Introduction

Dissolution of sulfide ores exposed to oxygen, water, and microorganisms results in acid production and environmentally detrimental acid mine drainage (AMD) (Nordstrom and Alpers 1999). Microorganisms that are able to develop under extreme conditions, especially iron-oxidizing chemolithotrophics are responsible for the solubilization of metals from sulfide minerals in acidic environments via direct action of their enzymes or indirectly through chemical action of their metabolic products (Southam and Beveridge 1992; Wulf-Durand et al. 1997; Wielinga et al. 1999).

Because of the limited types of substrates available in mining environments, the biotopes were initially expected to be extremely poor with respect to the diversity of the microbial flora. However, cultivation-based studies revealed a great diversity of the microbial community in AMD (Johnson 1998; Hallberg and Johonson 2001). The presence of various bacterial species including common prokaryotic chemolithotrophs other than Acidithiobacillus ferrooxidans, such as A. thiooxidans, A. caldus, and Leptospirillum ferrooxidans and so on, has been reported in mining environments. The presence of archaea including a group of sulfur and/or iron-oxidizers, such as Sulfolobus, Acidianus, Metallosphaera, Sulfurisphaera, and Ferroplasma has also been reported in acidic environments (Fuchs et al. 1995, 1996; Kurosawa et al. 1998; Edwards et al. 2000; Golyshina et al. 2000).

Cultivation-based analysis is not considered a suitable method for characterizing microbial communities (Ward et al. 1990; Delong 1992), and researchers have used methods based on the analysis of 16S rDNA sequences to study microbial communities of acidic environments (Edwards et al. 1999; Bond et al. 2000; Burton and Norris 2002; Simmons and Norris 2002; Baker and Banfield 2003).

To better understand the compositions and structures of microorganisms in different AMD environments, a PCR-based cloning approach was used in this work. Two AMD samples from Yunfu sulfide mine were studied. The Yunfu sulfide mine was in Guangdong province, China, which began to mine pyrite for sulfuric acid manufacture since 1988, for its high concentration of sulfur and low concentrations of other elements. The microbial communities in AMD samples from Yunfu sulfide mine have not been investigated to date yet.

Materials and methods

Sites description and samples collection

Samples were colleted from Yunfu sulfide mine, in Guangdong province, China. The mine had mainly produced pyrite for sulfuric acid manufacture since 1988, for its high concentration of sulfur and low concentrations of other elements.

There were three aqueous AMD samples collected from two separated sites in Yunfu sulfide mine. pH and temperature were same in two samples, which were 2.5 and 25.0°C, respectively. One of two samples was named as G1, from which the clones with 16S rDNA inserts were given the prefix GY. Similarly, the other sample was named as G2 and from which clones with 16S rDNA inserts were given the prefix G.

Ten-liter water sample was colleted from each position. Samples were processed within 24 h after collection. Two water samples were filtered through 0.22 μm hyper filtration membrane with vacuum pump, respectively. The sediments on the membrane were washed by sterile deionized water twice. Then the sediments were stored at −70°C for reservation. The filtered water samples were prepared for chemical analysis.

Chemical analysis of water samples

The element analysis of filtered water samples was carried out by Inductively Coupled Plasma-Atomic Emission Spectrometry (ICP-AES).

Twenty-nine elements were tested in each water sample. They were as following: Hg, As, P, Ni, Co, Cr, Be, Ti, W, Zn, In, Mg, Mn, Ca, S, Mo, Bi, Au, Fe, Si, Cu, Sn, Sb, Cd, Ga, Pt, Al, Ag.

DNA extraction and purification

Extraction of nucleic acids was according to procedure described by Zhou et al. (1996). Five gram of sediment was mixed with 13.5 ml extraction buffer (0.1 M phosphate [pH 8.0], 0.1 M EDTA, 0.1 M EDTA, 1.5 M NaCl, 1% CTAB) and 50 μl protenase K (10 mg/ml) in 50 ml centrifuge tube, then incubated at 37°C for 30 min. 1.5 ml of 20% SDS was added and mixed gently, then incubated at 65°C for 2 h. The mixture was centrifuged and the supernatant was transferred into a new 50 ml of centrifuge tube. The soil pellet was resuspended with extraction buffer, and 0.5 ml 20% SDS was added. The mixture was incubated at 65°C for 15 min, then centrifuged and the supernatant was collected and combined with the previous supernatant. The combined supernatant was extracted with chloroform. 2-Isopropanol was added to the supernatant collected and then mixed gently. The mixture was kept at the room temperature for an hour or overnight, then centrifuged. The pellet was washing with 70% ethanol and dissolved with 200–500 μl sterile water. By using combined methods that included grinding, freezing and thawing, and treatment with sodium dodecyl sulfate, various types of bacterial could been effectively lysed. The crude DNA was purified by using Wizard plus sv Minipreps DNA purification system (Promega Corporation, USA) and quantified by ethidium bromide-UV detection on an agarose gel.

