OUP user menu

Bacterial community analysis of Indonesian hot springs

Gillian C. Baker, Shabarni Gaffar, Don A. Cowan, Adrian R. Suharto
DOI: http://dx.doi.org/10.1111/j.1574-6968.2001.tb10700.x 103-109 First published online: 1 June 2001


We report the first attempts to describe thermophilic bacterial communities in Indonesia's thermal springs using molecular phylogenetic analyses. 16S rRNA genes from laboratory cultures and DNA directly amplified from three hot springs in West Java were sequenced. The 22 sequences obtained were assignable to the taxa Proteobacteria, Bacillus and Flavobacterium, including a number of clades not normally associated with thermophily.

  • 16S rRNA sequence analysis
  • Hyperthermophile
  • Hot-spring community
  • Proteobacterium
  • Bacillus
  • Flavobacterium

1 Introduction

Thermophiles are organisms growing at elevated temperatures [1] and typically associated with solar, geothermally, industrially or biologically heated environments [2]. These are generally and somewhat arbitrarily separated into three categories based on their cardinal growth temperatures: thermophiles (35–70°C), extreme thermophiles (55–85°C) and hyperthermophiles (75–113°C).

Indonesia is one of the most tectonically active areas in the world with over 70 active volcanoes [3] and a substantial number of geothermal regions [4]. Although Indonesia's diverse terrestrial hydrothermal biotopes harbour large communities of hyperthermophilic bacteria (Akhmaloka, personal communication), to date little effort has been made to characterise these hyperthermophilic communities. Published studies are based on in vitro culturing and classical microbial identification methods [5]. It is generally accepted that fewer than 1% of extant microbial ‘species’ have been cultured [6] and that the majority do not grow in conventional isolation media [7]. Moreover, the organisms found on isolation plates are those best adapted for growth on this ‘artificial’ medium, and are not necessarily those that are metabolically or numerically dominant in the field [8]. Investigations of microbial diversity using culture-independent methods provide a valid approach to the assessment of true microbial diversity [9,10].

The advent of new molecular techniques, especially PCR, has revolutionised microbial identification [11] and amplification of DNA directly from field samples, without the need for culturing, has facilitated the characterisation of natural microbial communities [10,12].

In this study we describe the classification of bacteria from three hot-spring sites in Indonesia using 16S rRNA sequences directly amplified from the field and from mixed laboratory cultures. Two of these sites, Cibuni and Domas, are highly acidic sulfur-rich springs of 70–90°C, and the third, Cimanggu, is of neutral pH with temperatures ranging from 50 to 70°C.

2 Materials and methods

2.1 In vitro culture

Six replicate 50-ml samples of five different media (Table 1) were prepared, sterilised and incubated overnight at 70°C. The media were then placed in a Thermos™ box and transported to the field. Samples were taken from three geographically distinct sites in West Java, Indonesia: Domas (82°C and 90°C, pH 2), Cibuni (90°C, pH 2) and Cimanggu (70°C, pH 7). At each site 1 ml of sediment-rich water recovered by sterile syringe was transferred to each of the five media flasks. The media were then immediately returned to the Thermos™ box to maintain their temperature. At each site a second set of media were sham-inoculated as a control for any possible contamination by airborne bacteria. All samples were returned to the laboratory within a period of hours and incubated at 70°C (Cimanggu), 80°C (Domas) or 85°C (Cibuni) with shaking at 150 rpm. After 2 days a 1-ml sample from each flask was transferred to fresh medium and incubated for a further 48 h at the specified temperatures.

