Quantification of Tinto River Sediment Microbial Communities: Importance of Sulfate-Reducing Bacteria and Their Role in Attenuating Acid Mine Drainage

Abstract

Tinto River (Huelva, Spain) is a natural acidic rock drainage (ARD) environment produced by the bio-oxidation of metallic sulfides from the Iberian Pyritic Belt. This study quantified the abundance of diverse microbial populations inhabiting ARD-related sediments from two physicochemically contrasting sampling sites (SN and JL dams). Depth profiles of total cell numbers differed greatly between the two sites yet were consistent in decreasing sharply at greater depths. Although catalyzed reporter deposition fluorescence in situ hybridization with domain-specific probes showed that Bacteria (>98%) dominated over Archaea (<2%) at both sites, important differences were detected at the class and genus levels, reflecting differences in pH, redox potential, and heavy metal concentrations. At SN, where the pH and redox potential are similar to that of the water column (pH 2.5 and +400 mV), the most abundant organisms were identified as iron-reducing bacteria: Acidithiobacillus spp. and Acidiphilium spp., probably related to the higher iron solubility at low pH. At the JL dam, characterized by a banded sediment with higher pH (4.2 to 6.2), more reducing redox potential (−210 mV to 50 mV), and a lower solubility of iron, members of sulfate-reducing genera Syntrophobacter, Desulfosporosinus, and Desulfurella were dominant. The latter was quantified with a newly designed CARD-FISH probe. In layers where sulfate-reducing bacteria were abundant, pH was higher and redox potential and levels of dissolved metals and iron were lower. These results suggest that the attenuation of ARD characteristics is biologically driven by sulfate reducers and the consequent precipitation of metals and iron as sulfides.

INTRODUCTION

The Tinto River is a natural acidic rock drainage (ARD) environment located in the Huelva region, in southwestern Spain. On its way from Peña de Hierro to the Atlantic Ocean, the Tinto River emerges from the core of the Iberian Pyritic Belt (IPB), one of the largest metal sulfide (pyrite, chalcopyrite, etc.) deposits on Earth. When exposed to air and water, the microbiologically enhanced oxidation of sulfides produces waters with low pH (pH 2.3) and high ferric iron (∼2 g liter⁻¹) and sulfate (∼6 g liter⁻¹) contents along the river course. Low pH and ferric iron levels facilitate metal solubilization; therefore, when these waters come into contact with metallic sulfides, they become highly metalliferous (Cu, ∼0.1 g liter⁻¹; Zn, ∼0.2 g liter⁻¹) (20).

ARD results in contamination of aquifer, with secondary effects like the death of fish, rodents, and livestock and reduced crop yields (36). These unpleasant consequences have led to efforts to remediate ARD environments (21). Besides the environmental concerns, there are biological, geochemical, and biotechnological reasons to analyze microbial populations in acidic environments and to understand their ecology. In ARD, acidophilic, chemolithotrophic, aerobic Bacteria and Archaea dissolve metallic sulfides by oxidizing the iron and sulfur components. Products resulting from these oxidation processes can be used as electron acceptors in dissimilatory reduction of ferric iron and sulfate under anaerobic conditions linked to the carbon and nitrogen cycles. So far, only ARD aquifers have been studied intensively (7, 8, 19, 22, 23); in contrast, little is known about the ARD sediments (17, 41).

