Application of cation-exchange membranes for characterisation and imaging ammonia-oxidising bacteria in soils

Abstract

A new approach, in which ammonia-oxidizing bacteria (AOB) are entrapped from soil onto cation-exchange membranes, was applied to identify terrestrial AOB by fluorescence in situ hybridization (FISH). An experimental hot spot of ammonia oxidation was developed by establishing a gradient of ammonium substrate (200 to <20 mg ) diffused through the cation-exchange membranes incubated in soil for 6 months. By this approach we were able to characterise and image indigenous AOB populations growing in heavily oil-polluted soil using FISH and sequence analysis of PCR-amplified 16S rRNA genes, respectively. The FISH results revealed that Nitrosospira-like AOB were dominant on the ammonium-enriched membranes incubated in the soil. Fourteen unique Nitrosospira-like 16S rRNA gene sequences belonging to clusters 2 and 3 were recovered from the soil-incubated membranes and from the soil, suggesting the importance of Nitrosospira-like AOB in the oil-polluted landfarming soil.

Cation-exchange membrane, Fluorescence in situ hybridization, 16S rRNA gene, Nitrosospira, Soil pollution

1 Introduction

In soil environments, the oxidation of ammonia to nitrite is primarily carried out by autotrophic ammonia-oxidizing bacteria (AOB). Therefore, AOB play a crucial role in the nitrogen cycle in soils [1]. Nitrification and AOB are considered to be sensitive to environmental pollutants [2–4]. However, the effects of soil pollution on the various AOB species are poorly known, because ecological studies of AOB in soils have been hampered by difficulties (low growth rate and biomass yield) related to the application of pure-culture methodologies. Cultivation-independent molecular biological techniques, such as specific amplification of 16S rRNA or amoA genes, using the polymerase chain reaction (PCR) combined with denaturing gradient electrophoresis (DGGE) and nucleotide sequencing, competitive- or real-time PCR, or fluorescence in situ hybridization (FISH) have expanded the potential opportunities for studying the distribution and diversity of AOB in soils [5–10].

Digital microscopy combined with FISH provides a convenient way to quantify and image AOB in habitats with large and actively growing AOB populations, such as sewage sludge, bioreactors or rice root biofilms [11–13]. In soil environments, where AOB often constitute a minor proportion of the total population of soil microorganisms [14], the applications of FISH are usually hampered by a high density of solid autofluorescense particles that interfere with the microscopy of low numbers of AOB target cells. The aim of the present study was to examine the AOB populations present at oil-polluted landfarming soil by applying a novel approach, in which AOB are entrapped from the soil particles onto cation-exchange membranes. To achieve this, an experimental hot spot of ammonia oxidation was developed on the surface of the membranes incubated in the soil in sealed bottles filled with aqueous ammonium salt. By this approach indigenous terrestrial AOB grown on the ammonium-enriched membranes were successfully characterised and imaged using sequence analysis of PCR-amplified 16S rRNA genes and FISH, respectively.

2 Material and methods

2.1 Study site

An oil-waste landfarming site located in southern Finland (60° 11′ N and 24° 60′ E) was used in the field experiment. Over 20 years of landfarming of oily waste products derived from the adjacent petroleum refinery has yielded an average oil concentration of 35–59 g kg⁻¹ soil. The soil had been fertilized for more than 10 years with high amounts of urea targeted to give a carbon: nitrogen ratio of 10:1 [15]. During the period of study, no oily waste sludges, lime or nitrogenous fertilizers were added to the soil. Table 1 shows the relevant physical and chemical characteristics of the applied (sandy loam) soil.

1

Physical and chemical parameters determined in the experimental soil

Parameter	Time of sampling
May 2001	November 2001
Moisture content (%)	22	28
Organic matter (%)	21.3	nd
	4.9	4.9
Total N (%)	0.44	0.41
–N (mg kg−1)	3	5.5
–N (mg kg−1)	155	73
P (mg kg−1)	16	15
K (mg kg−1)	84	83
Pb (mg kg−1)	56	nd
Hg (mg kg−1)	1.4	nd
Cr (mg kg−1)	82	nd
S (mg kg−1)	3090	nd

nd = no data.

