Content uploaded by Hazel Barton
Author content
All content in this area was uploaded by Hazel Barton on Feb 29, 2016
Content may be subject to copyright.
Available via license: CC BY-NC 4.0
Content may be subject to copyright.
Available online at www.ijs.speleo.it
International Journal of Speleology
Ofcial Journal of Union Internationale de Spéléologie
The Impact of Host Rock Geochemistry on Bacterial Community Structure
in Oligotrophic Cave Environments.
Hazel A. Barton1, Nicholas M. Taylor1, Michael P. Kreate2, Austin C. Springer2,
Stuart A. Oehrle3 and Janet L. Bertog2.
INTRODUCTION
Caves, with limited exception, form through the
erosional processes of water. By the time caves are
enlarged sufciently to allow human access, the water
has (generally) departed, leaving the cave exposed to
an oxygenated atmosphere (Klimchouk et al., 2000).
Without sunlight energy, the entry of nutrients into the
system becomes a function of the geology and depth
of the cave; signicant organic input is limited to the
entrance zone and areas fed by surface water entering
the system through faults and fractures (Klimchouk
et al., 2000). Due to extremely low biomass in these
environments and the difculty in extracting DNA
1.Department of Biological Sciences, Northern Kentucky
University, Highland Heights, KY 41099, United States of
America.
2.Department of Physics & Geology, Northern Kentucky
University, Highland Heights, KY 41099, United States of
America.
3.Department of Chemistry, Northern Kentucky University,
Highland Heights, KY 41099, United States of America.
Barton H. A., Taylor N. M., Kreate M. P., Springer A. C., Oehrle S. A. and Bertog J. L. 2007. The impact of host rock geochemistry on
bacterial community structure in oligotrophic cave environments. International Journal of Speleology, 36 (2), 93-104. Bologna (Italy).
ISSN 0392-6672.
Despite extremely starved conditions, caves contain surprisingly diverse microbial communities. Our research is geared toward
understanding what ecosystems drivers are responsible for this high diversity. To asses the effect of rock fabric and mineralogy,
we carried out a comparative geomicrobiology study within Carlsbad Cavern, New Mexico, USA. Samples were collected from
two different geologic locations within the cave: WF1 in the Massive Member of the Capitan Formation and sF88 in the calcareous
siltstones of the Yates Formation. We examined the organic content at each location using liquid chromatography mass spectroscopy
and analyzed microbial community structure using molecular phylogenetic analyses. In order to assess whether microbial activity
was leading to changes in the bedrock at each location, the samples were also examined by petrology, X-ray diffraction (XRD) and
scanning electron microscopy with energy dispersive X-ray spectroscopy (SEM-EDX). Our results suggest that on the chemically
complex Yates Formation (sF88), the microbial community was signicantly more diverse than on the limestone surfaces of the
Capitan (WF1), despite a higher total number of cells on the latter. Further, the broader diversity of bacterial species at sF88
reected a larger range of potential metabolic capabilities, presumably due to opportunities to use ions within the rock as nutrients
and for chemolithotrophic energy production. The use of these ions at sF88 is supported by the formation of a corrosion residue,
presumably through microbial scavenging activities. Our results suggest that rock fabric and mineralogy may be an important driver
of ecosystem function and should be carefully reviewed when carrying out microbial community analysis in cave environments.
Keywords: caves, geomicrobiology, geochemistry, phylogenetics, oligotrophy
Abstract:
Received 26 April 2007; Revised 14 May 2007; Accepted 22 May 2007
from chemically complex geologic samples, most
studies of microbial activity in cave environments
have tended to examine areas of measurable energy
input (Angert et al., 1998; Barton & Luiszer, 2005;
Bottrell et al., 1991; Culver, 1982; Groth et al.,
2001; Sarbu et al., 1996). In a recent study we used
molecular phylogenetics to determine what, if any,
microbial activity was occurring within an oligotrophic
cave without measurable energy input (Barton et al.,
2004). Our results suggested that a diverse microbial
ora subsisted in this oligotrophic environment, in
contrast to previous cultivation studies from similar
environments (Groth & Saiz-Jimenez, 1999; Groth et
al., 1999). Further, the microbial community within
this environment appeared to subsist by using barely
perceptible carbon and energy sources; these included
organics entering the system through percolation,
or the presence of volatile organic molecules within
the atmosphere (Barton et al., 2004). The presence
of bacterial phylotypes with identity to organisms
capable of carrying out iron oxidation suggested
the use of reduced iron as an energy source, while
International Journal of Speleology 36 (2) 93-104 Bologna (Italy) July 2007
94
a high proportion of nitrogen assimilating organisms
suggested a source for nitrogen (Laiz et al., 1999). A
broad phylogenetic distribution of bacterial species
has been identied by other investigators in similarly
oligotrophic environments (Chelius & Moore, 2004;
Osman et al., 2005).
Together, our data led us to hypothesize that
the large diversity of microorganisms found
in oligotrophic cave environments may reect
mutualistic interactions to support community
growth under such starved conditions (Barton &
Jurado, 2007); due to the complex nature of the
organic carbon and inorganic energy sources,
not one organism is capable of carrying out all
the energetically favorable reactions necessary to
support growth (Juttner, 1984; Laiz et al., 1999).
Rather, energetic restrictions allow certain reactions
to proceed only through a close interaction with
species that remove intermediates, allowing energy
conservation in what would otherwise be an
endothermic reaction (Schink, 2002). Mutualism has
been described for microbial communities carrying
out anaerobic ethanol fermentation, methane
oxidation and the breakdown of complex aromatic
compounds (Schink, 2002). Such interactions may
also be a central issue in the unculturability of most
microorganisms in the environment; mutualistic
interactions make many organisms recalcitrant to
cultivation, where appropriate growth conditions
may be dependent on specic interactions with
other species (Grotenhuis et al., 1991; Mohn &
Tiedje, 1992).
To further examine drivers of microbial diversity
in starved environments, this study attempts to
determine what role rock fabric and mineralogy
plays in community diversity under oligotrophic
conditions. This was done by carrying out a
comparative analysis of two microbial communities
within Carlsbad Cavern, New Mexico, USA, where
bacterial species exist on disparate geologic surfaces.
We used a molecular and geochemical approach
to examine community structure at each location
and compared microbial interactions with the rock
matrix of the cave.
MATERIALS AND METHODS
Sample sites and geology
Carlsbad Cavern was formed in the Capitan Reef complex
by hypogenic sulfuric acid speleogenesis, with a postulated
biogenic origin (Engel et al., 2004; Hill, 1990; Palmer, 2000).
The Carlsbad cave system is mostly within the Capitan
limestone, where the relatively impermeable iron-rich, silty
Yates Formation traps oxygenated groundwater and releases
it into the Capitan Formation (Palmer, 2000). Upper portions
of the cave are also located in the Yates Formation, which
contains numerous calcareous siltstones, sandstones and
secondary minerals. The cave is located is in a desert area,
does not contain any surface streams and is not prone to
ooding.
Samples for collection were identied within Carlsbad
Caverns based on a number of parameters, including
geologic location, altered bedrock or secondary
mineralization (Fig. 1). The rst site, WF1, is along the
Main Corridor and located in the limestone of the Massive
Member of the Capitan Formation (CaCO3), with an average
annual temperature of 12.5˚C (Forbes 2000) and a relative
humidity (RH) of 95%, measured using an RH300 Digital
Psychrometer (Extech Instruments, Waltham, MA). The
second site, sF88, is located within the Yates Formation,
with the sample collection site directly above a calcareous
siltstone bed (Fig. 1), with an average annual temperature of
16.3˚C and measured RH of 99% (Forbes 2000). The Yates is
comprised a ne-grained, laminated pisolitic dolomite in thin
beds, inter-layered with thin layers of calcareous red quartz
siltstones and ne-grained sandstones (Borer & Harris,
1991; Brown & Loucks, 1993; DuChene, 2000; Mutti &
Simo, 1993). The Yates is rich in magnesium and iron, with
its red color due to the presence of hematite (Fe2O3), which
is not generally detectable by scanning electron microscopy
energy dispersive X-ray spectroscopy (SEM-EDX) (Borer &
Harris, 1991). Three 5 g rock samples were collected from
each location using a sterilized Dremel drill tool and each
sample was preserved in an appropriate manner for the
subsequent tests: DNA extraction in 70% alcohol / -20°C;
chemical samples were collected in gamma-irradiated clean
tubes and stored at 4°C; rock samples were collected in 50
ml plastic tubes.
