Comparative microbial ecology study of the sediments and the water column of the Río Tinto, an extreme acidic environment


Correspondence: Antonio García-Moyano, Department of Biology, University of Bergen, PO Box 7803, N-5020 Bergen, Norway. Tel.: +47 5588187; fax: +47 55584450; e-mail:


Due to its highly metalliferous waters and low pH, the Rio Tinto has shown its potential for modelling both acid mine drainage systems and biohydrometallurgical operations. Most geomicrobiological studies of these systems have focused on the oxic water column. A sequence-based approach in combination with in situ detection techniques enabled us to examine the composition and structure of the microbial communities associated with the suboxic and anoxic sediments along the river course and to compare them with the planktonic communities inhabiting the water column. The results obtained with the different approaches were consistent and revealed some major patterns: higher cell density and higher richness (75 vs. 48 operational taxonomic units) in the sediments than in the water column. The microbial communities were related but the river sediments appear to be enriched in certain populations, some of which have not previously been reported in the Rio Tinto basin. The differences detected between sampling stations along the river correlate with certain environmental parameters (e.g. iron concentration gradient). The biological and geochemical data show the importance of the sediments as representing a phase of particular high diversity, probably related to key metabolic processes within both the iron and the sulfur cycles.


Several studies concerning the microbial ecology associated with acid mine drainage (AMD) have been carried in recent years (Bond et al., 2000; González-Toril et al., 2003ab; Johnson & Hallberg, 2003, 2005, 2009). The main interest of these studies was to increase our knowledge of the processes controlling metal dissolution and acid generation (Johnson & Hallberg, 2003). This is especially important in improving biohydrometallurgical operations and in searching for environmental strategies to reduce metal-laden acidic waters (Johnson & Hallberg, 2005). The characterization of anoxic sediments from AMD environments are underrepresented in comparison with their aquatic counterparts (but see Lu et al., 2010; Sánchez-Andrea et al., 2011; Sanz et al., 2011), because most studies of these habitats have focused on aerobic iron-oxidizing microorganisms and have suffered methodological limitations. However, it is clear that such sediments have to be considered to understand the integrated microbial ecology of these peculiar extreme systems.

The Rio Tinto has demonstrated its potential as a model system for AMD and biohydrometallurgical operations (González-Toril et al., 2003ab; Malki et al., 2006). Previous work at Rio Tinto focused on describing the microbial ecology involved in the iron cycle operating in the water column (González-Toril et al., 2003ab) and the diversity associated with macroscopic growth products (García-Moyano et al., 2008). Recently, studies of the sediment of a specific sampling site in the Tinto ecosystem have been reported (Sánchez-Andrea et al., 2011; Sanz et al., 2011), but a systematic study of the microbial diversity associated with the sediments along the river iron gradient was missing and is required for a complete assessment of the microbial diversity of this ecosystem and to understand the operation of the system at a global scale. The aim of the present study was therefore to study the diversity of the microbial communities intimately associated with the sediments and their relationship to the planktonic communities on the water column, using qualitative and quantitative culture-independent methodologies, and their correlation with variations of the physicochemical parameters along the river.

Materials and methods


Samples were collected in May 2005 at 11 different stations located along the course of the Rio Tinto (Fig. 1). Water samples for DNA extraction were taken in sterile 1.5-L bottles from the surface. A volume of water was collected in duplicate and then fixed in 4% formaldehyde (final concentration). Regardless of the sedimentation pattern only the very first 2 cm of sediment was sampled. Sediment samples were collected at the same spot, assembled with a 10-mL cut-tip syringe by pulling the emboli up while the syringe was submerged in the sediment. The small corex was then placed into a sterile 15-mL Falcon tube. A second sample was placed in a solution containing 4% formaldehyde in minimal Mackintosh medium (Mackintosh, 1978). All samples were kept on ice until processed and treated gently to avoid any detachment of the cells from the sediment particles. Once in the laboratory, all samples were filtered, washed and then stored at −20 °C as explained below.

Figure 1.

Schematic map of Rio Tinto showing the location of the sampling stations. The configuration of the river led to the differentiation of three areas (origin 1, origin 2 and middle course). In each area, the localization of all the stations (RT1–RT11) sampled in this study is shown. The lower frame shows the position of the river in the Iberian Peninsula, and the upper frame is a detailed map of the headwaters section of the river comprising origin 1 and origin 2.

Physicochemical parameters

Water flow in each station was measured with a Global Water Flow Probe FP111 (Geotech). Data regarding the depth of the water column in each station were also collected. Data for water temperature, conductivity (Orion 122 conductimeter; Orion Research), pH and redox potential (Crison 506 Ag/AgCl reference pHmeter) and oxygen concentration (Orion 810 oxymeter) were obtained in duplicate at each station. Data for redox potential and oxygen concentration were specifically measured within the sediment in those locations where allowed by the riverbed. Chemical analyses of dissolved heavy metals were carried out as described previously (González-Toril et al., 2006). Total iron content and oxidation state were determined by the 2-2′-bipyridyl colorimetric method (Easton, 1972).