2.2 Experimental setup

Glass bottles, volume 120 ml, were filled with 100 ml of NH₄HCO₃ (1 g l^−l, pH 7.9) or NaHCO₃ (1 g l^−l, pH 8.6) solution. The openings of the bottles were sealed with 7 cm²× 0.137 mm thick cation-exchange membrane (Selemion CMV, Asahi Glass, Tokyo, Japan), which is a homogeneous membrane with sulfonic acid groups as fixed charges [16]. It exhibits a pore size in the range of a few angstrom, thus enabling the transport of single and Na⁺ ions from the filling solutions in the bottles to the membrane outer surface. The bottles were placed at 15 ± 5 cm depth in the landfarming soil in early May 2001 with the membranes oriented downwards and pressed against the intact soil profile at an angle of 20°. This approach ensured minimum disturbance of the soil and direct contact between the membrane and the native soil. The soil adjacent to the membranes was sampled using a hand shovel before burying the bottles. The soil samples were homogenised by sieving with a 2-mm sieve (mesh) and stored in glass jars at 4 °C until analysis. In mid-November 2001, after incubation for 170 d, the bottles including the membranes and a profile of the adjacent topsoil were collected. At the time of collection a slight negative pressure was observed in the bottles, but both the membranes and the bottles were intact, and no major loss of the filling solution (<2% of the initial volume) was observed. However, when the concentrations of total-N, and were measured spectrophotometrically a loss of approx. 90% of the initial of the filling solution was observed. The remaining nitrogen in the bottles was recovered as .

2.3 Potential ammonia oxidation rate

Potential ammonia oxidation rate was determined (1) on cells growing on segments of 3 cm² of or Na⁺-enriched membranes incubated in the soil and (2) on the bulk landfarming soil sampled in May and November 2001. Cells were harvested from the or Na⁺ membranes in 3 ml of 1 mM phosphate-buffered saline (PBS) (g l⁻¹, NaCl, 8.0; KCl, 0.2; Na₂HPO_4, 0.2; NaH₂PO_4, 0.2; pH 7.1) by ultrasonication in an icebath for 25 min. Potential ammonia oxidation rate of the cell extracts (20 μl) was measured using 24-well microtitre plates filled with 2 ml of liquid mineral (LM) medium (17, pH 7.5] containing 0, 10, 100 and 200 mg and 10 mg KClO₃ l⁻¹. Allylthiourea-spiked (1 mM) samples served as negative controls. The plates were incubated for 24 h at room temperature in the dark, and at the end of the incubation aliquots of 500 μl were withdrawn from the wells and analysed spectrophotometrically for the presence of using sulphonamide and naphthyl-ethylene-diamide as reagents at 545 nm. Five grams (fresh weight) of soil were incubated in a 50 ml Falcon tube containing 20 ml of 1 mM PBS and 0, 28, 140 and 280 mg at room temperature in the dark for 24 h. KClO₃ (final concentration 10 mg l⁻¹) was added to the tubes to inhibit nitrite oxidation. To extract , 5 ml of 2 M KCl were added to the tubes after incubation. After centrifugation (5000 rpm, 15 min) the supernatant was filtered through Whatman 1 paper and analysed for as described above. Calibration was carried out with KNO₂ over a range of 0–100 μg l⁻¹. The potential ammonium oxidation activity was calculated from the slope of the nitrite accumulation curve. One way-ANOVA followed by Tukey’ s test was applied to test differences between potential ammonium oxidation rates at the four different concentrations of substrate. P < 0.05 is reported as statistically significant.

