Recently, a unique microbial community, growing in a whitish, macroscopically visible strings-of-pearls-like structure was discovered in the cold, sulfidic marsh water of the Sippenauer Moor near Regensburg, Bavaria, Germany. The pearls interior is predominated by microcolonies of the non-methanogenic SM1 euryarchaeon; the outer part of the pearls is mainly composed of Thiothrix. To screen sulfidic ecosystems for the distribution of such unique microbial communities, comparative microbial and geochemical analyses of cold, sulfidic springs of three geographically distinct locations in Bavaria, Germany, and Dalyan, Turkey, were performed. Here, we report on the discovery and study of another type of strings-of-pearls revealing a new microbial community structure. While the SM1 euryarchaeon is again the predominant archaeal constituent, the bacterial partner is the so-called IMB1 η-proteobacterium. Due to the predominance of the IMB1 η-proteobacterium, the strings-of-pearls reveal a fluffy and greyish macroscopical appearance. The phylogenetic survey revealed SM1 euryarchaeal relatives, designated as SM1 group, in all sites studied, indicating a widespread distribution of these archaea in terrestrial ecosystems.
A decade ago, the domain Archaea was considered to be confined to special environments [1,2]. A great diversity of hyperthermophilic archaea with optimal growth temperatures above 80 °C has been isolated from sulfur-rich, volcanic ecosystems . During the last years, however, archaea have been detected also as major constituents of less extreme and more common biotopes [4–6], especially owing to the use of the 16S rRNA gene as a molecular marker [7,8]. Despite their seeming omnipresence, their biology and ecological role in the biosphere remain largely obscure.
Springs with waters rich in sulfide are fairly common around the world and are in the focus of (micro)biologists for more than 150 years [9,10]. They were classified as areas of high bioorganic growth, as visible by the development of extended, (white) mats, vast populations of various genera of filamentous sulfur bacteria . During a recent microbial survey of cold (∼10 °C), sulfidic springs in the marsh “Sippenauer Moor” (SM) near Regensburg, Bavaria, we discovered a unique community of novel archaea and bacteria . In the sulfidic streamlet, these microbes grow in a string-of-pearls-like, macroscopically visible structure: tiny, whitish pearls (diameter up to 3 mm) are connected to each other by thin, white-coloured threads. The interior of each pearl is predominated by the non-methanogenic SM1 euryarchaeon, representing a deep phylogenetic branch within the 16S rRNA tree . The outer part of the pearls and their connecting threads are mainly composed of a single phylotype belonging to the filamentous sulfur bacterium Thiothrix (estimated up to approx. 95% of all bacterial cells) . The most obvious role of Thiothrix can be seen in the formation and maintenance of the three-dimensional structure of this microbial string-of-pearls community. Due to the constant coexistence of the SM1 euryarchaeon and Thiothrix, a syntrophic or even symbiontic relationship was proposed with a sulfur cycle or specific nutrient exchange occurring within single pearls .
Using their cold biotope as a “natural chemostat’, a novel cultivation method was designed and successfully applied to obtain larger quantities of microbial strings-of-pearls communities from nature. The exposure of large polyethylene nets allowed their fast, controlled and reliable growth and periodical harvesting once a week . From these microbial net populations, the SM1 euryarchaeon was specifically separated by gentle physical methods in the laboratory . This technique allowed us to gain new insights into the lifestyle of low-temperature archaea including the discovery of a novel cell surface structure, unique for prokaryotes, for which the name “hamus” (plural: “hami”) is proposed [14,15].
The ecology and geographical distribution of the SM1 euryarchaeon and microbial strings-of-pearls communities remained so far unknown. It was, therefore, of interest to investigate cold, sulfidic springs of distinct geographic areas for the occurrence of this unique microbial life form in a comparative survey, combining in situ observations with geochemical analyses, microscopical investigations and molecular studies. Here, we report on the discovery of a second biotope harbouring microbial strings-of-pearls communities, in which the SM1 euryarchaeon lives in close association with a so far unknown bacterial partner. Furthermore, SM1 euryarchaeal relatives were detected in all sites studied. Independently, sequences clustering within the SM1 group have been identified from habitats all over the world (Fig. 3), indicating a widespread distribution of these archaea in terrestrial (and marine) ecosystems.