Fig. 1. Prole (line plot facing north) of the Carlsbad Cavern cave system (approximately 48 km of passage is represented) with the corresponding
geologic units of the Capitan Reef complex overlain. The two sample locations (WF1 and sF88) are indicated by the lled circles. The entrance and
‘Big Room’ are designated. Courtesy of the Cave Resource Ofce, Carlsbad Caverns National Park.
Hazel A. Barton, Nicholas M. Taylor, Michael P. Kreate, Austin C. Springer, Stuart A. Oehrle and Janet L. Bertog.
International Journal of Speleology, 36 (2), 93-104. Bologna (Italy). July 2007
95
Chemical Analysis
Total organic carbon (TOC) was measured by
extracting crushed rock with dH2O and then
determining the TOC g-1 of rock material using a
Shimadzu TOC-VCSN analyzer at Waters Laboratory,
Western Kentucky University, KY. Analyses of sample
extracts for organic carbon were carried out using
high performance liquid chromatography-mass
spectrometry (HPLC/MS). The system consisted of an
Alliance 2695 HPLC system, 2996 photodiode array
detector and a ZQ single quadrapole mass spectrometer
(all equipment was from Waters Corp., Milford, MA).
Approximately 100mg of sample from each of the sites
was extracted in 1 ml of a 50/50 water and acetonitrile
solution by sonication (all solvents were of HPLC
grade or better). Samples were than allowed to stand
and settle prior to the top (clear) layer extracted and
analyzed by HPLC/MS. A standard gradient was run
using an Xterra MS C18 column (2.1X100mm 3.5µm
particle) using a formic acid and acetonitrile gradient
over 30 minutes.
DNA extraction
DNA extractions were carried out in a laminar-
ow hood, using aseptic techniques and aerosol
resistant tips to reduce the chance of contamination
from outside sources (Barton et al., 2006). Unless
stated otherwise, all chemicals were obtained from
Sigma-Aldrich (St. Louis, MO) and reagents used
were prepared from Fluka ultrapure DNase/RNase
Free water, followed by ltration through a 0.2 µm
cellulose lter to prevent contamination. In cases
where contaminating DNA may be introduced from
reagents, these were subjected to 3000 µJ cm-1 of
UV radiation using a Stratalinker 2400 (Stratagene,
La Jolla, CA). To extract the small amount of DNA
present in the rock, a modied bead beating method
was used.
To extract the DNA, approximately 0.5g of sample
was crushed using a ame-sterilized plattner’s mortar
and pestle (Humboldt Manufacturing, Norridge, IL).
To this 500µl 2X buffer EA [200 mM Tris (pH 8.0),
300 mM EGTA, 200 mM NaCl], 3 mg/ml lysozyme
and 10 µg/ml poly-dIdC were added, and incubated
at 37˚C for 30 min. Proteinase K (to 1.2 mg/ml) and
sodium dodecyl sulfate (SDS to 0.3% wt/vol) were
added, mixed gently and incubated at 50°C for 30
min. Subsequently, 200µl of 20% SDS and 500µl
phenol-chloroform-isoamyl alcohol (24:24:1) was
added before disruption using a Mini-bead beater
(Biospec, Bartlesville, OK) on low setting for 2 min
and high for 30 s. Samples were centrifuged at 13,000
x g in a micro-centrifuge for 3 min at 4°C to deposit
the sample debris; the supernatant (approximately
700-800µl) was then removed and the DNA by the
addition of 2 µg poly-dIdC, 0.3 M sodium acetate and
2 volumes of cold ethanol. Isolated DNA was further
puried by dialysis against 100 ml of 20 mM EGTA
at 4°C for 4 hours in a Silde-A-Lyzer mini dialysis
unit (3500 MWCO; Pierce, Rockford, IL) to remove
any remaining calcium carbonate. The concentration
of the nal DNA preparation was determined using
a Nanodrop ND-1000 spectrophotometer (Nanodrop
Technologies, Wilmington, DE).
Polymerase Chain Reaction (PCR) and cloning
To isolate individual 16S rRNA gene clones, PCR
amplication was used). The bacterial 16S rRNA gene
specic 27F (5’ – AGA GTT TGA TCC TGG CTC AG
– 3’) and universal 805R (5’ – GAC TAC CAG GGT ATC
TAA T – 3’) primers were used in reaction mixtures
containing 1 X PCR buffer (Perkin Elmer), 2.5 mM
MgCl2, 200 μM of each deoxynucleoside triphosphate,
300 nM of each forward and reverse primer, and
0.025 U of AmpliTaq Gold (Perkin Elmer) per μl.
Reaction mixtures were incubated on a Mastercycle
Gradient thermal cycler (Eppendorf Scientic) at 94˚C
for 12 min for initial denaturation and activation of
the AmpliTaq Gold. PCR was then carried out with 34
cycles of 94ºC 30 s, 58ºC 30 s, 70ºC 1 min 30 s, and
a nal extension period of 70ºC 2 min. PCR products
were quantied by electrophoresis using a 1.2% wt/
vol agarose gel containing ethidium bromide, puried
using a Qiagen PCR clean up kit (Qiagen, Valencia,
CA) and cloned into an pCR2.1-TOPO cloning vector
according to the manufacturers recommendations
(Invitrogen, Carlsbad, CA).
Screening of rDNA clones by restriction length
polymorphism (RFLP) and DNA sequencing
The 16S rRNA gene inserts were PCR re-amplied
using 100 ng T3 forward and T7 reverse primers
under standard conditions and amplied using 94ºC
for 4 min for initial denaturation, then 38 cycles of
94ºC 1 min, 52ºC 45 s, 72ºC 1 min, with an extension
period of 72ºC for 8 min. PCR products were then
digested using HindPI1 and MspI restriction enzymes
in NEB buffer 2 (New England Biolabs, Beverly, MA).
The restriction digest was incubated at 37ºC for 2
hours before being run on a 2% wt/vol SeaKem LE
agarose gel (FMC BioProducts) and visualized with
ethidium bromide staining with UV illuminescence.
The unique RFLP patterns were grouped visually
and a representative was selected for sequencing.
Sequencing was carried out using the Thermo
Sequenase Cycle Sequencing kit (USB, Cleveland,
Ohio) according to the manufacturer’s guidelines. For
areas that were problematic due to regions of high
GC content, a SequiTherm Excell II DNA sequencing
kit (Epicenter Technologies, Madison, WI) was used.
Sequencing was carried out using uorescently
labeled sequencing primers M13 and T7 on a Long
ReadIR 4200 DNA sequencer (Li-Cor, Lincoln, NE),
which achieved approximately an 800 base rRNA gene
insert in both the forward and reverse directions.
Phylogenetic Analysis
Sequences were compared to available databases
by use of the BLAST (Basic Local Alignment Search
Tool) network service [http://www.ncbi.nlm.nih.gov/
BLAST; Altschul et al., 1997] Partial sequences of the
16S rRNA gene were compiled using the AlignIR 2.0
Fragment Assembly and Contig Editor software (Li-Cor,
Inc). Compiled sequences were examined for chimeric
The Impact of Host Rock Geochemistry on Bacterial Community Structure in Oligotrophic Cave Environments.
International Journal of Speleology, 36(2), 93-104. Bologna (Italy). July 2007
96
sequences by use of the CHIMERA_CHECK program
[http://rdp.cme.msu.edu/html/analyses.html]
and by phylogenetic branching order discrepancies.