View this table:
Table 1

Media used for isolation of thermophilic bacteria

Medium 1820.1% yeast extract, 0.1% peptone, 0.31% KH2PO4, 0.25% (NH4)2SO4, 0.02% MgSO4.7H2O, 0.025% CaCl2, 1% trace element solution. pH 4.0
Medium 880.00013% (NH4)2SO4, 0.00028% KH2PO4, 0.0025% MgSO4.7H2O, 0.007% CaCl2.2H2O, 0.1% yeast extract, 1% trace element solution. pH 2.0
Medium 1620.25% yeast extract, 0.25% tryptone, 0.004% CaSO4.2H2O, 0.02% MgCl2.6H2O, 0.054% KH2PO4, 0.43% Na2HPO4.12H2O, 0.5% trace element solution. pH 7.2
Medium 740.4% yeast extract, 0.8% polypeptone, 0.2% NaCl. pH 7.0
Medium 130.1% yeast extract, 0.02% (NH4)2SO4, 0.05% MgSO4.7H2O, 0.025% CaCl2.2H2O, 0.06% KH2PO4, 0.1% glucose. pH 7.0
Trace element solution0.193% FeCl3.6H2O, 0.018% MnCl2.4H2O, 0.045% Na2B4O7.10H2O, 0.002% ZnSO4.7H2O, 0.0005% CuCl2, 0.0003% Na2MoO4.2H2O, 0.0004% VOSO4.5H2O, 0.0002% CoSO4.7H2O.

2.2 Field sampling

Samples of sediment-rich water were recovered from each of the three study sites: Cimanggu, Cibuni and Domas. The samples were removed aseptically with sterile 50-ml syringes and placed in pre-warmed sterilised 300-ml centrifuge bottles for return to the laboratory.

2.3 DNA extraction from field samples

The DNA extraction method used was based on the procedure of Klijn [13]. Field samples were centrifuged (5000×g) at 4°C for 20 min and the supernatant removed. Pellets were resuspended in 100 ml of ice-cold STE (NaCl, Tris–HCl, EDTA) buffer and re-centrifuged for a further 20 min. The supernatant was removed and the pellet resuspended in 5 ml of 10 mM Tris–HCl pH 7.6 and vortexed vigorously for 5 min. The sample was then passed through sterilised Whatman filter paper under vacuum. The filtrate was mixed with 8 mg ml−1 lysozyme and incubated at 30°C in a 50-ml centrifuge tube for 1 h. An equal volume of lysis buffer (2% SDS, 200 mM EDTA, protease K 0.8 mg ml−1) was added and the sample was incubated for 30 min at 55°C. The sample was then precipitated with 2 volumes of ethanol and centrifuged to produce a pellet. The pellet was dried, resuspended in 0.5 ml sterile water, extracted with phenol–chloroform and re-precipitated with ethanol. A negative control of sterile water was treated in the same manner as the field samples to check for airborne contamination of samples.

2.4 DNA extraction from cultured biomass

Approximately 109 cells from 2 ml of culture medium were pelleted by centrifugation and incubated in 200 μl of lysozyme (8 mg ml−1 in 10 mM Tris–HCl) for 1 h at 37°C. An equal volume of lysis buffer (2% SDS, 200 mM EDTA, proteinase K 0.8 mg ml−1) was added and the sample was incubated for 30 min at 55°C. The cell digest was then extracted in 200 μl phenol/chloroform (1:1) and precipitated with ethanol. The genomic DNA pellet was resuspended in 50 μl sterile water.

2.5 16S rRNA gene amplification and cloning

Approximately 200 ng genomic DNA was amplified by PCR (initial denaturation 3 min at 94°C; 30 cycles: 1 min at 94°C; 1 min at 48°C; 1 min at 72°C followed by a single period at 72°C for 10 min; 1 U of Sigma Taq polymerase) using bacterial 16S rRNA gene-specific primers [14]. 100-μl volumes of PCR products were electrophoresed on 1% agarose gels and purified using GFX™. The purified PCR products were ligated into a pGEM-T (Promega) vector and transformed into competent JM 109 Escherichia coli. Twenty colonies from each transformation were selected for restriction fragment length polymorphism (RFLP) analysis. White colonies were streaked out, picked and boiled in 40 μl TE for 3 min, then centrifuged at 10 000×g for 10 min. 10 μl of the supernatant was amplified by PCR with primers specific to the M13 sites in the pGEM-T vector. The PCR product was precipitated with 2 volumes of ethanol and 0.1 volumes of 3 M sodium acetate. The resulting pellet was washed with 70% ethanol, dried and resuspended in 5 μl sterile water. A 4-μl aliquot of this DNA was cut with HinfI for 2.5 h at 37°C. 1 μl of uncut DNA and 5 μl of the restriction enzyme products were electrophoresed on a 1.5% agarose gel against a 1-kb ladder. Samples with different restriction patterns were chosen for sequencing.

