Geobiology (2005), 3, 211227

Archaeal and bacterial communities in geochemically diverse hot springs of Yellowstone National Park, USA
Geomicrobiology in Yellowstone Blackwell O RIGINA Publishing L ART IC Ltd LE

D. R . M E YE R - D O M B A RD 1 , * , E . L . S H O C K 2 AN D J. P. AM EN D 1 , 3
1Department

of Earth and Planetary Sciences, Washington University, St. Louis, MO 63130, USA of Geological Sciences and Chemistry, Arizona State University, Tempe, AZ 85287, USA 3 Division of Biology and Biomedical Sciences, Washington University, St. Louis, MO 63130, USA *Present address: Department of Earth, Atmospheric, and Planetary Sciences, Massachusetts Institute of Technology, Cambridge, MA 02142, USA
2Departments

ABSTRACT
Microbiological and geochemical surveys were conducted at three hot springs (Obsidian Pool, Sylvan Spring, and Bison Pool) in Yellowstone National Park (Wyoming, USA). Microbial community structure was investigated by polymerase chain reaction (PCR) amplication of 16S rRNA gene sequences from DNA extracted from sediments of each hot spring, followed by molecular cloning. Both bacterial and archaeal DNA was retrieved from all samples. No Euryarchaea were found, but diverse Crenarchaea exist in all three pools, particularly afliating with deep-branching, but uncultivated organisms. In addition, cloned DNA afliating with the Desulphurococcales and Thermoproteales was identied, but the distribution of taxa differs in each hot spring. The bacterial community at all three locations is dominated by members of the Aquicales and Thermodesulfobacteriales, indicating that the knallgas reaction (aerobic hydrogen oxidation) may be a central metabolism in these ecosystems. To provide geochemical context for the microbial community structures, energy-yields for a number of chemolithoautotrophic reactions are provided for >80 sampling sites in Yellowstone, including Obsidian Pool, Sylvan Spring, and Bison Pool. This energy prole shows that the knallgas reaction is just one of many exergonic reactions in the Yellowstone hot springs, that energy-yields for certain reactions can vary substantially from one site to the next, and that few of the demonstrated exergonic reactions are known to support microbial metabolism. Received 14 July 2005; accepted 11 November 2005 Corresponding author: D. R. Meyer-Dombard. Tel: +1 (617) 253-7850; fax: +1 (417) 253-8298; e-mail: drmd@mit.edu.

INTRODUCTION
Yellowstones diverse geochemical environments host a variety of chemotrophic and phototrophic thermophilic communities (Brock, 1978; Barns et al., 1994, 1996; Hugenholtz et al., 1998; Pierson et al., 1999; Langner & Inskeep, 2000; Reysenbach et al., 2000; Jackson et al., 2001; Johnson et al., 2001; Langner et al., 2001; Blank et al., 2002; Boomer et al., 2002; Fouke et al., 2003; Papke et al., 2003; Inskeep et al., 2004; Jahnke et al., 2004; Macur et al., 2004; Spear et al., 2005). Since the early work of Thomas Brock, numerous thermophiles have been isolated from these springs, including members of Thermus, Thermocrinis, Sulfolobus, Thermomicrobium, Thermoanaerobium, Thermodesulfovibrio, Thermodesulfobacterium, Thermoanaerobacter, Anaerobranca, Thermobacteroides, Acidianus, and Sulphurococcus (Brock & Freeze, 1969; Brock et al., 1972; Jackson et al., 1973; Zeikus et al., 1979, 1983; Ben-Bassat
2005 The Authors Journal compilation 2005 Blackwell Publishing Ltd

& Zeikus, 1981; Wiegel & Ljungdahl, 1981; Henry et al., 1994; Engle et al., 1995). However, many of these organisms appear only infrequently in 16S rRNA libraries, which feature primarily sequences from novel, as yet uncultured, strains (Barns et al., 1994, 1996; Hugenholtz et al., 1998; Reysenbach et al., 2000). As an example, 16S rRNA signatures of 54 novel Bacteria and 36 novel Archaea, including members of the deeply branching Korarchaea, were identied in Obsidian Pool (Mud Volcano area) (Barns et al., 1994, 1996; Hugenholtz et al., 1998). Based on these molecular surveys, this pool appears to be one of the most genetically diverse sites on Earth, yet the vast majority of these organisms have not been cultured. The environmental factors that most directly inuence hot spring microbial diversity remain unresolved. Temperature has perhaps received the most attention, but other constraining factors may include pH, redox disequilibria, trace element

211

212

D. R. MEYER-DOMBARD et al.
2 H2S + 2O2 SO4 + 2H+ ,

levels, organic matter composition, solar radiation budgets and trophic relationships. It has been argued previously that geochemical energy-yields may be a key determinant of microbial community structure and diversity in thermal environments (Baross, 1995; Amend & Shock, 2001). Accordingly, geochemical energy-yields from dozens of redox reactions were evaluated in deep-sea hydrothermal systems (McCollom & Shock, 1997; Bach & Edwards, 2003), shallow submarine hydrothermal systems (Amend et al., 2003, 2004) and continental hot springs (Shock et al., 2005a,b; Spear et al., 2005). It was shown that in situ energy-yields depend much more on chemical composition (such as redox sensitive compounds) than on temperature, and the energy available to thermophiles from any given metabolism can vary substantially from one site to the next. In illustration, in >80 Yellowstone locations investigated by Shock et al. (2005a,b), the free energy from ferric iron reduction, represented by goethite + 0.5H2,g + 2H+ Fe2+ + 2H2O, (1)

(3)

resulting in acidic uids of elevated sulphate concentrations capable of considerable chemical weathering, which releases cations into solution (White et al., 1971). Acid-sulphate systems at Yellowstone include the Mud Volcano area (encompassing Obsidian Pool) and Fountain Paint Pots. In contrast, waterdominated systems (alkaline-chloride and carbonate) are characterized by substantial waterrock interaction and subsurface boiling in deep-heated reservoirs, but little interaction with reduced mantle-derived gases (White et al., 1971; Fournier, 1989). Many of the large geyser basins at Yellowstone, including the Lower Geyser Basin (encompassing Bison Pool), host alkaline-chloride and carbonate thermal features. Owing to the differing geochemical consequences of these processes, chloride and sulphate concentrations (as well as pH) can be used to chemically categorize thermal waters (Ellis & Mahon, 1964, 1967; Krauskopf, 1964), an example of which is depicted in Fig. 2. Bison Pool Bison Pool (Fig. 1A) is the unofcial name for a small hot spring in the Sentinel Meadow (Lower Geyser Basin) area

(catalysed by the Obsidian Pool bacterium Geothermobacterium ferrireducens (Kashe et al., 2002)) is strongly pH dependent and highly variable [8 to 55 kJ (mol e)1]. In contrast, the knallgas reaction, H2,g + 0.5O2,g H2O, (2)

(mediated by many species of Aquicales found in Yellowstone (Huber et al., 1998; Spear et al., 2005), yields 100 kJ (mol e)1 on average at all locations studied. To more directly investigate the effects of temperature, pH and other geochemical parameters on thermophile diversity, the microbial community structures at one location in each of three Yellowstone hot springs were determined and interpreted in light of energy-yields. Extensive geochemical analyses were performed over a period of several years at >80 sample locations to provide environmental context, and three individual springs were chosen for a focused microbiological survey. The three springs1 Obsidian Pool, Sylvan Spring, and Bison Pool (Fig. 1) are all 80 C, range in pH from 5.5 to 8.1, and differ substantially in their overall geochemical characteristics.

CHARACTERIZATION OF SAMPLING AREAS

Hot spring features, including those investigated in this study, are generally categorized in two classes vapor-dominated (acid-sulphate) systems and water-dominated (alkaline-chloride and carbonate) systems (Fournier, 1989). At Yellowstone, vapordominated systems are fed via fractures by reduced gases (e.g. H2 and H2S) from the 3 10 km deep magma body (Fournier & Pitt, 1985; Kennedy et al., 1985; Fournier et al., 1994). Approaching the surface, the sulphides are oxidized in accord with, for example,

The thermal feature referred to here as Sylvan Spring (Fig. 1C) of the northern Sylvan Springs group (Gibbon Geyser Basin) has been given multiple names in the literature, causing some confusion. Early observations (Allen & Day, 1935) described the region much as it appears today, with extensive chemical weathering, acidic pools, and sulphur depositing fumaroles, but these did not mention the pool shown in Fig. 1C. Researchers from the US Geological Survey (USGS) made note that this pool, the largest in the group, was not obvious on an aerial photograph taken prior to 1959 (J. Ball, pers. comm. September, 2004), indicating that the feature may have developed more recently. Whittlesey (1988) names this feature Dantes Inferno (Whittlesey 1998), placing a different Sylvan Spring 300 feet to the north. A sketch map in Brock (1978) misidentied Dantes Inferno as Sylvan Spring, and as a result, many refer to the location by this name. This naming is reinforced because the feature is the largest hot spring in the group (thus lending its name to the whole region). We have decided to adhere to the naming convention used by the majority of workers in this eld, calling the feature in Fig. 1C Sylvan Spring. The USGS team did not name this feature in their geochemical surveys of the area (Ball et al., 1998, 2001), and this decision should not confuse others when referencing these works. Shock et al. (2005b) consistently refer to the feature as Sylvan Spring, labelling it as such on detailed maps of the area. In addition, it should be noted that Bison Pool is an unofcial name for the feature whose GPS coordinates are given in the Methods section. While some researchers refer to this location as Rosette Geyser, an ofcial National Park designation of a hot spring in the area, our efforts to compare older maps and data (Allen & Day, 1935) with our maps and eld observations suggest that these features are different. Indeed, there are several dormant and extinct features in the general area, which may be Rosette Geyser. For our purpose, we will leave this matter unresolved, and provide GPS coordinates. 2005 The Authors Journal compilation 2005 Blackwell Publishing Ltd