DNA extraction

Water samples were filtered through a polycarbonate Millipore 0.22-μm pore-size filter using a water pump. Filters were washed twice in TE buffer [10 mM Tris–HCl (pH 8.0), 1 mM EDTA) and then frozen at −20 °C until processed. For DNA extraction, filters were cut into small pieces with a sterile scalpel, and introduced into a Lysing Matrix Tube from Fast DNA Kit for Soil (Q-Bio Gene Inc., CA). To disrupt cells, three 30-s pulses of the FastPrep instrument (Bio 101) at speed 4.0 were performed. To eliminate the pore water and the microorganisms suspended in it, sediment samples were filtered and the filters were washed as previously described in order to remove non attached cells. The drained and washed sediment was then transferred to another tube using a sterile spatula and kept at −20 °C until extraction. Approximately 500 mg of sediment was introduced into the Lysing Matrix Tube and then processed as described above. Genomic DNA was purified by passage through a GeneClean Turbo column (Q-Bio Gene Inc.). Integrity was analysed by 1% agarose gel electrophoresis and the yield was quantified spectrophotometrically.

16S rRNA gene amplification

Amplification of nearly full-length bacterial and archaeal 16S rRNA genes was carried out with the primer sets 8F/1492R (Lane, 1991) and 25F/1492R (Reysenbach et al., 1992). All primers were synthesized by Isogen Bioscience BV (Maarsen, the Netherlands). Each 50-μL reaction tube contained 20–30 ng template DNA, 1× PCR reaction buffer (Promega Biotech Iberica, Spain), 2.5 μM each dNTP (Amersham Biosciences, UK), 2.5 mM MgCl2, 1 mg mL−1 bovine serum albumin, 500 mM of the forward and reverse primers and 0.025 U μL−1 of DNA Taq polymerase (Promega Biotech). PCR reactions were performed in a Perkin Elmer Thermocycler with the following conditions: initial denaturation at 95 °C for 5 min, followed by 25 cycles of denaturation at 95 °C for 1 min, annealing at 49 °C for the bacterial primer set and 52 °C for the archaeal primer set, and extension at 72 °C for 1 min. A final extension step at 72 °C for 10 min was performed. Positive and negative controls were used. Positive amplification was checked by conventional electrophoresis in 1% agarose gel and compared with a known size marker.

16S rRNA gene library construction and sequence analysis

Amplified 16S rRNA gene products (> 1400 bp) were cloned using the TOPO Ta Cloning Kit (Invitrogen, CA) and sequenced using a Big-Dye sequencing kit (Applied Biosystems) following the manufacturer's instructions. Partial sequences were assembled and the nearly full-length sequence was then imported to a database of over 50 000 prokaryotic 16S rRNA gene primary structures by using the ARB software package aligning tool ( ) (Ludwig et al., 2004). The rRNA gene sequence alignments were corrected manually and alignment uncertainties were omitted in the phylogenetic analysis. Phylogenetic trees were generated using parsimony, neighbour-joining and maximum-likelihood analyses with a subset of 406 nearly full-length sequences (> 1400 bp). Filters, which excluded highly variable positions, were used. In all cases, the general tree topology and clusters obtained were stable. A consensus tree was generated.

An ARB-generated distance matrix was used as input file to mothur v1.12.0 (Schloss et al., 2009). Sequences were clustered into operational taxonomic units (OTUs) for every possible distance. Rarefaction analysis and the Chao1 nonparametric diversity estimator (Chao, 1984) were applied to the clone library to estimate how completely the library had been sampled and to extrapolate to total sequence diversity. Estimated steepness values (angle θ) of the line tangent to the rarefaction were also determined by using the last point. Community structure was studied by estimating the similarity between communities based on membership and structure (Schloss et al., 2009). Taxonomic assignments were based on the silva 104 database (Pruesse et al., 2007).

Sequences obtained in this study have been deposited in the EMBL sequence database under accession numbers JF737859JF737929 and JF807634JF807641.


