Microb Ecol
DOI 10.1007/s00248-009-9525-9
MICROBIOLOGY OF AQUATIC SYSTEMS
Habitat Heterogeneity and Associated Microbial Community
Structure in a Small-Scale Floodplain Hyporheic Flow Path
Jennifer L. Lowell & Nathan Gordon & Dale Engstrom &
Jack A. Stanford & William E. Holben &
James E. Gannon
Received: 16 December 2008 / Accepted: 22 April 2009
# Springer Science + Business Media, LLC 2009
Abstract The Nyack floodplain is located on the Middle
Fork of the Flathead River, an unregulated, pristine, fifthorder stream in Montana, USA, bordering Glacier National
Park. The hyporheic zone is a nutritionally heterogeneous
floodplain component harboring a diverse array of microbial
assemblages essential in fluvial biogeochemical cycling,
riverine ecosystem productivity, and trophic interactions.
Despite these functions, microbial community structure in
pristine hyporheic systems is not well characterized. The
current study was designed to assess whether physical
habitat heterogeneity within the hyporheic zone of the
Nyack floodplain was sufficient to drive bacterial β diversity
between three different hyporheic flow path locations.
Habitat heterogeneity was assessed by measuring soluble
reactive phosphorous, nitrate, dissolved organic carbon,
dissolved oxygen, and soluble total nitrogen levels seasonally at surface water infiltration, advection, and exfiltration
zones. Significant spatial differences were detected in
J. L. Lowell (*) : N. Gordon : W. E. Holben : J. E. Gannon
Microbial Ecology Program, Division of Biological Sciences,
The University of Montana,
32 Campus Dr. 4824,
Missoula, MT 59812-1002, USA
e-mail: [email protected]
J. L. Lowell : W. E. Holben
Montana—Ecology of Infectious Diseases Program,
The University of Montana,
Missoula, MT 59812-1002, USA
D. Engstrom
Department of Geosciences, The University of Montana,
Missoula, MT 59812-1002, USA
J. A. Stanford
Flathead Lake Biological Station, The University of Montana,
Polson, MT 59860, USA
dissolved oxygen and nitrate levels, and seasonal differences
were detected in dissolved oxygen, nitrate, and dissolved
organic carbon levels. Denaturing gradient gel electrophoresis (DGGE) and cell counts indicated that bacterial diversity
increased with abundance, and DGGE fingerprints covaried
with nitrate levels where water infiltrated the hyporheic
zone. The ribosomal gene phylogeny revealed that hyporheic
habitat heterogeneity was sufficient to drive β diversity
between bacterial assemblages. Phylogenetic (P) tests
detected sequence disparity between the flow path locations.
Small distinct lineages of Firmicutes, Actinomycetes, Planctomycetes, and Acidobacteria defined the infiltration zone
and α- and β-proteobacterial lineages delineated the
exfiltration and advection zone communities. These data
suggest that spatial habitat heterogeneity drives hyporheic
microbial community development and that attempts to
understand functional differences between bacteria inhabiting nutritionally heterogeneous hyporheic environments
might begin by focusing on the biology of these taxa.
Introduction
Fluvial systems are dynamic environments shaped by
interactions between channel-associated geomorphic processes, the alluvial aquifer, and riparian vegetation, all of
which drive lateral and vertical connectivity within a
floodplain [39, 48, 49]. The hyporheic zone is a unique
heterotrophic component of fluvial ecosystems, comprising
the interface between open surface water and groundwater
[17, 45, 50, 51]. It is within this spatially fluctuating
ecotone that the exchange of nutrients and biota between
surface water, groundwater, the alluvial aquifer, and the
riparian zone occurs [7]. The resulting mosaic of habitats
and nutrient gradients are capable of supporting a diverse
J. L. Lowell et al.
array of microorganisms important for biogeochemical
cycling between floodplain components [14–16, 23]. As a
result, productivity of the hyporheic zone can be orders of
magnitude greater than that in benthic sediments [10]
accounting for 76–96% of floodplain ecosystem respiration
[38], and playing a crucial role in nutrient supply and biotic
productivity in surface waters [21, 23]. Despite these
contributions from hyporheic biofilms, most studies
addressing microbial community structure in streams have
targeted benthic biofilms [2, 5, 9, 19, 27, 34, 35]. The few
existing studies of hyporheic microbial community structure have focused on contaminated systems [15], artificial
systems [3, 4, 6] and geographically distinct hyporheic sites
[16], leaving the drivers of hyporheic biofilm community
structure within natural floodplain ecosystems virtually
unknown.
In the current study, we measured nutrient and dissolved
oxygen levels to assess habitat heterogeneity along a 100-m
hyporheic flow path in a pristine northwestern Montana
floodplain. These values were compared to bacterial
abundance, 16S rRNA gene sequence data, and 16S
polymerase chain reaction (PCR)-DGGE patterns. We
hypothesized that bacterial assemblages in the hyporheic
zone would correlate to spatial habitat heterogeneity
creating β diversity (i.e., differences in communities
between heterogeneous environments) among bacterial
communities. A redundancy analysis and tests for correlation were used to relate differences in bacterial community
structure to environmental variability. Additionally, we
incorporated phylogenetic information into our community
analyses, accounting for the degree of sequence divergence
and relative sequence abundance, to make a quantitative
assessment of bacterial β diversity [32, 36]. Finally, we
discuss the characteristics of the bacterial taxa identified
using 16S rRNA gene sequence data and the implications
of these taxa for hyporheic zone productivity.
