Environmental Microbiology (2003)

5

(4), 256266
2003 Society for Applied Microbiology and Blackwell Publishing Ltd

Blackwell Science, LtdOxford, UKEMIEnvironmental Microbiology 1462-2912Blackwell Publishing Ltd, 20035Original Article

Microbial community dynamics in stream sedimentsS. D. Sutton and R. H. Findlay

Received 13 June, 2002; revised 4 October, 2002; accepted 18
October, 2002. *For correspondence. E-mail ndlarh@muohio.edu;
Tel. (

+

1) 513 529 5422; Fax (

+

1) 513 529 2431.

Present address:
Eli Lilly and Company, Lilly Research Laboratories DC6065,
Indianapolis, IN 46285, USA.

Sedimentary microbial community dynamics in
a regulated stream: East Fork of the Little
Miami River, Ohio

Susan D. Sutton

and Robert H. Findlay*

Department of Microbiology, Miami University, Oxford,
Ohio 45056, USA.

Summary
A eld study was conducted in the Lower East Fork
of the Little Miami River, a regulated stream in
Clermont county, Ohio, to determine how changes
in streamow, water temperature and photo-period
affect sediment microbial community structure. Sur-
face sediment cores were collected from sampling
stations spanning 60 river kilometers three to four
times per year between October 1996 and October
1999. During the nal year of the eld study, water
temperature, water depth, conductivity, total sus-
pended solids, dissolved organic carbon, instanta-
neous streamow velocity, sediment grain size and
sediment organic matter were determined. Total
microbial biomass was measured using the phospho-
lipid phosphate technique (PLP) and ranged between
2 and 134 nmol PLP g
--- -

1

dry weight sediment with a
mean of 25 nmol PLP g
--- -

1

. Microbial community
structure was determined using the phospholipid
fatty acid analysis and indicated seasonal shifts in
sedimentary microbial community composition.
January to June sedimentary microbial biomass
was predominately prokaryotic (60%



2), whereas
microeukaryotes dominated samples collected during
the late summer (55%



2.4) and fall (60%



2). These
changes were correlated with stream discharge and
water temperature. Microbial community structure
varied spatially about a reservoir with prokaryotic bio-
mass dominant at upstream stations and eukaryotic
biomass dominant at downstream stations. These
ndings reveal that sedimentary microbial communi-
ties in streams are dynamic responding to the sea-
sonal variation of environmental factors.
Introduction

Environmental factors such as climate, season, watershed
hydrology, geomorphology and land use are important
effectors of stream processes. Other abiotic factors such
as streamow velocity and water temperature can vary
dramatically throughout the year and inuence the sea-
sonal dynamics of both biotic and abiotic parameters.
Historically, investigators have focused on how these envi-
ronmental factors impact on sh, macroinvertebrates and
phototrophic communities in streams, and attempted to
use these organisms as bioindicators of stream health
(Ford, 1989; Biggs, 1996; 2000; Dyer

et al

., 1998).
Dissolved organic material is a primary biologic sub-
strate for stream metabolism (Kaplan and Newbold,
1993). Heterotrophic bacteria consume this resource
directly and they are therefore important mediators of
stream energy ow. Recently, researchers have begun to
investigate determinants of prokaryotic community com-
position and physiologic status in streams and rivers
(Guckert

et al

., 1992; Lemke and Leff, 1999; Brmmer

et al

., 2000) but most studies have focused on the effects
of pollution at small spatial and temporal scales. Informa-
tion on the long-term seasonal dynamics of microbial
communities that describes both prokaryote and microeu-
karyote components is lacking and, consequently, the
microbial ecology of lotic systems is poorly understood
(Leff, 1994).
Our objectives were to determine natural patterns of
seasonal variation in sedimentary microbial biomass and
community structure and evaluate the potential impact of
regulated ow in a mid-order stream in south-west Ohio.
During a three-year eld study (October 1996October
1999), we used phospholipid phosphate (PLP) and phos-
pholipid fatty acid (PLFA) analysis to quantify sedimentary
microbial biomass and identify patterns of variation in
community structure at six stations along a 60 km section
of the East Fork Little Miami River. This river has been
designated an exceptional warm-water habitat by the Ohio
Environmental Protection Agency; it is well oxygenated
and contains moderate levels (for Ohio) of dissolved nutri-
ents (Table 1). Three of the six stations were at distant
locations (greater than 5 km) downstream of an articial
reservoir (Harsha Lake) where mainstem streamow was