Water-fixed samples were filtered trough a 0.22-μm pore-size GTTP-type Millipore filter. Sediment particles were washed over a first filter to remove the fixative and the pore water. Drained sediment was then transferred and resuspended in minimal Mackintosh mineral medium (Mackintosh, 1978). A small aliquot of this mixture was then pipetted and filtered through a 0.22-μm GTTP-type Millipore filter and immersed in 0.8% low-melting-point agarose to avoid detachment of the sediment particles during incubation and washing. Hybridization and staining with DAPI (4′,6-diamidino-2-phenylindole) were carried out as previously described (Amann, 1995). Cy-3, Cy-5 and fluorescein isothiocyanate (FITC)-labelled probes for fluorescence in-situ hybridization FISH were provided by Bonsai Technology (Barcelona, Spain). Horseradish peroxidase-labelled probes and fluorochromes AlexaFluor488 and AlexaFluor534 were purchased from Molecular Probes (Invitrogen). Sequences and detailed information on the oligonucleotide probes is available at probeBase (Loy et al., 2007) and in previous studies (González-Toril et al., 2003ab; García-Moyano et al., 2007). Sediment samples were hybridized using the catalysed reporter deposition (CARD) method (Pernthaler et al., 2002). Citifluor mounting medium (Citifluor Ltd, UK) was added to preparations to avoid fluorescence fading. An Axioskop microscope (Zeiss, Germany) equipped with the proper filter set was used to visualize the FISH results. Cell counting was carried out when possible, as described previously (Kepner & Pratt, 1994). CARD-FISH results were further visualized under an LSM510 scanning confocal microscope (Zeiss) equipped with an Ar ion laser (458–514 nm) and two He/Ne lasers (543 and 633 nm). All images were recorded with a Planapochromat 63× (1.4; oil immersion) objective. Image processing was performed with the LSM510 package (v. 1.6).


Sampling sites and physicochemical analyses

Sampling locations were selected after a complete understanding of the hydrology of the river in order to cover most of its acidic course. The river basin can be divided into three different areas: origin 1, origin 2 and middle course (Fig. 1). Origin 1 corresponds to the river headwaters. Three different stations were sampled here: two sources with divergent physicochemical characteristics (RT1 and RT2) and their intersection (RT3). Two stations were selected for origin 2: RT6 corresponds to a source flowing from a hedge. After flowing for 500 m, the waters are retained in a 6-m-deep dam (RT8). The course of origin 1 and origin 2 merge close to the town of Nerva and flow together for almost 90 km until reaching the ocean at Huelva. In the middle course, three different stations were sampled (RT9, RT10 and RT11). A few kilometres beyond RT11, the river is under the tidal influence of the Atlantic Ocean, and therefore this area has not been taken into consideration here. At origin 1, there are several minor sources flowing from an ancient mining area; these sources have also not been included.

Data concerning the physicochemistry of the water column at each location are shown in Table 1. These parameters are not assumed to be identical in the sediments, but can be used as a reference, since they were not specifically measured within the sediments except for the dissolved oxygen concentration, which was measured in the stations where it was possible (Table 1). Dissolved oxygen concentrations were variable, with lower values in the sediments. pH remained roughly constant (average 2.8, SD 0.7) with the exception at station RT2 (pH 3.7). Standard redox potentials were always positive and over 500 mV, correlating with the iron oxidation state. Conductivity values were also variable and correlated positively with metal concentrations, especially iron (Table 1). Water temperature was variable and related to air temperature (data not shown).

Table 1. Physicochemical parameters measured at each station
Sampling stationDepth (m)Flow (L s−1)pHEhwater (mV)Ehsediment (mV)Conductivity (mS cm−1)DOwater (mg L−1)DOsediment (mg L−1)Fet (mg L−1)Fe2+ (mg L−1)Zn (mg L−1)Cu (mg L−1)Ni (mg L−1)S (mg L−1)Mn (mg L−1)
  1. The table shows the average values for depth, water flow, pH, standard redox potential (Eh), conductivity, dissolved oxygen (DO) and concentration of heavy metals measured in the water column and within the sediments of each location. Iron speciation determined by colorimetric methods is shown. n.a., not analysed; n.d., below the detection limit of the technique; s, shallow water.

RT1s32.0665n.a.30.53.7n.a.15 206236920.420.411.315 042.349.8

Iron concentration decreased along the course of the river. The highest concentrations (> 15 g L−1) were measured at origin 1, station RT1 (Table 1). Waters from origin 2 had lower iron concentrations. Along the middle course, iron concentration decreased progressively downstream due to the dilution effect of neutral tributary waters and the consequent hydrolysis which maintains the pH of the river at a constant value (González-Toril et al., 2003ab).