At the Tinto River, knowledge of the diversity and abundance of benthic microbial populations is limited to a methanogen population study (43) and a description of the microbial ecology of two contrasting sampling sites (the SN and JL dams) (42). In the latter study, an extensive survey of the Tinto River sediment microbiota was performed using two culture-independent approaches: denaturing gradient gel electrophoresis (DGGE) and 16S rRNA gene sequencing. The taxonomic affiliation of the identified Bacteria showed a high degree of diversity, falling into 5 different phyla: Proteobacteria, Firmicutes, Bacteroidetes, Acidobacteria, and Actinobacteria; meanwhile, all the Archaea were affiliated with the order Thermoplasmatales. Microorganisms involved in the iron (Acidithiobacillus ferrooxidans, Sulfobacillus spp., and Ferroplasma spp.), sulfur (Desulfurella spp., Desulfosporosinus spp., and Thermodesulfobium spp.), and carbon (Acidiphilium spp., Bacillus spp., Clostridium spp., and Acidobacterium spp.) cycles were identified. However, the abundance of these groups of organisms was unknown. In this study, the microbial communities of the Tinto River sediments were quantified for the first time and analyzed in relation to the physicochemical properties of the sediments with the objective of identifying the dominant and active organisms thriving in the sediments and inferring the main microbial activities taking place.

MATERIALS AND METHODS

Field site description and sampling.

The selected sediments are located in the Tinto River basin (Huelva). Samples were collected from two sampling sites (Fig. 1), the SN dam (37.72173°N, 6.557465°W) and the JL dam (37.691207°N, 6.560587°W), at different positions of the river. The SN dam is a spring feeding the main stream of the river, while JL is located in the main course of the river. Sediment cores (inner diameter, 7 cm; length, 45 cm) were taken with a sampler (Eijkelkamp Agrisearch equipment, The Netherlands). The cores were sliced at 5-cm intervals. The redox potential (E) and pH of the drill core samples were measured in situ with E and pH probes, connected to a Thermo Orion 290A potentiometer placed into the fresh sediment just after extraction of the core. Sediment samples for hybridization were fixed immediately in 4% (wt/vol) paraformaldehyde (PFA) in phosphate-buffered saline (PBS; 145 mM NaCl, 1.4 mM NaH₂PO₄, 8 mM Na₂HPO₄, pH 7.6) at 4°C for 4 h, washed twice with PBS, and then stored in 50% (vol/vol) ethanol-PBS at −20°C until further use.

Fig 1.

Map of Tinto River in Huelva (Spain) with the location of the studied dams. (Adapted from reference 16a with permission from Elsevier.)

Chemical analysis.

According to physicochemical parameters and visual appearance, the JL dam core was composed of two main layers, with black and brown colors, whereas the SN core looked homogeneous (Fig. 2a and e). The black (10 to 15 cm) and brown (25 to 30 cm) layers from the JL dam and two layers between 10 to 15 cm and 25 to 30 cm from the SN dam were excised and immediately stored in 15-ml tubes at 4°C until further processing in the laboratory (less than 1 week). Sediments were centrifuged at 13,000 × g for 5 min, and the supernatant was analyzed for the presence of heavy metals by inductively coupled plasma mass spectrometry (ICP-MS) and organic anions/volatile fatty acids (VFA) by ionic chromatography (IC) using an 861 Advanced Compact IC instrument. Water samples of both dams were also taken, kept at 4°C, and analyzed by total reflection X-ray fluorescence (TXRF) for S, P, Mn, Fe, Cu, Zn, As, and Ca; by ICP-MS for the rest of the elements; and by thermochemical oxidation (Thermostar LT200 of Hachlange) for total organic carbon (TOC).

Fig 2.

Depth profiles of pH and redox at JL (a) and SN (e), Sybr green I and domain-specific probes counts at JL (b) and SN (f), phylum-specific probes counts at JL (c) and SN (g), and genus-specific probes counts at JL (d) and SN (h).

Quantification of microorganisms. (i) Sybr green I direct counts.

Aliquots of fixed samples were sonicated to detach bacteria from the particles with a sterilized Sonic probe (Bandelin, Sonopuls HD200) for 20 s, 3 times at 20% intensity. Samples were kept on ice for 4 min to precipitate large sediment particles that could clog the filter, and then the supernatant was gently filtered through a black 0.2-μm polycarbonate membrane (Nuclepore; 25-mm diameter) (<0.2 bar). The filters were stained with Sybr green I in Moviol medium as described elsewhere (30). Samples were examined with an epifluorescence microscope (Axioplan 2; Carl Zeiss, Jena, Germany). A minimum of 1,000 Sybr green I-stained cells or at least 100 independent microscopic fields were counted.