Before further phylogenetic analysis, those sequences
displaying similar BLAST hits were directly compared
using the pairwise BLAST alignment tool [http://
www.ncbi.nlm.nih.gov/blast/bl2seq/bl2.html]. Any
sequences that demonstrated ≥ 98% identity toward
each other were considered representatives of the same
phylotype and grouped accordingly; the remaining
sequences screened for contaminants against
the LBC database (Barton et al., 2006). Sequence
alignments were carried out using the ARB Software
Package [http://mpi-bremen.de/molecol/arb], with
additional sequences from the Ribosomal Database
Project (Maidak et al., 2000). Sequence alignments
used for phylogenetic inference were minimized by
use of the Lane Mask, which removes hypervariable
regions of the 16S rRNA gene sequence from the
analysis for Bacterial data sets. Due to the size of the
16S rRNA clones isolated in this study (~800 bp), only
representative sequences from position 40 to 790 (E.
coli numbering) were used in subsequent phylogenetic
analyses. All presented dendrograms were constructed
by use of ARB with evolutionary distance (neighbor-
joining) and parsimonious (heuristic) algorithms.
The robustness of inferred topologies was tested by
bootstrap resampling (1000 replicates) of phylogenetic-
trees, calculated for both algorithms using PAUP*
software (Sinauer Associates Inc., Sunderland, MA).
The sequences obtained from the 16S rRNA gene
clones in this study were deposited in the Genbank
database, accession numbers DQ066600-DQ066618
and DQ228711-DQ228720.
Statistical Approaches
The statistical analyses were performed using
EstimateS version 7.5.0 (Colwell, 2005). Each clone
represented a separate sample without replacement,
and 100 randomizations were performed to obtain the
Chao2 estimator for each sample size (Chao et al.,
2005; Gotelli & Colwell, 2001; Hughes et al., 2001).
Using the singletons and doubletons calculated for
each sample collection by EstimateS, we used the
log transformation of Chao to calculate the 95%
condence intervals (Chao, 1987; Chao et al., 2005).
Rarefaction curves were plotted using SigmaPlot for
Windows Version 7.0 (SPSS Inc., Point Richmond,
CA).
Scanning Electron Microscopy (SEM)
Samples were xed prior to analysis in 4%
paraformaldehyde/PBS. Biological samples were
washed in 70% ethanol, and dehydrated in an ethanol
series to 100%. Samples were dried in a critical point
dryer using liquid CO2 before examination under
environmental SEM conditions using a FEI Quanta
200 ESEM with a Princeton Gamma Tech Avalon
Microanalysis system and environmental secondary as
well as backscatter electron detectors. Bulk chemistry
of the host rock was determined by crushing samples
and examining them by SEM-EDX using and EDAX
brand EDX mounted on a Philips XL30 TMP scanning
electron microscope.
Geologic analysis
Whole rock samples were separated into an outer
residue layer and interior bedrock, powdered using a
Spex Certiprep 8515 shatterbox and analyzed using
a Rigaku Ultima III XRD under air-dried conditions.
Samples were run from 2 (theta) to 70 (theta) with
a step size of 0.05 (theta) and a count time of 2
seconds. For petrographic analysis, whole rock
samples were sectioned, embedded in resin and cut
and ground to 0.03mm for thin section analysis by
Vancouver Petrographic Ltd. Samples were analyzed
using a Nikon E400Pol polarizing microscope.
RESULTS
We began by identifying two distinct areas in Carlsbad
Cavern that appeared to have limited organic carbon
input, but quite different geologic settings (Fig. 1).
WF1 is considered a chemically simple environment
predominately calcium carbonate in chemistry, while
sF88 is chemically complex, reecting shoal and
eolian deposition process that interlayered pisolitic
dolomite with crystalline volcanic and metamorphic
rock fragments. As a result, the Yates contains various
accessory minerals, including hematite, magnetite
(Fe3O4), tourmaline [(Ca,Na)(Li,Mg,Al)(Al,Fe,Mn)6(BO
3)3(Si6O18)(OH)4], zircon (ZrSiO4), rutile (TiO2), apatite
[Ca5(PO4)3(F,Cl,OH)] and epidote [Ca2(Al,Fe)3(SiO4)3(OH
)]. To conrm these chemistries, we carried out EDX
analysis of the host rock (Table 1), which conrmed
the limestone nature of WF1 and the presence of
quartz siltstone (SiO), magnesium, aluminum,
potassium and iron at sF88. The other elements
previously identied in the Yates (Li, Mn, B, etc.) were
below the minimal detectable limits of this technique.
At WF1 (Capitan), the rock on which microbial species
are growing does not appear to have undergone any
signicant rock fabric changes, while at sF88 (Yates)
varied colored corrosion residues were seen on the
surface. We counted the number of microbial cells at
each location by uorescent microscopy (Barton et
al., 2006). At WF1, there were 1.75 x 106 cells g-1 of
wall rock, while at sF88 we observed 6.48 x 105 cells
g-1 of material, which match the cell numbers seen
in similarly oligotrophic cave environments (Barton et
al., 2006). No eukaryotic predators were observed by
microscopy from either location, ruling out predation
as a driver of diversity (Hahn & Hoe, 2001).
Organic Chemistry and LC/MS analysis
Given the different lithology and mineralogy at
each site, for example the Yates in known to be rich
in organics and is a major reservoir for oil and gas
in the Delaware Basin (Borer & Harris, 1991), we
carried out a qualitative analysis to determine if
there was a difference in the type of organic material
present at each location. This would also rule out the
presence of organics as a result of surface spills from
the commercial facilities (sewage tanks, fuel tanks,
etc.) above the cave. Analysis was done of each of
International Journal of Speleology, 36 (2), 93-104. Bologna (Italy). July 2007
Hazel A. Barton, Nicholas M. Taylor, Michael P. Kreate, Austin C. Springer, Stuart A. Oehrle and Janet L. Bertog.
97
Fig. 2. Consensus dendogram of the 16S rRNA gene sequence phylotypes identied from the Carlsbad clone libraries WF1 (white boxes) and sF88
(black boxes). Phylogenetic analyses were carried out using both distance (Neighbor-Joining) and parsimonious (Heuristic) searches, with the
robustness of inferred topologies determined by bootstrap analysis (1000 replicates); the consensus dendogram of both methods is shown. Branch
points supported (bootstrap values >70%) in both phylogenetic analyses are indicated by closed circles, while marginal branch support (bootstrap
values >50% but >70%) in both analyses are show by an open circle. The bar indicates 10% sequence divergence.
The Impact of Host Rock Geochemistry on Bacterial Community Structure in Oligotrophic Cave Environments.
International Journal of Speleology, 36(2), 93-104. Bologna (Italy). July 2007
98
the sites using liquid chromatography coupled mass
spectrometry (LC/MS). A preliminary analysis of
the LC/MS peaks indicates that the organic matter
present is low (measured as 3 µg g-1 at sF88), although
slightly higher at WF1 (results not shown). The data
suggest that the organic material is of a phenolic and
aromatic nature, which is in accord with the structure
of the organic material commonly found in soils and
the postulated origin of much of the organic carbon
observed in caves (Saiz-Jimenez & Hermosin, 1999;
Sylvia et al., 1999).
Molecular Phylogenetic Analysis of WF1 versus
sF88 Communities
In order to determine whether the differences
in geologic chemistry affected the structure of the
microbial communities found at WF1 and sF88 we
carried out a molecular phylogenetic study. Due to the
difculty of obtaining DNA from calcium rich samples,
along with the small biomass associated with these
extremely starved environments, we have developed
a new extraction protocol for calcium rich samples
from caves (Barton et al., 2006). While this protocol
allows us to reproducibly obtain DNA from extremely
low-biomass environments, we still routinely obtain
less than 500 ng of community DNA/g of material. We
therefore use the 8F and 805R 16S rRNA gene primer
set for amplication, which provides the most reliable
PCR amplication at these low DNA concentrations.