Materials and Methods
Study Site
The Nyack Floodplain (hereafter, Nyack) located on the
Middle Fork of the Flathead River is a fifth-order
unregulated stream in Northwestern Montana (48°26′
53.77″ N, 113°49′28.28″ W). The valley encompassing
the Nyack borders Glacier National Park and the Bob
Marshall/Great Bear wilderness area and is considered a
nearly pristine river catchment. The Nyack is approximately
8 km long and 1 km wide, is situated under a riparian
forest, and contains 80% of regional vascular plant flora
and at least 70% of regional aquatic invertebrate species,
including 80 hyporheic species [22, 46]. Hyporheic flow
paths exist at two distinct spatial scales: at the large
floodplain scale (kilometers) and at the scale of individual
gravel bars (meters) [39]. Our study site was located on a
smaller scale flow path (~100 m) designated as the Movie
Road gravel bar (hereafter, MR) within the Nyack (48°26′
53.77″ N, 113°48′33.41″ W). Infiltration, river-dominated
lateral flow of groundwater (advection), and exfiltration
zones were previously identified by measuring hydraulic
head gradients through piezometer placements in the near
bank and river bottom around the gravel bar perimeter.
Water and Sediment Sampling
Groups of five 1-in. polyvinyl chloride (PVC) sampling
wells were installed using a Geoprobe model 5400 at each
of the three hydraulic zones (infiltration, advection, and
exfiltration) across the gravel bar (n = 3, N = 15). Each
group of five wells was placed perpendicular to the flow
path in order to obtain five replicate samples for each time
point and location (Fig. 1). Bulk sediment was collected
from the banks of the MR gravel bar and hand-sieved using
methods previously described [16]. The 1.7–2.36-mm size
sediment fraction was transported to the laboratory, washed
with Milli-Q water, autoclaved at 120°C for 120 min, and
baked overnight in a drying oven at 265°C. This procedure
was repeated three times on separate days. Sterile sediment
(10 g) was packed into acid-washed PVC tubes 9 cm
long × 2.1 cm in diameter and slotted with 0.15 cm wide
slots [16]. The ends of each tube were plugged with
autoclaved, size 16D silicone stoppers (Saint-Gobain,
Akron, OH) to retain the sediments. Four tubes of sediment
were connected end to end using acid-washed zip ties and
installed in each of the 15 1-in. wells. The tubes were
suspended within the top meter of the saturated zone using
60-pound-test fishing line attached to the well cap.
Following sampler placement in the wells, biofilms were
allowed to colonize the sediment for 20 weeks. This time
interval was selected based on prior experiments (performed at this site using the same techniques) that
compared microbial communities colonizing implanted
samplers to those on fresh sediment collected from adjacent
pits of equal depth and sampling time interval. DGGE
pattern matching analysis showed that, after 16–20 weeks
of incubation in situ, sediment community composition in
the samplers was highly similar to that on freshly taken
native sediment samples [24]. Following the colonization
period, one sediment sampler from each well was collected
every 16 weeks over a 64-week period, for a total of 60
sediment samples. Sediment samples were placed in sterile
“Twirl’Em” bags (Fisher Scientific, Pittsburg, PA) and
transported to the laboratory on wet ice. Samples to be used
in molecular analyses were immediately frozen at −70°C
before DNA extraction.
Hyporheic Flow Path Microbial Communities
nitrogen (STN) were measured during the spring and early
fall sampling times only. Water chemistries were measured
using the following calorimetric methods according to “The
standard methods for the examination of water and
wastewater, American Public Health Association guidelines” (http://www.apha.org): SRP, ascorbic acid colorimetry; NO3−, Cd reduction–azole colorimetry; and STN,
filtration–persulfate digestion–CD reduction–azole colorimetry. DOC was measured using a Tekmar–Dohrmann
carbon analyzer (Teledyne Tekmar, Mason, OH), and DO
was measured using an YSI Model 85 handheld DO meter
(Yellow Springs Instruments Inc., Yellow Springs, OH).
Direct Microscopic Enumeration of Sediment Biofilm
Bacterial Communities
Figure 1 The top panel shows the location of the Nyack floodplain
within MT. The bottom panel shows the MR gravel bar expanded
from the Nyack floodplain map. M1 A-E infiltration wells, M2 D-H
advection wells, M3 D-H exfiltration wells
Additionally, wells were purged for several minutes, and
two water samples per well were collected. The first were
collected in sterile 250-ml plastic bottles for inorganic
nutrient analyses and the second in sterile 250-ml glass
bottles for dissolved organic carbon (DOC) analyses.
Samples were either acidified with phosphoric acid or
filtered through 0.2-μm nylon membrane filters (Whatman,
Florham Park, NJ) and frozen for nutrient measurements.
Sampling dates were July 8, 2005 (summer), November 10,
2005 (late fall), May 8, 2006 (spring), and September 15,
2006 (early fall). Winter samples were not taken, as this
area receives very heavy snowfall.
Environmental Variables
To assess habitat heterogeneity, soluble reactive phosphorous (SRP), nitrate (NO3−), and DOC were measured for all
sampling times. Dissolved oxygen (DO) and soluble total
Bacteria in each sediment sample were enumerated (in
triplicate) by direct epifluorescent microscopy after staining
with 4′,6-diamidino-2-phenylindole (DAPI) [40]. Treatments were carried out in sterile 15-ml centrifuge tubes.
Two grams of sediment (wet weight) were suspended in
filter-sterilized formation water containing 0.01 M tetrasodium pyrophosphate [47] and fixed with formalin (3% final
concentration). Samples were mixed thoroughly for 1 min,
held on ice for 30 min, mixed again for 1 min, and
sonicated for 20 min using a Branson 3210R-DTH
ultrasonic cleaner (Fredricksburg, VA). After sonication,
samples were vortex-mixed for 1 min and centrifuged at
750×g for 10 min using a Sorvall RC 5B Plus (Thermo
Scientific, Waltham, MA) to remove large particulates [12].
Subsamples from each supernatant were transferred to
amber microcentrifuge tubes (Fisher Scientific, Waltham,
MA) and stained for 30 min with DAPI at a final
concentration of 10.0 mg l−1 [52]. Samples (1.0 ml) from
each tube were filtered through 0.2-μm black polycarbonate
membranes (Poretics Products, Kent, WA), and microscope
slides were prepared as previously described [52]. Filters
were examined microscopically at magnification ×1,000
using a Zeiss Axioskop 20, and ten fields per filter were
counted.