Microbial community dynamics in stream sediments

257

2003 Society for Applied Microbiology and Blackwell Publishing Ltd,

Environmental Microbiology

,

5

, 256266

regulated. In addition to seasonal dynamics, we deter-
mined the relative contribution of prokaryotes and
microeukaryotes to total viable biomass. During the later
sampling dates (September 1998 until October 1999) we
measured water temperature, water depth, conductivity,
instantaneous streamow velocity, total suspended solids,
dissolved organic carbon, sediment particle size and sed-
iment organic matter. Historical stream discharge data
were obtained from the U.S. Geological Survey. These
environmental parameters were investigated as proximal
determinants of the observed changes in sedimentary
microbial biomass and community composition.

Results

Determinants of microbial distribution and abundance

The above environmental parameters (eight measured
parameters plus mean 7-day discharge) were evaluated
for seasonal trends. Water temperature, conductivity, total
suspended solids and mean 7-day discharge showed sig-
nicant changes with sampling date. We grouped sam-
pling dates,

a posteriori

, into four seasonal groups based
on unique combinations of water temperature and mean
7-day discharge (Table 2). Subsequently, mean 7-day dis-
charge is referred to as streamow and the a

posteriori

groupings are referred to as seasonal sampling intervals.
Data for each measured environmental determinant
were grouped according to season and position (upstream
or downstream of the reservoir) and tested for signicant
differences using two-way

ANOVA

. Note that positional dif-
ferences for streamow could not be evaluated because
there is only one active gaging station (located in the
downstream reaches) located in the study area. Only
instantaneous streamow velocity and total suspended
solids showed signicant above/below reservoir effects.
During low and moderate ow months instantaneous
streamow velocity was 2599% higher at downstream
stations. During four of the ve sampling periods when
total suspended solids were measured, total suspended
solids were greater above the reservoir than below indi-
cating that the reservoir may improve water clarity. Envi-
ronmental determinant data (O

2

, total suspended solids,
nitrate/nitrite, total Kjeldahl nitrogen, dissolved ortho-
phosphorus and total phosphorus), available from a
Clermont County Ofce of Environmental Quality spon-
sored study of water quality in East Fork (Tetra Tech,
2001), were screened for above/below reservoir effects.
Comparisons of annual means at Stations EFRM 34.8
and 44.1 (upstream) and EFRM 12.7 and 4.0 (down-
stream) indicated that only total suspended solids varied
signicantly above and below the reservoir (suspended
solids were twofold greater above the reservoir). Nutrient
levels on the mainstem of EFLMR below the reservoir
increased with distance from the reservoir and encom-
passed the values reported for upstream stations.

Total microbial biomass

Total viable microbial biomass in East Fork sediments
ranged between 2 and 134 nmol PLP g

-

1

dry weight
(d.w.) with a mean of 25 nmol PLP g

-

1

d.w. A one-way

ANOVA

indicated that natural log biomass changed signif-
icantly with season (

P



<

0.001). Sedimentary microbial
biomass was signicantly greater in samples collected
between October and December than during any other
sampling interval (Table 2, Tukey HSD,

a =

0.05). Mean
biomass levels did not vary signicantly among samples
collected between January and September. These values

Table 1.

Nutrient levels at most downstream and upstream monitoring stations of the East Fork Little Miami River.

a

Station Dissolved oxygen (mg l

1

) Nitrite/nitrate (mg l

1

) Dissolved ortho-phosphate (mg l

1

)
EFRM0.5 199600 87 8.70

2.40

b

199800 47 1.79

0.79 199600 78 0.38

0.19
EFRM44.1
199700 67 7.31

1.14
199800 46 0.84

1.07 199700 67 0.87

0.06

a.

From Clermont County, Ohio, monitoring programme to characterize the surface water quality of the East Fork of the Little Miami River (Tetra
Tech, 2001).

b.

Sample period,

n

, mean

S.D.

Table 2.