Geomicrobiology in Yellowstone

213

Fig. 2 Sulphate and chloride concentrations (in ppm) measured in 19992003 from the Lower Geyser Basin (open squares), the greater Obsidian Pool area (open circles), and the Sylvan Spring area (open triangles). Corresponding lled symbols indicate the three key 2001 locations used in the focused microbiological survey (Bison Pool, square; Obsidian Pool, circle; Sylvan Spring, triangle).

trace element concentrations (e.g. B, Al, Rb, W) relative to the other two study areas. These geochemical signatures suggest that the Lower Geyser Basin thermal uids have undergone substantial waterrock interaction. The sampling location used for the microbiological survey reported here for Bison Pool (arrow, Fig. 1A) was 82.6 C at a pH of 8.1. Obsidian Pool The Mud Volcano region, home of the greater Obsidian Pool area (1 km2), has been dened as a vapor-dominated system (White et al., 1971; Zohdy et al., 1973). Gas-delivering faults associated with the Sour Creek resurgent dome intersect the area (Kennedy et al., 1985). However, the greater Obsidian Pool area also contains thermal features characteristic of waterdominated systems (open circles in Fig. 2). With few exceptions, the thermal waters contain intermediate to high levels of chloride (10 500 ppm), but sulphate concentrations range from the lowest (5 ppm) to the highest (>1000 ppm) values we observed. Our Obsidian Pool sample (arrow, Fig. 1B and lled circle in Fig. 2), 79.9 C, is characterized by moderate levels of chloride (20 ppm) and sulphate (100 ppm) with a near neutral pH of 6.5 (similar to hot water-dominated systems), but elevated cation concentrations (e.g. NH+ 4, Fe+2, Ca+2, Mg+2), characteristic of vapor-dominated systems. This sample can be interpreted as a diluted acid-sulphate uid.

Fig. 1 Photographs of (A) Bison Pool (pH, 8.1), (B) Obsidian Pool (pH, 6.5), and (C) Sylvan Spring (pH, 5.5), with arrows indicating the sampling locations. At the arrow in (B), the pool is 2 m across.

that features mound-forming siliceous sinter deposits. It can be seen in Fig. 2 that samples from the Lower Geyser Basin (open squares), including Bison Pool (lled square), exhibit high chloride and low sulphate concentrations. In addition, the Sentinel Meadow hot springs are alkaline (pH 79.5, where Bison Pool 8) and generally feature certain elevated
2005 The Authors Journal compilation 2005 Blackwell Publishing Ltd

214

D. R. MEYER-DOMBARD et al. is a rapid mixing of uids in the sedimentsurface interface due to the very active nature of these features. In addition, to provide environmental context, a broad geochemical survey was conducted between 1999 and 2003 at various other features in the Obsidian Pool, Sylvan Spring, and Sentinel Meadow geothermal systems. DNA extraction from sediment samples Because different DNA extraction methods have different recovery efciencies (Hugenholtz et al., 1998; Miller et al., 1999), ve methods were applied to each of the frozen samples: bead-beating using the Bio101 extraction kit (Q-BIOgenes FastDNA Spin Kit for Soil; Carlsbad, CA, USA), bead-beating (adapted from Bond et al. (2000)), freeze/thaw (adapted from Barns et al. (1994) and pers. comm. with C. Blank), freeze/thaw/grind (adapted from Hurt et al. (2001)), and chemical extraction (adapted from Miller et al. (1999)). To prevent damage to extracted nucleic acids, the extraction protocols were adjusted to suit the pH at each study area. For specic protocols, see Meyer-Dombard (2004). PCR of extracted community DNA

Sylvan Spring Like the greater Obsidian Pool area, the Sylvan Springs area (Gibbon Geyser Basin) is characteristic of both vapor-dominated and water-dominated systems. However, Sylvan Spring uids (open triangles in Fig. 2) are all high in sulphate (>100 ppm), but range in chloride levels from the lowest (<0.1 ppm) to the highest ( 600 ppm) measured. Numerous springs in a ravine formed by a NESW trending fault that appears to follow an anticlinal fold hinge are highly acidic (pH 1.93.2), and the surrounding rock is extensively weathered. Sylvan Spring proper (Fig. 1C and lled triangle in Fig. 2) and other features on the anks of the ravine exhibit both acid-sulphate and alkalinechloride signatures. They range in pH from 4.1 to 7.5 (Sylvan Spring 5.5), and, compared to most acidic features in the area, their sulphate concentrations are lower and their chloride concentrations are higher (Fig. 2). In addition, the Sylvan Spring sample used in the microbiological survey (81 C) contains 550 ppm SiO2(aq), 300 ppm Na+, and elevated trace elements (data not shown) typical of an alkaline-chloride system.

METHODS
Sample locations /collection Sediment and hydrothermal uid samples for a focused, three-pool survey of geochemistry and microbiology were collected in 2001 from Obsidian Pool (Mud Volcano Area), Sylvan Spring (North Sylvan group, Gibbon Geyser Basin), and Bison Pool (an unnamed feature in the Sentinel Meadow, Lower Geyser Basin). Specically, samples were taken at one location in each pool (arrows, Fig. 1) at the following coordinates: Obsidian Pool (443610.9N, 1105154.7E), Sylvan Spring (44 4155.9N, 110465.8E), Bison Pool (443410.9N, 1105154.7E). Samples for DNA extraction were retrieved at these locations from just below the sediment/water interface (top 5 cm). Sediment samples were collected with a sterile spatula, placed in 50 mL Falcon tubes, kept on ice in the eld, and stored (within 6 h) at 20 C. Sediments for total cell counts using DAPI (4,6diamidino-2phenylindole; Sigma, St. Louis, MO, USA) were stored in a 4% paraformaldehyde:phosphate-buffered-solution (PFM:PBS). The sediment was allowed to x for 3 6 h on ice, the xative was decanted, PBS: 96% ethanol (EtOH, 1 : 1 v/v) was added and samples were stored at 20 C. The depth to the interface between thermal uid and sediment was 510 cm in each case. Hydrothermal uid and gas at these three locations were collected concurrently with the microbiological samples (for details see Shock et al., 2005a). Due to the volume of sample needed for the extensive geochemical analysis, hydrothermal uid was collected from the 510 cm above the sediment interface. Ideally, these analyses would be done using pore uid from the upper 5 cm of sediment. This is, however, beyond the scope of this investigation: it is probable that there DNA was amplied with primer pairs 21F1391R (targeting archaea) and 27F-1492R (targeting bacteria) (Lane, 1991) using a Hybaid PCR Express thermalcycler. Cycling conditions for 21F1391R: preheat at 95 C for 5 min; 30 cycles at 95 C for 30 s, 55 C for 30 s, and 72 C for 1.5 min; extension at 72 C for 15 min. Cycling conditions for 27F 1492R: preheat at 95 C for 5 min; 35 cycles at 95 C for 1 min, 52 C for 1 min, 72 C for 1 min; and extension at 72 C for 5 min. The reaction mixtures included (per 20 L reaction) 1.8 L 25 mM MgCl2, 0.2 L dNTP mix (12.5 mol of each dATP, dGTP, dTTP and dCTP; Bioline, Randolph, CA, USA), 0.25 L 5 U L1 AmpliTaq Gold (Applied Biosystems, Foster City, CA, USA), 2.0 L 10 AmpliTaq Gold reaction buffer (100 mM Tris-HCl, pH 8.3, 500 mM KCl), 0.25 pmol L1 primer (IDT), and 13.75 L ddH2O. Molecular cloning of amplied community DNA Polymerase chain reaction (PCR) products from the ve extractions were combined and cloned using the Invitrogen TOPO TA cloning kit (cat. no. K4500-01; Carlsbad, CA, USA). This resulted in six libraries, one archaeal and one bacterial for each of the three samples. The efciency of the ve extraction methods was not evaluated in this study. Between 192 and 384 white colonies were picked from each library, and the plasmids were puried using the QIAGEN Miniprep Kit (cat. no. 27106; Valencia, CA, USA). The insert was amplied using M13F and M13R primers. The reaction mix included: 1.2 L 25 mM MgCl2, 0.2 L dNTP mix, 0.25 L 5 U L1 AmpliTaq
2005 The Authors Journal compilation 2005 Blackwell Publishing Ltd