(ii) CARD-FISH.

Samples were filtered following the same procedure as described above for the direct counts, except that white filters (0.2-μm pore size; Millipore GTTP) were used. In situ hybridizations with horseradish peroxidase (HRP)-labeled probes (biomers.net, Ulm, Germany) followed by catalyzed reporter deposition of fluorescently labeled tyramide were carried out as described elsewhere (38). Permeabilization was done with lysozyme (10 mg/ml) for 90 min at 37°C, followed by achromopeptidase (60 U/ml) for Gram-positive cells (44) and proteinase K for Archaea (15 μg/ml) (46). Probe sequences and formamide concentrations required for specific hybridization are given in Table 1. Hybridized samples were examined with an epifluorescence microscope (Axioplan 2). For each probe and sample, roughly 1,000 stained cells or at least 100 independent microscopic fields were counted. The signal obtained with probe NON338—a negative control—at each sampling site was subtracted from the specific counts. This signal was about 0.2% of the cell counts for each station.

Table 1.

Oligonucleotide probes used in this study

Probe name	Sequence (5′ → 3′)	Target microorganism(s)	% FAa	Reference
NON338	ACTCCTACGGGAGGCAGC		0	47
EUB338	GCTGCCTCCCGTAGGAGT	Bacteria	35	3
EUB338-II	GCAGCCACCCGTAGGTGT	Planctomyces	35	9
EUB338-III	GCTGCCACCCGTAGGTGT	Verrucomicrobiae (and others)	35	9
ALF968	GGTAAGGTTCTGCGCGTT	Alphaproteobacteria, except of Rickettsiales	20	35
ACD840	CGACACTGAAGTGCTAAGC	Acidiphilium spp.	10	12
BET42ab	GCCTTCCCACTTCGTTT	Betaproteobacteria	35	32
GAM42ac	GCCTTCCCACATCGTTT	Gammaproteobacteria	35	32
THIO820	ACCAAACATCTAGTATTCATCG	Acidithiobacillus spp.	10	37
DELTA495a	AGTTAGCCGGTGCTTCCT	Most Deltaproteobacteria and Gemmatimonadetes	30	27
cDELTA495a	AGTTAGCCGGTGCTTCTT	Competitor for DEL495a		31
DELTA495b	AGT TAG CCG GCG CTTCCT	Some Deltaproteobacteria	30	27
cDELTA495b	AGTTAGCCGGCGCTTC(T/G)T	Competitor for DEL495b		28
DELTA495c	AATTAGCCGGTGCTTCCT	Some Deltaproteobacteria	30	27
cDELTA495c	AATTAGCCGGTGCTTCTT	Competitor for DEL495c		28
SYNBAC824	GTACCCGCTACACCTAGT	Syntrophobacter spp.	10	4
DSU655	CGCTTGCTTTTCCCGAAC	Desulfurella spp.	35	This study
HGC69a	TATAGTTACCACCGCCGT	Gram-positive high GC content (Actinobacteria)	25	40
ARCH915	GTGCTCCCCCGCCAATTCCT	Archaea	35	45
LGC354a	TGGAAGATTCCCTACTGC	Gram-positive low GC DNA content (Firmicutes)	35	33
LGC354b	CGGAAGATTCCCTACTGC	Gram-positive low GC DNA content (Firmicutes)	35	33
CLIT135	GTTATCCGTGTGTACAGGG	Clostridium cluster XI	0	15
CLOST1	TTCTTCCTAATCTCTACGCA	Clostridium clusters I and II	30	24
NTR712	CGCCTTCGCCACCGGCCTTCC	Nitrospira group	35	10
NTR712c	CGCCTTCGCCACCGGTGTTCC	Competitor for NTR712		10
DSP648	CTCTCCTGTCCTCAAGAT	Desulfosporosinus, Desulfitobacterium, Dehalobacter spp.	30	18
SS_HOL1400	TTCGTGATGTGACGGGC	Acidobacteria	20	34
EUK516	ACCAGACTTGCCCTCC	Eukarya	30	3

^aFormamide concentration in the hybridization buffer to ensure specific detection of target organisms.