2.4 Enumeration of ammonia-oxidising bacteria

The most probable number (MPN) for ammonia-oxidising bacteria on the or Na⁺ membranes and in the soil were determined using 96-well microtitre plates [18] and LM-medium [17] containing 5, 50 and 100 mg of at pH 7.5. Cells growing on segments of 3 cm² of the or Na⁺ membranes were harvested in 3 ml PBS by ultrasonication in an icebath for 25 min. Five grams of soil (fresh weight) was extracted by shaking at 250 rpm with 45 ml of buffer (0.9% NaCl, 1 ml of 10.4% Na₅P₃O₁₀ and 100 μl 2% Tween 80) for 25 min. Cell extracts and soil suspensions were serially diluted twofold on three separate microtitre plates. Inoculated plates were sealed with water-saturated Whatman 3 MM paper, enclosed in plastic bags, and incubated for 10 weeks at room temperature in the dark. The presence of ammonium and nitrite oxidation was checked using a drop of diphenylamine reagent (0.2 g in 100 ml of concentrated sulphuric acid). The MPN counts presented in this study are expressed as the numbers of ammonia-oxidising bacteria cm⁻² membrane or g⁻¹ soil d.w. One way-ANOVA followed by Tukey’ s test was applied to test differences between MPN counts at the three different concentrations of . P < 0.05 is reported as statistically significant.

2.5 DNA extraction and PCR with 16S rDNA primers

DNA was extracted from the cells attached to 1 cm² of or Na⁺-enriched membrane using the Microbial DNA Isolation Kit (Mo Bio Laboratories, Solana Beach, CA, USA) following the manufacturer’ s suggested protocol. DNA was extracted directly from 0.5 g (fresh weight) of soil using a FastDNA SPIN Kit for soil (Bio 101, Inc., La Jolla, CA, USA) as specified by the manufacturer. Fragments of the 16S rRNA gene were amplified using nested PCR approach with primers described in Table 2. All primers used in this study were purchased from TAGC (Copenhagen, Denmark). For the first-round PCR, 1 μl of crude DNA extract was amplified using the β-AMO primers of McCaig et al. [19] in 50 μl total volume reactions with a PTC-100 thermocycler (MJ Research Inc., Waltham, MA, USA). The reaction mixture contained: 1× PCR buffer (10 mM Tris–HCl, pH 8.8, 1.5 mM MgCl₂, 150 mM KCl, 0.1% Triton X-100), 200 μM each nucleoside triphosphate, 20 pM of each primer and 0.5 μl Dynazyme II DNA Polymerase (Finnzymes Oy, Espoo, Finland). The reaction thermal profile was as follows: 94 °C for 5 min followed by 30 cycles of 40 s at 94 °C, 40 s at 55 °C, 120 s at 72 °C, and a final extension of 10 min at 72 °C. The PCR amplicons (1 μl) were used as template in nested DGGE-PCR using CTO primers [7]. The reaction mixture was as described above and the thermal profile was as follows: 94 °C for 5 min followed by 35 cycles of 30 s at 94 °C, 30 s at 55 °C, 45 s at 72 °C and a final extension of 10 min at 72 °C. Aliquots (5 μl) of the PCR products were electrophoresed (80 V, 30 min) in 1% agarose gels and visualized with ethidium bromide staining using standard electrophoresis procedures.

