ0050F01–BIOS–BRIC

297–450

93−96

Desulfomicrobium macestii

KJ0050F01

450

93

Desulfoarculus baarsii

KJ0052F01–anaerobic aquifer-BRIC

297–450

84−88

Desulfovibrio aminophilus

Planktonic–anaerobic aquifer-BRIC

297

80−81

Desulfobulbus rhabdoformis

BIOS–planktonic-BRIC

297

80

Desulfotomaculum luciae

Anaerobic aquifer-BRIC

297

96

Desulfofustis gylcolicus

BIOS–BRIC

297

84

Unknown SRB

BIOS–BRIC

297

Identification of SRB and acetogens by DGGE analysis

The DGGE fingerprinting revealed differences between the SRB communities present in groundwater and clay, unenriched and enriched groundwater, and clay and between the samples from the different temperature treatments (Fig. 4). For duplicate DNA extractions from the same samples, sequencing of the excised bands from different gels suggested that identical bands with the same mobility were due to the same sequence type. Band resolution for the unenriched clay was poor and gave poor sequence readouts (lanes B and I). The enriched clay samples generated stronger bands that could be successfully excised and sequenced. The fingerprints from the enriched 19°C clay samples generally indicated more bands (lane C) than for the samples recovered from the higher temperatures (lanes E and G) pointing to a more diverse community. Overall, the clostridia (e.g., band 7) present in the clay could not be found in groundwater (e.g., lane A) which revealed approximately ten bands in the gels while the Spirochaetes (e.g., bands 1 and 2) and D. aespoeensis (band 3) prominent in the groundwater (based on visual band intensity) were not identified in the clay. Bands that affiliated and had similar mobility patterns identical to those of Desulfosporosinus (bands 5, 8, 9, and 12) and S. hydroxybenzoicus (e.g., bands 4 and 6) were common in both clay and groundwater.
https://static-content.springer.com/image/art%3A10.1007%2Fs00253-008-1436-z/MediaObjects/253_2008_1436_Fig4_HTML.gif
Fig. 4

An example of a DGGE gel showing 16S rRNA gene PCR fingerprints of sulfate-reducing bacteria and acetogens. C and D and E and G represent the enriched clay samples from the LOT at 19°C and 35−55°C, respectively. B and I indicate lanes for the unenriched clay samples from the LOT at 35−55°C and the CRT at 67°C, respectively. Lane A represents unenriched groundwater sampled from borehole KJ0052F01 while lanes F (diluted to 10−3) and H and J (diluted to 10−4) represent the enrichments for groundwater from the same borehole

After maintenance of the groundwaters for a month in the SRB medium at different dilutions, the species in the samples changed with fewer bands visible in the gel (lanes F, H, and J) and with the number of bands inversely related to the dilution coefficient. Two prominent bands that were common in the various dilutions were identical to D. aespoeensis and Desulfosporosinus (bands 10, 11, 13, and 14). At the 10−4 dilution, only one dominant band was visible and had a 99% identity to D. aespoeensis (bands 13 and 14).

Aerophilic heterotrophic bacterial analysis

Replicate agar plates for aerobic heterotrophic bacteria cultures from the bentonite blocks operated at the same temperature generally revealed similar colony morphologies and pigmentations. The different isolates from both the LOT and CRT samples affiliated with major known bacterial phyla on the basis of their 16S rRNA gene sequences and could be grouped within the α-/δ-Proteobacteria and Actinobacteria (Fig. 3). Pseudomonas stutzeri was the most frequently detected Proteobacteria in the LOT samples albeit present in the CRT samples. A LOT isolate which had a 99% 16S rRNA gene sequence identity to a deep oil-bearing formation isolate (DQ224384) was further typed and deposited at the Culture Collection of the University of Göteborg Sweden (52604). Two P. stutzeri isolates were discovered in the CRT samples at 35–55°C that had a 100% sequence identity to a P. stutzeri species previously isolated from the aquifers at Äspö.