Geomicrobiology in Yellowstone Gold, 2.0 L 10 AmpliTaq Gold reaction buffer, 2 pmol L1 primer (forward and reverse), and 14.95 L ddH2O. Reactions were cycled as follows: preheat at 95 C for 5 min; followed by 25 cycles at 94 C for 30 s, 58 C for 30 s, and 72 C for 1 min; followed by a 15-min extension at 72 C. Products were initially screened using restriction fragment length polymorphism (RFLP) by HhaI (New England BioLabs, Ipswich, MA, USA) digestion. Digests were run on a 4% MetaPhor gel (Cambrex, Baltimore, MD, USA) to identify unique banding patterns. Following maximization of the rarefaction curves (data not shown), screening was done by short segment (700 bp) sequencing. Sequencing and phylogenetic inference At least one clone representing each unique RFLP banding pattern in each library was chosen for sequencing. A 700 bp fragment of the 16S rRNA gene was sequenced for 35 bacterial and 51 archaeal clones from Sylvan Spring, 49 bacterial and 67 archaeal clones from Bison Pool, and 50 bacterial and 77 archaeal clones from Obsidian Pool. Following this initial screening, the full 16S rRNA gene was sequenced for unique clones from the Bison Pool (22) Obsidian Pool (33), and Sylvan Spring (14) libraries. Plasmids with DNA inserts were directly sequenced using the primers 21F/958R/1391R and 27F/907R/1492R. Cycling conditions for sequencing reactions were: preheating step at 96 C for 3 min; 35 cycles at 96 C for 30 s, annealing temperature (Ta) for 15 s (where Ta = 55 C (21F), 52 C (27F, 958R), 45 C (907R, M13R), 57.3 C (1392R), 50 C (1492R), 48.3 C (M13F)), and 60 C for 4 min. Sequencing reactions contained 37.5 pmol L1 primer, 1.0 L BigDye Terminator premix version 3.1 (Applied Biosystems), and ddH2O to bring the nal volume to 5 L. Products were sequenced on an Aurora Genetic Analysis System (SpectruMedix, State College, PA, USA). Contiguous sequences were assembled using SEQUENCHER (version 4.1.4; Gene Codes Corporation, Ann Arbor, MI, USA), submitted to GenBank (accession numbers DQ243725 DQ243783), and compared to the NCBI (National Center for Biotechnology Information) database using BLAST (Altschul et al., 1997) to nd closest relatives. Approximately 12001400 nucleotide bases were aligned using the software BIOEDIT (version 5.0.9; http:/ /www.mbio.ncsu.edu/ BioEdit/ bioedit.html) and then manually adjusted using the predicted secondary structure of the molecule as a template. Chimeric analyses were largely performed by careful alignment comparisons, but the CHIMERACHECK (Cole et al., 2003) and BELLEROPHON (Huber et al., 2004) software were also used as guidelines. Phylogenetic inference of homologous positions was performed using neighbour joining (NJ), maximum parsimony (MP), and maximum likelihood (ML) with the software PAUP (version 4b.10; Sinhauer Associates, Sutherland, MA, USA). MP analyses were performed with random addition of taxa (1000 replicates). Long-branch attraction issues were
2005 The Authors Journal compilation 2005 Blackwell Publishing Ltd

215

addressed by considering branch lengths of >10% of the total number of nucleotides aligned as long branches, and the analysis was adjusted appropriately. In addition, the placement of taxa within monophyletic groups was tested by additionsubtractionreplacement experiments. For ML tree construction, evolutionary models were evaluated with a likelihood ratio test to nd the simplest model appropriate for the data (Huelsenbeck & Rannala, 1997), which was then used to construct the trees. This process was iterated three times with 1000 replicates in the nal iteration. All trees were evaluated using only unambiguously aligned nucleotides and bootstrapped (1001000 replicates) (Felsenstein, 1985). Taxa not uniquely afliated with a phylogenetic position are noted in the results. Geochemistry Analyses of water and gas samples collected between 1999 and 2001 are reported in Shock et al. (2005a,b). Temperature, pH, conductivity, and alkalinity were measured in the eld with hand-held meters (YSI 30, Orion 290 A plus). Portable spectrophotometers (Hach, Loveland, CO, USA) were used in the eld to measure dissolved oxygen, nitrate, nitrite, total ammonia, ferrous iron, total sulphide, and dissolved silica. Major ions and trace elements were analysed in the laboratory using ion chromatography (Dionex, Sunnyvale, CA, USA) and inductively coupled plasma mass spectrometry (Finnigan), respectively. Gases were collected in evacuated Giggenbach bottles containing known amounts of NaOH (Giggenbach, 1975; Giggenbach & Goguel, 1989; Giggenbach et al., 2001) and analysed by gas chromatography and established wet chemical methods (Giggenbach & Goguel, 1989) in the laboratory of Tobias Fischer, University of New Mexico. Results of gas analyses are not reported directly here, but these data were used in the energetic calculations. The compositional data were converted to chemical activities using the computer code EQ3 and used to compute values of the overall Gibbs free energy (Gr) as described in Amend & Shock (2001) and Shock et al. (2005a). Briey, values of Gr for >180 inorganic reactions were computed using the relation:
o G r = G r + RT ln Q r ,

(4)

o where G r stands for the standard state Gibbs free energy of reaction, R and T represent the gas constant and temperature (in Kelvin), respectively, and Q r denotes the activity product, which is expressed as:

Q r = a vi,r i

(5)

where a i stands for the activity of the ith component in the reaction, and i,r denotes the stoichiometric reaction coefcient. By convention, a negative value of Gr indicates an exergonic, or energy-yielding, reaction.

216

D. R. MEYER-DOMBARD et al.

RESULTS AND DISCUSSION

The archaeal and bacterial community structures in sediments from three hot springs in Yellowstone were determined and then interpreted in the context of in situ chemical composition and reaction energetics. The three sites (Obsidian Pool, Sylvan Spring, and Bison Pool) chosen for this comparative geobiological study of springs are of similar temperature (79.9 82.6 C), but variable chemistry (e.g. pH: 5.58.1; conductivity: 650 5050 S cm1). Whole cell counts in the three sediment samples are as follows: Obsidian Pool, 2.3 106 cells cm3; Sylvan Spring, 3.2 106 cells cm3; and Bison Pool, 2.1 106 cells cm3. Community structure analysis Archaeal and bacterial DNA was successfully amplied from all three sediment samples. PCR products from all extractions were pooled, and two libraries (archaeal and bacterial) were constructed for each sample. The assembled sequences of the 16S rRNA gene for each clone were submitted to GenBank and assigned accession numbers DQ243725DQ243783. All sequenced clones, their phylogenetic afliations, the closest GenBank match, percentage similarity to the GenBank data, and the numbers of clones screened by RFLP patterns and partial gene sequencing are recorded in Table 1. Following RFLP screening, rarefaction curves (data not shown) for ve of six libraries maximized at <40 plasmids; the exception was the Obsidian Pool archaeal library, which had not maximized even after screening 66 clones. The remaining clones analysed were screened using short segment sequencing. Analysis of RFLP banding patterns revealed 13 bacterial and 35 archaeal patterns at Obsidian Pool, 8 bacterial and 26 archaeal patterns at Sylvan Spring, and 16 bacterial and 32 archaeal patterns at Bison Pool. The clonal distribution differed between the RFLP and sequencing screening methods in some cases. For example, 12 of 67 clones screened using sequencing, but only 5 of 75 clones screened using RFLP were identied as YNP BP A23. The differences are attributed to difculties in identifying unique banding patterns using RFLP, and the consistency of conserved restriction sites. The distribution and diversity of clones representing unique taxa are shown in Fig. 3 as percentages of the total number of clones screened by partial (700 bp) 16S rRNA sequencing. The six pie charts represent the archaeal (Fig. 3A,C,E) and bacterial (Fig. 3B,D,F) libraries from Bison Pool, Obsidian Pool, and Sylvan Spring; each pie wedge is colour-coordinated such that taxa within each domain can be readily compared among the three sample locations. These results reect the distribution of clones within each library, but do not necessarily quantify the diversity in the natural environment. Figure 3 reveals that, although certain taxa are similar between all three sediment samples, the distribution of taxa between the hot springs varies greatly.

Phylogenetic analysis of archaeal clones Phylogenetic evaluation of archaeal sequences is shown in Fig. 4A; a subset of the uncultured crenarchaeal clones are highlighted in Fig. 4B. Except where noted below, neighbour-joining and maximum-likelihood analyses were in close agreement with the maximum parsimony tree depicted here. Samples from the current study are shown in bold and coded according to sample location. All three archaeal communities were dominated by Crenarchaeota, containing members of the Desulphurococcales, Thermoproteales, and uncultured Crenarchaeota. Sulfolobales were not observed in this study despite their reported presence at the Sylvan Springs (Brock et al., 1972; Brock, 1978). No Euryarchaeota and only one Korarchaeon (from Obsidian Pool) were found in the three samples. It can be seen in Figs 3 and 4 that the archaeal communities from the three pools are similar on the order level, but vary signicantly on the genus level, examples of which are discussed separately below. Clones afliating with the Desulphurococcaceae were found in all three pools, but no members of the Pyrodictiaceae were identied. The Desulphurococcaceae clones fell into ve groups, four of which (D-I to D-IV) may be novel clades, forming unassociated monophyletic groups that are not afliated with the other known genera of Desulphurococcaceae (Aeropyrum, Thermodiscus, Ignicoccus, Staphylothermus, Stetteria, and Sulfophobococcus). The remaining clones (SSp A84 and BP A49) fall within clades containing Desulphurococcus mobilis (96.2% similarity) and Thermosphaera aggregans (99.1% similarity). The novel phylotypes are generally well supported by the bootstrap analysis. However, the placement of group D-II may be affected by long-branch attraction artefacts; neighbourjoining and maximum-likelihood analyses (not shown) placed BP A73 as a separate branch within group D-IV. Further, group D-I may represent a separate monophyletic group related to Aeropyrum, Thermodiscus, and Ignicoccus, and group D-III may afliate most closely with Pyrodictium and Pyrolobus. Regardless of the exact placement of each clade, signicant diversity in the Desulphurococcaceae is evident in all three sediments. Differences in the distribution of D-I, D-II, D-III, and D-IV among the three sites are noteworthy. It can be seen in Fig. 3A,C,E, and Fig. 4 that group D-I is comprised entirely of Bison Pool clones, and group D-IV contains only samples from Obsidian Pool and Sylvan Spring. However, groups D-II and D-III are represented in all three pools. Note that these novel Desulphurococcales represent >65% of the Bison Pool library, compared to 11% of the Obsidian Pool and 23% of the Sylvan Spring libraries (Fig. 3). Within the Thermoproteales, close relatives of members of the Thermoproteaceae and Thermolaceae were found in this study. Clones from all three samples (ObP A130, ObP A166, SSp A81, and BP A6) form a clade with the previously identied Yellowstone clone pBA2 (Reysenbach et al., 2000). While closely related (99.5% similarity), short-branch attraction
2005 The Authors Journal compilation 2005 Blackwell Publishing Ltd

