Patricia L. Siering, Jessica M. Clarke, and Mark S. Wilson. Department of Biological Sciences, Humboldt State University, Arcata, California 95521, USA.
Geomicrobiology Journal, 23:129–141, 2006 c Taylor & Francis Group, LLC Copyright ISSN: 0149-0451 print / 1521-0529 online DOI: 10.1080/01490450500533916
Geochemical and Biological Diversity of Acidic, Hot Springs in Lassen Volcanic National Park Patricia L. Siering, Jessica M. Clarke, and Mark S. Wilson Department of Biological Sciences, Humboldt State University, Arcata, California 95521, USA
We used phylogenetic approaches and chemical/physical analyses to investigate the diversity of microorganisms in geochemically distinct acidic hot springs of Lassen Volcanic National Park (LVNP) in northern California. Geochemical composition was analyzed for twelve sites within LVNP, and 16S rRNA clone libraries were prepared and analyzed for three sites (DK12 −93.5◦ C, pH 1.2; USW5 −86.2◦ C, pH 2.2; and SW3 −82◦ C, pH 1.2) using universal and Archaeal-specific primers. Phylogenetic relationships of sequences were estimated using parsimony, maximum likelihood, and Bayesian methods. The geochemical composition of the features varied widely even among proximate sites having similar temperature and pH values. Acridine Orange Direct Counts indicated cell concentrations of 106 to 108 cells/ml in the sites analyzed, which is comparable to neutral pH, mesophilic environments. In contrast to similar studies of less-acidic hot springs in which members of the Bacterial order Aquificales have been shown to dominate, no Bacterial sequences were detected in any of the libraries. Using 97% sequence identity as a cutoff value, approximately 65 phylotypes were detected, many with low similarity (70◦ C), low chloride, low metal systems that seem to be characteristic of LVNP. The Lassen volcanic center in north central California is the southernmost active volcanic system in the Cascade Range. Lassen Peak, which formed 27,000 years ago in a large eruptive event, reawakened in May 1914 and continued erupting intermittently for a period of 4 years. Silicic volcanism at Lassen over the past 300,000 years has greatly altered the surrounding landscape and created the thermal features that today are characteristic of LVNP. Roaring fumaroles, mudpots, boiling pools, and steaming ground are present in various locations
129
130
P. L. SIERING ET AL.
throughout the LVNP vicinity. Within the park boundaries are several high-elevation, acid-sulfate, low chloride springs characteristic of vapor-dominated hydrothermal systems (Ingebritsen and Sorey 1985). These thermal acidic features represent some of the most extreme life-supporting environments on earth with temperatures ranging from 50◦ C–115◦ C, and pH from 0–3. Only a small number of studies have examined biology in the acidic, thermal features at Lassen Volcanic National Park (Gross et al. 2001; Whitaker et al. 2003), and these have focused on narrow taxonomic groups. Throughout the past several years, U.S. Geological Survey volcanologists have undertaken detailed investigations of geothermal features in LVNP (Muffler et al. 1982; Ingebritsen and Sorey 1985; Thompson 1985; Sorey and Colvard 1994). Previous geochemical data from various sites within LVNP indicate that geochemical composition of springs varied widely even among features in similar regions of the park with similar temperatures and pH values (Thompson 1985). Previous gas analyses (mid 1990s) of features in several locations within LVNP indicate the volcanic gases emitted from different thermal features vary significantly among different thermal regions within LVNP (C. Janik, U. S. Geological Survey, unpublished results); however, most features analyzed within a given region have similar gas compositions. Sulfur Works sites (SW) (upper and lower) were the lowest in average mole percent H2 S, CH4 , and H2 (0.41, 0.0025, and 0.047, respectively) when compared to features in Bumpass Hell (BH) (6.31, 0.079, and 0.90, respectively), Devil’s Kitchen (DK) (3.56, 0.109, and 0.4, respectively), and the Boiling Springs Lake thermal area (BSLTA) (6.95, 0.072, and 0.275, respectively) (C. Janik, U. S. Geological Survey, unpublished results). As is common for gases emitted from geothermal features, the mole percent dry weight CO2 was high in all sites [averages of 89.9 (BH), 98.7 (SW), 92.33 (DK), and 88.7 (BSLTA)]. The purpose of our study was to begin an investigation of the biologic and chemical diversity in various regions of LVNP. Geochemical composition was analyzed for twelve sites within LVNP boundaries. Acridine Orange Direct Counts (AODC) of