Direct Plate Counts from Sediment Biofilm Bacterial
Communities
Two grams of sediment (wet weight) from each well were
suspended in 8 ml of filter-sterilized formation water
containing 0.01 M tetrasodium pyrophosphate [47]. Samples were mixed thoroughly for 1 min, held on ice for
30 min, mixed again for 1 min, then sonicated for 20 min as
described above. After sonication, samples were vortexmixed for 1 min and allowed to settle for 15 min. Each
supernatant was serially diluted using filter-sterilized
formation water. Four dilutions (1:1,000–1:1,000,000) were
J. L. Lowell et al.
plated in triplicate on agar made with a filter-sterilized
mixture (1:9) of soil leachate and river water and 1% Noble
agar (Difco). Briefly, soil leachate was created by continuous agitation of floodplain soil/sediment (excavated from
the top meter of the water table) in Milli-Q water for 1 h on
a rotary shaker (150 rpm) followed by centrifugation
(5,000 rpm for 10 min) and filter sterilization (0.2 μm).
The leachate was analyzed for dissolved organic carbon
using the above method and diluted to 3 mg/L DOC (about
three times over in situ DOC levels). Plates were incubated
for 4 weeks at 10°C in the dark (in situ temperatures range
from 4 to 12°C). Colonies were counted on plates
containing 30–300 total colony-forming units (CFUs).
DNA Extraction from Sediment Biofilm Bacterial
Communities
Bacterial community DNA was recovered from 1.0 g of
sediment collected from each well in each season as
previously described [16, 53] with modification as follows:
Bead beating was performed using the Genogrinder 2000
(Clifton, NJ), followed by freezing in liquid nitrogen, and
then boiling for 2 min. The bead-beating and freeze–thaw
cycles were repeated four times before proceeding to the
extraction and precipitation steps [53]. Before molecular
analyses, DNA extractions were further purified using the
Promega Wizard® DNA cleanup system (Madison, WI).
Samples were quantified using a NanoDrop® ND-1000
UV-Vis spectrophotometer (Wilmington, DE) and stored at
−20°C for downstream analyses.
PCR Amplification and DGGE Analysis
PCR-DGGE was performed on individual DNA samples
and also on pooled replicate DNA samples from each
location and season for a total of 72 samples. Each 50-µl
reaction contained 1× PCR buffer with 1.5 mM MgCl2,
200 μM dNTPs, 0.8 μM of each general bacterial 16S
rRNA gene primer 536fc, which included a GC clamp for
DGGE [25] and 907r [28], 5 μg of bovine serum albumin
(Roche, Pleasanton, CA), 1.25 U HotStar Taq polymerase
(Qiagen, Valencia, CA), and 100 ng of template DNA.
Reactions were performed in a PTC-100 thermal cycler (MJ
Research, Waltham, MA) at 95°C for 5 min followed by 30
cycles of 95°C for 1 min, 56°C for 1 min, and 72°C for
1 min and a final elongation step of 72°C for 10 min. PCR
product was visualized on 1.5% agarose gels and quantified
using a FirstLight UV Illuminator (UVP Inc., Upland, CA)
gel image analyzer.
PCR amplicons from individual sediment DNA samples
were separated by DGGE on a Bio-Rad D-GENE System
(Bio-Rad Laboratories, Hercules, CA) to assess microbial
community variability across each group of wells (n=5) for
each time-point (season) and for each zone (location).
Amplicons from the pooled replicate DNA samples were
used for community comparisons on one DGGE gel to
avoid introducing gel-to-gel variability issues. A linear
gradient of 45–60% denaturant (100% denaturant equals
7 M urea and 40% w/v formamide) in an 8% poylacrylamide gel matrix was used. Gels were run at 70 V for 17 h at
60°C. Following electrophoresis, gels were stained for 1 h
at 37°C with SYBR® Gold (Molecular Probes, Eugene,
OR) and visualized using a FirstLight UV Illuminator
(UVP Inc., Upland, CA) as described above. Similarity
matrices were constructed from DGGE banding information using the Jaccard similarity coefficient function in
Bionumerics v. 4.61 (Applied Maths, Austin, TX).
Cloning and Sequencing of DGGE Bands
Major DGGE bands were excised from each lane in the
pooled sample gel, re-amplified, cloned, and sequenced.
Excised bands were placed in 100 µl of sterile water and
incubated at 4°C overnight to elute DNA from the gel
matrix. Re-amplification was performed according to the
above PCR conditions using 1 µl of the eluted DGGE band
DNA as template and an unclamped version of primer 536f
and 907r. PCR products were purified using the Qiagen
PCR purification kit (Valencia, CA), and TA cloned using
the pGEM®-T Easy Vector System and kit supplied JM109
cells (Promega, Madison, WI). Five colonies per each
cloned amplicon were chosen in order to detect heterogeneity in DGGE bands caused by the co-migration of
multiple PCR amplicons [43]. The selected colonies were
grown for 20 h at 37°C in Luria–Bertani broth containing
100 µg/ml of ampicillin after which plasmids were purified
with the 5-Prime Inc., Perfectprep Plasmid 96 Vac Kit
(Gaithersburg, MD) and submitted to Polymorphic DNA
Technologies, Inc. (Alameda, CA) for sequencing. Sequencing of DGGE bands yielded products of approximately 371 bp (following removal of primer sequences)
spanning the V4–V5 region of the 16S gene. Recent studies
have indicated that relatively short 16S fragments of the
V4–V5 region are sufficient for detecting taxon relationships with high jackknife support and that, by using this
region, it is possible to obtain the same classification
outcome (to the genus level) as from full-length sequences
[30, 37].
Statistical Analyses
Two-way analyses of variance (ANOVAs) were used to
detect variation in SRP, NO3−, DOC, DO, and STN levels
and bacterial counts. Similarity matrices were constructed
from DGGE fingerprints, and non-parametric ANOVAs
(Kruskal–Wallis) were used to assess bacterial community
Hyporheic Flow Path Microbial Communities
structure variability within replicate samples at each gravel
bar location. Samples from each time and location were
then pooled based on failure to detect significant differences in banding patterns within the replicates. Pooled
sample analysis also enabled us to run all samples on one
gel, thereby eliminating gel-to-gel variability [43]. Correlations between environmental variables, DAPIs, CFUs, and
DGGE similarity matrices were examined by Spearman’s
rho tests for correlation. The above statistical tests were
carried out using SPSS® v. 15.0 for windows (SPSS Inc.,
Chicago, IL).