2

Oligonucleotide primers and probes used in this study

Name	Specificity	Sequence (5′–3′)	Target site	References
PCR primers
β-AMOf	Betaproteobacterial AOB	TGGGGRATAACGCAYCGAAAG	141–161	McCaig et al. [19]
β-AMOr	Betaproteobacterial AOB	AGAACTCCGATCCGGACTACG	1301–1320	McCaig et al. [19]
CTO189f A/B-GC	Betaproteobacterial AOB	CCGCCGCGCGGCGGGCGGGGCGGGGGCACGGGGGGAGRAAAGCAGGGGATCG	189–207	Kowalchuk et al. [7]
CTO189f C-GC	Betaproteobacterial AOB	CCGCCGCGCGGCGGGCGGGGCGGGGGCACGGGGGGAGGAAAGTAGGGGATCG	189–207	Kowalchuk et al. [7]
CTO654r	Betaproteobacterial AOB	CTAGCYTTGTAGTTTCAAACGC	632–653	Kowalchuk et al. [7]
FISH probes
EUB338	Bacteria	GCTGCCTCCCGTAGGAGT	338–355	Schramm et al. [22]
BET42	Betaproteobacteria			Schramm et al. [21]
NSO190	Betaproteobacterial AOB	CGATCCCCTGCTTTTCTCC	190–208	Schramm et al. [21]
NSV443	Nitrosospira spp.	CCGTGACCGTTTCGTTCCG	444–462	Schramm et al. [21]
NSM156	Nitrosomonas spp.	TATTAGCACATCTTTCGAT	156–174	Schramm et al. [21]

2.6 Denaturing gradient gel electrophoresis analysis of PCR-products and reamplification of DNA fragments

Denaturing gradient gel electrophoresis (DGGE) was performed with the Dcode™ Universal Mutation System (Bio Rad, Hercules, CA, USA). PCR products (20–25 μl) were loaded onto 6% (w/v) polyacrylamide gels in 1× Tris-acetate-EDTA buffer (pH 7.4). The gels had denaturing gradients of 35–57% (100% denaturant contains 7 M urea and 40% formamide) and were run at 150 V, at 60 °C for 4.5 h. After electrophoresis, the gels were stained for 30 min with SYBER Green I (10⁻⁴ dilution, Molecular Probes, Eugene, OR, USA), rinsed with distilled water and photographed under UV transillumination. Single bands from DGGE gels were cut out, crushed in Eppendorf tubes containing 20 μl of distilled water and kept overnight at 4 °C. Three μl of the eluted DNA was used as a template for PCR amplification with CTO primers (without GC clamp) and the thermocycling program as described above for DGGE-PCR.

2.7 Sequencing and phylogenetic analysis of DNA fragments from polyacrylamide gels

Prior to sequencing, 45 μl of each PCR product were purified with the QIAquick PCR purification kit and sequenced using Big Dye terminators with the API Genetic Analyzer 373 (Applied Biosystems, Foster City, CA, USA). All sequence chromatograms were analysed with the Staden Package (University of Cambridge, UK). The sequences were compared with those available in the EMBL database using a BLAST server and aligned using the CLUSTAL W package at the European Bioinformatique Institute (EBI; URL http://www.ebi.ac.uk, Hinxton Hall, Cambridge, UK). Alignment of the sequences (404 bp) was then checked manually using the GENEDOC program (v2.6.002), and phylogenetic analyses were performed using the PHYLIP package (v.3.57C) as described earlier by Galand et al. [20]. DNADIST was used to calculate the genetic distances with the Kimura-2 model. The closest relatives of unknown sequences were estimated from the distance matrix data, using FITCH with global rearrangement of branches and randomised species input order (3 jumbles). The dendrogram was verified by maximum likehood and parsimony methods. The partial 16S rRNA gene sequences reported in this paper were deposited at GenBank and have the accession numbers AY792975–AY792988.

2.8 Fixation of cells

Two cm² segments of soil-incubated membrane were suspended in 1 ml of 4% (wt/vol) freshly prepared paraformaldehyde solution in PBS at 4 °C for 16 h. The cells were then collected by centrifugation (13 000 rpm, for 5 min) and washed twice with PBS to remove paraformaldehyde. The resulting pellet was resuspended in 0.2 ml of 50% ethanol in PBS (vol/vol) and stored at −20 °C.