total microbial numbers, and 16S rRNA gene clone libraries were prepared for three sites (DK12 −93.5◦ C, pH 1.2; USW5 −86.2◦ C, pH 2.2; and SW3 −82◦ C, pH 1.2), using a variety of primers. We sequenced several clones from each library and estimated the phylogenetic relationships of sequences using standard methods. This work established essential baseline data with which we can begin to investigate the interplay between biotic and abiotic components of these extreme ecosystems. METHODS Sampling For each sampling site, locations were determined with a Garmin 12XL Global Positioning Satellite system, and all sampling sites were photographically documented. Water and sediment samples were collected with a modified telescoping (to
5 m) aluminum paint rod to which we attached sterile 1 L Nalgene containers. We surface-sterilized the outside of the collection device with several rinses of 80% ethanol, followed by a rinse in sterile distilled, deionized water. Values for temperature, and temperature-corrected pH were recorded on site using a Thermo-Orion 290A Plus meter (Fisher Scientific, Pittsburgh, PA). Upon collection, samples were immediately processed for acridine orange direct counting (AODC) by addition of formalin (Sigma-Aldrich, St. Louis, MO) to a final concentration of 3.7%. Approximately 4–10 hours passed from the time of sample collection until completion of processing for nucleic acid extraction. This time period corresponds to the time required to carry the samples back to the trailhead (∼4 km), drive them to the lab facilities within the park, and filter the water samples (or dewater the sediment samples), prior to freezing. Geochemical Analyses Samples for geochemical analyses were aseptically collected in polypropylene bottles, and stored at 4◦ C for approximately 48 hours prior to processing prior by centrifugation and filtration through a 0.45 µm filter. Filtered water samples were stored at 4◦ C until analyses were completed. Geochemical analyses were performed by Dale Counce at Los Alamos National Laboratory, Los Alamos, NM. Cations were determined by Inductively Coupled Plasma Mass Spectroscopy (ICP-MS). Anions were determined by ion chromatography. Acridine Orange Direct Counts (AODC) All fixed samples were stored at 4◦ C until counts were performed. Quantitative direct counts of total microbial numbers in water and sediment slurry samples were performed using previously described AODC methods (Siering and Ghiorse 1997a; Colwell et al. 1999). Direct counts of microorganisms contained in hot spring water were also attempted by collecting cells (from formalin-fixed samples) directly on Irgalin black stained Nucleopore filters as previously described (Paul 1982; Cochran and Paul 1998; Wilson et al. 1999a). Nucleic Acid Extractions Planktonic cells contained in water samples were filtered onto 0.2 µm Durapore Millipore filters (typically 50–500 ml/filter), and sediments were dewatered by filtration or centrifugation, prior to freezing and nucleic acid extraction. Filters containing collected cells, and dewatered sediment aliquots of 0.75 ml volumes were stored at −80◦ C in individual 2 ml screw capped microcentrifuge tubes until nucleic acids were extracted. Nucleic acids were extracted from water and sediment methods by a modification of previously reported methods (Siering and Ghiorse 1997b; Wilson et al. 1999a; Miller 2001), as detailed next. Nucleic acids from cells contained in water and sediment samples were collected by successive chemical and physical lysis treatments. Frozen sediment samples were thawed and
GEOCHEMICAL AND BIOLOGICAL DIVERSITY OF ACIDIC, HOT SPRINGS
neutralized on ice in 1.0 ml 0.5 M Tris (pH 8.0), pelleted in a microcentrifuge, resuspended in 800 µl proteinase K-lysis buffer (200 µg/ml proteinase K, 0.1 M NaCl, 10 mM EDTA, 10 mM Tris pH 8.0, 2% SDS), and incubated at 50◦ C for 30 minutes. Filters containing planktonic cells were not neutralized prior to proteinase K lysis. Tubes were microcentrifuged for 30 seconds, and supernatants were combined with 400 µl 24:1 chloroform:isoamyl alcohol (C/IAA), and stored on ice during the subsequent lysing steps. The remaining sediments in the tubes were combined with 400 µl lysis buffer, 300 µl C/IAA, and 1– 1.5 g sterile zirconia-silica beads (0.1 mm) (Bio Spec Products. Bartlesville, OK). Cells were lysed by a brief cell disruption (1–2 minutes, 2500 rpm) in a Mini-bead beater (Bio Spec Products. Bartlesville, OK). Following bead beating, samples were microcentrifuged, and supernatants were combined with the proteinase K lysate generated in the previous step. Typically, 1–3 