Bray–Curtis distance measures were generated from
DGGE band presence/absence data, as recommended by
[8]. We used distance-based redundancy analysis (DBRDA) [29] to test the hypotheses that sample location
contributed to a significant proportion of the total
variance in the bacterial community profiles and that
bacterial community structure as determined by DGGE
varied with nutrient levels. For these analyses, distance
calculations and principal coordinate analysis (PCA) were
performed using the PrCoord utility packaged with
Canoco (ver. 4.5).
combined communities. The weighted UniFrac option was
chosen as a quantitative measurement of β diversity,
accounting for the relative number of times each sequence
was observed in each habitat [32]. The UniFrac lineagespecific analysis tool was then used to determine which
phylogenetic lineages contributed significantly to β diversity detected by weighted UniFrac [31]. Additionally, the P
test was used to examine whether each location harbored
distinct phylogenetic lineages (i.e., covaried with the tree
phylogeny) and which taxa spatially delineated bacterial
communities [36]. The β-proteobacterial lineage was
identified as a significant contributor to community differences. Therefore, β diversity measures and P tests were
repeated with all sequences in the tree, the proteobacterial
sequences alone, and with each major β-proteobacterial
group removed from the tree as follows: the Comamonadaceae and each of two groups of Oxalobactereaceae. The
tree was also re-evaluated with the α-proteobacterial group
removed. Elimination of each group individually allowed
us to further assess how potential lineage differences
between the three flow path locations influenced β
diversity.
Sequence Analysis
Results
Sequence electropherograms were examined and trimmed
using the SeqMan option in Lasergene v. 5.01. Only unique
sequences were kept from each of the five clones sequenced
per DGGE band. Sequence data were submitted to the
Ribosomal Database Project (RDP) Chimera Check (http://
35.8.164.52/cgis/chimera.cgi?su=SSU) to identify artifactual sequences, and those that appeared chimeric were
excluded from downstream analyses. The RDP Sequence
Match tool was used to search for the nearest neighbors of
the remaining sequences. The 191 retained sequences were
aligned using the nearest alignment space termination
aligner [11], and a maximum parsimony tree was constructed using ARB parsimony [33]. The tree was rooted
with Thermotoga maritima (GenBank accession number
AE000512). The resulting phylogeny was exported for use
in UniFrac to measure β diversity among the infiltration,
advection, and exfiltration zone bacterial communities [31,
32]. This tool allows comparisons of β diversity using
genetic diversity, phylogenetic distances and the degree of
divergence between taxa to determine microbial community
structure according to environment or habitat [32]. For this
study, the infiltration, advection, and exfiltration zones were
considered disparate habitats based on significant chemical
profile differences. Each sequence in the phylogenetic tree
was assigned to the habitat from which the sample came,
and UniFrac was used to measure the phylogenetic
distances between communities by comparing the genetic
diversity of each community to the total diversity of the
Environmental Variables
Measurements were taken for SRP, NO3−, and DOC for
each location and season across the gravel bar to assess
habitat heterogeneity. SRP levels did not differ significantly. Spatially, mean NO3− levels were higher at infiltration
and decreased significantly across the flow path (p<0.001)
except in the late fall. Seasonally, NO3− levels were higher
in the spring and late fall (p<0.001). Levels of DOC
followed a similar seasonal trend (p<0.001; Fig. 2).
In addition to the above variables, DO and STN were
measured during spring and early fall. Mean levels of each
showed similar trends to those of NO3−, although DO levels
did not drop as dramatically as NO3− levels between the
infiltration and advection zones during the spring (p<0.05).
Seasonally, DO and all but exfiltration STN levels varied
significantly (p<0.001; Fig. 2).
Plate Counts and Direct Microscopic Enumeration
Direct microscopic enumeration consistently detected
higher numbers of cells in the infiltration zone and in the
spring at all locations (p<0.01). Plate counts correlated well
with direct microscopic enumeration and indicated that the
number of culturable organisms were also higher in the
infiltration zone (p < 0.01). Direct count values were
consistently ~1.5 orders of magnitude higher than plate
J. L. Lowell et al.
Dissolved Organic Carbon
2.5
250
2
200
nitrate (ug/L)
DOC (mg/L)
Figure 2 Seasonal and spatial
nitrate fluctuations compared to
cell counts and other nutrient
measurements across the hyporheic flow path. Cell enumeration methods, nutrient
abbreviations, and statistical
differences are described in the
text. I infiltration, A advection,
E exfiltration
1.5
1
0.5
150
100
50
0
0
spring
summer
DOC I
early-fall
DOC A
late-fall
spring
Soluble Total Nitrogen
350
DO (mg/l)
300
250
200
150
100
50
0
spring
STN I
9
8
7
6
5
4
3
2
1
0
NO3 A
late-fall
NO3 E
Dissolved Oxygen
DO I
STN E
early-fall
DO A
DO E
Colony Counts
DAPI Counts
1.00E+10
cfu/g sediment
1.00E+10
cells/gm sediment
early-fall
spring
early-fall
STN A
summer
NO3 I
DOC E
400
STN (ug/l)
Nitrate
1.00E+09
1.00E+08
1.00E+07
1.00E+06
1.00E+09
1.00E+08
1.00E+07
1.00E+06
1.00E+05
1.00E+05
spring
DAPI I
summer early-fall late-fall
DAPI A
count values, except during the spring when DAPI and
plate counts were comparable (Fig. 2).
DAPI E
spring
CFUs I
summer early-fall late-fall
CFUs A
CFUs E
from all locations and all seasons except early fall, and two
were generated only from the summer infiltration zone
sample.