Seasonal changes in water temperature, steamow velocity and discharge and microbial biomass in the East Fork.
Seasonal sampling interval
Water temperature
(

C)
Instantaneous velocity
(ms

-

1

)
Mean 7-day discharge

a

(m

3

s

-

1

)
Microbial biomass
(PLP g

-

1

)
JanApril 7.3

b

0.39 36 17
MayJune 23 0.11 45 25
JulySept. 20 0.12 1.8 20
OctDec. 9 0.11 2.2 34

a.

Mean 7-day discharge prior to sampling determined from measurements collected at the USGS Perrintown gaging station.

b.

Mean value for all samplings occurring during the designated seasonal sampling interval between the years 1996 and 1999.

258

S. D. Sutton and R. H. Findlay

2003 Society for Applied Microbiology and Blackwell Publishing Ltd,

Environmental Microbiology

,

5

, 256266

are comparable to stream sediments studied by Bott and
Kaplan (1985) and, for the most part, fall within the total
range of values reported for freshwater sediments (7
464 nmol PLP g

-

1

d.w., Dobbs and Findlay, 1993).

Multivariate analysis of microbial community structure

We used principal component analysis (PCA) to examine
patterns of variation among PLFA proles of sediments
collected throughout the three-year eld study (177 sam-
ples). Scatter plots of sample factor scores revealed gen-
eral patterns of seasonal variation in community structure
for PC1, PC3 and to a lesser extent for PC2 (Fig. 1).
Samples with positive PC1 scores (Fig. 1) were enriched
in PLFA with high PC1 loadings (

>

0.5). These fatty acids
(10me16:0, a17:0, br17:1a, 17:0, 17:1

w

6/cy17:0, cy19:0,
18:1

w

7c, 18:2a, i17:0) are typically present in the mem-
branes of heterotrophic bacteria. Conversely, samples with
negative PC1 scores were enriched in the polyenoic PLFA
that had component loadings

<-

0.5. With few exceptions,
the latter group is restricted to microeukaryote membranes
and is typical of phototrophic microeukaryotes (Fig. 1).
Comparison of PC1 scores by seasonal sampling inter-
val conrmed that the major component of variation in
community structure was seasonal and reected changes
in the balance between prokaryotic and eukaryotic biom-
ass (Fig., 2, Tukeys HSD,

a



=

0.05). Seventy-nine per
cent of the samples collected between January and June
had positive PC1 scores indicating that prokaryotes were
disproportionately abundant in these samples compared
to samples with highly negative PC1 scores. Seventy-two
per cent of samples collected between July and Decem-
ber had negative PC1 scores indicating that these sam-
ples were more enriched in microeukaryotes compared to
samples with highly positive PC1 scores.
Although patterns of seasonal variation on PC2 were
less pronounced between January and September, there
was a signicant shift in community composition between
the July to September and October to December sampling
intervals (Fig. 2, panel 2). Phospholipid fatty acid with 18
carbons in the aliphatic chain (18:0, 18:1

w

9c and 18:3

w

6)
were enriched in samples with highly negative PC2 scores
(Fig. 1). PLFA with high PC2 loadings (16:1

w

7c, 16:1

w

9c,
i15:0, a15:0, 16:1

w

5c and 14:0) were enriched in samples
with highly positive PC2 scores (Fig. 1). Patterns of vari-
ation in community structure along PC2 were not related
to sample station position nor were they correlated with
water temperature.
Ninety-three per cent of samples collected between
January and April had negative PC3 scores while 81%
of samples collected between May and September had
positive PC3 scores (Fig. 2). PLFA with high PC3 load-
ings (saturated 14-, 15- and 16-carbon fatty acids; Fig. 1)
were generally more predominant in samples collected
between May and September than in samples collected
between January and April. PLFA with low PC3 loadings
included 18:1

w

7c and an unidentied dienoic 18-carbon
fatty acid and were enriched in samples with highly neg-
ative PC3 scores (predominantly JanuaryApril samples).
Sediment samples collected during the October to
December sampling interval had intermediate PC3
scores. This pattern of variation was moderately positively
correlated with water temperature (

R



=

0.685).

Fig. 1.