2.9 Fluorescence in situ hybridization

All hybridizations were performed as described earlier by Schramm et al. [21,22]. Oligonucleotide probes labelled with carbocyanine (CY3) were purchased from TAGC (Copenhagen, Denmark). The specific nucleotide sequence for each probe is summarized in Table 2. For hybridisations, fixed cell suspensions (2 μl) were loaded onto a gelatin-coated (0.1% (w/v) gelatin, 0.01% KCr(SO₄)₂) Teflon 8 well slide. The samples were air-dried and dehydrated by an ethanol series of 50%, 80% and 96% (vol/vol), respectively, for 3 min each. A 9-μl aliquot of hybridisation buffer (0.9 M NaCl, 20 mM Tris–HCl (pH 7.4)), 0.01% sodium dodecyl sulphate (SDS) and from 5% to 30% formamide ((vol/vol), [22]) and 1 μl of oligonucleotide probe (50 ng) were placed on each spot of fixed cells and incubated at 46 °C for 2 h in an equilibrated 50 ml Falcon-tube. After hybridisation the slides were transferred to prewarmed (48 °C) Falcon-tubes containing 50 ml of washing buffer (20 mM Tris–HCl, pH 7.4), 5 mM EDTA, 0.01% SDS and 56–225 mM NaCl [21] and incubated at 48 °C for 15 min. The washing buffer was removed by rinsing the slides twice with distilled water. The slides were then air-dried, stained with the universal DNA stain 4,6-diamidino-2-phenylindole (DAPI, 10 μg ml⁻¹) for 10 min in the dark in an icebath, rinsed again with distilled water and dried. The slides were finally mounted with Citifluor antifadent (Citifluor Ltd., Canterbury, UK) and the preparations were examined using a Leitz Diaplan epifluorescence microscope (EFM; Leica, Wetzlar, Germany) equipped with specific filter sets for 340–380 nm excitation and 430 nm (DAPI) or 550 nm (Cy3) emission. The image analysis software, ImageJ (freely available at http://rsb.info.nih.gov/ij/) was used to analyze the EFM images for detecting and eliminating background fluorescence. Four different membranes were analysed, and a minimum of 15 fields were counted per membrane. The counting of Cy3-labelled cells (>300 per field of view) revealed that the bacteria detached from the membranes were unevenly distributed among the solid aggregates, which limits the statistical accuracy of the procedure [23], and therefore comparisons of the probe-specific cell counts are only tentative.

3 Results

3.1 Potential ammonia oxidation and MPN counts

An experimental hot spot of ammonium oxidation was developed in a landfarming soil that had been used for over 20 years for disposal of oil-refinery waste sludges. The hot spot was constructed by incubation of aqueous NH₄HCO₃ (200 mg N l^−l) filled bottles sealed with cation-exchange membranes in the soil from May to November 2001. Bottles filled with NaHCO₃ solution served as controls. At the time of collection, the residual concentration of ammonia in the NH₄HCO₃ filled bottles was <20 mg , confirming that the membranes functioned properly for 170 d in the soil. Potential ammonia oxidation rate and MPN counts of AOB were determined on the soil-incubated membranes and in the landfarming soil. Potential ammonia oxidation rates in Fig. 1 show that the cells harvested from the enriched membranes responded readily on increasing concentrations of substrate up to 100 mg . No such increase in the potential ammonia oxidation activity was found on the Na⁺ membranes. The MPN counts (Fig. 2) showed that numbers of AOB (0.4 ± 0.3 × 10⁵ cm⁻² of membrane) were also high on the membranes up to a substrate concentration of 100 mg . In contrast, the MPN counts (Fig. 2) in the landfarming soil decreased (log-transformed data, Tukey, P < 0.05) when the concentration of substrate was increased. The same pattern was seen in the potential ammonium oxidation activity in the landfarming soil (Fig. 3), which responded to the increasing concentrations of substrate with a declining activity. The results suggested that indigenous AOB preferring low concentrations of ammonium prevailed in the landfarming soil, whereas AOB preferring mostly high concentrations of ammonium were enriched on the membranes. However, specific activities of AOB, calculated from potential ammonium oxidation rate and MPN-data in Figs. 1–3, indicate that AOB on the membranes preferring concentrations of ≤10 mg were highly active, with a rate of 8.4 pg N oxidized MPN-countable AOB⁻¹ h⁻¹ compared to that in the soil (1.1 pg N oxidized MPN-countable AOB⁻¹ h⁻¹, ≤28 mg ).