2Materials and methods
2.1Collection and preparation of samples
Before sampling, the equipment was treated with 5% HClO4 to get rid of DNA impurities. From all sampling sites (Table 1), biomaterial with a macroscopic appearance comparable to microbial strings-of-pearls communities was preferentially sampled and processed as described .
Table 1. Description of the sampling sites, located in the southern part of Germany and in Dalyan, Turkey
Water temperatures were measured using a GTH 1150 digital thermometer (Greisinger Electronic GmbH, Regenstauf, Germany). Oxygen concentrations, pH and conductivities were determined using the multifunction-measurement instrument Multi-LineP4 (WTW, Weilheim, Germany). Sulfide content was quantified by the use of a colorimetric assay (hydrogensulfide-test; Merck KG, Darmstadt, Germany). Chemical analyses (Table 2) were performed at Blasy-Busse GmbH (Eching am Ammersee, Germany). The exact position of the sampling sites was determined with an eTrex Vista personal navigator (Garmin international, Olathe, KA).
Table 2. Chemical analyses of spring waters from Sippenauer Moor (SM) and Islinger Mühlbach (IM)
2.3FISH and oligonucleotide probes
For domain-specific FISH, three bacteria-specific probes and three archaea-specific probes were employed as mixtures in hybridizations. EUB338/I , EUB 338/II and EUB 338/III  were used for bacteria, while ARCH915 , ARCH344 and ARCH1060  were used for archaea (Table 3). Furthermore, the SM1 euryarchaeal probe SMARCH714 , the Thiothrix probe TN1  and the newly designed IMB1-specific probe IMB1–87 (5′-GCAAGCTTCCTCTCATCG-3′) were employed (Table 3). The archaeal oligonucleotides were labelled with rhodamine green, the bacterial probes with CY3. Whole cell hybridizations were carried out as recently described . After the FISH procedure, each sample was stained with 10 μl of DAPI (2 mg/l; prepared in washing buffer).
2.4DNA isolation, PCR, cloning, RFLP, sequencing and phylogenetic analyses
Cell lysis and bulk DNA extraction were performed as described . Extracted DNA was used as template for the PCR amplification of archaeal 16S rRNA gene sequences with the archaea-specific forward primers 109aF  or 345aF and the reverse primer 1119aR (Table 3) . PCR amplification of bacterial 16S rRNA gene sequences was performed with the forward primer 9f  and the reverse primer 1392r (Table 3) . PCR, cloning, restriction fragment length polymorphism (RFLP) and sequencing was done as described elsewhere . Clones with unique RFLP patterns were chosen for sequencing. Phylogenetic analyses were done according to Rudolph et al. . For the alignment, approximately 11,000 homologous full and partial sequences (>1300 nucleotides) available in public databases (ARB Project [24,25]) were used. For all tree reconstructions, maximum parsimony, distance-matrix (Jukes–Cantor correction) and maximum-likelihood (fastDNAml) methods were applied as implemented in the ARB software package.
2.5Nucleotide sequence accession number
The archaeal and bacterial 16S rRNA gene sequences were deposited in the EMBL nucleotide sequence database. The accession numbers are given in Figs. 3–5.
2.6Cloning, sequencing and comparison of intergenic spacer regions
For the amplification of the 16S–23S rDNA intergenic spacer region (ISR), the primers 1044aF  and 64R (23S; ) were used (Table 3). Four different samples from the Sippenauer Moor and the Islinger Mühlbach (IM) were analysed: microbial strings-of-pearls communities and specifically enriched archaeal net populations (see results; ). The PCR products obtained were cloned as described above. Thirty clones per sample were analysed by RFLP and those with unique restriction pattern were sequenced and compared, using the alignment program ClustalX 1.81 .