While the subsequent rDNA product is short (~800 bp
in length), it still provides sufcient information for
statistically signicant phylogenetic placement (Nei et
al., 1998).
Clone libraries were created for both sF88 and WF1
locations by ligating the PCR product into the plasmid
vector and transforming into chemically competent E.
coli cells; for sF88 144 clones were isolated, while at
WF1, 96 representative clones were used. The 16S
rRNA gene sequence of each clone library was screened
by RFLP analysis to identify unique phylotypes. The
nal clone libraries contained 49 unique phylotypes
for sF88 and 38 unique phylotypes for WF1 and were
groups into operational taxanomic units (OTUs),
demonstrating >98% identity for tree building. The
sequences of these phylotypes were compared with
the NCBI database and the closest cultivated relative
was identied (Table 2).
Surprisingly, many of the phylotypes we identied
shared a greater degree of identity with previously
cultivated species than in a past cave study (Barton et
al., 2004); however, this may simply reect the increase
in size of the 16S rRNA gene sequence database. In order
to conrm the identity of these identied phylotypes, they
were phylogenetically aligned using the ARB sequence
analysis program, followed by statistical analysis of the
resultant dendogram using the PAUP* software program.
The consensus tree for each location conrmed the
phylogenetic placement of the identied species (Fig. 2).
Interestingly, the distribution of phylotypes identied
correlates well with previous cave environments and
similarly starved locations (Barton et al., 2004; Chelius
& Moore, 2004; Osman et al., 2005). It is also interesting
to note that many of the sF88 phylotypes share the
closest identity to 16S rRNA gene sequences identied
in the WF1 library, suggesting the shared ancestry of
the two communities within this cave environment.
While the identity of unique phylotypes at sF88 and
WF1 demonstrates a similar distribution among the
bacterial divisions, the dendogram does not reect the
relative abundance of the phylotypes identied in each
location, which are represented in Fig. 3. It is interesting
that this comparative pie-chart demonstrates a much
greater species distribution within the chemically
complex sF88 environment, even while there is a higher
absolute number of bacterial cells at the WF1 site. In
order to determine whether there were statistically
signicant differences in community structure between
each site, we created rarefaction curves using the Chao
2 non-parametic estimator to determine true species
richness (Fig. 4). These rarefaction curves did indeed
suggest that there were differences in the absolute
diversity between the two microbial communities;
however, the 95% condence levels suggest that the
sample sizes need to be increased to determine the
signicance of these differences.
Geologic Samples: Thin-sections and X-ray powder
diffractrometry (XRD)
One of the most striking differences between WF1
and sF88 sites was the presence of a corrosion residue
Fig.3. Pie-charts representing species distribution at each location
within Carlsbad Cavern. The distribution of all divisions are
represented by each cloned phylotype from the WF1 (88 clones) and
sF88 (144 clones) libraries.
Fig.4. Rarefaction curves of observed operational taxonomic units
(OTUs) represented by individual phylotypes identied within this
study. Each rarefaction curve was calculated from the variance of the
number of OTUs drawn in 100 randomizations at each sample size.
International Journal of Speleology, 36 (2), 93-104. Bologna (Italy). July 2007
Hazel A. Barton, Nicholas M. Taylor, Michael P. Kreate, Austin C. Springer, Stuart A. Oehrle and Janet L. Bertog.
99
at the more starved sF88 site. The two localities were
compared for mineralogic alteration using thin-section
petrography, XRD and SEM-EDX. Petrologically, both
localities were quite different, reecting the difference in
composition between the Capitan and Yates formations
(Fig. 5A and 5B). SEM images conrmed a calcite
bedrock with an apparent microbial biolm at WF1,
but no signicant corrosion surface (Fig. 5A and 5C).
At sF88, however, the surface of the rock underwent a
number of mineralogical and crystallographic changes,
resulting in a poorly consolidated corrosion residue
(Fig. 5B and 5D). This corrosion residue is comprised
of dolomite recrystallized into coarser crystals, with
ne-grained clay minerals and other opaque minerals
present between them.
Microbial species often change the chemical nature
of the environment on which they live through
catabolic processes (Baneld & Nealson, 1997). In
order to determine if such transformations were
occurring at WF1 or sF88 we carried out an SEM-
EDX analysis of insoluble particulate matter in the
rock. In order to identify such minerals, we extracted
the rock in each location with 1M hydrochloric acid
to remove the overwhelming carbonate minerals and
examined individual particulate grains. At WF1 this
material is comprised of clay particles (representative
EDX spectra in Fig. 5E and 5F) normally associated
with the Massive Member of the Capitan Formation
while at sF88 this material comprised of iron oxides
and elemental iron (representative EDX spectra in Fig.
5G and 5H). These iron forms are too small and/or
amorphous to be detected in our XRD analyses (Fig.
6), suggesting a biogenic origin. These results support
the theory that iron oxidation may be one mechanism
of energy production at the sF88 location, indicating
that the microorganisms living in these environments
may be acquiring energy from the host rock itself.
In order to compare the changes in geologic structure
at the two localities, samples were analyzed using
comparative XRD (Fig. 6). At each location, the sample
collected was broken down into two components; the
interior bedrock and the surface layers. The interior
rock was >1 cm away from any observable surface
feature (as demonstrated by the thin-section analysis)
Fig.5. Petrographic thin-sections, SEM and EDX analysis of rock samples from WF1 and sF88. Samples were collected and subjected to thin
sectioning and petrographic analysis at WF1 (A) and sF88 (B). The interior bedrock is mineralogically similar to the outer layer at WF1 (A), while
the corrosion residue lls the entire eld-of-view of the sF88 sample (B). SEM analysis reveals a similar surface mineralogy at WF1 (C), where
individual calcite crystals covered with a biolm material are seen on the surface. A ne, powdery residue predominates on the surface at sF88 (D).
EDX analyses of acid extracts at each location demonstrate the predominance of clay particles at WF1 (E and F) and iron oxides and elemental
iron at sF88 (G and H)
The Impact of Host Rock Geochemistry on Bacterial Community Structure in Oligotrophic Cave Environments.
International Journal of Speleology, 36(2), 93-104. Bologna (Italy). July 2007
100
and was considered representing the bedrock
mineralogy. Each layer was clearly marked in hand
sample, allowing for segregation of the layers prior to
powdering for XRD. The results (Fig. 6) demonstrated
that, in agreement with our thin-section analysis,
that there were no signicant changes in the
mineral structure of the surface rock at WF1, when
compared with the host-rock matrix; however, there
was an accumulation of clay sized particles. We have
seen such clays accumulating at other sites within
Carlsbad Cavern that demonstrate biogenic activity
(Bertog et al., unpublished results). Heating of the
sample to 350˚C for 30 min led to the loss of the clay
peak, suggesting kaolinite. At sF88 there appeared
to be more signicant mineral changes; dolomite in
the bedrock had been removed with an increase in
the relative abundance of the non-soluble bedrock
material, such as quartz and other silicates (Fig. 6)
as would be expected if the calcareous cements of
the Yates Formation had been dissolved. SEM images
of sF88 conrmed a crystallographic change in the
corrosion residue, with a ne (<1µm) powdery residue
(Fig. 5D).
DISCUSSION
The majority of caves contain little available carbon,
making them an ideal environment in which to study
oligotrophic microbial interactions and geochemical
processes on exposed surfaces (Laiz et al., 1999). Such
geomicrobial activity is thought to be indicated by the
presence of corrosion residues: areas of fabric and
mineralogical change in the bedrock, characterized
by a color change and softening or powdering of the
rock (Boston et al., 2001; Canaveras et al., 2001;
Northup et al., 2003). While these residues do contain
an observable microbial population, a mechanism of
formation remains to be determined (Canaveras et al.,
2001; Northup et al., 2003).