2.6 Sequencing and sequence analysis

Transformants chosen by RFLP analysis were purified using Qiagen mini-prep kit and sequenced initially with internal primers 357F and 803F [15]. Sequences were aligned using Seqman in Lasergene and novel operational taxonomic units (OTUs) were sequenced with the internal sequencing primers 1114F, 519R and 909R [15] and primers specific to the pGEM-T (Promega) vector, SP6 and T7. The sequence fragments were assembled and edited in Lasergene Seqman and the consensus sequences were compared with other 16S rRNA genes in the GenBank using NCBI BLAST [16]. Consensus sequences were also checked for chimeric sequences using the Ribosomal Database Project CHECK CHIMERA [17]. Each OTU sequence was aligned with closely related sequences identified from the BLAST search and representative in-group and out-group taxa using Clustal in Lasergene MegAlign. Nucleotide substitution rates were estimated for edited alignments using TREECON [18]. Distances based on the Jukes–Cantor function were used to infer neighbour-joining trees and analyses were bootstrapped (n=100) in TREECON.

3 Results

Bacterial PCR products were obtained from amplification of directly extracted DNA from two hot springs at Domas (82°C and 90°C, pH 2) and at Cimanggu (70°C, pH 7). Amplification products were also obtained from Cimanggu samples grown in media 13, 182 and 88, from Domas samples grown in medium 162 and from Cibuni samples grown in medium 74. The eight RFLP types obtained clustered within three distinct taxonomic groups: Proteobacteria, Flavobacterium and Bacillus.

3.1 Proteobacteria

Twelve clones from directly extracted DNA and samples isolated in media had phylogenetic affinities with two groups within the γ sub-division of the Proteobacteria (Table 2).

View this table:
Table 2

Taxonomic affinities of 16S rRNA gene sequences for clones derived from Indonesian hot springs

Sequence typeClone
RFLP type 1Domas 90 direct clone D2
Proteobacteria γ sub-division
RFLP type 1Cimanggu medium 88 clone B
Domas 82 direct clone A3
Domas 82 direct clone A4
Domas 82 direct clone C1
Domas 82 direct clone D3
Domas 90 direct clone A2
Domas 90 direct clone B2
RFLP type 2Cibuni medium 74 clone A
Cimanggu direct clone A
Domas 82 direct clone A1
Domas 82 direct clone A2
Domas 82 direct clone D2
RFLP type 1aDomas 82 medium 162 clone A1
Domas 82 medium 162 clone A4
RFLP type 1bDomas 82 medium 162 clone A5
Domas 82 medium 162 clone B5
RFLP type 2aDomas 90 direct clone C1
RFLP type 2bCimanggu medium 13 clone A
Cimanggu medium 182 clone A

Six clones from DNA isolated from the 90°C and 82°C pools of Domas and a Cimanggu isolate grown in medium 88 had identical sequences in the region primed with 803F and 357F. A 1228-bp sequence fragment from the Cimanggu sample grown in medium 88 was submitted to GenBank under accession number AF229452. A phylogenetic tree inferred from a matrix of 1161 bp shows that the Cimanggu medium 88 sample is related to Frateuria aurantia (Fig. 1), but differs from this species by 2.5% of the 1161 bases compared.

Figure 1

Neighbour-joining tree based on an alignment of 1161-bp 16S rRNA gene sequence fragments.