Clone library analysis

Positive amplification and cloning results were obtained with a bacterial-specific primer set in all the samples from water and sediments. Archaeal amplicons were only obtained and subsequently cloned in samples from RT9. A total of 487 clones were obtained from water and sediment samples. Chimeric assemblies were detected in 48 clones, which were discarded from subsequent analysis. The remaining clones were clustered into OTUs, based on 97% sequence similarity. The water library contained 173 clones grouped in 48 different OTUs, while the 218 clones in the sediment library grouped in 75 OTUs. Regarding the community structure, 13 OTUs appeared to be common to both the sediment and the water libraries at the same level of divergence. Rarefaction curves were obtained for the two libraries (Fig. 2). A second-order polynomial function was adjusted (R2 > 0.97) for each curve. The tangent lines intersecting the last point of the curves were calculated. The slopes of these lines (θ = 1.1° and −0.9° for the sediments and the water column, respectively) indicate that both curves approach saturation. Comparison of the distribution of OTUs (based on a 3% cut-off value) between the water column and the sediments along the river is shown in Fig. 3. Estimated richness based on the Chao1 index resulted in 89 OTUs [95% confidence interval (CI) 72–180] for the water library and 177 (95% CI 96–240) for the sediments library.

Figure 2.

Rarefaction curves for water column and sediment clone libraries. Sequences were resampled and sorted into OTUs 1000 times, based on a 0.03 cut-off value.

Figure 3.

Cluster analysis comparison between the water column and sediment samples at each station. OTUs were obtained at a cut-off value of 0.03.

Phylogenetic assemblages

The phylogenetic position of representative clones obtained in this study (Table 2) is shown in Fig. 4. Concerning Bacteria, 54% of the clones from the water column were affiliated to the chemolithotrophic iron-oxidizing bacteria of the genera Leptospirillum and Acidithiobacillus, which were retrieved in samples from most of the stations in all three areas. Heterotrophic members of the class Alphaproteobacteria, belonging to the genera Acidiphilium and Acidisphaera, were preferentially detected in water samples from origin 2 (stations RT6 and RT8) and the middle course (RT9, RT10 and RT11). Clones identified as members of the genus Ferrovum were retrieved from stations RT2 and RT6. Several clones close to uncultured bacterium WJ2 (Gammaproteobacteria) were obtained from RT6 water samples. Within the phylum Firmicutes, several clones clustering within Clostridiales were retrieved from RT9 water samples. Clones related to the actinobacteria Ferrimicrobium were detected in the water column in samples from stations RT3 and RT8. Moreover, some sequences related to the bacterium KP1 (Acidobacteria) were obtained in samples from station RT8.

Figure 4.

Phylogenetic affiliation of 16S rRNA gene sequences from Río Tinto. Representative clones from both the water column and the sediments for different phylotypes were selected. A maximum-parsimony tree was selected as a consensus tree after reconstructing the phylogeny by using different algorithms, substitution models and filters. Related isolates or environmental sequences were selected for comparison purposes. Clones from Rio Tinto are tagged with the sampling station number and the letter W for water (in blue) or S for sediment (in red). Bar, 10% estimated phylogenetic divergence.

Table 2. Closest blastn relative of representative 16S rRNA gene clones retrieved from Rio Tinto samples
StationCloneaClosest relative by blastnAccession no.Percentage identity
  1. a

    Clones are tagged with W or S at the end of the name when retrieved from water samples or sediments, respectively.