At the geochemically simple WF1 site (Capitan), the
microbial community is dominated by members of
the Actinobacteria, a broad class of high G+C, gram-
positive bacteria found predominantly within soil. Of
these, representative phylotypes of the Pseudonocardia
appear to be the most abundant, representing half of
all the identied Actinobacteria and over 80% of the
total community of bacteria found at this location
(Fig. 3). While the identity between the predominant
phylotype and the next closest cultivated species is
only 97%, we can postulate on a general function of
this species in the environment (Achenbach & Coates,
2000; Pace, 1997). Members of the Pseudonocardia
are aerobes that demonstrate a wide metabolic range
for the degradation of complex plant matter, such
as cellulose, suggesting that the community at WF1
is primarily using soil detritus for growth (Dworkin,
2002). Interestingly, other phylotypes identied at
WF1 share similarity to Acinetobacter johnsonii, able
to mobilize phosphate from inorganic sources, and
Comamonas spp., which degrade a number of nitrogen-
containing aromatic compounds, with the release of
usable nitrate and ammonia (Dworkin, 2002; Itoh &
Shiba, 2004). Both of these groups similarly display
saprophytic lifestyles and are routinely found in
the environment under nutrient limiting conditions
(Dworkin, 2002).
In contrast to the relatively simple microbial
diversity identied at WF1, the clone library generated
at the more geochemically complex sF88 site (Yates)
was more diverse, with representatives from the
Alpha-, Beta- and Gammaproteobacteria (Fig. 3);
very similar in structure to other oligotrophic cave
environments (Barton et al., 2004; Chelius & Moore,
2004). Among the phylotypes identied, there was
signicant representation by members of the genera
Brevundimonas, Massilia and Stenotrophomonas;
26%, 18% and 17% respectively. Representative
Brevundimonas spp. are from the Caulobacter family,
which are oligotrophic organisms able to adapt to
extremely starved environments (Dworkin, 2002; Li
et al., 2004). Members of the genus Massilia are able
to utilize a large number of carbohydrates and other
complex organic molecules as carbon and energy
sources, with the subsequent production of acids
(Dworkin, 2002). Such activity may explain some of
the signicant structural changes observed in the
Yates host rock, where the calcareous cements of
the siltstones are easily dissolved by acids, leading
to the formation of the observed corrosion residues.
The large number of phylotypes representative of
Stenotrophomonas and Delftia spp. identied at sF88
(Table 2) is less easily explained, as members of these
genera carry out denitrication reactions, with the
conversion of ammonia to nitrous oxide (Dworkin,
2002). These organisms also play an important role
in the denitrication of complex organic compounds,
such as nitrobenzene. Such denitrication activity
Fig.6. X-ray powder diffractometry (XRD) of the samples at WF1
and sF88. The samples were segregated into an interior (bedrock)
and outer layer for a comparative analysis of mineralogical changes.
The analysis at WF1 reveals the predominantly calcite nature of
the bedrock, with an accumulation of clay particles in the surface
layer. At sF88 there is a dramatic change in the mineral structure,
with removal of the soluble dolomite and an enrichment of insoluble
silicates. Peaks correspond to: C = calcite; Q = quartz; D = dolomite; *
= unconsolidated endolite peak; and G = signature of glass support.
International Journal of Speleology, 36 (2), 93-104. Bologna (Italy). July 2007
Hazel A. Barton, Nicholas M. Taylor, Michael P. Kreate, Austin C. Springer, Stuart A. Oehrle and Janet L. Bertog.
101
is difcult to explain in the context of nitrogen
starvation, unless a key energy conservation activity
is nitrate reduction and/or the reduction of nitrogen-
containing aromatic compounds.
One interesting observation through the SEM-
EDX analyses was the selective enrichment of iron
oxides within the corrosion residues observed at
sF88, while our clone library does not demonstrate
the presence of any ‘classic’ iron-oxidizing species
(Table 2). One explanation may be that the iron-
oxidation that is observed on the surface of the rocks
could be the direct result of autoxidation, wherein
reduced iron within the host rock is exposed to the
oxygenated atmosphere of the cave through microbial
processes (Ehrlich, 2002). Nonetheless, it would be
surprising if the oligotrophic community at sF88 did
not harness Fe(II) as an electron donor before its loss.
The absence of well-known iron-oxidizing species may
reect the need for a more exhaustive phylogenetic
examination of this site (Ehrlich, 2002; Ley et al.,
2006). As with the WF1 community, nitrogen and
phosphorous must be growth limiting factors at sF88.
It is then hardly surprising that phylotypes related
to the nitrogen assimilating species Herbaspirillum
frisingense and Janthinobacterium agaricidamnosusm
were found. Interestingly Acinetobacter spp., which
can also mobilize inorganic phosphate, were identied
at sF88 and may provide an important clue for
nutrient acquisition in these starved ecosystems (Van
Groenestijn et al., 1988).
In attempting to understand the mechanisms
that support the often surprising levels of microbial
diversity in very starved cave environments, our
results suggest that community structure may be
greatly affected by the chemical nature of the rock
on which these organisms grow. In the case of the
WF1 community, which grows on limestone, the rock
has little potential for additional energy sources. As
a result, the community appears to rely more heavily
on heterotrophic growth from allochthonous energy
sources. The clay particles seen with XRD at this
site may be due to the production of organic acids by
microbial species, utilizing these reduced compounds
for growth, which leads to the accumulation of
these insoluble particles. At sF88, the geochemical
complexity of the rock may provide additional energy
sources, allowing species to use chemolithotrophic
mechanisms to conserve energy. The trace elements
available at sF88 could also prove essential to the
growth of microbial species, allowing the formation
of co-enzymes critical in intermediate metabolism.
Indeed, we have known for decades that many cell
types cannot grow without the addition of specic
mineral supplements (Conway de Macario et al.,
1982; Morgan, 1958; Roth et al., 1996). It is therefore
no surprise that the geochemistry of the bedrock can
impact both the microbial species capable of growth
as well as the types of energy conservation reactions
observed. The necessity for trace elements and
inorganic energy sources in growth is apparent at the
sF88 site, where microbial metabolic transformation
has led to extensive mineralogic alterations of the
Yates rock fabric and the formation of a corrosion
residue. Our results suggest that such variations in
geochemistry may have a profound affect on microbial
community structure in cave environments. As a
result, care should be taken when choosing sample
sites for microbial study within caves, as the geologic
setting may add unforeseen complexity to analyses or
complicate the interpretation of comparative studies.
Not only does this study hint at the high microbial
diversity in caves, in which niche biogeochemistry
may be an important driver of species diversity
(Begon et al., 1998), it also emphasizes the need for
a thorough understanding of the geologic conditions
when studying such environments.
ACKNOWLEDGEMENTS
The authors wish to thank Brad Lubbers for
excellent technical assistance, Karl Hagglund and
Brenda Racke for assistance with the SEM and
EDX analyses, Matthew Zacate for assistance in
establishing the ARB database and running the PAUP
software, Michael Queen for excellent assistance in
interpretation of the geology of Carlsbad Caverns,
and Harvey DuChene and an anonymous reviewer
for critical comments that signicantly improved the
manuscript. We would also like to thank the staff, in
particular Paul Burger, at the Cave Resource Ofce at
Carlsbad Caverns National Park for their invaluable
assistance with sample collection.
This work was supported in part by the Kentucky
EPSCoR Program, the Kentucky Academy of Science,
the Center for Integrative Natural Science and
Mathematics (CINSAM) at NKU, and the National Park
Service. Infrastructure support was provided, in part,
by the National Institutes of Heath KY INBRE program
(5P20RR016481-05). NMT and MPK were additionally
supported by NKU SURG awards.
REFERENCES
Achenbach L.A. & Coates J.D., 2000 - Disparity between
bacterial phylogeny and physiology. ASM News,
66: 714-715.
Altschul S.F., Madden T.L., Schaffer A.J., Zhang J., Zhang Z.,
Miller W. & Lipman D.J., 1997 - Gapped BLAST and PSI-
BLAST: a new generation of protein databae search programs.