The survival and viability of the microorganisms in the clay determined by cultivability was close to exclusion at temperatures as high as 67–110°C. From samples run at ≥19°C and ≤55°C and compacted to the same density, viable non-spore-forming aerophilic microorganisms affiliated to the gram-negative bacteria could be isolated on the agar plates but not on the plates for the samples that were exposed to temperatures ≥67°C. Up to 82.6% of the total bacterial 16 S rRNA gene sequences identified in the clay were gram-positive of which 42.9% were derived from cultures. This indicated that quite a substantial portion of the microorganisms we report in the clay survived in a viable and culturable state. The 17.4% of the total number of bacterial 16S rRNA gene sequences recovered from the clay were related to the gram-negative P. stutzeri and Devosia riboflavina and mostly found at 19°C (Table 2).

Discussion

This study presents microbiological results from one of the longest SNF disposal experiments ever simulated in igneous bedrock. We explored the phenotypic (gram-stain biotype inferred from genetic typing), genotypic (16S rRNA and DsrAB gene distribution), and numeric differences between microorganisms common in groundwater and those present in bentonite to evaluate the interactions of subsurface microorganisms native to an igneous rock aquifer with the bentonite under repository conditions. Although the saturation of the clay with groundwater was not a homogeneous process at all fronts (Sandén et al. 2007), at full saturation of a block, an estimated 1.8 × 107 number of groundwater microorganisms could be introduced into 2.0-g cm−3 dense bentonite. This is because averagely 106 cells per milliliter are frequently counted in the aquifers (Pedersen 2001). Therefore, due to a large advantage in numbers and diversity as opposed to that in the bentonite (Pedersen et al. 2000b; Masurat 2006), the gram-negative mesophilic groundwater microorganisms will be expected to dominate in the clay samples.

Sulfate-reducing bacteria in the clay and groundwater samples

Cultivability of SRB was taken to represent viability and sulfide production to indicate activity. While sulfide production was sufficient to predict the presence of SRB in the ≥67°C samples, it also implied that even though present they were not prominent. Referring to the DGGE gel (Fig. 4), a faint band present in lane I that was obtained from the ≥67°C unenriched clay sample had the same mobility as other bands found related to Desulfosporosinus. Previous SRB enrichment studies on commercial Wyoming MX-80 bentonite indicated the production of detectable amounts of sulfide in the enrichments despite a low SRB diversity and numbers; identification was generally difficult (Masurat 2006). Therefore, the detection of hydrogen sulfide conveniently establishes their presence and activity even when they cannot be identified.

DGGE is a powerful tool that offers possibilities of detecting small differences between microbial communities as well as changes between differentially treated samples (Muyzer and Smalla 1998). We showed by the DGGE analysis that at the 10−4 dilution D. aespoeensis predominated in the groundwater. At the −450-m depth at the Äspö HRL, approximately104 SRB per milliliter are common (Haveman and Pedersen 2002). This implied that at the 10−4 dilution an estimated one sulfate-reducing bacterium was present as inoculum. The results therefore suggest that the most likely widespread and most predominant SRB in the groundwater in contact with the buffer appeared to have failed to establish in the clay. Desulfosporosinus was discovered in the groundwater before and after enrichment as well as in the enriched clay samples representing only approximately <5% of the total SRB genera known to occur in the aquifers. Members of this genus form spores and are common in most environments where SRB have been previously identified including groundwater aquifers (Robertson et al. 2000). The detection of this organism adds new information to our overall knowledge of the SRB lineages occurring in granitic aquifers and capable of surviving in repository bentonite at temperatures of up to 55°C and possibly ≥67°C.

Acetogens in the clay and groundwater samples

More than two decades of research at the Äspö HRL is yet to reveal acetogens related to clostridia in the groundwaters existing in the aquifers. Previous analyses that were focused on acetogens reported the presence of mainly lineages related to the Eubacterium (Pedersen 2001) which we could not identify in the clay samples. We showed here that while the clostridia were commonly identified in the clay, this was not the case for the groundwaters irrespective of the analytical method that was used. Albeit these microorganisms have never been found in commercial Wyoming MX-80 bentonite, they might be part of the autochthonous flora in the clay given that they are closely related to the bacilli commonly found in bentonite (Pedersen et al. 2000a, b). We further present new information that suggested that Sedimentibacter species form part of the heterotrophic acetogenic microbial community present in the aquifers at the Äspö HRL and that they could survive in the repository bentonite (Table 3 and Fig. 4).