1

Potential ammonia oxidation rate on and Na⁺-enriched membranes incubated in oil-polluted landfarming soil for 170 days from May to November in 2001. Means ± standard errors are shown (n= 3). Different letters indicate significant differences (Tukey, P < 0.05).

2

MPN counts of AOB on and Na⁺-enriched membranes (cm⁻²) incubated in soil for 170 days and in oil-polluted landfarming soil (g⁻¹). Means ± standard errors are shown (n= 3). Different letters indicate significant differences (Tukey, P < 0.05).

3

Potential ammonia oxidation rate in oil-polluted landfarming soil determined in May and November 2001. Means ± standard errors are shown (n= 3). Different letters indicate significant differences (Tukey, P < 0.05).

3.2 Analyses of AOB diversity using DGGE and nucleotide sequencing

The compositions of the AOB populations on the and Na⁺ membranes and in the oily landfarming soil were characterised using DGGE analysis of partial 16S rRNA gene sequences amplified from the extracted DNA in a nested PCR approach. In the DGGE analysis, between two and seven distinct bands were detected with AOB-specific β-AMO and CTO primers from the membranes and from the soil (Fig. 4). No amplification product was obtained from the Na⁺ membranes using the same primers. The DGGE banding patterns obtained from the membranes were the same as those from the oil-polluted soil, but not all bands were present in each of the parallel samples. In the soil the banding patterns obtained in samples taken in May and November 2001 were similar, indicating the presence of stable AOB populations. Representative bands were excised from DGGE gels, reamplified with PCR using CTO-primers and sequenced. Nine AOB-like sequences (S73–S81) from the soil and five sequences (M82–M90) from the NH₄–N membranes were recovered. A database search revealed high similarity (98–99.8%) with known Nitrosospira sequences. The sequences (404 bp) were aligned, and alignments formed a basis for construction of the phylogenetic tree shown in Fig. 5. Since the topology of the tree was independent of the algorithms used, only the Fitch-Margoliash tree is shown. All Nitrosospira-like sequences recovered from the -enriched membranes and from the oily landfarming soil grouped with clusters 2 and 3 as defined by Stephen et al. [21]. Two identical Nitrosospira-like sequences were recovered from the soil (S75) and from the NH₄–N membrane (M87) in cluster 2. Similarly, two sequences in cluster 2 were recovered from the soil, one sampled in May (S74) and the other in November (S78). Furthermore, identical sequences grouping with Nitrosospira cluster 3 were found in soil taken in May (S73, S77), in November (S81) and from the membranes (M88). The results from phylogenetic analyses of the recovered sequences showed that Nitrosospira clusters 2 and 3 dominated the AOB populations in the oily landfarming soil, and that both of these clusters were also present in the AOB community enriched on the membranes.

4

DGGE analysis of PCR-amplified DNA extracted from oil-polluted landfarming soil and from membranes incubated in the soil. 16S rRNA gene fragments were analysed on a denaturing gel with 35–57% of denaturant. Lanes 1 and 8 are reference Nitrosospira sp. B6 (cluster 2). Lanes 2 and 3 are soil extracts sampled in May and November 2001, respectively. Lanes from 4 to 7 are four individual membrane extracts sampled in November 2001.

5

Fitch-Margoliash tree showing relationships of partial 16S rRNA gene sequences (404 bp) recovered from enriched membranes from oil-polluted soil and other selected sequences belonging to beta- and gamma-proteobacteria. The sequences recovered in this study are presented in boldface. The scale bar indicates 10 mutations per 100 sequence positions.