RT1RT1-ant05-b02-WAcidithiobacillus ferrivorans NO37TAF37602099
RT1-ant05-a05-WLeptospirillum ferrooxidans DSM 2705TAF35683799
RT1-ant07-a04-SAcidithiobacillus ferrivorans NO37TAF37602099
RT1-ant07-e06-SBacterium WJ2 (Gammaproteobacteria)AY09603292
RT1-ant07-d01-SAcidisphaera rubrifaciens HS-AP3TD8651296
RT1-ant07-c06-SFerrimicrobium acidiphilum T23TAF25143699
RT1-ant07-a06-SLeptospirillum ferrooxidans 49879AF35683299
RT2RT2-ant09-f06-WFerrovum myxofaciens’ PSTREF13350898
RT2-ant07-b08-SAcidocella sp. NO-12AF37602196
RT2-ant07-a11-SBacterium Lie1-3 (Neisseriaceae)DQ98455891
RT2-ant07-f11-SAcidithiobacillus ferrivorans NO37TAF376020100
RT2-ant07-a10-SBacterium WJ2 (Gammaproteobacteria)AY09603299
RT2-ant07-a07-SUncultured (Xanthomonadales)EF44626999
RT2-ant07-d11-SUncultured (Cyanobacteria-clade WD272)FJ22831999
RT3RT3-ant11-a06-WLeptospirillum ferrooxidans 49879AF35683299
RT3-ant11-d05-WFerrimicrobium acidiphilum T23TAF25143699
RT3-ant08-d02-SFerrimicrobium acidiphilum T23TAF25143699
RT3-ant08-a03-SLeptospirillum ferrooxidans 49879AF35683299
RT6RT6-ant09-d12-WAcidisphaera rubrifaciens HS-AP3TD8651292
RT6-ant09-d10-WBacterium WJ2 (Gammaproteobacteria)AY09603299
RT6-ant09-a11-WFerrovum myxofaciens’ PSTREF13350899
RT6-ant09-b11-WBacterium CS11 (Acidimicrobiaceae)AY76599999
RT6-ant09-a10-WLeptospirillum ferrooxidans SyAF35683999
RT6-ant10-e12-SDesulfosporosinus sp. LauIIIAJ30207897
RT8RT8-ant10-h06-WAcidiphilium rubrum ATCC 35905TD30776100
RT8-ant10-b03-WFerrimicrobium acidiphilum T23TAF25143699
RT8-ant11-a09-WLeptospirillum ferrodiazotrophum’ UBA1EF06517898
RT8-ant10-a01-WLeptospirillum ferrooxidans SyAF35683999
RT8-ant02-d03-WBacterium CH1 (Acidobacteria)DQ35518497
RT8-ant03-h07-WAtopococcis tabaci CCUG48253TAJ63491799
RT8-ant03-e05-WUncultured (Chloroflexi-clade JG37-AG-4)FJ22823999
RT8-ant03-c06-WUncultured (Gamma-clade RCP1-48)HQ73069398
RT8-ant03-g05-WUncultured (Acidimicrobiaceae)AB25479598
RT8-ant02-b02-SAcidiphilium sp. NO17AF37602699
RT8-ant02-e02-SAcidisphaera rubrifaciens HS-AP3TD8651296
RT8-ant02-f03-SBacterium WJ2 (Gammaproteobacteria)AY09603292
RT8-ant02-b01-SLeptospirillum ferrooxidans SyAF35683999
RT9RT9-ant02-h08-WAcidiphilium sp. DC2EF55624296
RT9-ant02-a09-WAcidithiobacillus ferrooxidans YTWDQ06211699
RT9-ant02-a12-WUncultured bacterium (Clostridiales)EF40151297
RT9-ant02-b12-WUncultured bacterium (Clostridiales)EF40420099
RT9-ant02-f08-WLeptospirillum ferrooxidans SyAF35683999
RT9-ant05-c07-SMesorhizobium sp. E14AB26516099
RT9-ant04-g08-SAcidocella sp. NO12AF37602199
RT9-ant04-c09-SAcidithiobacillus ferrooxidans YTWDQ06211699
RT9-ant05-h10-SUncultured bacterium (Xanthomonadales)DQ46319497
RT9-ant05-g12-SAcidiferrobacter thiooxydans m-1T (Gammaproteobacteria)AF38730199
RT9-ant05-d09-SBacterium WJ2 (Gammaproteobacteria)AY09603296
RT9-ant05-a09-SClostridium quinii DSM 6736TX7674598
RT9-ant05-f08-SLeptospirillum ferrooxidans SyAF35683999
RT9-ant14-h08-SFerroplasma acidiphilum DSM 12658TAJ22493699
RT9-ant14-b09-SUncultured archaeon (Thermoplasmata)DQ30325496
RT10RT10-ant10-b05-WAcidiphilium sp. DC2EF55624299
RT10-ant10-e05-WUncultured bacterium (Rickettsiales)AJ86790397
RT10-ant10-c05-WAcidithiobacillus ferrooxidans YTWDQ06211699
RT10-ant06-b06-SAcidiphilium sp. NO17AF37602698
RT10-ant06-a01-SAcidocella sp. NO12AF37602199
RT10-ant06-a05-SAcidithiobacillus ferrooxidans YTWDQ06211699
RT10-ant06-e06-SBacterium WJ2 (Gammaproteobacteria)AY09603298
RT10-ant06-b03-SUncultured bacterium (Xanthomonadales)EF446269100
RT10-ant06-a03-SUncultured bacterium (Clostridiales)EF40156999
RT10-ant06-e02-SFerrimicrobium acidiphilum T23TAF25143699
RT10-ant06-a06-SLeptospirillum ferrooxidans SyAF35683999
RT11RT11-ant10-b09-WAcidiphilium sp. NO17AF37602699
RT11-ant10-e08-WAcidiphilium sp. DC2EF55624299
RT11-ant04-a04-SAcidiphilium rubrum ATCC 35905TD3077699
RT11-ant06-c11-SAcidiphilium sp. NO17AF37602696
RT11-ant06-d09-SAcidisphaera rubrifaciens HS-AP3TD8651296
RT11-ant06-a11-SBacterium WJ2 (Gammaproteobacteria)AY09603299
RT11-ant06-e11-SUncultured bacterium (Xanthomonadales)EF44626999
RT11-ant06-h10-SBacterium KP3 (Acidobacteria)AY76599299
RT11-ant06-c10-SBacterium WJ7 (Acidobacteria)AY09603497
RT11-ant06-e10-SUncultured bacterium (Acidimicrobiaceae)DQ90607695
RT11-ant04-c06-SBacterium CS11 (Acidimicrobiaceae)AY765999100
RT11-ant06-a09-SLeptospirillum ferrooxidans DSM 2705TAF35683799
RT11-ant06-a07-SUncultured bacterium (Planctomycetes)DQ90607898