Three clones from the Domas 82°C pool, the Cimanggu 70°C pool and a Cibuni sample isolated in medium 74 also exhibited identical sequences in the region primed with 357F and 803F. A 1328-bp fragment of the Cibuni medium 74 sequence was submitted to GenBank under accession number AF229453. A phylogenetic tree inferred from a matrix of 1161 bp shows that this sequence is closely related to Pantoea ananas (Fig. 1), differing by only 0.17%.

3.2 Flavobacteria

One clone from a sample directly amplified from the Domas 90°C pool showed strong phylogenetic affinities with Flavobacterium meningosepticum. A 828-bp sequence fragment compared showed no sequence differences between F. meningosepticum and the Domas isolate (Fig. 2). A 1042-bp sequence fragment of the Domas clone was submitted to GenBank under accession number AF228796.

Figure 2

Neighbour-joining tree based on an alignment of 828-bp 16S rRNA gene sequence fragments.

3.3 Bacilli

Seven 16S clones from DNA directly extracted from Domas and Cimanggu cultures isolated in growth media had phylogenetic affinities with two groups of Bacillus (Table 2). The first group contained four clones from Domas isolates grown in medium 162. These fell into two distinct sub-groups which differed by 0.7% of 982 bp. A 1485-bp sequence fragment of clone A4 from group 1a and a 982-bp sequence fragment of clone A5 from group 1b were submitted to GenBank under accession numbers AF228798 and AF228797 respectively. A phylogenetic tree inferred from a 982-bp fragment showed that clones A4 and A5 were most closely related to each other, but formed a sister group to Bacillus megaterium and B. simplex (Fig. 3).

Figure 3

Neighbour-joining tree based on an alignment of 943-bp 16S rRNA gene sequence fragments.

The second group contained three clones that fell into two sub-groups which differed from each other at 1% of 296 bp sequenced. A phylogenetic tree based on this 296-bp fragment indicated that the sample amplified directly from Domas (GenBank accession number AF228795) was most closely related to Bacillus thermoleovorans and the Cimanggu samples isolated in media 13 and 182 were related to Bacillus caldovelox (GenBank accession numbers AF228801 and AF228800). These relationships were not supported by significant bootstrap values but all three clones were part of a closely related clade of thermophilic Bacilli supported by a bootstrap value of 98% (Fig. 4).

Figure 4

Neighbour-joining tree based on an alignment of 296-bp 16S rRNA gene sequence fragments.

4 Discussion

The phylogenetic affinities of sequence types amplified directly from Indonesian hot springs and from cultures grown at 70–85°C were, in part, unusual. While it is not surprising that moderately and extremely thermophilic Bacilli [19] were identified in the neutral 70°C Cimanggu hot spring, the absence of Gram-negative aerobes such as Thermus, which frequently dominate silicaceous neutrophilic thermal pools [20,21], is unexpected. We emphasise, however, that our results are not intended to represent a complete phylotypic analysis of these sites.

Due caution must be exercised in interpreting phylogenetic community data [22], since inadequate universality of primers, amplification of contaminant DNA, poor efficiency amplification from low-abundance sequences [23] and generation of chimeric products can all result in erroneous community analyses [24,17]. The presence of several unusual Bacterial sequence types, not normally associated with hot springs, is, however, interesting.

The presence of the same proteobacterial sequences from directly amplified DNA and from laboratory cultures grown at high temperatures is strong evidence for the in situ existence of these taxa. Hydrothermal springs may become contaminated by mesophilic organisms from soil and surface water, and spores (but not vegetative cells) might be expected to survive in the extreme conditions of the spring. There is at least the potential for DNA from these bodies to be amplified by PCR if few other competing DNA species were present. The identification of normally mesophilic organisms in thermophilic sites by direct PCR would normally be attributed to this type of contamination. However, the proteobacteria identified in this study were also present in laboratory cultures grown at 70–85°C. Mesophilic contaminant cells incapable of growth at high temperatures would quickly become dominated by a vast excess of thermophilic genomes and the likelihood of trace mesophilic contaminants being amplified by PCR after culturing is very small indeed. A further explanation for the presence of these unusual sequence types is possible contamination of the DNA extraction and PCR processes. While this cannot be ruled out entirely, negative controls were used throughout the study, none of which demonstrated any indication of contamination. We thus conclude that those sequences obtained from both culturing and direct amplification are true thermophiles.