Nuc. Acids Res., 25: 3389-3402.
Angert E.R., Northup D.E., Reysenbach A.-L., Peek A.S., Goebel
B.M. & Pace N.R., 1998 - Molecular phylogenetic analysis of a
bacterial community in Sulphur River, Parker Cave, Kentucky.
Am. Mineral., 83: 1583-1592.
Baneld J.F. & Nealson K.H., 1997 - Geomicrobiology: Interactions
between microbes and minerals. Rev. Mineral., Washington,
D.C. Mineralogical Society of American, 448 p.
Barton H.A. & Jurado V., 2007 - What’s Up Down There: Microbial
Diversity in Starved Cave Environments. Microbe, 2: 132-138.
Barton H.A. & Luiszer F., 2005 - Microbial Metabolic Structure
in a Suldic Cave Hot Spring: Potential Mechanisms of
Biospeleogenesis. J. Cave Karst Stud., 67: 28-38.
Barton H.A., Taylor M.R. & Pace N.R., 2004 - Molecular
Phylogenetic Analysis of a Bacterial Community in
an Oligotrophic Cave Environment. Geomicrobiol. J.,
21: 11-20.
The Impact of Host Rock Geochemistry on Bacterial Community Structure in Oligotrophic Cave Environments.
International Journal of Speleology, 36(2), 93-104. Bologna (Italy). July 2007
102
Sample Site
WF1 sF88
Element Average Wt%aSD Average Wt% a SD
C 12.32 ± 2.79 ND -
O 41.25 ± 0.72 49.11 ± 12.32
Mg 9.43 ± 0.31 1.18 ± 0.62
Ca 36.16 ± 4.44 ND -
Si 0.84 ± 0.64 33.55 ± 15.21
Al ND - 7.74 ± 3.98
K ND - 2.66 ± 1.23
Fe ND - 5.77 ± 4.25
Total 100% 100%
ND=none detected a=the average was taken from two separate analysis for 50 seconds at 2 kV (take of angle 38,3° count min-1
Table 1. EDX Bulk Elemental Analyses at WF1 and sF88
Phylogenetic Group Clone Clones
Group Closest identied relative % Sequence ID NCBI Accession
Number
sF88
Alphaproteobacteria NMTsF8 26/32 Brevundimonas nasdae 99% DQ066606
NMTsF27 5/32 Brevundimonas vesicularis 94% DQ066612
NMTsF32 1/32 Caulobacter subvibrioides 99% DQ066613
Betaproteobacteria NMTsF1 18/29 Massilia sp. 98% DQ066600
NMTsF3 6/29 Deltia tsuruhatasis 99% DQ066602
NMTsF7 2/29 Acidovorax sp. 98% DQ066605
NMTsF13 2/29 Herbaspirillum frisingense 100% DQ066607
NMTsF44 1/29 Janthinobacterium agaricidamnosusm 99% DQ066617
Gammaproteobacteria NMTsF2 17/31 Stenotrophomonas sp. 98% DQ066601
NMTsF6 1/31 Uncultured Acinetobacter sp. 99% DQ066604
NMTsF16 1/31 Uncultured bacteria clone FS117-02 89% DQ066608
NMTsF19 7/31 Cellvibrio ostraviensis 98% DQ066609
NMTsF26 4/31 Xanthomas retroexus 99% DQ066611
NMTsF36 1/13 Pseudomonas borealis 99% DQ066614
Cytophagales NMTsF40 1/1 Uncultured Bacteroidetes bacterium 98% DQ066616
Actinobacteria NMTsF4 2/7 Nocardioides sp. 98% DQ066603
NMTsF23 2/7 Rhodococcus sp. 100% DQ066610
NMTsF38 1/7 Mycobacterium gordonae 99% DQ066615
NMTsF47 2/7 Curtobacterium sp. 99% DQ066618
WF1 Library
Alphaproteobacteria NMT-WF15 1/1 Methylobacterium aquaticum 99% DQ228717
Betaproteobacteria NMT-WF23 1/1 Comamonas sp. 98% DQ228718
Gammaproteobacteria NMT-WF13 11/13 Uncultured bacterial mud-clone 90% DQ228716
NMT-WF31 2/13 Acinetobacter johnsonii 99% DQ228719
Actinobacteria NMT-WF1 34/65 Pseudonocardia sp. 97% DQ228711
NMT-WF4 14/65 Bacterium Chibacore 1500 90% DQ228712
NMT-WF7 4/65 Crossiella equi 97% DQ228713
NMT-WF9 1/65 Saccharothrix cryophillis 97% DQ228714
NMT-WF10 11/65 Actinomyces sp. 91% DQ228715
NMT-WF33 1/65 Actinobacterium sp. 93% DQ228720
Phylogenetic Group Clone Clones
Group Closest identied relative % Sequence ID NCBI Accession
Number
sF88
Alphaproteobacteria NMTsF8 26/32 Brevundimonas nasdae 99% DQ066606
NMTsF27 5/32 Brevundimonas vesicularis 94% DQ066612
NMTsF32 1/32 Caulobacter subvibrioides 99% DQ066613
Betaproteobacteria NMTsF1 18/29 Massilia sp. 98% DQ066600
NMTsF3 6/29 Deltia tsuruhatasis 99% DQ066602
NMTsF7 2/29 Acidovorax sp. 98% DQ066605
NMTsF13 2/29 Herbaspirillum frisingense 100% DQ066607
NMTsF44 1/29 Janthinobacterium agaricidamnosusm 99% DQ066617
Gammaproteobacteria NMTsF2 17/31 Stenotrophomonas sp. 98% DQ066601
NMTsF6 1/31 Uncultured Acinetobacter sp. 99% DQ066604
NMTsF16 1/31 Uncultured bacteria clone FS117-02 89% DQ066608
NMTsF19 7/31 Cellvibrio ostraviensis 98% DQ066609
NMTsF26 4/31 Xanthomas retroexus 99% DQ066611
NMTsF36 1/13 Pseudomonas borealis 99% DQ066614
Cytophagales NMTsF40 1/1 Uncultured Bacteroidetes bacterium 98% DQ066616
Actinobacteria NMTsF4 2/7 Nocardioides sp. 98% DQ066603
NMTsF23 2/7 Rhodococcus sp. 100% DQ066610
NMTsF38 1/7 Mycobacterium gordonae 99% DQ066615
NMTsF47 2/7 Curtobacterium sp. 99% DQ066618
WF1 Library
Alphaproteobacteria NMT-WF15 1/1 Methylobacterium aquaticum 99% DQ228717
Betaproteobacteria NMT-WF23 1/1 Comamonas sp. 98% DQ228718
Gammaproteobacteria NMT-WF13 11/13 Uncultured bacterial mud-clone 90% DQ228716
NMT-WF31 2/13 Acinetobacter johnsonii 99% DQ228719
Actinobacteria NMT-WF1 34/65 Pseudonocardia sp. 97% DQ228711
NMT-WF4 14/65 Bacterium Chibacore 1500 90% DQ228712
NMT-WF7 4/65 Crossiella equi 97% DQ228713
NMT-WF9 1/65 Saccharothrix cryophillis 97% DQ228714
NMT-WF10 11/65 Actinomyces sp. 91% DQ228715
NMT-WF33 1/65 Actinobacterium sp. 93% DQ228720
Table 2. Summary of the unique phylotype groups identied in the sF88 and WF1 clone libraries.
aSequences were compared against the NCBI GenBank database using a standard BLAST search (08/04; Altschul et al. 1997)
International Journal of Speleology, 36 (2), 93-104. Bologna (Italy). July 2007
Hazel A. Barton, Nicholas M. Taylor, Michael P. Kreate, Austin C. Springer, Stuart A. Oehrle and Janet L. Bertog.
103
Barton H.A., Taylor N.M., Lubbers B.R. & Pemberton A.C.,
2006 - DNA extraction from low-biomass carbonate rock:
an improved method with reduced contamination and
the low-biomass contaminant database. J. Microbiol.