Community Structure Analyses by DGGE
Replicate samples (n=5) from each flow path location were
tested using DGGE to examine differences in community
composition within each zone. Significant differences in
DGGE patterns within replicates at each location and at
each time-point were not detected (p value range=0.23–
0.95). Therefore, each group of five replicates was pooled
for comparison on one DGGE gel. DB-RDA indicated that
the infiltration communities differed significantly from
advection and exfiltration communities (p=0.002). Overall,
NO3−, SRP, and DOC explained 29.8% of the community
variance (p=0.26), while NO3− levels alone accounted for
13.4% of the variance (p=0.06) and were associated with
the infiltration zone samples. The community variance
attributed to SRP and DOC were insignificant and
accounted for 7.1% (p=0.79) and 6.8% (p=0.82) respectively (Fig. 3). Although statistically significant seasonal
community clusters were not detected, we observed a
seasonal shift in three major DGGE bands. One was present
Environmental Variables and Microbial Community
Relationships
Bacterial counts and mean DGGE similarity values were
compared with environmental variables to assess nutrient
level influences on bacterial abundance and community
structure. Spearman rho tests identified significantly positive seasonal and spatial correlations between NO3− levels
and colony and DAPI counts (p<0.05), except in late fall.
DOC levels did not correlate with bacterial counts. DAPI
counts correlated negatively with DGGE similarity in all
seasons (p<0.05) except the late fall. This indicated that
zones with increased numbers of cells commonly had more
variability in their DGGE patterns.
Spring and early fall bacterial counts were also compared to DO and STN. Spring plate count values correlated
positively with DO (p<0.01) but spring DAPI values did
not. Both DAPI values and plate count values correlated
positively with STN in the spring (p<0.05). Bacterial
Hyporheic Flow Path Microbial Communities
Figure 3 PCA of DGGE pattern dissimilarity for bacterial community structure between hyporheic locations in each season. Environmental variables were introduced as supplemental data and passively
projected onto the plot to aid in the visualization of correlation
detected by DB-RDA. Vectors are of relative importance to sample separation. Symbols represent locations; circles = infiltration, diamonds =
advection, squares = exfiltration. Abbreviations indicate the seasons in
which samples were taken: SP spring, SU summer, EF early fall, LF late
fall. The nutrient abbreviations are defined in the text
counts did not correlate with any of the environmental
variables in the early fall
Sequencing and Phylogenetic Analyses
Excised DGGE bands were cloned and sequenced to
examine the influence of location (habitat heterogeneity)
across the gravel bar on microbial community structure. β
diversity measures based on sequence phylogeny indicated
that communities from each location along the flow path
were unique. When pair-wise comparisons were made to
test for β diversity, 16S rRNA gene sequences from
infiltration differed from advection sequences (p=0.05),
and those from advection differed from exfiltration sequences (p=0.03), indicating that communities at each location
harbored less within-location genetic diversity than all
samples combined. The P test was also significant for
pair-wise comparisons between infiltration and advection
sequences (p=0.05) and between advection and exfiltration
(p=0.03). When all three locations were compared, the
infiltration community was disparate from the rest of the
gravel bar (p=0.01). This indicated that communities from
each location harbored distinct phylogenetic lineages.
Our phylogeny based on all sequences yielded six major
clades. The Acidobacteria, Actinomycetes, α-Proteobacteria, two groups of β-Proteobacteria (Comamonadaceae
and Oxalobacteraceae), and the γ-proteobacteria. Lineagespecific analyses indicated that the β-proteobacterial lineage was contributing significantly to community differences
between the flow path locations. To assess the influence of
the β-proteobacterial sequence distribution, β diversity
measurements and P tests were repeated using all Proteobacteria alone and with each β-proteobacterial group
independently removed from the tree. When the Comamonadaceae group was removed from the phylogeny, β
diversity between infiltration and advection was still
detected (p<0.05), but P test values became insignificant.
When the Oxalobacteraceae group was removed, all
significance was lost for the infiltration and advection
pair-wise comparison, suggesting that the Oxalobacteraceae were delineating the communities in the advection
samples. The effects of removing α-Proteobacteria from
the phylogeny were also assessed. Removal of this group
eliminated community differences between the advection
and exfiltration samples suggesting that the α-Proteobacteria
were contributing heavily to the exfiltration community.
When only Proteobacteria were left in the tree, measures of
β diversity and P test values became insignificant for the
infiltration zone. This loss of significance suggested that the
non-proteobacterial clades removed from the tree contributed to the differentiation of the infiltration zone community.
Discussion
The chemical gradients observed across MR were consistent with the theories that complex hydrological processes
in fluvial hyporheos influence habitat heterogeneity [17, 45,
50, 51] and that hydrologic flux through the hyporheic zone
leads to changes not achievable with faster surface water
velocities [17]. Decreasing nutrient levels from the infiltration to exfiltration zones on MR were attributed to
increased water subsurface residence time and contact time
with hyporheic sediment microbial communities [17].
Seasonal variation measured across MR reflected terrestrial
nutrient inputs similar to those in other lotic systems and
were consistent with spring snowmelt and fall defoliation
[1, 10, 42]. The observed β diversity among our three flow
path locations suggested that biofilm communities in the
hyporheic zone may be subject to niche diversification and
habitat filtering in ways similar to those observed for
bacterioplankton from other aquatic environments, as well
as soils, and plants [20, 26, 44]. Biofilm sloughing and
bacterial movement through lotic systems surely provides
translocation and mixing of certain community members,
but spatial differences in genetic diversity and major
phylogenetic lineages indicated that the microbial taxa
found in each of our hyporheic locations biofilms originated from different pools of diversity [36]. While habitat
J. L. Lowell et al.
attributions for 16S rRNA gene sequence presence should
be approached with caution, the detection of certain
bacterial groups might reflect how steep chemical and
substrate concentration gradients contribute to seasonal
effects on communities and select for unique phylogenetic
lineages within this hyporheic flow path.