additional bead beatings in lysis buffer/C/IAA were performed; following each treatment, supernatants were combined into a single tube containing the lysates from each sequential round of lysis. Beads were rinsed by 2–3 washes in ice-cold sterile TE, and these rinses were combined with the crude lysates. The crude lysates were extracted with phenol/chloroform, concentrated by butanol extractions, further purified on Sepharose CL4b columns (Miller 2001), and precipitated overnight in ethanol using standard methods. Nucleic acids were resuspended in 2 mM Tris (pH 7.8), 0.2 mM EDTA, and stored at −80◦ C. Nucleic acids were extracted from multiple aliquots (4–6) from a given site, and equimolar amounts were pooled prior to preparation of clone libraries. PCR Amplification, Cloning, and Sequencing SSU ribosomal RNA genes were amplified with the universal primers 515F and 1391R (Reysenbach and Pace 1995) or Archaeal-specific 1100R (Reysenbach and Pace 1995) (with 515F) using Promega MasterMix (Promega Corporation, Madison, WI), and standard conditions. Amplicons from four PCRs were pooled, purified with Wizard PCR purification systems and cloned into the pGEM-T Easy cloning vector (Promega, Madison, WI), according to manufacturer’s recommendations. Plasmids were isolated from the confirmed clones using Wizard Plus SV minipreps (Promega). The sequencing was either done in-house by the authors, or submitted to CSUPERB Microchemical Core Facility (San Diego, CA) or Macrogen (Seoul, Korea). In-house sequencing was performed using the Licor 4200 Long ReadIR automated sequencing system according to the manufacturer’s recommendations. Approximately 232 clones were sequenced in this study; 13 clones that represented potential chimeras (see below) were not included in the final analyses (Table 3). Sequence Analysis and Phylogeny Estimation In order to identify unique phylotypes within the clone libraries, sequences were initially aligned to one another using
131
Clustal W, and genetic distances were estimated using Clustal Dist; both programs were accessed through the Biology Workbench http://workbench.sdsc.edu. Unique phylotypes were defined as those clones (or clusters of clones) with less than 97% identity to all other clones (or clusters). Sequences that differed by less than 3% to all other sequences within a cluster were considered to represent the same phylotype. For each phylotype containing multiple representatives, 2 clones were chosen for phylogenetic analysis (the most divergent sequence within the cluster, and the most conserved). If all sequences within a cluster differed by ≤0.5%, a single consensus sequence was used for phylogenetic estimation. Sequences representing each phylotype were compared with those in the NCBI database http:www.ncbi.nlm.nih.gov using gapped blast analyses (Altschul et al. 1997) via the Discontiguous MegaBlast program. Sequences in the database showing the highest identities to particular phylotypes were downloaded for phylogenetic analyses. We attempted to identify potential chimeric sequences using the Chimera Check program available from the Ribosomal Database Project (RDP) (Cole et al. 2003) and Bellerophon (Huber et al. 2004); obvious chimeras were excluded from further analyses. Rarefaction analyses were performed using Analytic Rarefaction 1.3 (Holland 2003). For phylogenetic analyses, sequences were aligned to the most similar sequences extracted from the NCBI and the RDP databases. Sequences were retrieved from the RDP in an aligned format. Alignments of sequences generated in this work and those retrieved from NCBI were done manually using conserved primary and secondary structural features and the RDP alignments as a guide. Distance matrix, parsimony, maximum likelihood, and Bayesian methods were compared for phylogenetic inference. The general approach for computing distance matrix and parsimony trees are reported elsewhere (Siering and Ghiorse 1996). Maximum likelihood trees were determined using Fast DNA ML (Olsen et al. 1994) and the Maximum likelihood option in PAUP 4.0 b10 software package (Swofford 2002). Nonparametric bootstrap analyses were performed using the PAUP software package. Bayesian analyses were performed using Mr. Bayes 3.0 (Huelsenbeck 2000) applying the GTR model of nucleotide substitution with gamma-distributed rate variation across sites. A minimum of six independent runs of 1 million generations each were generated for each data set. Various degrees of “burnin” were employed to assess the point of likelihood convergence. Various “burn-in” values were used for consensus tree construction in PAUP 4.0 b10.