3.3 Imaging of membrane-associated bacteria with FISH

To identify and visualize AOB with FISH, cells were harvested from four individual soil-incubated -enriched membranes and analysed with five different Cy3-labelled oligonucleotide probes. A set of 16S rRNA-targeted hierarchical probes was used to identify the domain of Bacteria (EUB338), the Betaproteobacteria (BET42a), betaproteobacterial ammonia-oxidizing bacteria (NSO190), Nitrosospira spp. (NSV443) and Nitrosomonas spp. (NSM156), respectively. The results showed that 60 ± 11% of the cells that were stained with DAPI also gave bright positive hybridization signals with EUB338, indicating favourable probe penetration. On the membranes, more than 32% of the DAPI-stained cells hybridised with probe BET42a and approximately 21% gave maximum signals with the NSO 190 probe. These numbers suggest that betaproteobacterial AOB represented a major fraction of the bacteria that had grown on the membranes during incubation in the oily landfarming soil. The probes NSV443 and NSM156 were used to differentiate between the genera of Nitrosospira and Nitrosomonas within the betaproteobacterial ammonia-oxidizing. Fig. 6 shows an example of cells grown on the membrane hybridized with NSV443-probre. A mean of 11 ± 5% of the DAPI-stained cells exhibited staining with the NSV443 probe. No membrane-associated AOB could be detected with the Nitrosomonas sp.-specific NSM156 probe (detection limit 1% of DAPI-stained cells). The hybridisation results support the sequencing data, suggesting that Nitrosospira-like AOB were dominant on the surfaces of membranes incubated in the soil.

6

AOB cells grown on membranes identified by FISH. Bacteria were detached from the membranes after incubation for 170 days in oil-polluted soil and hybridized with Nitrosospira sp. specific NSV-probe. Identical microscopic fields show DAPI-signal (A) and the signal of Cy3-labelled NSV-probe (B). Bar 20 μm.

4 Discussion

In the present study, we utilized the essential property of cation-exchange membranes to selectively leach ammonium for activating and attracting AOB from heavily oil-polluted soil onto the membrane surfaces. In soils, natural hot spots of ammonification are present, which may activate indigenous AOB [25]. Addition of substrate (for example via hot spots) may also reduce the strong attachment of AOB to soil particles [26]. An experimental hot spot of ammonia oxidation was developed by establishing a gradient of substrate from 200 to <20 mg on the soil-incubated membranes during in situ exposure of 6 months (one growing season). The present results indicated that the AOB that accumulated on the -enriched membranes had up to eight times higher specific activities per cell (based on the calculations using data for potential ammonia oxidation rate and MPN-counts) than AOB in the soil. Thus, by applying the ammonium gradient it was possible to induce the activity of AOB in the oil-polluted landfarming soil in a manner similar to that found by applying ammonium salt to agricultural soils [10,27,28], and to entrap AOB from the soil particles onto the surfaces of membranes.

The 16S rRNA gene sequences recovered in the present study from the -enriched membranes and from the oily landfarming soil were Nitrosospira-like sequences belonging to clusters 2 and 3 [24]. Nitrosospira-like sequences belonging to clusters 2 and 3 are frequently retrieved from agricultural soils and grasslands [28,29] and from acid or forest soils [24,30,31], respectively, but no Nitrosospira-like sequence in clusters 2 and 3 has been previously reported from oil-polluted soils. Nitrosospira-like sequences in cluster 3 were reported mostly from soils near pH 7, whereas sequences found in cluster 2 favour acid conditions [32]. Furthermore, Nitrosospira cluster 3 strains are known to grow well in culture media containing high levels of [33]. In the present study, the applied landfarming soil was acid (pH 4.9) and low in (<6 mg kg⁻¹), but it had a 10-year history of heavy urea-fertilisation [15], which possibly can explain the presence of cluster 3 in the soil.