PCR and fractionation of 16S rDNA genes

Bacterial 16S rDNA genes were amplified with the primer set was 1492R (5′-CGGCTACCTTGTTACGACTT-3′), and 27F (5′-AGAGTTTGATCCTGGCTC AG-3′) (Lane 1991). A gene amp (Biometra, T-Grandient, Genman) was used to incubate reactions through an initial denaturation at 94°C for 2 min, followed by 35 cycles of 94°C for 40 s, 55°C for 30 s, and 72°C for 1 min, and completed with an extension period of 10 min at 72°C. Products from the amplification reactions of expected size (about 1,500 bp) were pooled and purified before ligation later.

PCR amplification of archaeal 16S rDNA genes was carried out following the PCR reactions described as above with two different sets of archaea-specific primers, which were as follows: S-D-Arch-0025-a-S-17 (5′-CTGGT TGATCCTGCCAG-3′) (Robb et al. 1995) or S-D-Arch-0344-a-S-20 (5′-ACGGGGCGCAGCAGGCGCGA-3′) (Weisburg et al. 1991) with S-*-Univ-1517-a-A-21 (5′-ACGGCTACCTTGTTACGACTT-3′) (Raskin et al. 1994) to yield 1,500 or 1,120 bp PCR products, respectively.

Cloning, RFLP, and sequencing

The purified PCR products were ligated to the vector PGEM-T (Promega Corporation), and used to transform DH5α competent host cells. About 120 white colonies were randomly selected from each library. The transformation efficiency in our study was around 5 × 108 cfu/μg DNA. And we used an external control to track efficiency, offered by the PGEM-T vector system (Promega company). For restriction fragment length polymorphism (RFLP) and sequencing, the inserted fragments were amplified with the vector-specific T7 and SP6 primers. These unpurified PCR products were digested with two restriction endonucleases AfaI and MspI (TaKaRa Biotechnology Co, Ltd.), incubated at 37°C for 3 h. The restricted fragments were separated by gel electrophoresis in 3.0% agarose with ethidium bromide staining and observed on UV illumination. RFLP patterns were identified and grouped, and representative clones were selected for nucleotide sequencing.

Phylogenetic analysis

Phylogenetic affiliations of the partial sequences were initially estimated using the program BLAST (Basic alignment search tool) (Bond et al. 2000). Similarity of partial sequences was determined using ARB (a software environment for sequence data) (Strunk and Ludwig 1995). The initial phylogenetic trees were based on all available sequences and were constructed by using the DNA distance program Neighbor-Joining with Felsenstein Correction in ARB (Smith et al. 1994). Based on the initial phylogenetic results, appropriate subsets of 16S rDNA sequences were selected and subjected to a final phylogenetic analysis with CLUSTAL X.

Statistical methods

The rarefaction analysis was performed with SigmaPlot software. An exponential model, y = a × [1 − exp(−b × x)], was used with SigmaPlot 8.0 nonlinear regression software to fit the clone distribution data.

Nucleotide sequence accession numbers

Sequences have been submitted to GenBank with accession numbers are as follow: DQ480487 (G77), DQ480488 (GY28), DQ480486 (G74), DQ480485 (G51-19), DQ480484 (G31), DQ480483 (G24), DQ480481 (G5), DQ480482 (G71), DQ480479 (G28), DQ480480 (G44), DQ480478 (G51-12), DQ480476 (G27), DQ480477 (G66), DQ480475 (G52), DQ480474 (G8).

Results

Biogeochemical properties of two AMD samples

Although pH and temperature in two AMD samples both were 2.5 and 25.0°C, respectively, elements’ concentration of them was very different by analysis of ICP-AES. Sample G2 had higher elements’ concentration than those in sample G1, except element copper. The data of 29 elements’ concentration are shown in Table 1.