3.1Location and chemistry of the sampling sites
For comparative ecological studies, a total of seven cold (10–13 °C), sulfidic springs of three geographically distinct locations in Southern Germany, were selected for sampling. The areas were located in the vicinity of the Bavarian cities of Regensburg, Garmisch-Partenkirchen and Passau (Table 1; Fig. 1). The springs selected were characterized by a neutral pH, a sulfide concentration between 0.5 and 4 mg/l and a low salt content, typical for fresh water environments (Table 1). For comparative purposes, we also sampled a sulfidic spring near Dalyan, Turkey, which revealed a higher temperature of 38 °C and salinity comparable to half-strength seawater (Table 1). From sulfidic surface water of all sampling sites, whitish biomaterial with a macroscopic appearance reminiscent to the characteristic string-of-pearls community was preferentially collected (Table 1; Fig. 2).
In our initial studies, microbial community structure of the different strings-of-pearls-like samples was investigated by the use of domain-specific bacterial and archaeal probe mixtures and the specific SM1 euryarchaeal hybridization probe. FISH studies revealed that most samples consisted almost exclusively of bacteria, with filamentous morphotypes predominating. In addition, archaeal rod-shaped and coccoid cells were rarely detected in most of the samples (estimated to represent less than 0.1% of the total microbial population), which did not hybridize with the SM1-specific probe. In contrast, samples from the Islinger Mühlbach revealed the existence of large amounts of small cocci (approx. 50% of the total population), which hybridized with the archaeal and the SM1 probe.
3.3Archaeal 16S rRNA gene sequence analyses
In a FISH-independent study, the archaeal 16S rRNA gene pool of all samples was analysed (Table 1). Comparative phylogenetic analysis revealed that most of the sequences belonged to the euryarchaeal branch. Within this kingdom, they clustered together with the SM1 euryarchaeal sequence , but showed significant phylogenetic distances to each other (Fig. 3). The designation “SM1 group” is proposed for this new sequence cluster (Fig. 3). Furthermore, we detected 16S rRNA gene sequences related to the Methanomicrobiaceae (sampling sites of Bad Höhenstadt, Bad Gögging, Pilzweg and Sittling; Table 1) and even crenarchaeal sequences from the Pilzweg sample (Table 1). Together with clone sequences derived from soil, gold mine fissure water and from a termite gut, they form a new, separate sequence cluster within the crenarchaeota (Fig. 4).
Interestingly, all 37 clones obtained from the Islinger Mühlbach sample showed the same restriction pattern. The 16S rRNA gene sequence of four clones was determined, which showed the same base composition. Furthermore, it was identical to the SM1 euryarchaeal sequence from the microbial string-of-pearls community from the Sippenauer Moor (; Fig. 3). This result is in agreement with the FISH studies, indicating that the Islinger Mühlbach also harbours microbial strings-of-pearls communities. Therefore, it was of interest to perform a detailed comparative analysis of the communities from the Islinger Mühlbach (IM) and from the Sippenauer Moor (SM) [12–14].
3.4Comparative analysis of microbial strings-of-pearls communities from the Sippenauer Moor and Islinger Mühlbach
About 20 springs arise in the SM, most of them containing sulfide as a characteristic chemical compound . In contrast, the IM consists of a single sulfidic spring, which emanates from a metal lined drill hole with a diameter of 18.5 cm. The spring forms a small streamlet with high bioactivity; after a distance of about 30 cm from the rim of the drill hole differently coloured microbial mats and streamers become visible. In addition, microbial assemblages with a string-of-pearls-like morphology are floating in the water current (Fig. 2). From both sites, an extensive chemical spring water analysis was performed; the data are given in Tables 1 and 2.
Field observations over a certain time period showed a similar string-of-pearls morphology and growth behaviour in the Islinger Mühlbach compared to the Sippenauer Moor . However, in contrast to the whitish colour of the SM strings-of-pearls, the IM strings-of-pearls communities had a greyish appearance. Furthermore, single SM pearls were very compact in structure, while IM pearls appeared fluffier.
Periodical archaeal 16S rRNA gene sequence analyses of IM pearls verified our first phylogenetic analysis: almost all derived sequences belonged to a single phylotype that was identical to the SM1 euryarchaeal sequence from the SM microbial string-of-pearls communities. In line with this result, FISH studies of the inner part of about 100 IM pearls, sampled within about two years, revealed the existence of microcolonies of coccoid cells, hybridizing with the archaeal- and simultaneously with the SM1 euryarchaeal hybridization probe (Fig. 6(a)–(c)). In the IM pearls, these SM1 euryarchaeal cocci were grouped in a three-dimensional arrangement with defined distances between each other; this archaeal architecture is very similar to the one observed in the SM pearls . To analyse the phylogenetic relationship of the SM and IM euryarchaeota in more detail, the archaeal 16S–23S rDNA intergenic spacer region sequences from SM pearls and IM pearls were determined. We found that the sequences were identical, indicating that the same single archaeal phylotype SM1 exists in pearl communities of different biotopes.