We can be less confident about those sequences (e.g. Bacillus and Flavobacterium) obtained only from direct extraction or from culturing, as it is impossible to completely rule out mesophile contamination of the hot springs or contamination of laboratory cultures with thermophilic spores.

The first Proteobacterial sequence type showed phylogenetic affinities with Frateuria which were supported by a bootstrap value of 100%. The only known Frateuria species is a chemoorganotrophic acidophile that oxidises ethanol to acetic acid. This organism is known to grow on MYP medium at 30°C and is not known to exhibit thermophily [25]. The related organism identified by this study grew in a medium of pH 2.0 and was directly extracted from hot springs of this pH. It is clearly also an acidophile and may be a new species of Frateuria or a closely related taxon. The clade containing Frateuria, Xanthomonas and our isolate also contains an unidentified bacterium from a hydrothermal vent, suggesting that members of this generally mesophilic group do exhibit thermophily. Further phenotypic description of our isolate description is required in order to confirm its taxonomic assignment.

The second Proteobacterial sequence type was closely related to P. ananas and only differed in sequence from this organism by 0.17% of 1116 bp. A taxonomic position within the genus Pantoea was supported by a bootstrap value of 95%. The genus Pantoea lies within the taxon Enterobacteria [26] and is not reported to have any thermophilic representatives. Based on sequence comparisons, our organism is probably a thermophilic species of the genus Pantoea.

Only one sequence type grouped within the Flavobacteria. This sequence type was derived from direct extraction of community DNA and could possibly be a mesophilic contaminant. Flavobacterium species have previously been found in anoxic rice paddy soil [27], a habitat found in close geographic proximity to the hydrothermal site from which this sample was derived, and are widely distributed in soil and water [24]. Our isolate clustered with the species F. meningosepticum, with a bootstrap value of 100%. F. meningosepticum is a pathogen causing meningitis in humans and is not known as a thermophile.

The first group of Bacillus sequences clustered within a clade containing B. megaterium and B. simplex. Thermostable enzymes have been isolated from strains of B. megaterium (e.g. [28,29]), but it is not recorded as a hyperthermophilic species. Our isolates thus potentially represent new extremely thermophilic strains of Bacillus.

The second group of Bacillus sequences clustered with the known extreme thermophiles Bacillus caldovelox and B. thermoleovorans, although a close relationship with these species was not supported by particularly high bootstrap values. The clade containing these and other thermophilic Bacilli is well defined but the relationships between strains and species within it are poorly supported. It is thus impossible to state with any confidence within which species our isolates belong.

Indonesia's metafauna and flora have a high degree of endemism, attributed to the separation of Gondwanaland into isolated islands during the Cretaceous period as a result of tectonic plate movements. The ancestors of modern plants and animals became isolated on islands such as those in the Indonesian archipelago and evolved into species that are found nowhere else on earth. The results from these preliminary studies suggest that Indonesia's thermoacidophilic microbial communities are also possibly unique, and certainly deserving of further detailed investigation.


The work described in this paper was carried out under the auspices of Phase II of the UK-Indonesia Environmental Biotechnology Project (1994–1999) funded by the Department for International Development (DFID). The authors would like to thank the British Council and the International Institute for Biotechnology for facilitating the collaboration established during this project. We also wish to thank Endang Srieatimah and Wawan Kosasih for DNA sequencing and Drs David Ward and Mary Bateson for their valuable advice during the project.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
  29. [29].
View Abstract