Several factors indicated that higher nutrient levels may
have contributed to higher species richness and greater
seasonal community fluctuations in the infiltration zone
compared to the other two locations. First, P tests revealed
that the infiltration zone was defined by several different
taxa with sequences from the Firmicutes, Planctomycetes,
Acidobacteria, and Actinomycetales distributed throughout
the tree (Table 1). Second, while cell counts correlated with
NO3− levels at all locations, NO3− levels only varied with
DGGE pattern dissimilarity at the infiltration zone. Third,
seasonally unique DGGE bands common to all locations
yielded greater sequence heterogeneity from the infiltration
zone as compared to the other two locations. One band
common to all locations and all seasons (except early fall)
returned sequences most similar to Arthrobacter globiformis at all locations. This same band also returned Acidobacteriaceae, Rhizobiales, Acidovorax, and Polaromonas
like sequences from the infiltration samples, while two
unique summer infiltration bands yielded close matches to
the Undibacterium, Rhodobacter, Hyphomicrobium, and
Clostridium genera. Seasonal hyporheic nutrient inputs are
affected by river stage and temporal changes in hydrologic
flux within the floodplain [17, 18, 50, 51], which in turn
influence microbial community structure [14]. Community
dynamics may be more exaggerated in the infiltration zone
where nutrient levels and redox potentials are the highest,
and pulses of resources and bacterioplankton from surface
waters are more likely to enter the hyporheic zone.
β-Proteobacterial Oxalobacteraceae and α-Proteobacteria delineated the advection and exfiltration zone communities respectively (Table 1). Species in the family
Oxalobacteraceae are metabolically diverse; however, our
Table 1 Sequence matches providing statistical support to unique microbial communities for MR hyporheic flow path locations
Location
Closest match
Percent identity
GenBank accession number
Infiltration
Clostridium perfringens; 13 (Firmicutes)
Clostridium perfringens; ATCC 13124 (Firmicutes)
Curtobacterium flaccumfaciens; P 259/26 (Actinomycetes)
Curtobacterium herbarum (T); P 420/07 (Actinomycetes)
uncultured Holophaga sp.; JG37-AG-31 (Acidobacteria)
uncultured Gp6 (Acidobacteria)
uncultured unclassified Planctomycete; DEL75 (Planctomycetes)
uncultured Planctomyces; ANTLV2_C10 (Planctomycetes)
Janthinobacterium sp. An8
Janthinobacterium lividum
99.7
99.7
99.5
99.5
99.5
94.5
96.6
98.2
99.2–99.5
100
AB045289
AB075767
AJ310414
AJ310413
Advection
Exfiltration
Massilia cf timonae
uncultured Massilia sp.
uncultured Herbaspirullum sp.
Herbaspirillum rhizosphaerae UMS-40
uncultured Betaproteobacterium
uncultured bacterium P3OB-60
uncultured Betaproteobacterium JG36-GS-10
Hyphomicrobium sp. PMC
Hyphomicrobium sp. LAT3
uncultured Hyphomicrobium sp.
uncultured forest soil bacterium (Rhoplanes sp.)
Rhizobium radiobacter
Brevundimonas sp. Tibet-IX23
uncultured Alphaproteobacterium; Gitt-KF-194 (unclassified Rhizobiales)
uncultured sludge bacterium H6
uncultured Alphaproteobacterium; GuBH2-AD-10
uncultured Alphaproteobacterium; JG37-AG-107
94.2
95.6
99.5
99.0
97.1
97.9
99.0
97.6
97.6
99.2
93.7
98.9
98.8
98.2
94.8
98.9
99.0
AKYG420
AJ616291
DQ521510
AJ551147
Y08846
AY157762
DQ537532
AY662009
DQ188985
AY921721
AF414584
AJ582037
AF279787
AY934489
DQ336968
AY913592
AY851692
DQ177489
AJ532683
AF234689
AJ519651
AJ518775
Sequences from the advection zone all belong to the Betaproteobacterial Oxalobacteraceae and those from the exfiltration zone belong to the
Alphaproteobacteria. Infiltration zone group associations are indicated in parentheses.
Hyporheic Flow Path Microbial Communities
closest matches were to strict aerobes in the genera
Janthinobacterium and Massilia, known to utilize a wide
array of carbon sources like organic acids, mono- and
disaccharides, and some amino acids. It is theorized that the
generally porous nature of alluvial aquifers typically
supports more heterotrophic aerobic respiration [10] and
that continuous delivery of carbon to the microbial
biofilms, presumably due to high flow rates, provides
carbon levels capable of supporting high levels of aerobic
microbial growth in spite of uniformly low DOC levels
(<2 mg L−1) that were measured across the flow path. It is
likely that variations in DOC quality affect the types of
organisms present in the hyporheic zone [17, 18] and
allochthonous DOC sources, like stream benthic algae
metabolites that are readily assimilated by hyporheic
microbes [17], are surely important for supporting aerobic
respiration.
The α-proteobacterial lineage from the exfiltration zone
was most closely related to genera within the Rhizobiales
order, including Hyphomicrobium, Rhodoplanes, Rhizobium, and Brevundimonas (Table 1). These genera have
previously been identified in the hyporheic zones of other
freestone rivers in northwestern MT [16] and may have
important implications for microbial communities living in
low DO environments like that of the MR exfiltration zone.
Prior studies have indicated that productivity bursts (e.g., as
manifested in algal blooms) are often seen where water
upwells from the subsurface, indicating that elevated levels
of inorganic N and P are returned to the open channel as a
result of high rates of hyporheic metabolism [17]. In low
DO environments, denitrifiers, like certain Hyphomicrobium species, may help to maintain the observed
productivity associated with hyporheic exfiltration zones.
In addition to surface water inputs, the high abundance
of plant-associated Herbaspirillum and Rhizobium like
sequences suggests that riparian vegetation and associated
roots may provide significant vertical inputs to the
hyporheic microbial community. Microorganisms from
these genera are known for their ability to fix N2 in
symbiosis with plants [13, 41] and may have important
roles in nitrogen inputs into the hyporheic zone.