Geomicrobiology in Yellowstone

217

Table 1 Inventory of clones analysed, the associated gure number, afliated phylogenetic group, closest GenBank match and % similarity (based on BLAST search), and screening results (RFLP and short-segment sequencing) Clone* (type representative) YNP BP A23 YNP BP A10 YNP BP A78 YNP BP A73 YNP SSp A61 YNP ObP A84 YNP BP A60 YNP BP A32 YNP BP A9 YNP ObP A110 YNP BP A22 YNP SSp A24 YNP SSp A51 YNP ObP A25 YNP BP A49 YNP SSp A84 4a YNP ObP A130 4a YNP BP A6 YNP ObP A166 YNP SSp A81 YNP SSp A50 YNP SSp A10 YNP BP A17 YNP BP A81 YNP ObP A97 YNP ObP A136 YNP ObP A150 YNP BP A89 YNP ObP A31 YNP SSp A70 YNP ObP A62 YNP ObP A68 YNP ObP A5 YNP ObP A23 YNP ObP A16 YNP ObP A153 YNP ObP A17 YNP BP A103 YNP ObP A10 YNP ObP A42 YNP ObP B89 YNP ObP B32 YNP BP B97 YNP BP B71 YNP BP B86 YNP SSp B90 YNP SSp B60 YNP ObP B24 YNP BP B68 YNP BP B72 YNP SSp B24 YNP ObP B80 YNP BP B73 YNP ObP B48 YNP ObP B45 YNP ObP B34 4a 4a 4b Fig 4a Phylogenetic Group D-I (unknown Desulfurococcales) Closest GenBank match (% similarity) Aeropyrum pernix (AB078022) (94%) A. camini (AB109559) (94%) Staphylothermus marinus (X99560) (92%) Stetteria hydrogenophilia (Y07784) (94%) S. marinus (92%) S. marinus (92%) S. marinus (93%) S. marinus (95%) Sulfophobococcus zilligii X98064 (91%) (AB075802) (91%) Pyrodictium occultum (M21087) (93%) Caldococcus noboribetus (D85038) (94%) Desulfurococcus mobilis (M36474) (89%) D. mobilis (89%) Thermosphaera aggregans (X99556) (94%) D. mobilis (90%) Pyrobaculum organotrophicum (AB063647) (92%) Thermoproteus neutrophils (AB009618) (98%) P. organotrophicum (93%) P. organotrophicum (92%) Vulcanisaeta distributa (AB063630) (97%) Thermofilum pendens (X14835) (96%) T. pendens (94%) Clone pJP41 (L25301) (93%) Clone pJP41 (87%) Clone pJP41 (87%) Clone pJP33 (L25300) (99%) Clone pJP33 (99%) Clone pJP96 (U63338) (99%) Clone pISA9 (AB019732) (86%) Clone HAuD-LA19 (AB113630) (97%) Clone SUBT-13 (AF361212) (98%) Clone 19b-34 (UAR294865) (91%) Clone pJP89 (L25305) (99%) Clone pISA7 (AB019733) (85%) Isolate S10TFL (X99564) (89%) Clone pISA7 (87%) Clone pUWA2 (AB007307) (86%) Isolate S10TFL (X99564) (84%) Clone pJP78 (U63344) (98%) Thermocrinis ruber (L09672) (97%) Thermocrinis ruber (L09672) (97%) Thermocrinis ruber (L09672) (97%) Thermocrinis ruber (L09672) (97%) Thermocrinis ruber (L09672) (98%) Isolate YNP-SS1 (AF507961) (99%) Isolate YNP-SS1 (AF507961) (99%) Isolate YNP-SS1 (AF507961) (99%) Clone SHA-2 (AJ306740) (83%) Thermodesulfobacterium hveragerdicum (X96725) (89%) Clone SRI-27 (AF25595) (99%) Geothermobacterium ferrireducens (AF411013) (98%) G. ferrireducens (99%) Clone SRI-93 (AF255596) (98%) Clone SRI-27 (98%) Nesterenkonia sp. (AY574575) (96%) Clonal distribution 12/67 : 5/75 12/67 : 4/75 12/67 : 1/75 1/67 : 3/75 1/51 : 1/88 2/77 : NA 6/67 : 4 /75 2/67 : 1/75 1/67 : 1/75 1/77 : 1/89 5/67 : 1/75 1/51 : 1/88 1/51 : 1 /88 3/77 : 2 /89 2/67 : 2 /75 4/51 : 16 /88 1/77 : 2 /89 7/67 : NA 5 /77 : 3 /89 2/51 : 4 /88 2/51 : 3/88 3/51 : 1/88 2/67 : 1/75 1/67 : 1/75 4 /77 : NA 4 /77 : 2/89 2/77 : 8/89 1/67 : 14/75 7/77 : 5/89 2/51 : 8/88 1/77 : NA 1/77 : NA 1/77 : 2/89 4 /77 : 3/89 7/77 : 8/89 7/77 : 8/89 1/77 : 2/89 1/67 : NA 1/77 : 1 /89 2/77 : 5 /89 4/50 : 1/66 10/50 : NA 11/49 : 6/83 11/49 : 23/83 6/49 : 14/83 22/35 : 1/86 22/35 : 3/86 1/50 : 3/66 4/49 : 2/83 4/49 : 4/83 1/35 : 4/86 3/50 : 3/66 2/49 : NA 1/50 : 3 /66 5/50 : 3 /66 2/50 : 1/66

4a

D-II (unknown Desulfurococcales)

4a

D-III (unknown Desulfurococcales)

4a

4a

D-IV (unknown Desulfurococcales) Desulfurococcales/ Thermosphaera Desulfurococcales/ Desulfurococcus Thermoproteales/ Thermoproteus

Thermoproteales/ Vulcanisaeta Thermoproteales/ Thermofilum UC-I (Uncultured Crenarchaeota)

4a

UC-II (Uncultured Crenarchaeota)

4a 4a

UC-IV (Uncultured Crenarchaeota) UC-V (Uncultured Crenarchaeota)

4a 5

Korarchaeota Aquicales/ Thermocrinis

Aquicales/ Hydrogenothermus Unknown Aquicales Thermodesulfobacteriales Geothermobacterium Thermodesulfobacteriales Actinobacteriales

5 5

5 5

2005 The Authors Journal compilation 2005 Blackwell Publishing Ltd

218

D. R. MEYER-DOMBARD et al.

Table 1 Continued Clone* (type representative) YNP ObP B94 YNP ObP B118 YNP ObP B70 YNP BP B75 YNP SSp B57 YNP ObP B67 Fig 5 5 5 5 Phylogenetic Group Clostridiales Proteobacteria beta-Proteobacteria gamma-Proteobacteria Closest GenBank match (% similarity) C. thermocellum (L09173) (91%) S. xenophaga (AY611716) (99%) Tepidimonas sp. (AY324139) (96%) Stenotrophomonas sp. (AJ293468) (99%) Stenotrophomonas sp. (AJ293468) (99%) Stenotrophomonas sp. (AJ293468) (99%) Clonal distribution 1/50 : 1/66 1/50 : NA 1/50 : 3 /66 1/49 : 1/83 3/35 : 1/86 1/50 : 1/66

*YNP, Yellowstone National Park; BP, Bison Pool; ObP, Obsidian Pool; SSp, Sylvan Spring; A no., archaeal clone no.; B no., bacterial clone no. As seen in Figs. 4 and 5. Based on BLAST search. Percentage similarity determined by BLAST search and distance matrices (PAUP). Number of clones like sample/number of clones screened by sequencing : number of clones like sample/number of clones screened by RFLP.

Fig. 3 Diversity and distribution of plasmids within archaeal and bacterial clone libraries for Bison Pool (A, B), Obsidian Pool (C, D), and Sylvan Spring (E, F). Pie wedges indicate the percentages of each taxon in the total library, as determined by screens of partial (700 bp) 16S rRNA sequences (see Table 1). Colours are coordinated for all archaeal libraries (A, C, E) and all bacterial libraries (B, D, F), but not across domain boundaries.

artefacts prevented condent placement of each individual clone. These Thermoproteus-like organisms are ubiquitous in all three hot springs, representing 1217% of the archaeal clone libraries (Fig. 3A,C,E). Within the Thermolaceae, clones related to Thermolum pendens (94.4% similarity) are only represented in Sylvan Spring and Bison Pool. However, an earlier study found evidence of this family at Obsidian Pool (Barns et al., 1994, 1996), suggesting a temporally dependent community structure. Many of the archaeal clones recovered from the three sediments cluster within the uncultured Crenarchaeota a group

of organisms known only by their gene sequences. No naming system has been unanimously adopted for these phylotypes, and the clades represented in Fig. 4 are simply labelled UC (uncultured Crenarchaeota) IV, which also incorporates the terrestrial hot spring crenarchaeal group (THSC) (Takai & Sako, 1999; Takai et al., 2001). The uncultured Crenarchaeota feature prominently in the Obsidian Pool library (69%), but less so in the Bison Pool (7%) and Sylvan Spring (12%) libraries (Fig. 3A,C,E). A phylogenetic analysis of group UC-I is depicted in Fig. 4, utilizing clones from groups UC-II, III, and IV as an
2005 The Authors Journal compilation 2005 Blackwell Publishing Ltd

Geomicrobiology in Yellowstone

219

Fig. 4 Maximum parsimony phylogenetic inference from 16S rRNA gene sequences (1000 random addition replicates) of the Crenarchaeota (A) and a subgroup (UC-1) of the uncultured Crenarchaeota (B). Bootstrap values indicate 100 parsimony replicates. Organisms from this study in bold, with * for Bison Pool, + for Sylvan Spring, and ^ for Obsidian Pool. Bar indicates 10 changes.