Table 1 Twenty-nine elements’ concentration in two AMD samples G1 and G2
Full size table

RFLP analysis of 16S rDNA clone libraries

Two sets of archaea-specific primers used to amplify archaeal 16S rDNA from two AMD samples both failed, while the bacterial-specific 16S rDNA primer set 27F and 1492R worked.

Consequently, the primer set 27F and 1492R was used to amplify bacterial 16S rDNA gene from two AMD samples in Yunfu sulfide mine. The PCR products formed a single band approximately 1,500 bp in length. After T-A cloning, 120 clones containing 16S rDNA inserts were obtained from each sample. The profiles of RFLP in two samples are shown in Figs. 1, 2.

Fig. 1
figure 1

Restriction fragment length profiles of 16S rDNA fragments amplified from sample G1 in Yunfu sulfide mine. The 16S rDNA fragments were amplified using the primer set 27F and 1492R, digested with the restriction endonucleases AfaI and MspI, and then analyzed by 3.0% ararose gel electrophoresis. The clones tested were shown in (a–e). M: 100 bp DNA ladder plus

Full size image
Fig. 2
figure 2

Restriction fragment length profiles of 16S rDNA fragments amplified from sample G2 in Yunfu sulfide mine. The 16S rDNA fragments were amplified using the primer set 27F and 1492R, digested with the restriction endonucleases AfaI and MspI, and then analyzed by 3.0% ararose gel electrophoresis. The clones tested were shown in (a–g). M: 100 bp DNA ladder plus

Full size image

The rarefaction analysis was used in RFLP analysis. The results are shown in Fig. 3. Nonlinear regression suggested that saturations were at 70 clones and 40 clones for samples G2 and G1, respectively. The result also suggested that the clones tested in the experiment were sufficient to detect the level of microbial communities’ diversity and infer the level of distribution within communities of two samples.

Fig. 3
figure 3

Evaluation of the representation of clones obtained from two AMD samples in Yunfu sulfide mine. G1: one AMD sample in Yunfu sulfide mine, G2: the other AMD sample in Yunfu sulfide mine

Full size image

The RFLP analysis revealed extensive diversity of 16S rDNA for two AMD samples. A total of 15 OTUs (unique RFLP patterns) were obtained. There were 14 OTUs in sample G2; 5 OTUs were detected in sample G1. The distributions of OTUs, which were ranked in the order of abundance in each sample, were shown in Fig. 4.

Fig. 4
figure 4

Distributions of OTUs in clone libraries in two AMD samples. G1: one AMD sample in Yunfu sulfide mine, G2: the other AMD sample in Yunfu sulfide mine. Clones GY10 and G66: the RFLP patterns were same, with 99% similarity to Acidithiobacillus ferrooxidans strain QXS-1; clones GY8 and G28: the RFLP patterns were same, with 99% similarity to Acidithiobacillus ferrooxidans strain YTW; clones GY81 and G8: the RFLP patterns were same, with 99% similarity to uncultured bacterium clone 69-8H; clone GY21: the RFLP patterns were same, uncultured bacterium Tui3-12; GY28: with 96% similarity to Acinetobacter sp. An9; clone G52: with 99% similarity to Acidithiobacillus albertensis strain DSM 14366; clone G27: with 93% similarity to Nitrosospira sp. III7; clone G51-12: with 99% similarity to uncultured Proteobacterium clone EV221H2111601SAH95; clone G44: with 99% similarity to Leptospirillum ferrooxidans strain CF12; clone G5: with 99% similarity to uncultured bacterium clone EV818SWSAP4

Full size image

The RFLP patterns of clones GY10, GY8, GY81, GY21, and GY28 represented 48.2, 33.7, 8.4, 8.4, and 1.2% of the total clone populations in AMD sample G1, respectively. In the other AMD sample G2, the RFLP patterns of clones G8, G52, G27, G66, G51-12, G28, G44, and G5 represented 31.5, 17.6, 10.2, 10.2, 10.2, 5.6, 3.7, and 3.7% of the total clone populations, respectively. Some RFLP patterns in the two AMD samples were overlapped with each other, such as OTUs GY10 and G66, OTUs GY8 and G28, and so on. The percentage of overlapped OTUs between samples G1 and G2 was 42.1%. There were only five OTUs detected in sample G1, four of which could be also detected in sample G2.