Two independent bacterial phylogenetic analyses of IM pearls revealed one dominant clone sequence, IMB1, which showed identity to the recently obtained bacterial sequence sipK119 from SM pearls . The IMB1/sipK119 sequence clustered within the η-proteobacteria together with clone sequences LKC I recently obtained from microbial mats from Lower Kane Cave, Wyoming (Fig. 5) . In the IM pearls, we furthermore identified a clone sequence, which belonged to the genus Thiothrix with closest sequence similarity to the sipK4 sequence derived from SM pearls .
For a more quantitative analysis of the bacterial composition of IM pearls, phase contrast microscopy was combined with FISH studies, using domain- and Thiothrix-specific probes and the newly designed IMB1-specific hybridization probe (designated IMB1–87). Phase contrast microscopy revealed that the outer part of the pearls was mainly composed of filamentous microorganisms (approx. 95% of all morphotypes). The predominant filament type had a wavy form, a diameter of about 0.5 μm and did not reveal any cell inclusions (Fig. 6(a)). DAPI stain showed that the filaments consisted of a series of short rods with a cell length of about 1–2 μm (Fig. 6(c)). In addition, rosette-forming, thicker (ø 2 μm), straight filaments containing large amounts of refractile globules, most likely sulfur, were sometimes detected in the IM pearls. FISH analysis showed that both morphotypes belonged to the bacterial domain. The specific probes attributed the wavy filaments to the η-proteobacterium IMB1 (Fig. 6(b)), the thicker filaments to the genus Thiothrix. Periodical quantitative FISH analysis of approximately 100 pearls sampled within about two years showed that the outer part of the pearls was formed almost exclusively by IMB1 η-proteobacterium (Fig. 6(b); estimated approx. 90% IMB1, approx. 2%Thiothrix; approx. 8% bacteria with different morphologies). This demonstrates that IMB1 is the bacterial counterpart of SM1 euryarchaeon in the IM pearls, together forming a second type of microbial string-of-pearls community.
3.5In situ cultivation experiments of IM microbial strings-of-pearls communities
We used a recently developed in situ cultivation method to investigate the ability of IM microbial strings-of-pearls communities to grow on chemically defined material in their biotope . For these experiments, polyethylene nets were fixed in the sulfidic water current of the Islinger Mühlbach at a distance of approximately 50 cm from the rim of the borehole. The growth behaviour of the IM microbial communities observed over time was comparable to the one already reported for the SM microbial string-of-pearls . From the harvested IM microbial net population, the archaeal population could be selectively separated by physical methods as described recently . Sequencing of the archaeal 16S rRNA gene and the ISR of this archaeal fraction revealed in both cases sequence identity to the SM1 euryarchaeal fraction, obtained from net populations of the Sippenauer Moor . In line with these results, Immuno-FISH studies showed a strong and specific reaction of the IM euryarchaeota with the SM1-specific antibodies .
During comparative studies of cold, sulfidic springs of Southern Germany and Dalyan, Turkey (Table 1), we discovered a second biotope (Islinger Mühlbach, IM) harbouring microbial strings-of-pearls communities. Their growth morphology is similar to the recently described microbial string-of-pearls community from the Sippenauer Moor (SM). In the IM pearls interior, FISH revealed the existence of archaeal cocci, which hybridized strongly with the euryarchaeal SM1-specific probe. Furthermore, 16S rRNA gene sequence analyses showed that these archaea belonged to a single phylotype, identical to the SM1 euryarchaeon from SM pearls. This is an unanticipated result, considering the geographical and spatial separation of the two sampling sites.