The sediments revealed a higher level of diversity than the water column samples, at all sampling stations except RT6 and RT8 (Table 2, Fig. 3). Similarly to the water column samples, clones affiliated to chemolithotrophic iron-oxidizing bacteria of the genera Leptospirillum and Acidithiobacillus were retrieved from samples of most of the stations in the three areas and were again predominant in the clone libraries (Table 2). Clones related to the heterotrophic, iron-oxidizing gammaproteobacterium isolate WJ2 (Xanthomonadaceae) were detected in sediments from most of the sampling stations. Heterotrophic members of the class Alphaproteobacteria, belonging to the genera Acidiphilium, Acidocella and Acidisphaera, were mainly detected in the middle course stations. A higher level of diversity than that in the water column was also detected within the phylum Actinobacteria, where clones related to the heterotrophic, iron-reducing genus Aciditerrimonas, the heterotrophic iron-oxidizing, facultative iron-reducing genus Ferrimicrobium and the heterotrophic, iron-oxidizing acidobacterium isolate KP1 (AY765991) were identified. Within the phylum Firmicutes, some clones related to the genus Clostridium were identified in sediment samples from stations RT9 and RT10, and clones related to Desulfosporosinus were detected in the anoxic sediments of station RT6. Interestingly, the diversity observed within the phylum Acidobacteria, with sequences related to different isolates such as the heterotrophic, iron-reducing bacteria KP3 (AY765992), CH1 (DQ355184) and WJ7 (AY096034) were obtained from origin 2 (RT8) and middle course samples.

Sequences corresponding to groups not previously detected in the Rio Tinto ecosystem were also retrieved from RT2 sediment samples, including those related to the Neisseriaceae bacterium Lie1-3 (GU199448), the clade WD272 within the phylum Cyanobacteria and the Xanthomonadales bacterium CN40 (FN554395). The last-named was also detected in stations from the middle course (RT9, RT10 and RT11). Sequences within the Ectothiorhodospiraceae related to Acidiferribacter thiooxydans were recovered from station RT9. Sequences clustering with clade JG37-AG-4 (Chloroflexi) were obtained from water samples at RT8. Archaeal sequences were found only in the sediments and were obtained only at station RT9; all the sequences retrieved clustered within the class Thermoplasmata (Table 2).

Hybridization analysis

Total cell counting retrieved an average 105 cells mL−1 in the water column. Water from station RT8 showed the highest cell number (4 × 105 cells mL−1), while RT2 showed the lowest cell number (about 3.5 × 104 cells mL−1). Identification and quantification by hybridization was performed using FISH for water samples and CARD-FISH for sediments. No statistically valid quantification was acquired for sediment samples. Nevertheless, the main specific probes for each group were evaluated in terms of relative presence or absence in order to compare them with the water column samples. Sediment cell numbers were clearly higher than each respective water sample, with an average of 106 cells mL−1. A summary of comparative cell numbers of sediments and the corresponding water column is provided in Table 3 and some representative examples of fluorescence hybridization are shown in Fig. 5. Most of the cells detected by DAPI staining were also detected with the specific probe for the Bacteria domain (EUB338I-III) in both water and sediment sampless.

Figure 5.

Fluorescence micrographs for two hybridized samples. Epifluorescence micrographs of bacteria from Rio Tinto. (a) Laser scanning confocal photomicrograph of a sediment sample hybridized by CARD-FISH with a specific probe for the genus Acidithiobacillus (green) and with a specific probe for Acidiphilium (red). Autofluorescent eukaryotes (probably Euglenophyta) are in orange. (b1, b2) FISH. (b1) DAPI-stained cells from station RT9. (b2) Same field as b1 showing cells hybridized with probe LEP636 (Cy3-labelled) specific for Leptospirillum ferrooxidans. Scale bars, 5 μm.

Table 3. Comparative quantification using fluorescence in situ hybridization (%) of the main groups of microorganisms present in the water column (FISH) and the sediments (CARD-FISH)
  1. Relative estimation of the hybridization has been used for the sediments according to the following code: (−) below detection limit (< 1%); (+) testimonial or present in low numbers (1–10%); (++) significant abundance (11–30%); (+++) remarkable presence (31–60%); (++++) dominant (> 60%).

  2. a

    Also known as Ntspa712.