outgroup. Group UC-I is comprised of sequences similar to previously identied clones pJP33, pJP96, and pJP41 from Obsidian Pool (Barns et al., 1994, 1996). The placement of the root of this clade relative to the other UC groups is problematic due to long branches, the lack of informative characters in the 16S rRNA, and the paucity of sequences from similar organisms to support the topology. When isolated in
2005 The Authors Journal compilation 2005 Blackwell Publishing Ltd

this monophyletic analysis, Obsidian Pool and Bison Pool clones maintain a topology with high condence numbers overall, although one node is not well supported. UC-I features clones from both the Obsidian Pool (33%) and Bison Pool (10%) libraries, but not from Sylvan Spring. Groups UC-II through V are shown on the phylogenetic tree in Fig. 4, together with representatives from a submarine

220

D. R. MEYER-DOMBARD et al. Obsidian Pool sediments, representing 12 major crenarchaeal clades, and one group of Korarchaeota. However, independent attempts to amplify DNA from another Obsidian Pool sample and from Queens Laundry (Sentinel Meadow in the Lower Geyser Basin) using archaeal-specic primers were unsuccessful (Graber et al., 2001; Blank et al., 2002). Figure 6A shows the distribution of clones (relative to the total library size) for each major group of Crenarchaea found in Obsidian Pool in this study and in Barns et al. (1994). The 12 crenarchaeal groups found by Barns and colleagues were also identied here. However, only eight were retrieved from our Obsidian Pool sample, and four of those were also found in the Sylvan Spring or Bison Pool samples. The remaining four were present exclusively in the Sylvan Spring and/or Bison Pool samples. In addition, four clades (D-II, D-IV, UC-II, UC-V) from this study were not identied by Barns et al. (1994, 1996). Phylogenetic analysis of bacterial clones

Fig. 4 Continued

sulphide chimney (Finn) on the Juan de Fuca Ridge (Schrenk et al., 2003), subterranean hot springs in Iceland (Marteinsson et al., 2001), hydrate-bearing deep marine sediments (Reed et al., 2002), sediments from the Okinawa Trough (Takai & Horikoshi, 1999), and other clones from Obsidian Pool (Barns et al., 1994, 1996). Phylotype UC-II was found in Obsidian Pool and Sylvan Spring, phylotype UC-III was absent from all three samples, phylotype UC-IV was present only in Obsidian Pool, and phylotype UC-V was found in Obsidian Pool and Bison Pool. Groups UC-II through IV are supported by high bootstrap values. However, the placement of UC-V was difcult to constrain due to long-branch affects; in analyses with the ve UC groups plus the Korarchaeota, each UC-V clone pulled away from the groupings shown in Fig. 4, but without the Korarchaeota, the clones enjoyed high bootstrap values in their shown topology in all analyses. Despite the consequent lower bootstrap value, this phylotype may represent a new, deeply branching clade within the uncultured Crenarchaeota, which is thus far unique to Yellowstone at Obsidian Pool and Bison Pool. Obsidian Pool is also home to the deepest-branching archaeal kingdom the Korarchaeota (Barns et al., 1994, 1996). In this study, the Korarchaeota are represented only by the Obsidian Pool clone ObP A42, which is 93.3% similar to clone pJP78 (Barns et al., 1994, 1996). Notably, clone pJP27-like organisms the previously identied Korarchaeota at Obsidian Pool (Barns et al., 1994, 1996) were not identied in this study. Independent verication of the presence and identity of Korarchaeota from a hydrothermal environment has not been previously shown. Only a few previous studies at Yellowstone have targeted the Archaea in systems geochemically similar to those studied here. Barns et al. (1994, 1996) identied 36 archaeal sequences in

A phylogenetic inference from bacterial sequences is shown in Fig. 5. It can be seen that most of the bacterial sequences afliate with the Aquicales, but a number of clones also fall within the Thermodesulfobacteriales. The Aquicales clones are numerically well-represented in all three libraries (Fig. 3B,D,F). Members of the Aquicaceae were only identied in Bison Pool and Obsidian Pool libraries. Within the Aquicaceae, several clones, including ObP B89 and seven others that were not included in the phylogenetic tree (due to short-branch artefacts), represent apparent subspecies of Thermocrinis ruber, the pink lament-forming hyperthermophiles identied in a number of Yellowstone locations (Reysenbach et al., 2000; Jackson et al., 2001; Jahnke et al., 2001; Blank et al., 2002; Eder & Huber, 2002; Kato et al., 2004; Zhang et al., 2004; Spear et al., 2005). Organisms belonging to the family Hydrogenothermaceae within the Aquicales were found at Sylvan Spring (>85% of the library) and Obsidian Pool, but not at Bison Pool (Fig. 3B,D,F). Two clones from Bison Pool (BP B68 and BP B72) branch well outside the established Aquicaceae and Hydrogenothermaceae clades and may represent a new phylotype within the Aquicales. Alternatively, these clones may represent novel members of the genus Desulphurobacterium; a more denitive placement of the root of this clade will depend on new cultured representatives from these groups. Clones from Sylvan Spring (SSp B60 and SSp B90) and Obsidian Pool (ObP B24) are similar to clones found by other studies (pBB and OPB13) that form a group in the genus Sulphurihydrogenibium. Cultured organisms in this group are found in other terrestrial environments and have been shown to oxidize sulphur, hydrogen, arsenite, and selenite with a variety of electron acceptors (Nakagawa et al., 2005). Clones from all three sample libraries fall within the order Thermodesulfobacteriales (Fig. 5), but feature most prominently in the Obsidian Pool sample (Fig. 3D). Clones SSp B24, ObP B80, and BP B73 are 98% similar to Geothermobacterium
2005 The Authors Journal compilation 2005 Blackwell Publishing Ltd

Geomicrobiology in Yellowstone

221

Fig. 5 Maximum parsimony phylogenetic inference from 16S rRNA sequences (1000 random addition replicates) of the bacteria. Bootstrap values indicate 100 parsimony replicates. Organisms from this study in bold, with * for Bison Pool, + for Sylvan Spring, and ^ for Obsidian Pool. Bar indicates 10 changes.

ferrireducens, a thermophilic Fe(III) reducing organism recently isolated from Obsidian Pool (Kashe et al., 2002). In addition, clones ObP B48 and ObP B45 from Obsidian Pool may represent a new phylotype within the Thermodesulfobacteriales. Besides the strictly thermophilic Aquicales and Thermodesulfobacteriales, Stenotrophomonas-like clones were found at all three locations, and clones clustering with the Actinobacteriales, Clostridiales, and subdivisions of the Proteobacteria were retrieved from Obsidian Pool sediments. These clones are included in Table 1, but not in the phylogenetic tree in Fig. 5. While few archaeal gene surveys have been carried out in these three hydrothermal areas, several investigations have targeted bacterial diversity there. For example, at Queens Laundry, also in the Sentinel Meadow area, Blank et al. (2002) found almost exclusively Thermocrinis ruber-like organisms and low percentages of clones that afliated with Thermodesulfobacterium and the Thermotogales. These ndings are in close agreement with our results at Bison Pool (see Fig. 6B). In addition, Fervidobacterium strain BP1 (belonging to the Thermotogales) was isolated from Bison Pool
2005 The Authors Journal compilation 2005 Blackwell Publishing Ltd

(Meyer-Dombard, 2004), although sequences belonging to the Thermotogales were not found in our Bison Pool bacterial library. It should be noted that Queens Laundry and Bison Pool are only 0.5 km apart in Sentinel Meadow, and that they feature very similar chemical compositions. A key difference in the samples is that the DNA from Queens Laundry was extracted from siliceous sinter at the edge of the pool, and the DNA from Bison Pool was extracted from water saturated sediment. This geochemical difference may explain the subtle community variations. One of the best studied sites in Yellowstone is Obsidian Pool. In addition to the archaeal libraries of Barns et al. (1994, 1996), Hugenholtz et al. (1998) identied 54 novel bacterial sequences from an Obsidian Pool sediment sample, including 12 novel clades. Surprisingly, only 4 of the 54 sequences and none of the novel groups were similar to clones from this study. A further study at Obsidian Pool was conducted by Graber et al. (2001), in an investigation of bacterial diversity down a high temperature gradient in an outow channel. In a 76 C sample, the Aquicales dominated, with minor representation by the Thermotogales. At 70 C, Planctomycetales,

222

D. R. MEYER-DOMBARD et al.

Fig. 6 Comparison of diversity and distribution in archaeal clone libraries at Obsidian Pool (A), bacterial clone libraries at Sentinel Meadow (B), and bacterial clone libraries at Obsidian Pool (C) from this investigation and several studies reported in the literature. Bars depict percentages of the total library represented by each taxon, as determined by 16S rRNA sequencing. Error bars in (1) result from ambiguous reporting in source.

2005 The Authors Journal compilation 2005 Blackwell Publishing Ltd

Geomicrobiology in Yellowstone
Table 2 Chemolithoautotrophic reactions considered in Figs 7 and 8 Reaction Number 1 2 3 4 5 6 7 8 9 10 11

223

Reactions H2(g) + 0.5O2(aq) H2O 3Fe+2 + 0.5O2(aq) + 3H2O Magnetite + 6H+ 3Fe+2 + 0.25O2(aq) + 1.5H2O Goethite + 2H+ S0 + 1.5O2(aq) + H2O SO42 + 2H+ H2S(aq) + 2O2(aq) SO42 + 2H+ CH4(g) + 2O2(aq) CO2(g) + 2H2O CO(g) + 0.5O2(aq) CO2(g) Pyrite + 2H+ + 0.5O2(aq) 2S0 + Fe+2 + H2O S0 + H2(g) H2S(aq) 4H2(g) + SO42 + 2H+ H2S(aq) + 4H2O Goethite + 0.5H2(g) + 2H+ Fe+2 + 2H2O

Electrons transferred 2 2 1 6 8 8 2 2 2 8 1

Fig. 7 Values of overall Gibbs free energy [Gr, in kJ (mol e)1] as a function of pH for eight aerobic oxidation reactions at the Lower Geyser Basin, the greater Obsidian Pool area, and the Sylvan Spring area. Geochemical data to compute values of Gr taken from Shock et al. (2005a,b). Numbers in legend refer to reactions given in Table 2. (Bison Pool, pH 8.1; Obsidian Pool, pH 6.5; Sylvan Spring, pH 5.5.)