Our findings are consistent with other analyses of lotic
freshwater systems, with most sequences closely related to
Gram-negative heterotrophic bacteria known to inhabit
aquatic environments. This study is the first to indicate
that hyporheic microbial communities exhibit significant β
diversity between nutritionally heterogenous flow path
locations, suggesting that habitat heterogeneity in the form
of spatial chemical and nutrient differences is a driver of
microbial community composition in this system.
Acknowledgements Funding was provided by the National Science
Foundation Microbial Observatory Program (grant MCB-0348773).
We thank Scott Relyea, Phil Matson, and the laboratory staff from the
Flathead Lake Biological Station for technical support. Dan Mummey
from the University of Montana provided insightful discussion,
comments, and analytical support. We also thank the John Dalimata
family for access to the study site.
References
1. Allan JD (1995) Stream ecology: structure and function of
running waters. Kluwer, San Diego, USA
2. Anderson-Glenna MJ, Bakkestuen V, Clipson NJ (2008) Spatial
and temporal variability in epilithic biofilm bacterial communities
along an upland river gradient. FEMS Microbiol Ecol 64:407–418
3. Battin TJ, Kaplan LA, Denis Newbold J, Hansen CM (2003)
Contributions of microbial biofilms to ecosystem processes in
stream mesocosms. Nature 426:439–442
4. Battin TJ, Kaplan LA, Newbold JD, Cheng X, Hansen C (2003)
Effects of current velocity on the nascent architecture of stream
microbial biofilms. Appl Environ Microbiol 69:5443–5452
5. Battin TJ, Wille A, Sattler B, Psenner R (2001) Phylogenetic and
functional heterogeneity of sediment biofilms along environmental
gradients in a glacial stream. Appl Environ Microbiol 67:799–807
6. Besemer K, Singer G, Limberger R, Chlup AK, Hochedlinger G,
Hodl I, Baranyi C, Battin TJ (2007) Biophysical controls on
community succession in stream biofilms. Appl Environ Microbiol 73:4966–4974
7. Boulton AJ, Findlay S, Marmonier P, Stanley EH, Valett MH
(1998) The functional significance of the hyporheic zone in
streams and rivers. Annu Rev Ecol Syst 29:59–81
8. Bray RJ, Curtis JT (1957) An ordination of the upland forest
communities of southern Wisconsin. Ecol Monogr 27:325–349
9. Brummer IHM, Felske A, Wagner-Dobler I (2003) Diversity and
seasonal variability of beta-proteobacteria in biofilms of polluted
rivers: analysis by temperature gradient gel electrophoresis and
cloning. Appl Environ Microbiol 69:4463–4473
10. Craft JA, Stanford JA, Pusch M (2002) Microbial respiration
within a floodplain aquifer of a large gravel-bed river. Freshw Biol
47:251–261
11. DeSantis TZ Jr, Hugenholtz P, Keller K, Brodie EL, Larsen N,
Piceno YM, Phan R, Andersen GL (2006) NAST: a multiple
sequence alignment server for comparative analysis of 16S rRNA
genes. Nucleic Acids Res 34:W394–W399
12. dos Santos Furtado AL, Casper P (2000) Different methods for
extracting bacteria from freshwater sediment and a simple method
to measure bacterial production in sediment samples. J Microbiol
Methods 41:249–257
13. Elbeltagy A, Nishioka K, Sato T, Suzuki H, Ye B, Hamada T,
Isawa T, Mitsui H, Minamisawa K (2001) Endophytic colonization
and in planta nitrogen fixation by a Herbaspirillum sp. isolated from
wild rice species. Appl Environ Microbiol 67:5285–5293
14. Ellis BK, Stanford JA, Ward JV (1998) Microbial assemblages
and production in alluvial aquifers of the Flathead River,
Montana, USA. J North Am Benthol Soc 17:382–402
15. Feris KP, Ramsey PW, Frazar C, Rillig M, Moore JN, Gannon JE,
Holben WE (2004) Seasonal dynamics of shallow-hyporheic-zone
microbial community structure along a heavy-metal contamination
gradient. Appl Environ Microbiol 70:2323–2331
16. Feris KP, Ramsey PW, Frazar C, Rillig MC, Gannon JE, Holben
WE (2003) Structure and seasonal dynamics of hyporheic zone
microbial communities in free-stone rivers of the western United
States. Microb Ecol 46:200–215
17. Findlay S (1995) Importance of surface-subsurface exchange in
stream ecosystems—the hyporheic zone. Limnol Oceanogr
40:159–164
J. L. Lowell et al.
18. Findlay S, Sobczak WV (1996) Variability in removal of
dissolved organic carbon in hyporheic sediments. J North Am
Benthol Soc 15:35–41
19. Fischer H, Pusch M, Schwoerbel J (1996) Spatial distribution and
respiration of bacteria in stream-bed sediments. Arch Hydrobiol
137:281–300
20. Fisher MM, Klug JL, Lauster G, Newton M, Triplett EW (2000)
Effects of resources and trophic interactions on freshwater
bacterioplankton diversity. Microb Ecol 40:125–138
21. Grimm NB, Fisher SG (1984) Exchange between interstitial and
surface water: implications for stream metabolism and nutrient
cycling. Hydrobiologia 111:219–228
22. Harner MJ, Stanford JA (2003) Differences in cottonwood growth
between a losing and a gaining reach of an alluvial floodplain.
Ecology 84:1453–1458
23. Hendricks SP (1993) Microbial ecology of the hyporheic zone: a
perspective integrating hydrology and biology. J North Am
Benthol Soc 12:70–78
24. Holben WE (2005) Hyporheic zone microbial communities:
examination of microbial community structure and function in a
Pristine Freestone River in the Northwest United States. Bacterial
Genetics and Ecology 8th Symposium (BAGECO-8), Lyon, France
25. Holben WE, Williams P, Saarinen M, Sarkilahti LK, Apajalahti
JHA (2002) Phylogenetic analysis of intestinal microflora indicates a novel Mycoplasma phylotype in farmed and wild salmon.