In total, we found 14 unique Nitrosospira-like sequences: four from the landfarming soil and four from the membranes affiliated to cluster 3, five sequences from the soil and one from the membrane affiliated to cluster 2. The recovered Nitrosospira-like sequences indicated that the AOB species attached to the enriched membranes were the true indigenous AOB species present in the oily landfarming soil. Four out of the five Nitrosospira-like sequences retrieved from the AOB populations of the membranes grouped with cluster 3, which was suggested to prefer high concentrations of ammonium substrate [33]. The MPN counts implied also that the fraction of bacteria preferring high ammonium was higher on the membranes compared to the landfarming soil, where the low ammonium-preferring bacteria prevailed. These two observations indicate that there was a shift in the AOB populations from the soil to the enriched membranes, in which Nitrosospira cluster 3 possibly became more abundant compared to cluster 2. However, since the 16S rRNA gene similarities among closely related Nitrosospira clusters are high [34,35], and because the MPN assay suffers from well-accepted problems [1,36] resulting from selective nature of laboratory media and incubation conditions, validity of this assumption is questionable.

Methods based on isolation of total DNA from soil and its use as a template for the PCR involves limitations such as insufficient recovery of nucleic acids and biases in PCR amplification and sequencing [37]. Therefore, we applied FISH in our membrane assay to identify AOB cells directly on the experimental hot spot of ammonia oxidation. The FISH results demonstrated that Nitrosospira-like bacteria were the predominant Proteobacteria, representing 11% of the DAPI-stained cells that accumulated on the membranes, thus confirming the results from the sequencing analysis. No Nitrosomonas-like bacteria were detected on the membranes (detection limit >1% of DAPI-stained cells), nor any Nitrosomonas-like sequences on the soil-incubated membranes or in the soil. The present results indicate that there was no significant selection for Nitrosomonas sp. over Nitrosospira sp., as has been reported to occur in ammonium-rich habitats [38,39]. This does not however exclude the occurrence of AOB from the genus Nitrosomonas in the applied soil or on the membranes, since microorganisms at low relative abundance could remain undetected by the PCR-based or FISH methods used in this study.

Formation of aggregates and the presence of surfaces were reported to be essential for the activity of AOB [25]. FISH analysis of the membranes revealed that AOB mainly associated with microcolonies attached to solid particles. The specific activities per MPN-countable cells indicated that the AOB on the aggregates were highly active compared to the AOB in the soil. Generally, the growth of viable cells increases cellular rRNA content [5,40], which in turn may increase the sensitivity of FISH staining and could explain success in visualizing the membrane-associated AOB by FISH against relatively high background fluorescence. FISH was successfully used for quantification of AOB in situ in various habitats [41]. Application of cation-exchange membranes however, does not allow direct quantification of the original AOB numbers in the soil, because various AOB populations in the soil may adhere and grow in different ways on the membrane surfaces. Nevertheless, the cation-permeable membranes tested in this study may prove useful for imaging AOB in soils by reducing the number of interfering autofluorescent soil particles in the specimen [14,42]. Further, the present study showed that in situ cultivation of AOB on membrane surfaces could be a valuable tool for entrapping and enriching AOB from habitats difficult for isolation purposes. Beyond the present study, the application of cation-exchange membranes in combination with the supply of specific substrates via soil-incubated reservoirs may also be applicable for the investigation of other microorganisms in soils, such as thiosulfate oxidizing [43], chromium reducing [44] or uranium reducing bacteria [45].

Acknowledgements

We thank Wilfried Blümke (GKSS, Geesthacht, Germany) for providing the cation-exchange membranes and Jenni Hultman (University of Helsinki, Lahti, Finland) for technical help with sequencing. This work was supported by the Neste Foundation (JK, MR) and the Academy of Finland (MR, Academy researcher grant, MSS, project no. 177321).