Rhodothermus, and several phototrophic orders (including the green nonsulphur bacterium Chloroexus) were recovered; neither high-temperature Rhodothermus, nor Chloroexus was found in the present study (Fig. 6C). Further, the Aquicales clones from these three studies exhibited signicant sequence variability. Note that samples in the three studies were retrieved from different locations in Obsidian Pool, and that the collections spanned an 8-year period (1993 2001). The differences in the bacterial communities at Obsidian Pool may be due to temporal variations in geochemistry, compositional variations in microenvironments, or the resolution afforded by the applied molecular techniques. This is the rst molecular study to examine microbial diversity in the Sylvan Springs area. However, Bridge & Johnson (1998) and Johnson et al. (2001, 2003) cultured novel species of Sulfobacillus, Actinobacterium, and Methylobacterium from the same geothermal basin. The metabolic capabilities of these organisms include dissimilatory oxidation of ferrous iron and tetrathionate, as well as reduction of ferric iron. These Bacteria were isolated from undisclosed low temperature locations more acidic than our Sylvan Spring sample. However, these species add to the overall known microbial diversity of the region, which also includes members of the archaeal order Sulfolobales (Brock et al., 1972). Geochemical energy and microbial diversity The three hot spring sediment samples targeted in this study are similar in temperature (80 C), but signicantly different
2005 The Authors Journal compilation 2005 Blackwell Publishing Ltd

in chemical composition. We showed above that Obsidian Pool, Sylvan Spring, and Bison Pool share certain archaeal and bacterial taxa, but also that they feature variations in community composition. It is beyond the reach of this communication to denitively assert which of the numerous geochemical factors present in these extreme ecosystems most directly control hot spring community structure. The diverse geochemistry in Yellowstone hydrothermal systems results in a complex structure of energy regimes, and a rst order investigation of our gene surveys in the context of in situ energy-yields for chemolithoautotrophic reactions can help interpret thermophilic ecosystem dynamics (Shock et al., 2005a). Values of Gr of >180 reactions were calculated for >80 thermal features at Yellowstone (Shock et al., 2005a,b) using data collected between 1999 and 2001; here, we highlight only some of the most exergonic examples in the Obsidian Pool area, Sylvan Springs area, and Sentinel Meadow (Fig. 7), with particular attention paid to the three key springs. Values of Gr (normalized per electron transferred) of eight aerobic oxidation reactions in tens of sample locations are plotted as functions of pH in Fig. 7. The reactions, listed in Table 2, represent the oxidation of methane, hydrogen, carbon monoxide, ferrous iron, and reduced sulphur. It can be seen in this gure that Gr values of some of the reactions are relatively insensitive to changes in pH, yielding large but nearly invariable amounts of energy [80120 kJ (mol e)1] over the pH range 29. Only the energy-yields for Fe2+ and pyrite oxidation (reactions 2, 3 and 8) vary substantially with changing pH, increasing from 0 to 40 kJ (mol e)1 to >100 kJ (mol e)1, and decreasing from 80 kJ (mol e)1 to 50 kJ (mol e)1, respectively, over the pH range of 2 to >8. One of the most exergonic reactions is the knallgas reaction (reaction 1 in Table 2), yielding 100 kJ (mol e)1 at all Yellowstone springs investigated here. This reaction represents the central catabolic process of many of the Aquicales, and it was recently proposed as the dominant source of energy for primary productivity in Yellowstone hot springs (Spear et al., 2005). This high energy-yield for reaction (1) may be responsible for the

224

D. R. MEYER-DOMBARD et al. yields signicantly more energy at low pH [50 kJ (mol e)1] than at circumneutral pH [15 kJ (mol e)1]. At the in situ pH (6.5) of Obsidian Pool uids, Geothermobacterium ferrireducens and other organisms mediating reaction (11) obtain only 15 kJ (mol e)1. These ndings suggest that 1020 kJ (mol e)1 may be a sufcient amount of energy to support Geothermobacterium and chemolithoautotrophic strains of Desulphurococcales and Thermoproteales. This is consistent with minimum energy requirements observed for chemolithoautotrophs in Cape Lookout Bight, North Carolina (Hoehler et al., 2001). It is also plausible, however, that organisms relying on reactions (911) thrive at greater depths in Yellowstone thermal sediments, where conditions are likely more reducing and, hence, these reactions more exergonic. Further studies are necessary that explicitly target metabolic activity and rate measurements of certain redox reactions, especially along redox gradients with depth in thermal sediments. Many of the organisms in Yellowstone hot springs are known only through clone libraries in this and other investigations (Barns et al., 1994, 1996; Takai & Horikoshi, 1999; Marteinsson et al., 2001; Reed et al., 2002; Schrenk et al., 2003). These organisms, identied here as groups UC-I to UC-V and D-I to D-IV, represent 3090% of the phylotypes found in this study at Obsidian Pool, Sylvan Spring, and Bison Pool, but their metabolisms and energy sources are completely unknown. It is therefore difcult to use their phylogenetic positions to help interpret the microbial communities in a geochemical context. However, our abilities to culture the unculturables, and thereby elucidate their metabolisms, may be aided by designing growth media that target the most exergonic reactions. Recall that several of the reactions represented in Fig. 7 (and most of the >180 reactions considered by Shock et al. (2005a,b)) are not currently known to support any thermophiles. It seems unlikely that these large and numerous energy sources are left untapped by the usually resilient and metabolically versatile organisms in extreme environments.

Fig. 8 Values of overall Gibbs free energy [Gr, in kJ (mol e)1] as a function of pH for ve reactions that represent metabolisms mediated by known organisms closely afliated with clones from this study. Data to calculate values of Gr taken from Shock et al. (2005a,b). Numbers in legend refer to reactions given in Table 2. (Bison Pool, pH 8.1; Obsidian Pool, pH 6.5; Sylvan Spring, pH 5.5.)

apparent ubiquity of the Aquicales at Yellowstone, including at our three key sites (Reysenbach et al., 2000; Jackson et al., 2001; Jahnke et al., 2001; Blank et al., 2002; Eder & Huber, 2002; Kato et al., 2004; Zhang et al., 2004; Spear et al., 2005). However, it can also be seen in Fig. 7 that aerobic respiration reactions with other electron donors are equally, if not more, exergonic. Interestingly, of the eight reactions featured in Fig. 7, only two (reactions 1 and 4) are known to be used by phylotypes (the Aquicales and Sulphurihydrogenibium) identied in this study, and most exergonic reactions analysed in Shock et al. (2005a,b) are not known to support microbial metabolism at Yellowstone. Values of Gr for ve reactions that are known metabolic strategies of taxa similar to the clones identied in this study are shown in Fig. 8 (energy-yields for reactions (1) and (4) catalysed by certain Aquicales are replotted from Fig. 7). For example, some members of the archaeal orders Desulphurococcales and Thermoproteales can obtain energy by the reduction (with H2) of elemental sulphur (reaction 9) and sulphate (reaction 10). Further, Geothermobacterium ferrireducens isolated from Obsidian Pool by Kashe et al. (2002) (close relatives of which were found in all three springs) makes a living by reducing Fe(III); this metabolism is represented here by the reduction of ferric iron in the hydroxide mineral goethite (FeOOH) (reaction 11). Note that the reduction of sulphur species (reactions 9, 10) yields signicantly less energy [535 kJ (mol e)1] than the aerobic reactions (1, 4), which yield >80 kJ (mol e)1, and both sets of reactions are only weakly dependent on pH. Reaction (11), on the other hand,

CONCLUSIONS
The archaeal and bacterial communities in three Yellowstone sediment samples were investigated within the context of geochemical signatures of their greater geothermal regions. Variability in the thermophilic community structures of these features is argued to be a reection of geochemical conditions and the resulting differences in available energy. Examining the geochemistry of hydrothermal uids in tandem with the biological community reveals the complexity of the metabolic framework and thermophilic community structure as a function of geochemistry. Even relatively minor differences in pH and other geochemical factors, which translate directly to varying energy-yields from chemolithotrophic reactions, may be responsible for the observed differences in community structure in the thermal environments at Obsidian Pool, Sylvan Spring, and Bison Pool. Within the Bacteria, clones
2005 The Authors Journal compilation 2005 Blackwell Publishing Ltd

Geomicrobiology in Yellowstone similar to the alkaline-loving Thermocrinis are found only in the circum-neutral Obsidian Pool and alkaline Bison Pool. The Sylvan Spring Aquicales are most similar to the genera Sulphurihydrogenibium (made up of terrestrial organisms) (Nakagawa et al., 2005) and Hydrogenothermus, Persephonella, and Desulphurobacterium (primarily marine genera), some of which tolerate mildly acidic environs. It is possible, though not established (Jahnke et al., 2001; Spear et al., 2005), that the apparent ubiquity of the Aquicales within these Yellowstone environments is due to the abundance of energy available via aerobic hydrogen oxidation. Conversely, all three pools feature relatives of Geothermobacterium ferrireducens and other members of the Thermodesulfobacteriales, despite relatively modest energy-yields from the reduction of Fe(III). This raises questions about the in situ activities of these taxa. In the archaeal domain, the Desulphurococcales are well represented in Sylvan Spring, Bison Pool, and Obsidian Pool; the Thermoproteales have a larger presence in Sylvan Spring and Obsidian Pool than in Bison Pool; and the greatest diversity of uncultured Crenarchaeota is found in Obsidian Pool. Because the majority of clones representing the archaeal communities in these features does not afliate with any cultured organisms, one can only speculate about their modes of metabolism, though geochemical context is essential in developing a model of community dynamics. Future studies focusing on culturing representatives of these groups and metagenomic studies targeting metabolic activity will further link microbial presence and geochemical variation.