Phylogenetic analysis

To determine the phylogenetic diversity, representative OTUs that occurred more than once in libraries, as well as representatives of the unique OTUs, were fully sequenced. The results of similarity among the sequences are shown in Table 2. Sequences’ comparison showed that the clones in two AMD samples had high similarity, which was 72–98% similar.

Table 2 Similarity matrices based on 16S rDNA sequences from two AMD samples
Full size table

The phylogenetic analysis in two samples was established with a bootstrap neighbor-joining method with the sequences. The phylogenetic tree is shown in Fig. 5. Of note is that there were 15 OTUs with RFLP analysis, corresponding to 12 kinds of particular organisms. The accuracy of RFLP screening was only about 80% evaluated by sequencing representative clones, which may be due to the comparatively simplicity of microbial community in AMD, and this method has been commonly used for it is comparatively accurate and fast.

Fig. 5
figure 5

Phylogenetic tree based on comparative analysis of 16S rDNA sequence data from 15 OTUs and their close relatives. The sequences obtained in this study are indicated in bold

Full size image

These 16S rDNA sequences fell into four putative phylogenetic divisions. They were Nitrospira, α-Proteobacteria, β-Proteobacteria, and γ-Proteobacteria four families. In sample G2, α-Proteobacteria, β-Proteobacteria, and γ-Proteobacteria, Nitrospira four families were all detected, proportions of which were 0.9, 48.1, 37.1, and 13.9%, respectively. In sample G1, there were only two families detected, which were β-Proteobacteria (16.9%) and γ-Proteobacteria (83.1%). The distributions of the four families divided by phylogenetic analysis in two AMD samples are shown in Fig. 6.

Fig. 6
figure 6

Distributions of four families divided by phylogenetic analysis in two samples. G1: one AMD sample in Yunfu sulfide mine, G2: the other AMD sample in Yunfu sulfide mine. Nit: Nitrospira family, Alp: α-Proteobacteria family, Bet: β-Proteobacteria family, Gam: γ-Proteobacteria family. a Proportions of four families in AMD sample G1, b proportions of four families in AMD sample G2

Full size image

γ-Proteobacteria family was a predominant one in two AMD samples. There were two groups in the family. One group was clustered with the genus Acidithiobacillus, including OTUs G52, G28, and G66. The other group was divided into two sub-groups. One with OTUs GY28 and G24 was clustered with Acinetobacter sp. The other with OTUs G51-19 and G77 was clustered with uncultured bacterium.

β-Proteobacteria family was the other predominant family in two AMD samples. It also fell into two groups. One group including OTUs G27, G71, G5, and G8 was clustered with the genus Gallionella. The other group was clustered with Pseudomonastestosterone, just including one OUT, G31.

Nitrospira family was only detected in sample G2, in which there were two OTUs G44 and G51-12. The two OTUs were both affiliated with the genus Leptospirillum. G44 was affiliated with L. ferrooxidans, while G51-12 was affiliated with the bacterium of the group III of Leptospirillum.

The last family was α-Proteobacteria family, which was also just detected in AMD sample G2. Only one OTU G74 was detected in this family. It had 99% similarity to Hyphomicrobium vulgare strain IFAM MC-750.

Discussion

Currently, AMD environments have been set as model systems for analysis of biogeochemical interactions and microbial community structure and function (Hallberg and Lindström 1994; Baker and Banfield 2003). And any AMD system has certain microbial niches for variations in environmental data (Druschel et al. 1999; Bond et al. 2000).

In this work, culture-independent molecular methods were employed, including 16S rDNA clone library analysis and sequence determination. Microbial community structures in two AMD samples from Yunfu sulfide mine have been identified, which have not been investigated to date. According to the RFLP analysis, there were 15 OTUs detected in two samples. They fell into four putative families, which were α-Proteobacteria, β-Proteobacteria, γ-Proteobacteria, and Nitrospira four families. The organisms of genuses Acidithiobacillus and Gallionella, which were in γ-Proteobacteria and β-Proteobacteria, respectively, were dominant in two AMD samples we studied.