^bUsed with unlabeled GAM42a as competitor.

^cUsed with unlabeled BET42a as competitor.

Total cell numbers in the SN core were too low to allow specific cell counting. Thus, a density gradient centrifugation was carried out to increase the cell density for quantification of less abundant bacterial groups. Cell separation was performed by Histodenz (Sigma-Aldrich) density centrifugation. Subsamples (1 ml) were placed in 2-ml tubes; 1 ml of Hystodenz solution (60% [wt/vol] in PBS, density of 1.3) was carefully placed underneath using a syringe needle to avoid mixing. Centrifugation was performed at 14,000 × g for 90 min at 4°C. The supernatant above the Hystodenz layer was treated as explained above for the rest of the samples. Although there was significant cell loss during the procedure (30% compared to untreated samples), this method was shown to maintain representative ratios between communities (48); thus, it could be used for relative counts.

Design, evaluation and application of a Desulfurella sp. probe.

The high hybridization numbers shown for Deltaproteobacteria probe in this study, the abundant presence of Desulfurella clones in the former study of Tinto sediments (42), and the lack of a genus-specific probe for Desulfurella made it necessary to design a new probe for this genus. This was done using the probe design tool in the ARB software package (29). The target position of DSU655 probe (655 to 673) was chosen to be in a highly accessible region according to Behrens et al. (6) and to be fully specific for the genus Desulfurella (checked with the ARB Probe Match tool). Using the described CARD-FISH protocol, the stringent hybridization conditions of Desulfurella probe were determined with positive-control cells of Desulfurella kamchatkensis and negative-control cells of Syntrophobacter wolinii (DSM 2805) and Acidobacterium capsulatum (DSM 11244) cultures. Hybridization stringency was optimized by increasing formamide concentrations starting with 10% in 5% steps. A 35% formamide concentration was optimal, resulting in high signal intensity of D. kamchatkensis and no signal for the negative controls.

SEM.

Sediment samples were studied by scanning electron microscopy (SEM) as described elsewhere (2). Samples were fixed by immersion in glutaraldehyde (2.5%) for 2 h, washed twice in sodium cacodylate buffer (0.2 M, pH 7.1), and dehydrated in a graded series (10, 30, 50, 70, 90, and 100%) of ethanol-water mixtures for 20 min each. After dehydration, the samples were critical point dried and mounted on stubs. After gold shadowing, the samples were examined with a Phillips XL30 EDAX DX4i SEM.

RESULTS

Physicochemical properties of the sediments.

The redox potential and pH values of the sediment profiles were significantly different (P value of <0.001 with the Mann-Whitney U test) at the two sites (JL and SN) (Fig. 2a and b). JL cores were banded and varied from reducing (blackish) to oxidizing (brownish) zones, with a redox ranging from −246 to 50 mV and a pH range from 6.2 to 4.2. SN cores did not show strong variations with depth, the pH remained between 2.3 and 2.5, and the redox potential remained between +248 and +312 mV.

Comparing the black and brown layers of the JL dam, arsenic, heavy metals, and sulfate concentrations in the interstitial water were higher in the brown layers. In the black layers, there was a strong decrease (higher than 90%) for Co, Ni, Cu, Zn, and As and close to 50% for Fe, Mn, and sulfate (Table 2). A concentration of 79 mg liter⁻¹ of acetate, an intermediary of the anaerobic digestion pathways and a substrate for some methanogens and sulfate reducers, was measured in the black band, and the concentration in the brown band was as low as 1 mg l⁻¹. On the other hand, there was no stratification with depth at the SN dam, so the average of both studied layers is presented in Table 2. Compared to in JL, heavy metal concentrations varied strongly in SN: Co and Ni were under the detection limit, while the concentrations of Zn and Cu were much higher.