225

ACKNOWLEDGEMENTS
This work was funded largely by a NASA Graduate Student Research Program (GSRP) fellowship (NGT5-50348); also NSF-LExEN (OCE-9817730), and NASA Astrobiology Institute (Carnegie Institution). Extensive discussion with J. Ball (USGS) regarding naming of thermal features was very helpful and much appreciated. Phylogenetic advice provided by C. Blank.

REFERENCES
Allen ET, Day AL (1935) Hot Springs of the Yellowstone National Park. Waverly Press, Inc, Baltimore, MD. Altschul S, Madden T, Schaffer A, Zhang J, Zhang Z, Miller W, Lipman D (1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Research 25, 33893402. Amend JP, Rogers KL, Meyer-Dombard DR (2004) Microbially mediated sulfur-redox: Energetics in marine hydrothermal vent systems. In Sulfur Biogeochemistry Past and Present (eds Amend JP Edwards KJ, Lyons TW). GSA Special Paper 379, pp. 1734. Amend JP, Rogers KL, Shock EL, Inguaggiato S, Gurrieri S (2003) Energetics of chemolithoautotrophy in the hydrothermal system of Vulcano Island, southern Italy. Geobiology 1, 37 58. Amend JP, Shock EL (2001) Energetics of overall metabolic reactions of thermophilic and hyperthermophilic Archaea and Bacteria. FEMS Microbiology Reviews 25, 175 243. 2005 The Authors Journal compilation 2005 Blackwell Publishing Ltd

Bach W, Edwards KJ (2003) Iron and sulde oxidation within the basaltic ocean crust-extent, processes, timing, and implications for chemolithoautotrophic primary biomass production. Geochimica et Cosmochimica Acta 67, 38713887. Ball JW, Nordstrom DK, Jenne EA, Vivit DV (1998) Chemical Analyses of Hot Springs, Pools, Geysers, and Surface Waters from Yellowstone National Park, Wyoming, and Vicinity, 197475. US Geological Survey, US Department of the Interior, Boulder, CO, p. 45. Ball JW, Nordstrom DK, McCleskey RB, Schoonen MAA, Xu Y (2001) Water Chemistry and On-Site Sulfur-Speciation Data for Selected Springs in Yellowstone National Park, Wyoming, 1996 98. US Geological Survey, US Department of the Interior, Boulder, CO, p. 41. Barns SM, Delwiche CF, Palmer JD, Pace NR (1996) Perspectives on Archaeal diversity, thermophily and monophily from environmental rRNA sequences. Proceedings of the National Academy of Sciences, USA 93, 91889193. Barns SM, Fundyga RE, Jeffries MW, Pace NR (1994) Remarkable Archaeal diversity detected in a Yellowstone National Park hot spring environment. Proceedings of the National Academy of Sciences, USA 91, 16091613. Baross JA (1995) Isolation, Growth, and Maintenance of Hyperthermophiles. In Archaea, A Laboratory Manual: Thermophiles (eds Robb FT, Place AR). Cold Spring Harbor Press, New York, pp. 1523. Ben-Bassat A, Zeikus JG (1981) Thermobacteroides acetoethylicus gen. nov. and sp. nov., a new chemoorganotrophic, anaerobic, thermophilic bacterium. Archives of Microbiology 128, 365370. Blank CE, Cady SL, Pace NR (2002) Microbial composition of near-boiling silica-depositing thermal springs throughout Yellowstone National Park. Applied and Environmental Microbiology 68, 51235135. Bond PL, Smriga SP, Baneld JF (2000) Phylogeny of microorganisms populating a thick, subaerial, predominantly lithotrophic biolm at an extreme acid mine drainage site. Applied and Environmental Microbiology 66, 38423849. Boomer SM, Lodge DP, Dutton BE, Pierson B (2002) Molecular characterization of novel red green nonsulfur bacteria from ve distinct hot spring communities in Yellowstone National Park. Applied and Environmental Microbiology 68, 346355. Bridge TAM, Johnson DB (1998) Reduction of soluble iron and reductive dissolution of ferric iron-containing minerals by moderately thermophilic iron-oxidizing Bacteria. Applied and Environmental Microbiology 64, 21812186. Brock TD (1978) Thermophilic Microorganisms and Life at High Temperatures. Springer-Verlag, New York. Brock TD, Brock KM, Belly RT, Weiss RL (1972) Sulfolobus: a new genus of sulfur-oxidizing bacteria living at low pH and high temperature. Archives of Microbiology 84, 5468. Brock TD, Freeze H (1969) Thermus aquaticus gen. n. and sp. n., a nonsporulating extreme thermophile. Journal of Bacteriology 98, 289297. Cole JR, Chai B, Marsh TL, Farris RJ, Wang Q, Kulam SA, Chandra S, McGarrell DM, Schmidt TM, Garrity GM, Tiedje JM (2003) The Ribosomal Database Project (RDP-II): previewing a new autoaligner that allows regular updates and the new prokaryotic taxonomy. Nucleic Acids Research 31, 442443. Eder W, Huber R (2002) New isolates and physiological properties of the Aquicales and description of Thermocrinus albus sp. nov. Extremophiles 6, 309318. Ellis AJ, Mahon WAJ (1964) Natural hydrothermal systems and experimental hotwater/rock interactions (Pt I.). Geochimica et Cosmochimica Acta 28, 13231357.

226

D. R. MEYER-DOMBARD et al.
division level bacterial diversity in a Yellowstone hot spring. Journal of Bacteriology 180, 366376. Hurt RA, Qiu X, Wu L, Palumbo AV, Tiedje JM, Zhou J (2001) Simultaneous recovery of RNA and DNA from soils and sediments. Applied and Environmental Microbiology 67, 44954503. Inskeep WP, Macur RE, Harrison G, Bostick BJ, Fendorf S (2004) Biomineralization of As(V)-hydrous ferric oxyhydroxide in microbial mats of an acid-sulfate-chloride geothermal spring, Yellowstone National Park. Geochimica et Cosmochimica Acta 68, 31413155. Jackson CR, Langner HW, Donahoe-Christiansen J, Inskeep WP, McDermott TR (2001) Molecular analysis of microbial community structure in an arsenite-oxidizing acidic thermal spring. Environmental Microbiology 3, 532542. Jackson TJ, Ramaley RF, Meinschein WG (1973) Thermomicrobium, a new genus of extremely thermophilic bacteria. International Journal of Systematic Bacteriology 23, 2836. Jahnke L, Eder W, Huber R, Hope JM, Hinrichs K-U, Hayes JM, Des Marais DJ, Cady SL, Summons RE (2001) Signature lipids and stable carbon isotope analyses of Octopus Spring hyperthermophilic communities compared with those of Aquicales representatives. Applied and Environmental Microbiology 67, 51795189. Jahnke L, Embaye T, Hope J, Turk KA, Van Zuilen M, Des Marais DJ, Farmer JD, Summons RE (2004) Lipid biomarker and carbon isotopic signatures for stromatolite-forming, microbial mat communities and Phormidium cultures from Yellowstone National Park. Geobiology 2, 3147. Johnson DB, Body DA, Bridge TAM, Bruhn DF, Roberto FF (2001) Biodiversity of acidophilic moderate thermophiles isolated from two sites in Yellowstone National Park and their roles in the dissimilatory oxio-reduction of iron. In Thermophiles: Biodiversity, Ecology and Evolution (eds Reysenbach AL, Voytek M, Mancinelli R). Kluwer Academic/Plenum Publishers, New York, NY, pp. 2339. Johnson DB, Okibe N, Roberto FF (2003) Novel thermo-acidophilic bacteria isolated from geothermal sites in Yellowstone National Park: physiological and phylogenetic characteristics. Archives of Microbiology 180, 6068. Kashe K, Holmes DE, Reysenbach AL, Lovley DR (2002) Use of Fe(III) as an electron acceptor to recover previously uncultured hyperthermophile: isolation and characterization of Geothermobacterium ferrireducens gen. nov., sp. nov. Applied and Environmental Microbiology 68, 17351742. Kato K, Kobayashi T, Yamamoto H, Nakagawa T, Maki Y, Hoaki T (2004) Microbial mat boundaries between chemolithotrophs and phototrophs in geothermal hot spring efuents. Geomicrobiology Journal 21, 9198. Kennedy BM, Lynch MA, Reynolds JH, Smith SP (1985) Intensive sampling of noble gases in uids at Yellowstone: I. Early overview of data; regional patterns. Geochimica et Cosmochimica Acta 49, 12511261. Krauskopf KB (1964) The possible role of volatile metal compounds in ore genesis. Economic Geology 59, 2245. Lane DJ (1991) 16S/23S rRNA sequencing. In Nucleic Acid Techniques in Bacterial Systematics (eds Stackebrandt E, Goodfellow M). John Wiley and Sons, Chichester, UK, pp. 115175. Langner HW, Inskeep WP (2000) Microbial reduction of arsenate in the presence of ferrihydrite. Environmental Science and Technology 34, 31313136. Langner HW, Jackson CR, McDermott TR, Inskeep WP (2001) Rapid oxidation of arsenite in a hot spring ecosystem, Yellowstone National Park. Environmental Science and Technology 35, 33023309. 2005 The Authors Journal compilation 2005 Blackwell Publishing Ltd