Patterns of seasonal variation in sedimentary microbial com-
munity composition in the East Fork.
A. Scatter plot of sample scores for principal components 1 (abscissa)
and 2 (ordinate).
B. Scatter plot of sample scores for principal components 1 (abscissa)
and 3 (ordinate). PC1 explained 32.3% of the variance and compo-
nents two and three explained an additional 17.9 and 17.4% respec-
tively. Solid squares represent samples collected between January and
April, circles represent samples collected in May and June, triangles
represent samples collected between July and September and
open diamonds represent samples collected between October and
December. Identied fatty acids had component loadings

>

| 0.5 | and
exerted strong inuence on the pattern of variation among samples
along the respective component axes.
A
B
4 3 2 1 0 1 2 3
4
3
2
1
0
1
2
3
4
4
3 2 1 0 1 2 3
3
2
1
0
1
2
3

Microbial community dynamics in stream sediments

259

2003 Society for Applied Microbiology and Blackwell Publishing Ltd,

Environmental Microbiology

,

5

, 256266

Changes in microbial community structure by season

During January to June sedimentary microbial biomass
was predominately prokaryotic (60%, Fig. 3A), while
microeukaryote biomass was dominant in samples col-
lected between July and December. Total microeukaryote
biomass more than doubled between JulySeptember
and OctoberDecember while there was no signicant
seasonal variation in total prokaryote biomass (Fig. 3B).
Unique community shifts occurred between January
April and MayJune samples. Although samples collected
between January and April were predominantly prokary-
otic, 20:5

w

3/20:4

w

6 ratios were relatively high suggesting
a seasonal peak in diatom abundance (Fig. 4, panel 3).
MayJune and JulySeptember samples showed a sig-
nicant decrease in 20:5

w

3/20:4

w

6 ratios indicating a
decreased importance of diatoms during this period.
Conversely, the microeukaryote component as a whole
showed a proportional increase during these warm-water
months (Fig. 4 panel 3 versus Fig. 3A).
Many samples collected during the MayJune sampling
interval were enriched in saturated PLFA. These included
16:0 and 14:0, which are broadly distributed among many
types of microorganism, as well as two fatty acids that are
broadly distributed among prokaryotes (15:0 and 17:0).
Other inuential PLFA included i15:0 and i16:0. Together,
this group of PLFA represents Gram-positive prokaryotes
and facultative Gram-negative bacteria (Kaneda, 1977;
Parkes and Taylor, 1983; Edlund

et al

., 1985; Guckert

et al

., 1985). These data suggest that during warm-water
months there was a persistence (or perhaps emergence)

Fig. 2.

Seasonal patterns of variation in sedimentary microbial com-
munity structure determined by PLFA analysis. Mean factor scores for
samples grouped according to seasonal sampling interval. The top
panel is a summary for PC 1, the middle panel for PC 2 and the bottom
panel for PC3. Vertical error bars represent the 95% condence inter-
val around the means. Letter designations at the top of each panel
correspond to the bar below and indicate homogenous subsets iden-
tied by Tukeys HSD (

a



=

0.05). Bars identied by different letters are
signicantly different, whereas those with the same letter are not.
1.5
1.0
0.5
1.0
0.5
1.0
1.5
Jan.Apr. MayJune JulySept. Oct.Dec.

Fig. 3.

Seasonal variation in microeukaryote and prokaryote biom-
ass in East Fork sediments. Samples are grouped according to sea-
sonal sampling interval. In A, bars represent the mean percentage
total biomass (as g Carbon). In B, bars represent the mean microbial
biomass (calculated as mg Cg

-

1

d.w.) of the two phylogenetic groups
in each sampling interval. In A and B, open bars represent the total
microeukaryote biomass and solid bars represent the prokaryotic
component. Vertical error bars represent the 95% condence interval
around each mean. Carbon biomass was estimated from PLP and
PLFA as described in Dobbs and Findlay (1993).