Meth., 66: 21-31.
Begon M., Harper J.L. & Townsend C.R., 1998 - Ecology:
individuals, populations and communities. Cambridge,
Mass., Blackwell Scientic Publications, 1068 p.
Borer J.M. & Harris P.M., 1991 - Lithofacies and Cyclicity
of the Yates Formation, Permian Basin: Implications
for Reservoir Heterogeneity. Am. Assoc. Petrol. Geol.
Bull., 75: 726-779.
Boston P.J., Spilde M.N., Northup D.E., Melim L.A.,
Soroka D.S., Kleina L.G., Lavoie K.H., Hose L.D.,
Mallory L.M., Dahm C.N., Crossey L.J. & Schelble R.T.,
2001 - Cave Biosignature Suite: Microbes, Minerals
and Mars. Astrobiol., 1: 25-55.
Bottrell S.H., Smart P.L., Whitaker F. & Raiswell R.,
1991 - Geochemistry and isotope systematics of
sulphur in the mixing zone of Bahamian blue holes.
Appl. Geochem., 6: 97-103.
Brown A.A. & Loucks R.G., 1993 - Inuence of Sediment
Type and Depositional Processes on Stratal Patterns in
the Permian Basin-Margin Lamar Limestone, McKittrick
Canyon, Texas. Am. Assoc. Petrol. Geol. Bull.,
57: 435-474.
Canaveras J.C., Sanchez-Moral S., Soler V. & Saiz-
Jimenez C., 2001 - Microorganisms and microbially
induced fabrics in cave walls. Geomicrobiol. J.,
18: 223-240.
Chao A., 1987 - Estimating the population size for
capture-recapture data with unequal catchability.
Biometrics, 43: 783-791.
Chao A., Chazdon R.L., Colwell R.K. & Shen T.-J., 2005
- A new statistical approach for assembling similarity
of species composition with incidence and abundance
data. Ecol. Lett., 8: 148-159.
Chelius M.K. & Moore J.C., 2004 - Molecular
Phylogenetic Analysis of Archaea and Bacteria in
Wind Cave, South Dakota. Geomicrobiol. J., 21: 123-
134.
Colwell R.K., 2005, EstimateS: Statistical estimation of
species richness and shared species from samples.
Version 6.01b http://viceroy.eeb.uconn.edu/
estimates
Conway de Macario E., Macario A.J. & Wolin M.J., 1982
- Specic antisera and immunological procedures
for characterization of methanogenic bacteria. J.
Bacteriol., 149: 320-328.
Culver D.C., 1982 - Cave Life: Evolution and Ecology.
Cambridge, MA, Harvard University Press, 200 p.
DuChene H.R., 2000 - Bedrock Features of Lechuguilla
Cave, Guadalupe Mountains, New Mexico. J. Cave
Karst Stud., 62: 109-119.
Dworkin M., 2002 - The Prokaryotes: An evolving
electronic resource for the microbiological community.,
Springer-Verlag, New York.
Ehrlich H.L., 2002 - Geomicrobiology: New York, Marcel
Dekker, Inc., 768 p.
Engel A.S., Stern L.A. & Bennett P.C., 2004 - Microbial
contributions to cave formation: New insights into
sulfuric acid speleogenesis. Geol., 32: 369-372.
Forbes J.R., 2000 - Geochemistry of Carlsbad Cavern
Pool Waters, Guadalupe Mountains, New Mexico. J.
Cave Karst Stud., 62: 127-134.
Gotelli N.J. & Colwell R.K., 2001 - Quantifying
biodiversity: proceedures and pitfalls in the measurement
and comparison of species richness. Ecol. Lett.,
4: 379-391.
Grotenhuis J.T.C., Smit M., Plugge C.M., Yuansheng
X., Van Lammeren A.A.M., Stams A.J.M. & Zehnder
A.J.B., 1991 - Bacteriological composition and structure
of granular sludge adapted to different substrates. Appl.
Environ. Microbiol., 57: 1942-1949.
Groth I. & Saiz-Jimenez C., 1999 - Actinomycetes in
Hypogean Environments. Geomicrobiol. J., 16: 1-8.
Groth I., Schumann P., Laiz L., Moral-Sanchez
S., Canaveras J.C. & Saiz-Jimenez C., 2001 -
Geomicrobiological study of the Grotta dei Cervi, Porto
Badisco, Italy. Geomicrobiol. J., 18: 241-258.
Groth I., Vettermann R., Schuetze B., Schumann P.
& Saiz-Jimenez C., 1999 - Actinomycetes in Karstic
caves of Northern Spain (Altamira and Tito Bustillo). J.
Microbiol. Meth., 36: 115-122.
Hahn M.W. & Hoe M.G., 2001 - Grazing of protozoa
and its effect on populations of aquatic bacteria. FEMS
Microbiol. Ecol., 35: 113-121.
Hill C.A., 1990 - Sulfuric acid speleogenesis of Carlsbad
Cavern and its relationship to hydrocarbons, Delaware
Basin, New Mexico and Texas. Am. Assoc. Petrol. Geol.
Bull., 74: 1685-1694.
Hughes J.B., Hellmann J.J., Ricketts T.H. & Bohannan
B.J.M., 2001 - Counting the uncountable: Statistical
approaches to estimating microbial diversity. Appl.
Environ. Microbiol., 67: 4399-4406.
Itoh H. & Shiba T., 2004 - Polyphosphate synthetic
activity of polyphosphate:AMP phosphotransferase
in Actinetobacter johnsonii 210A. J. Bacteriol.,
186: 5178-5181.
Juttner F., 1984 - Dynamics of the volatile organic
substances associated with Cyanobacteria and Algae
in a eutrophic shallow lake. Appl. Environ. Microbiol.,
47: 814-820.
Klimchouk A.B., Ford D.C., Palmer A.N. & Dreybrodt W., 2000
- Speleogenesis: Evolution of Karstic Aquifers. Huntsville,
AL., National Speleological Society, 528 p.
Laiz L., Groth I., Gonzalez I. & Saiz-Jimenez C., 1999 -
Microbiological study of the dripping water in Altamira Cave
(Santillana del Mar, Spain). J. Microbiol. Meth., 36: 129-
138.
Ley R.E., Harris J.K., Wilcox J., Spear J.R., Miller S.R., Bebout
B.M., Maresca J.A., Bryant D.A., Sogin M.L. & Pace N.R.,
2006 - Unexpected diversity and complexity of the Guerrero
Negro hypersaline microbial mat. Appl. Environ. Microbiol.,
72: 3685-95.
Li Y., Kawamura Y., Fujiwara N., Naka T., Liu H., Huang X.,
Kobayashi K. & Ezaki T., 2004 - Sphingomonas yabuuchiae
sp. nov. and Brevundimonas nasdae sp. nov., isolated
from the Russian space laboratory Mir. Int. J. Syst. Evol.
Microbiol., 54: 819-825.
Maidak B.L., Cole J.R., Lilburn T.G., Parker C.T., Saxman P.R.,
Stredwick J.M., Garrity G.M., Li B., Olsen G.J., Pramanik
S., Schmidt T.M. & Tiedje J.M., 2000 - The RDP (Ribosomal
Database Project) continues. Nuc. Acids Res., 28: 173-174.