Positive hybridizations were obtained with the group-specific probes for Alphaproteobacteria, Betaproteobacteria and Gammaproteobacteria (ALF968, BET42a and GAM42a). Percentages detected with the GAM42a probe were also close to those obtained with the THIO1 probe, specific for the genus Acidithiobacillus. There was good correlation between the number of cells detected for Acidithiobacillus in the water column and sediments. High hybridization values were also obtained with the probe for the phylum Nitrospira (NTR712). Most of the hybridized cells showed vibrio or spirillum shape, characteristic of the genus Leptospirillum. Most of the cells that hybridized with probe NTR712 also hybridized with one of the specific probes for the genus Leptospirillum, most of them with LEP636, specific for Leptospirillum ferrooxidans, the main species of this genus detected in the Rio Tinto basin (García-Moyano et al., 2008). In this case there was also good correlation between the hybridization values of the water column and the sediments.

Quantification using the Archaea domain-specific probe ARCH915 was not applied as this probe matches nonspecifically with the genus Acidithiobacillus (González-Toril, 2002). Therefore, quantification was performed with specific probes for members of the class Thermoplasmata (FER656 and TMP654) that cover the archaeal diversity identified in the clone libraries. The highest value, about 3%, was obtained in RT9 in accordance with results from the water column. Some hybridization was observed for samples from stations RT1 and RT8.


Ferric iron plays an important role in the Rio Tinto ecosystem, controlling not only pH due to its buffering capacity (González-Toril et al., 2003ab) but also redox potential and the concentration of other ionic metals resulting from the oxidation of metal sulfides (Sand et al., 2001). Despite the apparent homogeneity of the river in terms of pH, some important differences can be observed along its 92-km course, for instance the existence of an iron concentration gradient as a result of its hydrolysis with the merging of neutral tributary waters along its course (Table 1). In fact, this iron gradient seems to be responsible for the differences in microbial populations that appear in the Tinto basin.

The river can be divided into three diverse areas: origin 1, origin 2 and middle course. The origin 1 area captures water from different sources, RT1 and RT2, which are dissimilar in terms of their physicochemical properties (Table 1) and probably influence the different microbial populations they host (Table 2). These differences could be a consequence of the geochemistry of the bedrock where they are sourced (Fernández-Remolar et al., 2003). Site RT1 resembles a bioleaching reactor fed with metal sulphides, where the concentration of ferric iron is kept high to facilitate mineral dissolution. Environmental conditions (temperature, pH, aeration) in a bioleaching tank are spatially uniform and constant over time (Rawlings & Johnson, 2007). Consequently, microbial diversity is generally limited to a small number of microorganisms (Wakeman et al., 2008). Moreover, once established, the composition of the consortium may remain relatively constant (Rawlings & Johnson, 2007). This sampling site has always presented high ferric iron concentrations with no seasonal variations (González-Toril et al., 2003ab; Aguilera et al., 2006). These high ferric-containing headwaters are mainly inhabited by specialized chemolithotrophic, iron-oxidizing bacteria Acidithiobacillus ferrooxidans and L. ferrooxidans. These bacteria comprise at least 70% of the planktonic biomass and are also present in the sediments at higher numbers (Table 3). These two microorganisms are adapted to the high ionic strength and oxidative stress (> 15 g L−1 Fe3+ at station RT1), as can be inferred from metatranscriptomic analysis of L. ferrooxidans (Parro et al., 2007). As most of the biomass at this site corresponds to these two bacteria, we infer that iron oxidation is the main metabolic pathway carried out in this area (Parro et al., 2007). Reduced iron from the dissolution of pyrite in the subsurface can be quickly oxidized when the electron acceptor is available. As in stirred tanks, some mixotrophic or heterotrophic acidophiles are also present (Johnson & Hallberg, 2003) although their numbers were lower than for autotrophic iron oxidizers, both in the water column and in the sediments (Table 3). Moreover, RT3 (comprising water from RT1 and RT2) appears to be completely integrated with RT1 based on shared OTUs (Table 2, Fig. 3), pointing to the dominance of RT1 over RT2 and the probable inhibitory effect of metallic enrichment on some populations (Table 1).