Table 2.

Heavy metals, metaloids, sulfate, and volatile fatty acid concentrations detected in the porewater sediments and percentage of reduction in JL layers

Sample	Concn (mM)a
	Formate	Acetate	Sulfate	Fe	Co	Ni	Cu	Zn	As	Al	Mn
SN	1.1 × 10−2 ± 2.1 × 10−3	0.1 ± 2.1 ×10−2	20.4 ± 2.3	92.2 ± 2.1	UDL	UDL	1.0 ± 0.2	11.2 ± 0.4	1.1 × 10−2 ± 9 × 10−4	43 ± 3.5	2.9 ± 0.2
JL brown	1.3 ×10−2 ± 3.2 × 10−3	6.6 × 10−3 ± 3.2 × 10−4	25.5 ± 3.1	100.2 ± 1.2	1.3 × 10−1 ± 5 × 10−3	4.4 × 10−2 ± 8 × 10−4	7.9 × 10−3 ± 3 × 10−4	1.9 ± 0.1	1.4 × 10−2 ± 2.5 × 10−3	33 ± 1.4	4.1 ± 1.4
JL black	1.1 × 10−2 ± 2.1 × 10−3	1.3 ± 1.5 × 10−2	12.5 ± 1	52.2 ± 3.3	1.2 × 10−2 ± 2.1 × 10−3	1 × 10−3 ± 2.4 × 10−4	3.1 × 10−4 ± 1 × 10−5	6.1 × 10−3 ± 4.1 × 10−4	1.3 × 10−3 ± 4.5 × 10−4	3.3 × 10−2 ± 4 × 10−3	2.1 ± 0.1
% reduction in JL black			51.1	47.9	90.9	97.7	96.0	99.7	90.6	99.9	48.3

^aUDL, under the detection limit.

Chemistries of the overlaying waters at both dams were similar in pH and some metals (Table 3) but differed for K, Mn, Cd, P, S, and Fe. S and Fe are the elements that the oxidation of pyrite (the main substrate in the IPB) put into solution in ferric iron and sulfate forms. Both elements were less abundant at the JL dam, probably due to a dilution along the river of the concentrated ARD waters present at the SN dam.

Table 3.

Element concentration, TOC, redox potential, and pH values of SN and JL overlaid waters

Sample	pH	E (mV)	TOC (mg l−1)	Concn (mM)a
				Na	Mg	Al	K	Mn	Fe	Cu	Zn	As	Ca	Cd	S	P
SN	2.5	400	9 ± 1.7	0.62 ± 0.1	8.3 ± 0.3	14.5 ± 0.7	0.23 ± 0.01	UDL	41.30 ± 2.4	0.43 ± 0.04	1.19 ± 0.3	0.07 ± 2.1 × 10−3	1.81 ± 0.07	2.7 × 10−3 ± 8 × 10−4	85.8 ± 6.4	2.3 ± 0.7
JL	2.8	365	20 ± 4	0.24 ± 0.02	12.3 ± 0.4	13.6 ± 0.4	UDL	26.07 ± 1.3	16.34 ± 1.7	1.94 ± 0.07	1.41 ± 0.05	0.04 ± 3.7 × 10−3	0.98 ± 0.05	UDL	54.3 ± 4.6	UDL

^aUDL, under the detection limit.

Total cell counts with the Sybr green I method.

In comparison to DAPI (4′,6-diamidino-2-phenylindole) staining, the use of the highly sensitive Sybr green I dye in combination with the Moviol embedding medium that lowered photobleaching and the use of black filters, which have less background, significantly improved the signal-to-noise ratio. This enabled reliable counting of total cell numbers.