Phylogenetic Group Clone Clones
Group Closest identied relative % Sequence ID NCBI Accession
Number
sF88
Alphaproteobacteria NMTsF8 26/32 Brevundimonas nasdae 99% DQ066606
NMTsF27 5/32 Brevundimonas vesicularis 94% DQ066612
NMTsF32 1/32 Caulobacter subvibrioides 99% DQ066613
Betaproteobacteria NMTsF1 18/29 Massilia sp. 98% DQ066600
NMTsF3 6/29 Deltia tsuruhatasis 99% DQ066602
NMTsF7 2/29 Acidovorax sp. 98% DQ066605
NMTsF13 2/29 Herbaspirillum frisingense 100% DQ066607
NMTsF44 1/29 Janthinobacterium agaricidamnosusm 99% DQ066617
Gammaproteobacteria NMTsF2 17/31 Stenotrophomonas sp. 98% DQ066601
NMTsF6 1/31 Uncultured Acinetobacter sp. 99% DQ066604
NMTsF16 1/31 Uncultured bacteria clone FS117-02 89% DQ066608
NMTsF19 7/31 Cellvibrio ostraviensis 98% DQ066609
NMTsF26 4/31 Xanthomas retroexus 99% DQ066611
NMTsF36 1/13 Pseudomonas borealis 99% DQ066614
Cytophagales NMTsF40 1/1 Uncultured Bacteroidetes bacterium 98% DQ066616
Actinobacteria NMTsF4 2/7 Nocardioides sp. 98% DQ066603
NMTsF23 2/7 Rhodococcus sp. 100% DQ066610
NMTsF38 1/7 Mycobacterium gordonae 99% DQ066615
NMTsF47 2/7 Curtobacterium sp. 99% DQ066618
WF1 Library
Alphaproteobacteria NMT-WF15 1/1 Methylobacterium aquaticum 99% DQ228717
Betaproteobacteria NMT-WF23 1/1 Comamonas sp. 98% DQ228718
Gammaproteobacteria NMT-WF13 11/13 Uncultured bacterial mud-clone 90% DQ228716
NMT-WF31 2/13 Acinetobacter johnsonii 99% DQ228719
Actinobacteria NMT-WF1 34/65 Pseudonocardia sp. 97% DQ228711
NMT-WF4 14/65 Bacterium Chibacore 1500 90% DQ228712
NMT-WF7 4/65 Crossiella equi 97% DQ228713
NMT-WF9 1/65 Saccharothrix cryophillis 97% DQ228714
NMT-WF10 11/65 Actinomyces sp. 91% DQ228715
NMT-WF33 1/65 Actinobacterium sp. 93% DQ228720
Phylogenetic Group Clone Clones
Group Closest identied relative % Sequence ID NCBI Accession
Number
sF88
Alphaproteobacteria NMTsF8 26/32 Brevundimonas nasdae 99% DQ066606
NMTsF27 5/32 Brevundimonas vesicularis 94% DQ066612
NMTsF32 1/32 Caulobacter subvibrioides 99% DQ066613
Betaproteobacteria NMTsF1 18/29 Massilia sp. 98% DQ066600
NMTsF3 6/29 Deltia tsuruhatasis 99% DQ066602
NMTsF7 2/29 Acidovorax sp. 98% DQ066605
NMTsF13 2/29 Herbaspirillum frisingense 100% DQ066607
NMTsF44 1/29 Janthinobacterium agaricidamnosusm 99% DQ066617
Gammaproteobacteria NMTsF2 17/31 Stenotrophomonas sp. 98% DQ066601
NMTsF6 1/31 Uncultured Acinetobacter sp. 99% DQ066604
NMTsF16 1/31 Uncultured bacteria clone FS117-02 89% DQ066608
NMTsF19 7/31 Cellvibrio ostraviensis 98% DQ066609
NMTsF26 4/31 Xanthomas retroexus 99% DQ066611
NMTsF36 1/13 Pseudomonas borealis 99% DQ066614
Cytophagales NMTsF40 1/1 Uncultured Bacteroidetes bacterium 98% DQ066616
Actinobacteria NMTsF4 2/7 Nocardioides sp. 98% DQ066603
NMTsF23 2/7 Rhodococcus sp. 100% DQ066610
NMTsF38 1/7 Mycobacterium gordonae 99% DQ066615
NMTsF47 2/7 Curtobacterium sp. 99% DQ066618
WF1 Library
Alphaproteobacteria NMT-WF15 1/1 Methylobacterium aquaticum 99% DQ228717
Betaproteobacteria NMT-WF23 1/1 Comamonas sp. 98% DQ228718
Gammaproteobacteria NMT-WF13 11/13 Uncultured bacterial mud-clone 90% DQ228716
NMT-WF31 2/13 Acinetobacter johnsonii 99% DQ228719
Actinobacteria NMT-WF1 34/65 Pseudonocardia sp. 97% DQ228711
NMT-WF4 14/65 Bacterium Chibacore 1500 90% DQ228712
NMT-WF7 4/65 Crossiella equi 97% DQ228713
NMT-WF9 1/65 Saccharothrix cryophillis 97% DQ228714
NMT-WF10 11/65 Actinomyces sp. 91% DQ228715
NMT-WF33 1/65 Actinobacterium sp. 93% DQ228720
The Impact of Host Rock Geochemistry on Bacterial Community Structure in Oligotrophic Cave Environments.
International Journal of Speleology, 36(2), 93-104. Bologna (Italy). July 2007
104
Mohn W.W. & Tiedje J.M., 1992 - Microbial reductive
dehalogenation. Microbiol. Rev., 56: 482-507.
Morgan J.F., 1958 - Tissue Culture Nutrition. Bacteriol.
Rev., 22: 20-45.
Mutti M. & Simo J.A., 1993 - Stratigraphic Patterns and
Cycle-Related Diagenesis of Upper Yates Formation,
Permian, Guadalupe Mountains. Am. Assoc. Petrol.
Geol. Bull., 57: 515-534.
Nei M., Kumar S. & Takahashi K., 1998 - The optimization
principle in phylogenetic analysis tends to give incorret
topologies when the number of nucleotide or amino acids
used is small. Proc. Ntl. Acad. Sci. USA., 95: 12390-
12397.
Northup D.E., Barnes S.M., Yu L.E., Spilde M.N., Schelble
R.T., Dano K.E., Crossey L.J., Connolly C.A., Boston
P.J., Natvig D.O. & Dahm C.N., 2003 - Diverse microbial
communitiens inhabiting ferromanganese deposits in
Lechuguilla and Spider Caves. Environ. Microbiol.,
5: 1071-1086.
Osman S., Stuecker T., Newcombe D. & Venkateswaran
K., 2005 - Microbial burden and community proles of
the genesis curation laboratory. Abstract 105th General
Meeting, Am. Soc. Microbiol., Atlanta, GA.
Pace N.R., 1997 - A molecular view of microbial diversity
and the biosphere. Science, 276: 734-740.
Palmer A.N., 2000 - Hydrochemical Interpretation of Cave
Patterns in the Guadalupe Mountains, New Mexico. J.
Cave Karst Stud., 62: 91-108.
Roth J.R., Lawrence J.G. & Bobik T.A., 1996 - Cobalamin
(coenzyme B12): synthesis and biological signicance.
Ann. Rev. Microbiol., 50: 137-181.
Saiz-Jimenez C. & Hermosin B., 1999 - Thermally asisted
hydrolysis and methylation of dissolved organic matter
in dripping waters from the Altamira Cave. J. Anal.
Appl. Pyrol., 49: 337-347.
Sarbu S.M., Kane T.C. & Kinkle B.K., 1996 - A
chemoautotrophically based cave ecosystem. Science,
272: 1953-1955.
Schink B., 2002 - Synergistic interactions in the microbial
world. Anton. Leeuw., 81: 257-261.
Sylvia D.M., Fuhrmann J.J., Hartel P.G. & Zuberer D.A.,
1999 - Principles and Applications of Soil Microbiology.
Upper Saddle River, NJ, Prentice Hall, 550 p.
Van Groenestijn J.W., Vlekke G.J.F.M., Anink D.M.E.,
Deinema M.H. & Zehnder A.J.B., 1988 - Role of
cations in accumulation and release of phosphate by
Acinetobacter strain 210A. Appl. Environ. Microbiol.,
54: 2894-2901.
International Journal of Speleology, 36 (2), 93-104. Bologna (Italy). July 2007
Hazel A. Barton, Nicholas M. Taylor, Michael P. Kreate, Austin C. Springer, Stuart A. Oehrle and Janet L. Bertog.