Microb Ecol 44:175–185
26. Horner-Devine MC, Bohannan BJ (2006) Phylogenetic clustering
and overdispersion in bacterial communities. Ecology 87:S100–
S108
27. Hullar MAJ, Kaplan LA, Stahl DA (2006) Recurring seasonal
dynamics of microbial communities in stream habitats. Appl
Environ Microbiol 72:713–722
28. Lane DJ, Pace B, Olsen GJ, Stahl DA, Sogin ML, Pace NR (1985)
Rapid determination of 16S ribosomal RNA sequences for
phylogenetic analyses. Proc Natl Acad Sci U S A 82:6955–6959
29. Legendre P, Anderson MJ (1999) Distance-based redundancy
analysis: testing multi-species responses in multi-factorial ecological experiments. Ecol Monogr 69:1–24
30. Liu Z, Lozupone C, Hamady M, Bushman FD, Knight R (2007)
Short pyrosequencing reads suffice for accurate microbial community analysis. Nucleic Acids Res 35:e120
31. Lozupone C, Hamady M, Knight R (2006) UniFrac-an online tool
for comparing microbial community diversity in a phylogenetic
context. BMC Bioinformatics 7:371
32. Lozupone CA, Hamady M, Kelley ST, Knight R (2007)
Quantitative and qualitative beta diversity measures lead to
different insights into factors that structure microbial communities. Appl Environ Microbiol 73:1576–1585
33. Ludwig W, Strunk O, Westram R, Richter L, Meier H,
Yadhukumar BA, Lai T, Steppi S, Jobb G, Forster W, Brettske
I, Gerber S, Ginhart AW, Gross O, Grumann S, Hermann S, Jost
R, Konig A, Liss T, Lussmann R, May M, Nonhoff B, Reichel B,
Strehlow R, Stamatakis A, Stuckmann N, Vilbig A, Lenke M,
Ludwig T, Bode A, Schleifer KH (2004) ARB: a software
environment for sequence data. Nucleic Acids Res 32:1363–1371
34. Lyautey E, Jackson CR, Cayrou J, Rols JL, Garabetian F (2005)
Bacterial community succession in natural river biofilm assemblages. Microb Ecol 50:589–601
35. Manz W, Wendt-Potthoff K, Neu TR, Szewzyk U, Lawrence JR
(1999) Phylogenetic composition, spatial structure, and dynamics
of lotic bacterial biofilms investigated by fluorescent in situ
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
hybridization and confocal laser scanning microscopy. Microb
Ecol 37:225–237
Martin AP (2002) Phylogenetic approaches for describing and
comparing the diversity of microbial communities. Appl Environ
Microbiol 68:3673–3682
Morales SE, Cosart TF, Johnson JV, Holben WE (2009) Extensive
phylogenetic analysis of a soil bacterial community illustrates
extreme taxon evenness and the effects of amplicon length, degree
of coverage, and DNA fractionation on classification and
ecological parameters. Appl Environ Microbiol 75:668–675
Naegeli MW, Uehlinger U (1997) Contribution of the hyporheic
zone to ecosystem metabolism in a prealpine gravel-bed river. J
North Am Benthol Soc 16:794–804
Poole GC, Stanford JA, Frissell CA, Running SW (2002) Threedimensional mapping of geomorphic controls on flood-plain
hydrology and connectivity from aerial photos. Geomorphology
48:329–347
Porter KG, Feig YG (1980) The use of DAPI for identifying and
counting aquatic bacteria. Limnol Oceanogr 25:943–948
Pueppke SG, Broughton WJ (1999) Rhizobium sp. strain NGR234
and R. fredii USDA257 share exceptionally broad, nested host
ranges. Mol Plant–Microb Interact 12:293–318
Pusch M, Fiebig D, Brettar I, Eisenmann H, Ellis BK, Kaplan LA,
Lock MA, Naegeli MW, Traunspurger W (1998) The role of
micro-organisms in the ecological connectivity of running waters.
Freshw Biol 40:453–495
Sekiguchi H, Tomioka N, Nakahara T, Uchiyama H (2001) A
single band does not always represent single bacterial strains in
denaturing gradient gel electrophoresis analysis. Biotechnol Lett
23:1205–1208
Shade A, Jones SE, McMahon KD (2008) The influence of habitat
heterogeneity on freshwater bacterial community composition and
dynamics. Environ Microbiol 10:1057–1067
Stanford JA, Ward JV (1993) An ecosystem perspective of
alluvial rivers: connectivity and the hyporheic corridor. J North
Am Benthol Soc 12:48–60
Stanford JA, Ward JV, Ellis BK (1994) Ecology of the alluvial
aquifers of the Flathead River, Montana. In: Gibert J, Danielopol
DL, Stanford JA (eds) Groundwater ecology. Academic, San
Diego, pp 367–390
Velji MI, Albright LJ (1985) Microscopic enumeration of attached
marine bacteria of seawater, marine sediment, fecal matter, and
kelp blade samples following pyrophosphate and ultrasound
treatments. Can J Microbiol 32:121–126
Ward JV (1997) An expansive perspective of riverine landscapes:
pattern and process across scales. Gaia 6:52–60
Ward JV, Tockner K (2001) Biodiversity: towards a unifying
theme for river ecology. Freshw Biol 46:807–819
Wondzell SM, Swanson FJ (1996) Seasonal and storm dynamics
of the hyporheic zone of a 4th-order mountain stream. 1.
Hydrologic processes. J North Am Benthol Soc 15:3–19
Wondzell SM, Swanson FJ (1996) Seasonal and storm dynamics
of the hyporheic zone of a 4th-order mountain stream. 2. Nitrogen
cycling. J North Am Benthol Soc 15:20–34
Yu W, Dodds WK, Banks MK, Skalsky J, Strauss EA (1995)
Optimal staining and sample storage time for direct microscopic
enumeration of total and active bacteria in soil with two
fluorescent dyes. Appl Environ Microbiol 61:3367–3372
Yu Z, Mohn WW (1999) Killing two birds with one stone:
simultaneous extraction of DNA and RNA from activated sludge
biomass. Can J Microbiol 45:269–272