γ-Proteobacteria was a ubiquitous family in two AMD samples, with which clones clustered were mostly affiliated with the genus Acidithiobacillus. The proportions of clones affiliated with Acidithiobacillus in samples G1 and G2 were 81.9 and 33.4%, respectively.

Acidithiobacillus spp. has been extensively studied. Acidithiobacillus is widely considered to be the microorganism that controls the rate of generation of AMD and A.ferrooxidans has been used as model microbe in bioleaching research (Lane 1991; Boon and Heijnen 1993; Fowler et al. 1999; Kelly and Wood 2000; Baker and Banfield 2003). Scherenk et al. (1998) used FISH probes to show that Acidithiobacillus spp. were abundant in environments with pH >1.5 and temperature <30°C at the Richmond mine. Acidithiobacillus spp. was also found to be the dominant iron-oxidizer stream draining the King’s copper mine (pH 2.5) in Roeros, Norway (Johnson et al. 2001). In our study, we also found the genus Acidithiobacillus was dominant in the two AMD samples we studied. The reason should be that the environments in two samples were suitable for the growth of Acidithiobacillus. The pH and temperature in two samples were 2.5 and 25.0°C, respectively, which were around the optimum of pH and temperature for the growth of Acidithiobacillus. And concentration of poisonous ions was very low in two samples for Acidithiobacillus. Previous researches reported that Acidithiobacillus were very susceptive with concentration of arsenic ion, especially arsenic ion (III) (Goebel and Stackebrandt 1994). The concentrations of arsenic in samples G1and G2 were only 2.5 and 5.9 mg/l, respectively, which were much lower than the maximal withstanding concentration of element arsenic for Acidithiobacillus.

β-Proteobacteria family was the other ubiquitous family in two samples. The clones in the family were mostly affiliated with the genus Gallionella. The proportions of clones affiliated with Gallionella in samples G1 and G2 were 16.9 and 47.2%, respectively. Organisms of genus Gallionella were neutrophilic iron-oxidizing and chemolithotrophic, which lived in a relatively poor-nutrition environment containing reduced iron. And oxidization of iron might be mainly caused by organisms of Gallionella in low-oxygen environments (Halbach et al. 2001). Interestingly, organisms of Gallionella, neutrophilic microbe, were dominant in two AMD samples we studied with pH around 2.5. The same result recently appeared in the study by Bruneel et al. that Gallionlla was dominant in Fe–As-rich AMD waters of Carnoulès, France (Bruneel et al. 2006).

Nitrospira family was only detected in sample G2. The clones clustered `with the family were all affiliated with the genus Leptospirillum. The organisms of Leptospirillum have been reported to be dominant in AMD environments (Druschel et al. 1999; Bond et al. 2000; Hippe 2000; Coram and Rawlings 2002), but they were not dominant in the two AMD samples we studied. The reason might be that the environments in the two samples were not suitable for the genus Leptospirillum. First of all, nutrition in two samples was too poor for the growth of Leptospirillum. Leptospirillum spp. just fed on iron and could not use sulfur as energy resource. In two samples studied, the concentrations of element iron were only 1.07 g/l (G1) and 4.0 g/l (G2). Then the temperature and pH in two samples were not also suitable for the growth of Leptospirillum. Reports of previous studies by Schrenk et al. (1998) with respect to Leptospirillum suggested that Leptospirillum spp. might share more natural distribution in low pH and high temperature. Oligonucleotide probe-based studies within the Richmond deposit indicated that Leptospirillum strains often dominated microbial communities growing at temperatures up to 50°C and in solutions with pH values as low as 0.5 (Edwards et al. 1999; Druschel et al. 1999; Bond et al. 2000).

Some other microorganisms related to the AMD environments were also detected in two AMD samples. Although their presence with low number, they were also significant in AMD environments. For example, Hyphomicrobium sp. in α-Proteobacteria family was detected in AMD sample G2, which could live in acidic environment and had contribution to the AMD generation by oxidizing iron to get energy (Pelczar et al. 1993).

In addition, the PCR amplification of archaeal 16S rDNA genes from these two AMD samples have also been performed with two sets of archaea-specific primers described as above. But there were no PCR products detected with these two primer sets. The same experimental procedures have been carried out on some other AMD samples, and several archaeal lineages, reported from AMD environments, have been detected (data not shown here). Therefore, it suggested that there might have no achaea in microbial communities in the two AMD samples from Yunfu sulfide mine.