In contrast to the headwaters, origin 2 and the middle course of the river resemble a mineral heap. Their more heterogeneous conditions in terms of flow, aeration, pH, temperature, nutrient availability, etc., result in more variable environments compared with the relatively homogeneous bioleaching tanks (Rawlings & Johnson, 2007). As a consequence, spatial and temporal variations in microbial populations are more likely to appear. Downstream ferric iron is consumed to compensate for the dilution effect of tributary wasters and subsequent precipitation. This is also reflected in the lower conductivity values (Table 1). The microbial diversity detected in the sediments of these areas is characterized by the presence of new bacterial and archaeal groups that were not detected in origin 1. There is a remarkably high number of sequences related to members of the classes Alphaproteobacteria (Acidiphilium, Acidocella, Acidisphaera), Betaproteobacteria (order Nitrosomonadales, ‘Ferrovum myxofaciens’), Gammaproteobacteria (bacterium WJ2), Acidobacteria, Firmicutes, Actinobacteria and Thermoplasmata. Some of these groups of microorganisms also appear in the water column, as shown by both cloning and hybridization (Tables 2 and 3) but at much lower cell numbers (Fig. 5). Higher microbial diversity translates into a higher metabolic variety, especially due to the appearance of numerous mixotrophic, heterotrophic and iron-reducing microorganisms, cohabiting with the chemolithotrophic iron oxidizers already mentioned, but which show totally different adaptations in this part of the river (Parro et al., 2007). Oxygen concentrations tend to be lower in the sediment compared with the water column as it is consumed very efficiently by aerobic respirers and as a result of its diffusion limitation at low pH. Finally, ferric iron, as iron oxides, is available due to precipitation from the buffer reaction. The development of anaerobic conditions favours the reduction of ferric iron coupled either to the oxidation of organic matter or reduced sulfur compounds and sulfate-reducing activities (Moreau et al., 2010). Although photoreduction of ferric iron might have a significant contribution in the upper layer of the water column (McKnight et al., 1998; Gammons et al., 2008), it is dependent on radiation, which is efficiently filtered by ferric iron (Kimball et al., 1992). Other factors might be involved in the variability of measured rates of bacterial Fe(III) reduction in anoxic sediments (Lu et al., 2010), although they are probably more stable over time compared with the photoreduction rates. Although further ecophysiological studies are needed, the data suggest that most of the reduced iron in this area is biologically generated, facilitating the operation of a complete biological iron cycle in the sediments.

Nearly all the microorganisms detected and identified in this study are, in one way or another, related to the iron cycle. Most were previously detected and/or isolated in AMD sites (Johnson & Hallberg, 2003) or biohydrometallurgical operations (Rawlings, 2005). Nonetheless, some microorganisms have been identified for the first time in Rio Tinto (e.g. members of Actinobacteria, Firmicutes, Acidobacteria, Cyanobacteria, Planctomycetes and Chloroflexi). The diversity detected by hybridization techniques correlates exceptionally well with the cloning data. At most of the stations, at least 50% of the water column OTUs are shared in the sediment samples. In other words, most of the populations residing in the water column can also be tracked in the sediments. Communities from the water column and sediments seem to be intimately associated, along with biofilms and other biological structures previously studied (García-Moyano et al., 2008). However, the sediment microorganisms comprise some exclusive bacterial and archaeal groups, for example the wide range of bacteria belonging to the class Betaproteobacteria in RT2. Most of the clones were retrieved from the sediments and were related to the order Nitrosomonadales (Ferrovum-like) and the family Neisseriaceae. Other groups such as the sulfate-reducing Desulfosporosinus and several known fermenters or metal-reducing bacteria (e.g. Clostridium) within the phylum Firmicutes, typically found in AMD, have been detected (Han & Kim, 2009; Lee et al., 2009; Lu et al., 2010; Moreau et al., 2010; González-Toril et al., 2011). Finally, members of the archaeal Thermoplasmata, some unidentified Acidobacteria, the iron-oxidizing bacterium Acidiferribacter thiooxydans, some members of Xanthomonadales and Planctomycetes also appear exclusively in the sediments. This diversity might be responsible for the higher prokaryotic biomass observed in this compartment. Moreover, the well-known chemolithotrophic iron-oxidizing bacteria (genera Leptospirillum and Acidithiobacillus) appear also in higher cell numbers in the sediments than in the water column (Table 3). This observation clearly highlights the importance of this compartment as a cluster where active iron and sulfur cycles are operating.

Due to the fast flow of the river at most of the locations the characteristics of the water column are reasonably homogeneous. In contrast, the sediments appear to be geochemically stratified (Sánchez-Andrea et al., 2011; Sanz et al., 2011), allowing the development of diverse communities that might be involved in specialized anaerobic metabolism, such as metal (mainly iron) and sulfate reduction and/or fermentation. The resemblance of Rio Tinto with some AMD and biohydrometallurgical systems is of note. These specific niches where specialized metabolic processes take place is not only of biotechnological interest (bioleaching) but also has important environmental implications (redissolution of metals or natural attenuation processes).


This study was supported by the Spanish Ministry of Science and Innovation (grants CGL2008-02298/BOS and CGL2011-22540). We thank Marina Postigo and Mari Paz Martín for the sequencing and ICP-MS analysis.