The average of Sybr green I-stained cells of the JL core (Fig. 2b) was 5.5 × 10⁸ cells g⁻¹ wet weight of sediment, whereas in the SN dam sediments (Fig. 2f), the average was 1.4 × 10⁷, which is 40-fold lower. A common feature among depth profiles of both sites is that cell numbers in both dams are higher in the 15-to-20-cm layer. After that, depth profiles followed a general trend observed in many sediments of a drastic decrease with depth (25), from 1.4 × 10⁹ cells g⁻¹ (wet weight) in the 15-to-20-cm layer to 1.3 × 10⁸ in the 25-to-30-cm layer in the JL dam (10-fold lower) and from 2.4 × 10⁷ to 3.7 × 10⁶ in the SN dam (6-fold lower).

SEM.

Scanning electron microscopy (SEM) analysis of sediments revealed high densities of cells attached to the sediment particles in the black layer of site JL. The colonization of the brownish layer of JL and of sediment from site SN was scarce. This corroborated the direct counts obtained with the Sybr green I method.

CARD-FISH.

The sites investigated here had already been characterized in an earlier study by comparative 16S rRNA gene analysis, applying both fingerprinting by DGGE and comparative sequence analysis (42). FISH-based quantifications of particular taxa had first been hampered by high background due to the autofluorescence of sediment particles. These problems were solved in this study by applying the improved whole-cell hybridization technique CARD-FISH enabling the quantification of higher and lower taxa in a nested approach.

(i) Domain-specific probes.

First, oligonucleotide probes for the domains Bacteria, Archaea, and Eukarya were applied to the sediment samples. In all cases, Bacteria (>98%) dominated over Archaea (<2%) and Eukarya (below the detection limit) (Fig. 2b and f). As for the Sybr green I total cell counts, great differences were found between the two sampling sites. In the JL sediment, the average bacterial counts with probes EUB338 I to III (Fig. 2b) was 5.4 × 10⁸ cells g⁻¹ (wet weight) of sediment, ranging from 1.61 × 10⁹ at 15 cm to 2.12 × 10⁷ at 45 cm, whereas in the SN dam (Fig. 2f), the average was 8.9 × 10⁶, ranging from 1.6 × 10⁷ at 15 cm to 2.2 × 10⁶ at 45 cm.

(ii) In-depth analysis of the bacterial community.

An earlier study of Tinto sediments (42) had retrieved sequences belonging to five phyla of the domain Bacteria: Proteobacteria of the classes Alpha-, Beta-, Gamma-, and Deltaproteobacteria; Acidobacteria; Nitrospira; Actinobacteria; and Firmicutes. The selection of the phylum- and class-specific probes used in this study was done according to these previous findings.

Depth profiles of the respective bacterial groups are shown in Fig. 2c and g. In JL dam sediments, Deltaproteobacteria was the most abundant group along the sediment core, as detected by probe DEL495a-c (average of 6.6 × 10⁷ cells g⁻¹ [wet weight]). This class of Proteobacteria showed a strong variation with depth with two population maxima. The first peak (3.3 × 10⁸ cells g⁻¹ [wet weight]) was detected within the uppermost 5 cm, and after a sharp decline, a second population maximum (1.1 × 10⁸ cells g⁻¹ [wet weight]) was found in the most reduced layer (15 to 20 cm). The second most abundant group was Firmicutes, detected with a probe mix consisting of LGC354a, LGC354b, CLIT135, CLOSTI, and DSP648, with an average along the profile of 4.3 × 10⁷ cells g⁻¹ (wet weight). Alphaproteobacteria as detected by probe ALF968 averaged 4.2 × 10⁷ cells g⁻¹ (wet weight), Gammaproteobacteria as detected by probe GAM42a averaged 4.0 ×10⁷ cells g⁻¹ (wet weight), and Betaproteobacteria as detected with probe BET42a averaged 2.6 × 10⁷ cells g⁻¹ (wet weight). In contrast, Acidobacteria (SS_HOL1400 probe) and Actinomicetales (HGC69a probe) appeared only in low numbers, 5.9 × 10⁵ and 7.7 × 10⁶ cells g⁻¹ (wet weight), respectively. Signals for probe NTR712 (Nitrospira group) were below the detection limit.