Ellis AJ, Mahon WAJ (1967) Natural hydrothermal systems and experimental hotwater/rock interactions (Pt II.). Geochimica et Cosmochimica Acta 31, 519 538. Engle M, Li Y, Woese CR, Wiegel J (1995) Isolation and characterization of a novel alkalitolerant thermophile, Anaerobranca horikoshii gen. nov., sp. nov. International Journal of Systematic and Evolutionary Microbiology 45, 454461. Felsenstein J (1985) Condence limits on phylogenies: an approach using the bootstrap. Evolution 39, 783791. Fouke BW, Bonheyo GT, Sanzenbacher E, Frias-Lopez J (2003) Partitioning of bacterial communities between travertine depositional facies at Mammoth Hot Springs, Yellowstone National Park, USA. Canadian Journal of Earth Sciences 40, 15311548. Fournier RO (1989) Geochemistry and dynamics of the Yellowstone National Park hydrothermal system. Annual Reviews of Earth and Planetary Sciences 17, 13 53. Fournier RO, Pitt AM (1985) The Yellowstone magmatichydrothermal system, U.S.A. In Transactions of the Geothermal Councils International Symposium on Geothermal Energy, International Volume (ed. Stone C). Geothermal Resources Council, Davis, CA, pp. 319327. Fournier R, Thompson J, Hutchinson R (1994) The geochemistry of hot spring waters at Norris Geyser Basin, Yellowstone National Park. Geothermal Resources Council Transactions 18, 177179. Giggenbach WF (1975) A simple method for the collection and analysis of volcanic gas samples. Bulletin of Volcanology 39, 132145. Giggenbach WF, Goguel RL (1989) Collection and Analysis of Geothermal and Volcanic Water and Gas Discharges. Chemical Division of Department of Scientic and Industrial Research, Petone, New Zealand Report CD 2401, p. 81. Giggenbach WF, Tedesco D, Sulistiyo Y, Caprai A, Cioni R, Favara R, Fischer TP, Hirabayashi J-I, Korzhinsky M, Martini M, Menyailov I, Shinohara H (2001) Evaluation of results from the fourth and fth IAVCEI eld workshop on volcanic gases, Vulcano Island, Italy, and Java, Indonesia. Journal of Volcanology and Geothermal Research 108, 157172. Graber JR, Kirshtein J, Speck M, Reysenbach AL (2001) Community structure along a thermal gradient in a stream near Obsidian Pool, Yellowstone National Park. In Thermophiles: Biodiversity, Ecology and Evolution (eds Reysenbach AL Voytek M, Mancinelli R). Klewer Academic/Plenum Publishers, New York, pp. 8189. Henry EA, Devereux R, Maki JS, Gilmour CC, Woese CR, Mandelco L, Schauder R, Remsen CC, Mitchell R (1994) Characterization of a new thermophilic sulfate-reducing bacterium Thermodesulfovibrio yellowstonii, gen. nov. and sp. nov. its phylogenetic relationship to Thermodesulfobacterium commune and their origins deep within the bacterial domain. Archives of Microbiology 161, 62 69. Hoehler TM, Alperin MJ, Albert DB, Martens CS (2001) Apparent minimum free energy requirements for methanogenic and sulfate-bacteria in an anoxic marine sediment. FEMS Microbiology Ecology 38, 3341. Huber R, Eder W, Heldwein S, Wanner G, Huber H, Rachel R, Stetter KO (1998) Thermocrinis ruber gen. nov., sp. nov., a pink-lament forming hyperthermophilic Bacterium isolated from Yellowstone National Park. Applied and Environmental Microbiology 64, 3576 3583. Huber T, Faulkner G, Hugenholtz P (2004) Bellerophon; a program to detect chimeric sequences in multiple sequence alignments. Bioinformatics 20, 23172319. Huelsenbeck J, Rannala B (1997) Phylogenetic methods come of age: testing hypotheses in an evolutionary context. Science 276, 227232. Hugenholtz P, Pitulle C, Hershberger KL, Pace NR (1998) Novel

Geomicrobiology in Yellowstone
Macur RE, Langner HW, Kocar BD, Inskeep WP (2004) Linking geochemical processes with microbial community analysis: successional dynamics in an arsenic-rich, acid-sulphate geothermal spring. Geobiology 2, 163 177. Marteinsson VT, Hauksdottir S, Hobel CFV, Kristmannsdottir H, Hreggvidsson GO, Kristjansson JK (2001) Phylogenetic diversity analysis of subterranean hot springs in Iceland. Applied and Environmental Microbiology 67, 4242 4248. McCollom T, Shock EL (1997) Geochemical constraints on chemolithoautotrophic metabolism by microorganisms in seaoor hydrothermal systems. Geochimica et Cosmochimica Acta 61, 43754391. Meyer-Dombard DR (2004) Geochemical constraints on microbial diversity in Yellowstone National Park, Thesis, Doctor of Philosophy. Department of Earth and Planetary Sciences, Washington University in St. Louis, St. Louis, MO, p. 210. Miller DN, Bryant JE, Madsen EL, Ghiorse WC (1999) Evaluation and optimization of DNA extraction and purication procedures for soil and sediment samples. Applied and Environmental Microbiology 65, 4715 4724. Nakagawa S, Shtaih Z, Banta A, Beveridge TJ, Sako Y, Reysenbach A-L (2005) Sulfurihydrogenibium yellowstonense sp. nov., an extremely thermophillic, facultatively heterotrophic, sulfuroxidizing bacterium from Yellowstone National Park, and emended descriptions of the genus Sulfurihydrogenibium, Sulfurihyrogenibium subterraneum, and Sulfurihydrogenibium azorense. Geobiology 55, 2263 2268. Papke RT, Ramsing NB, Bateson MM, Ward DM (2003) Geographical isolation in hot spring cyanobacteria. Environmental Microbiology 5, 650659. Pierson BK, Parenteau MN, Grifn BM (1999) Phototrophs in high-iron-concentration microbial mats: physiological ecology of phototrophs in an iron-depositing hot spring. Applied and Environmental Microbiology 65, 5474 5483. Reed DW, Fujita Y, Delwiche ME, Blackwelder DB, Sheridan PP, Uchida T, Colwell FS (2002) Microbial communities from methane hydrate-bearing deep marine sediments in a forearc basin. Applied and Environmental Microbiology 68, 3759 3770. Reysenbach AL, Ehringer M, Hershberger K (2000) Microbial diversity at 83 C in Calcite Springs, Yellowstone National Park: another environment where the Aquicales and Korarchaeota coexist. Extremophiles 4, 61 67. Schrenk MO, Kelly DS, Delaney JR, Baross JA (2003) Incidence and diversity of microorganisms within the walls of an active deep-sea sulde chimney. Applied and Environmental Microbiology 69, 35803592. Shock EL, Holland M, Meyer-Dombard D, Amend JP (2005a) Geochemical sources of energy for microbial metabolism in

227

hydrothermal ecosystems: Obsidian Pool, Yellowstone National Park, USA. Geothermal Biology and Geochemistry in Yellowstone National Park (eds Inskeep WP McDermott TR). Thermal Biology Institute, Montana State University, pp. 95112. Shock EL, Holland ME, Meyer-Dombard DR, Amend JP, Osburn GR, Fisher T (2005b) Quantifying Inorganic Sources of Energy in Hydrothermal Systems, Yellowstone National Park, USA. In Preparation. Spear JR, Walker JJ, McCollom TM, Pace NR (2005) Hydrogen and bioenergetics in the Yellowstone geothermal ecosystem. Proceedings of the National Academy of Sciences, USA 102, 25552560. Takai K, Horikoshi K (1999) Genetic diversity of Archaea in deep-sea hydrothermal vent environments. Genetics 152, 12851297. Takai K, Komatsu T, Inagaki F, Horikoshi K (2001) Distribution of archaea in a black smoker chimney structure. Applied and Environmental Microbiology 67, 36183629. Takai K, Sako Y (1999) A molecular view of archaeal diversity in marine and terrestrial hot water environments. FEMS Microbiology Ecology 28, 177188. White DE, Mufer LJP, Truesdell AH (1971) Vapor-dominated hydrothermal systems compared with hot-water systems. Economic Geology 66, 7597. Whittlesey L (1988) Wonderland Nomenclature: A History of the Place Names of Yellowstone National Park. Montana Historical Society Press. Wiegel J, Ljungdahl LG (1981) Thermoanaerobacter ethanolicus gen. nov., sp. nov., a new, extreme thermophilic, anaerobic bacterium. Archives of Microbiology 128, 343348. Zeikus GJ, Dawson MA, Thompson TE, Ingvorsen K, Hatchikian EC (1983) Microbial ecology of volcanic sulphidogenesis: isolation and characterization of Thermodesulfobacterium commune gen. nov. and sp. nov. Journal of General Microbiology 129, 11591169. Zeikus GJ, Hegge PW, Anderson MA (1979) Thermoanaerobium brockii gen. nov. and sp. nov., a new chemoorganotrophic, caldoactive, anaerobic bacterium. Archives of Microbiology 122, 4148. Zhang CL, Fauke BW, Bonheyo GT, Peacock AD, White DC, Huang Y, Romanek CS (2004) Lipid biomarkers and carbon-isotopes of modern travertine deposits (Yellowstone National Park, USA): implications for biogeochemical dynamics in hot-spring systems. Geochimica et Cosmochimica Acta 68, 31573169. Zohdy AAR, Anderson LA, Mufer LJP (1973) Resistivity, self-potential, induced-polarization and surveys of a vapor-dominated geothermal system. Geophysics 38, 11301144.

2005 The Authors Journal compilation 2005 Blackwell Publishing Ltd