It has been reported that sometimes bacterial strains with distinct physiologies have identical 16S rRNA genes [29,30] but show sequence variations in the highly variable 16S–23S rDNA intergenic spacer region. Therefore, we further analysed the SM1 euryarchaeal intergenic spacer region from IM- and SM pearls, which showed to be identical and thus were in line with the 16S rRNA gene sequencing data. In agreement with the molecular data is the characteristic three-dimensional growth behaviour of the SM1 euryarchaeon in microcolonies, already described for the SM pearls  (Table 4).
Table 4. Comparison of the microbial composition of single pearls from microbial string-of-pearls communities thriving in Sippenauer Moor and Islinger Mühlbach
Archaeal 16S–23S rDNA ribosomal intergenic spacer region
Sequence identity to SM1 euryarchaeon
Bacterial 16S rRNA gene sequence analysis
Thiothrix sipK4 predominantb
IMB1 η-proteobacterium predominant
SM1 euryarchaeon predominanta
SM1 euryarchaeon predominant
Morphology and arrangement of the SM1 euryarchaeon
Microcolonies of small cocci, three-dimensionally arranged in the pearls interiora
Microcolonies of small cocci, three-dimensionally arranged in the pearls interior
Thiothrix sp. predominantb
IMB1 η-proteobacterium predominant
Phase contrast microscopy of the predominant pearls bacterium
Rosette-forming, straight filaments (Ø 2 μm) with large amounts of refractile globules, consisting of amorphous sulfurb
Filamentous rods with a wavy form (Ø 0.5 μm) without visible cell inclusions
However, as already visible to the naked eye, the IM microbial strings-of-pearls communities differ significantly from their SM counterparts in their greyish colour and fluffy appearance. This morphological peculiarity can be attributed to the predominance of a novel filamentous bacterium (IMB1) belonging to the η-proteobacteria, which forms the outer part of the IM pearls. This is in remarkable contrast to the SM string-of-pearls, in which Thiothrix is the specific and constant partner of the SM1 euryarchaeon.
This finding raises the question of what is the trigger responsible for this bacterial exchange? The most obvious difference in geochemistry of the two sites is oxygen concentration, which is about 10-fold higher in the Sippenauer Moor (Tables 1 and 2). In line with these data is the dominance of Thiothrix, known to grow preferentially at higher oxygen tensions. Most likely the IMB1 η-proteobacterium is better adapted to the more micro-aerophilic conditions in the Islinger Mühlbach and would therefore outcompete Thiothrix in the formation of strings-of-pearls communities. These data are consistent with a recent microbial survey of bacterial mats in sulfidic springs in Lower Lane Cave, Wyoming. Phylogenetically close relatives of the IMB1 η-proteobacterium are dominant and outcompete Thiothrix at the low oxygen concentrations . So far, metabolic properties of the newly discovered IMB1 η-proteobacterium are unknown. However, by its ability to replace Thiotrix as the specific partner of the SM1 euryarchaeon, similar metabolic features can be envisaged: the oxidation of sulfide under very low oxygen tensions or even anaerobically, also suggested for uncultivated close relatives from the Lower Kane Cave . These data strengthen the idea of an internal sulfur cycle also in microbial strings-of-pearls communities from the Islinger Mühlbach, as already proposed for SM pearls .
Phylogenetic analyses revealed that cold sulfidic springs with a constant temperature of approximately 10 °C harbour novel groups of cren- and euryarchaeota. Most of the euryarchaeal clone sequences obtained clustered together with the SM1 euryarchaeon from the SM and IM strings-of-pearls. In consideration of the phylogenetic distances, the sequence-predicted archaea are related at different taxonomic levels. Together with some other environmental sequences and sequences of “rice cluster V” they form a new sequence cluster called “SM1 group”. We propose this designation, because SM1 is the first biologically identified representative of this group. Interestingly, many SM1 group sequences have been obtained from low-salt environments, some of them containing sulfide as a characteristic feature. Moreover, SM1 group sequences were derived from biotopes distributed all over the world, indicating their broad ecological significance (Fig. 3) [20,31–35]. Members of the SM1 group have been identified in sulfidic surface- and subsurface biotopes with higher salinity (Fig. 3), which suggests that this group may have a significant role also in marine environments [12,34].
We are indebted to the Government of Bavaria, Germany, for a sampling permit. Financial support from the Deutsche Forschungsgemeinschaft (Hur 711/2) is gratefully acknowledged.