The depth profiles of SN dam sediments did not show the variation found in JL. In this core, numbers decrease homogeneously with depth. Gammaproteobacteria dominated (average 1.7 × 10⁶ cells g⁻¹ [wet weight]) followed by Alphaproteobacteria with an average along the profile of 1.1 × 10⁶ cells g⁻¹ (wet weight). The rest of the groups were close to the detection limit, around 10⁵ cells g⁻¹ (wet weight), in decreasing order: Actinomycetales, Deltaproteobacteria, Firmicutes, Betaproteobacteria, and Acidobacteria.

(iii) Genus-specific probes.

All genus-specific probes applied in this study were targeting genera for which sequences had been frequently retrieved in an earlier study (42). This included the deltaproteobacterial genera Syntrophobacter (probe Synbac824) and Desulfurella (probe DSU655), the genus Desulfosporosinus (probe DSP648) within the phylum Firmicutes, the alphaproteobacterial genus Acidiphilium (probe ACD638), and the gammaproteobactrial genus Acidithiobacillus (Thio820).

The genus-specific probing confirmed the great differences between the two sampling sites (Fig. 2d and h). In JL, the most abundant genus detected was Syntrophobacter. Cells were rod-shaped and had a size of about 2 by 1 μm. The average along the profile was 4.7 × 10⁷ cells g⁻¹ (wet weight), but just in the first layer the signal counts were 3.6 × 10⁸, representing 36% of the total Bacteria counts. Syntrophobacter spp. are propionate-degrading syntrophic bacteria that use sulfate as the electron acceptor. They typically occur in environments with neutral pH. The other abundant genera were Desulfurella (3.1 × 10⁷ cells g⁻¹ [wet weight], mostly rods of a size of 2 by 1 μm), and Desulfosporosinus (2.8 × 10⁷ cells g⁻¹ [wet weight], curved rods of a size of 2 to 4 by 0.7 μm). The latter genus is known to encompass spore-forming sulfate-reducing bacteria. Both genera showed higher numbers at those layers with higher pH and reducing redox potential (5 to 10 and 15 to 20 cm) (Fig. 2a and d). At SN, Acidithiobacillus spp. dominated with an average of 1.5 × 10⁶ cells g⁻¹ (wet weight) along the profile. The hybridized cells were coccoid rods with a size of 1 by 0.7 μm. Acidithiobacillus spp. derive energy from the oxidation of reduced sulfur compound and/or ferrous iron, or the reduction of ferric iron under anoxic conditions. Finally, Acidiphilium spp. were also present, with average numbers of 4.2 × 10⁵ cells g⁻¹ (wet weight) (rods of a size of 2 to 3 by 1 μm).

DISCUSSION

We used a nested-probe approach for the quantification of the microbial community composition in sediments of Tinto River. The generally low abundance of Archaea agreed with similar results found in the sediments of mine tailing dumps (23). In a previous study of Río Tinto sediments (42), the archaeal diversity had been restricted to two genera: Ferroplasma spp. and Thermoplasma spp. Although Sanz et al. (43) had described the presence of methanogens in the sediments of the Tinto basin, our data suggest that Archaea play a minor role in the ecology of these sediments. Both Ferroplasma and Thermoplasma are usually found in bioleaching operations at lower pH (around 2) and at much higher iron and sulfate concentrations than those prevailing in Río Tinto sediments (13). Although algae, ciliates, flagellates, amoebae, and fungi have been reported in the water column and the biofilms of the Río Tinto (1, 5), no Eukarya probe signal was detected by CARD-FISH; therefore, Bacteria seem to fully dominate the anoxic sediment layers of Tinto River.