Accession Numbers The SSU rRNA gene sequences analyzed in this study were deposited to the GenBank database under the following accession numbers: DK12a clones-AY779769-AY779782, DK12b clones-AY779829-AY779856, USW5b clones-AY779797AY779828, and SW3 clones-AY77983-AY779796.
132
P. L. SIERING ET AL.
RESULTS Site Characterization Approximately 40 thermal features were photographically documented, described, and sampled for temperature and pH determinations from June–August in 3 successive years (data not shown). There were a wide range of temperatures and pH with the majority of the features surveyed being ≥70◦ C, with pH values of 1–3. Some of the features examined included: clear, small ephemeral pools (present intermittently during summer months); rapidly boiling frying pans; high-viscosity muds; fumarole-fed springs; and a large acidic lake. Descriptions for sites from which SSU rRNA clone libraries were analyzed can be found in Table 1. All sites were sampled on multiple occasions during 1999–2004. As would be expected in this area of active volcanism, many of the features within LVNP were dynamic in terms of their presence, appearance, and sample consistency. Temperatures and pH values were more variable in the smaller features, and many tended to be cooler (and of higher pH) in the early season (June) vs. later in the summer (August). Chemical analyses were done for 12 of the features surveyed. Thermal features of similar temperature and pH values were geochemically distinct when located in the same region of the park [compare sites within BH and USW regions (Table 2)]; however, there do not appear to be region-specific differences in chemical composition of site waters. Community Analysis All sites investigated thus far have cell concentrations ranging from 2 × 106 to 5 × 108 cells/ml (Table 1). These values are
comparable to mesothermic, neutral pH, aquatic and sediment environments. A diversity of rods, cocci, amorphous and lobed sphere morphologies were observed in all samples (data not shown). In this study, clone libraries were generated from environmental nucleic acid extracts from SW3, DK12, and USW5 (Table 1) using the universal 515F primer with the universal 1391R or the Archaeal 1100R primers (Table 3). Attempts to identify chimeric sequences from the clone libraries using Chimera Check (Cole et al. 2003) and Bellerophon (Huber et al. 2004) were problematic and only led to a few (1–5) potential chimeras being identified in each library. This difficulty is to be expected since most of the sequences generated from these libraries represent potentially novel species, and many of the clones appear closely related to one another. For each library, multiple phylotypes were defined, and representative clones from each phylotype were compared to sequences in the databases as described in Methods (Table 3). Rarefaction analyses (Figure 1) indicated that the USW5 and DK12 libraries were sampled adequately to detect the majority of highly abundant phylotypes; however, extensive additional sequencing efforts would likely result in additional phylotypes being identified as the curves had not reached a complete plateau. It is not surprising that the SW3 library (in which only 26 clones were sequenced) was incompletely sampled (Figure 1). SW3 is an ephemeral pool that is typically dry by mid-July, and we have no geochemical analyses available on this feature. Freezer failures unfortunately resulted in the loss of SW3 samples and nucleic acids after these analyses were underway. However, since many of the sequences generated from this single sampling of
TABLE 1 Site descriptions for samples used to construct 16S rRNA gene clone libraries analyzed in this study Sample name (abbreviation) Sulfur Works 3 (SW3)
Devil’s Kitchen 12 (DK12)
Upper Sulfur Works 5 (USW5)
Upper Sulfur Works 3 (USW3) Boiling Springs Lake (BSL)
Description Small (