Aerophilic heterotrophic bacteria in clay samples

A P. stutzeri that had been previously identified in commercial Wyoming MX-80 bentonite (Pedersen et al. 2000b) was not found to align with any of the new sequences recovered in this study. An analysis of the microorganisms in an SNF buffer test in Canada revealed the presence of various P. stutzeri species (Stroes-Gascoyne et al. 1997). Those findings and those of this study indicated that P. stutzeri species possess exceptional properties that could guarantee their survival in commercial and repository bentonite exposed to mesophilic and thermophilic conditions. Our analysis further suggested that the P. stutzeri species commonly found in the granitic groundwaters at the Äspö HRL (Johnsson et al. 2006) could possibly survive in a viable and culturable state in the repository clay. B. megaterium which was present in the clay had previously been identified in some groundwater samples around the Äspö HRL (Pedersen 2001). However, in the latter study the bacilli community was not as diverse as those found in the clay in this study. Furthermore, previous studies showed that the bacilli dominate in MX-80 bentonite (Pedersen et al. 2000a, b). This might indicate that conditions in the clay favor their selection as opposed to the conditions in the aquifers. Heterotrophs such as P. stutzeri, Microbacterium, bacilli, and clostridia are known to influence the mobility of radioactive metals in the environment in different ways (Jonhsson et al. 2006; Madden et al. 2007; Essén et al. 2007; John et al. 2001; Nedelkova et al. 2007; Pollmann et al. 2006; Merroun et al. 2005). Given that several viable lineages within the above taxa were found surviving in the clay suggested potential bioremediation alternatives in the case of a canister failure under the extreme repository conditions.

In conclusion, the conditions in which the clay was treated appeared to retard the overall success of the predominantly gram-negative mesophilic groundwater microorganisms. The spore-forming gram-positive bacteria commonly found in the clay predominated, suggesting that the non-spore-forming microorganisms and gram-negative microorganisms predominant in granitic groundwaters might not form a prominent portion of the bacterial community that will develop in bentonite under repository conditions. The complete absence of the gram-negative microorganisms in the bentonite samples that ran at ≥67°C in contrast to their presence in the lower temperature blocks implied that their survival was most likely controlled by temperature and not by compaction density or the porosity of the bentonite. The results further indicated that temperature might to be a more effective tool for controlling bacterial survival as a whole in the clay than a bentonite compaction density of 2.0 g m−3. We showed that some bacteria prefer a clay habitat within a repository setting (e.g., the clostridia) while others only occur in groundwater (e.g., D. aespoeensis); interestingly, some bacterial lineages showed a potential to coexist with both the bentonite and groundwater (e.g., Sedimentibacter and Desulfosporosinus). Although SRB appeared not to predominate in the clay, they were sufficiently active to mediate sulfate reduction with the resultant production of hydrogen sulfide. Therefore, when making models to meet the long-term safety requirements envisaged for a successful geologic SNF repository, the activity of SRB must be taken into account due to the long geological time scales required for the radioactivity to return to background levels. Through this study, it became increasingly obvious which microbial lineages might survive the longest in an SNF repository which included the subsurface habiting lineages Desulfosporosinus, P. stutzeri, B. megaterium, and S. hydroxybenzoicus. Our data further suggested key microorganisms that could be useful for the modeling of biological processes such as sulfide production, bioremediation, and the transportation and accumulation of radionuclides in SNF repositories if they accidentally failed.

Acknowledgements

This work was funded by grants from the Swedish Research Council and the Swedish Nuclear Fuel and Waste Management Company to Professor Karsten Pedersen of Göteborg University Sweden. We are deeply indebted to Dr. Sara Eriksson and the staff of Microbial Analytics AB Sweden for the culture preparations.

Copyright information

© Springer-Verlag 2008