When genus-specific probes were applied, a pronounced variability was detected between the two sampling sites. In SN sediments, where parameters such as pH and redox potential were similar to water column values, Bacteria related to the iron cycle prevailed, likely reflecting the higher availability of iron at low pH and oxidative redox potential. This included the genera Acidiphilium (ACD638 probe) and Acidithiobacillus (Thio820 probe), which both are capable of iron reduction under anoxic conditions. No members of the Leptospirillum genus, Nitrospira-like bacteria (Ntr712 probe), were detected. The common iron oxidizers in ARD environments are members of the aerobic genus Leptospirillum and the facultative anaerobe bacterium Acidithiobacillus ferrooxidans, which can oxidize iron and sulfur in aerobic conditions and reduce iron in anoxia. Our results confirm the dominance of the versatile Acidithiobacillus ferrooxidans in the sediments over the strictly aerobic Leptospirillum spp. that usually dominate in extremely acidic oxic ARD environments (7, 13).

At the JL dam, where the iron solubility is lower due to higher pH and reducing redox potential, microbes of the sulfur cycle were most abundant. Here, sulfate-reducing bacteria (SRB) of the genera Desulfosporosinus (DSP 648 probe) and Syntrophobacter (Syn824 probe) as well as sulfur reducers of the genus Desulfurella (DSU655 probe) predominated, showing the importance of sulfur cycling in these sediments. Syntrophobacter-affiliated 16S rRNA gene sequences (42) (HQ730668, HQ730666) have been retrieved in several acidic environments before (11, 17, 26), suggesting that not-yet-cultured relatives of Syntrophobacter might be well adapted to acidic conditions in contrast to the general neutral pH environments where they thrive.

The analyses of both brown and black layers in the JL sediments showed a significant reduction in the last ones for dissolved metals, iron, and sulfate in the black layers, which run parallel to the SRB abundance. Desulfurella and Desulfosporosinus abundance peaked perfectly with those black layers where the pH values and acetate concentration increased and the redox potential and the levels of the sulfate and the dissolved metals decreased. Our results suggest that the attenuation of ARD characteristics in the reduced black layers is biologically driven by SRB. The dissimilative reduction of sulfate, driven by identified sulfate- and sulfur-reducing bacteria, leads to a consumption of protons, thus increasing alkalinity, and to sulfide formation. In addition, there is a sulfide-mediated metal precipitation: the sulfide reacts with the dissolved iron and other heavy metals which can form black amorphous precipitates (FeS) and metallic sulfides (MS). These results agree with previous studies where remediation in AMD associated with SRB metabolism was observed (14, 16, 21, 39).

In summary, the CARD-FISH quantifications performed in this study underline the great differences between microbial communities of the two Río Tinto sediments studied that had been previously indicated by DGGE and comparative sequence analysis. These differences correlate with their physicochemical characteristics, which would correspond to different hydrology at both sites. In the JL dam, the location, thickness, and values of pH and redox potential of the oxidized and reduced zones vary temporally, but stratification has been routinely observed. In contrast, sediments in SN have been more homogeneous. Different interpretations can be put forward. On the one hand, the two sampling sites have different hydrology. Site SN is a stream located close to the source of the river with a constant water flow containing a high concentration of heavy metals, mainly iron. Therefore, the sediment characteristics are close to those of the water column strongly buffered by ferric iron. In contrast, the site JL is located 5 km downstream, with an increasing flow due to the addition of neutral tributaries which lower the iron concentration, albeit with a strong seasonal variability. One of the tributaries upstream of JL contains the waters from the municipal wastewater treatment plant of Nerva. This significantly increases the total organic content (TOC) of the waters. The combination of both factors could explain the differences detected in both types of sediments. At the JL dam, with seasonal variability, higher TOC, and lower iron concentration, microbial reduction processes such as sulfate reduction increase the pH. This effect would take place firstly in microniches and then expands to macroscopic conditions. That would explain the banded sediment at the JL dam, which would correspond to the fluctuating conditions of the input water. Within them, SRB communities reduce the dissolved sulfates to sulfides that precipitate iron and heavy metals, performing a local natural bioremediation of Tinto River.