260

S. D. Sutton and R. H. Findlay

2003 Society for Applied Microbiology and Blackwell Publishing Ltd,

Environmental Microbiology

,

5

, 256266

of a specic group of prokaryotes that co-exist with newly
developing microeukaryote populations.
Samples collected during warm-water months were
enriched with 18-carbon PLFA (18:0, 18:1

w

9c and
18:3

w

6). While 18:0 is broadly distributed among micro-
organisms, 18:1

w

9c has been detected in the membranes
of bacteria, marine microfauna and microeukaryotic pho-
toautotrophs (Dobbs and Findlay, 1993; Findlay

et al

.,
1990a). The last of this group, 18:3

w

6, is part of the

w

6 animal series of PLFA (Dobbs and Guckert, 1988;
Findlay and Dobbs, 1993) and has been detected in
fungal membranes (White

et al

., 1997). Given the
increasing importance of microeukaryotes during warm-
water sampling intervals and

w

3/

w

6 ratios at approxi-
mately 2 (Fig. 4, panel 4) we concluded that the developing
microeukaryote populations were largely heterotrophic.
Cyclopropyl physiologic status ratios (cy19:0/18:1

w

7c)
varied signicantly with season. Sediment samples col-
lected between May and September had signicantly
higher mean cy19:0/18:1

w

7c ratios than samples
collected between January and April or October and
December (Fig. 4, panel 2).

Trans/cis

ratios showed sim-
ilar patterns of seasonal variation (Fig. 4, panel 1) with
relatively low values detected between January and April
and October through December and higher values
between May and September. These increases coincide
with higher water temperature. Twenty-nine of 177 sedi-
ment samples had 16:1

w

7t/16:1

w

7c ratios greater than
0.1 suggesting that these communities were under stress
or starvation conditions (White

et al

., 1997). Most of these
sediment samples (23 of 29) were collected between June
and September.
Yearly maximum values for microeukaryote biomass
observed during the October December sampling inter-
val coincided with signicant increases in the

w

3/

w

6 and
20:5

w

3/20:4

w

6 ratios. These increases were indicative of
an increase in the importance of phototrophic microeu-
karyote biomass (Fig. 4, panels 3 and 4).

Effects of regulated ow

An underlying source of variation in sedimentary microbial
community structure was the enrichment of microeukary-
otes at stations downstream of Harsha Lake. A two-way

ANOVA

of PC1 scores with season and relative position to
reservoir as factors indicated signicant position, season
and season

position interaction effects (

P



<

0.001,

P



<

0.001 and

P



=

0.005). The season effect was related
to ow and temperature changes as discussed above. The
position effect was indicated when we observed that sed-
iment samples collected upstream of Harsha Lake had
signicantly higher PC1 scores than downstream sam-
ples, suggesting that the reservoir inuenced microbial
community structure in this river system. The interaction
effect was due to one signicant deviation from parallelism
that occurred during June 1997 when streamow prior to
sampling was highest. On this sampling date, downstream
samples had PC1 factor scores greater than upstream
samples (Fig. 5). The position effect was independent of
season, with the one exception noted above. On average,
prokaryotic biomass represented 56% of total microbial
biomass at stations upstream of Harsha Lake whereas
prokaryotes contributed only 45% of the total microbial
biomass at downstream stations. PLFA proles from
downstream samples were enriched in fatty acids indica-
tive of phototrophic microeukaryotes.

Fig. 4.

Seasonal patterns of community structure and physiologic
stress ratios. Samples are grouped according to seasonal sampling
interval. The top two panels show changing patterns in w3 to w6-
phospholipid fatty acid ratios, which indicate the general importance
of phototrophic prokaryotic physiologic stress markers. The lower two
panels indicate changing patterns in microeukaryotes. Bars represent
the means and vertical error bars indicate the 95% condence inter-
vals. Letter designations at the top of each panel correspond to the
bar below and indicate homogenous subsets identied by Tukeys
HSD (a = 0.05).
Jan.Apr. MayJune JulySept. Oct.Dec.
Microbial community dynamics in stream sediments 261
2003 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 5, 256266
Discussion
Effects of regulated ow
Regulated ow can lead to diminished seasonal and diur-
nal temperature uctuations, elevation of winter water
temperatures and a depression of summer water temper-
ature (Winget, 1984). Streamow in regulated rivers devi-
ates from the natural hydrologic cycle with increased ow
stability during drier summer months, increased ows in
the fall associated with the draw-down of lake levels prior
to the onset of winter rains and periodic ushing ows in
the spring that may be required if spring rains cause the
reservoir to ll. Although other investigators have
observed changes in water quality parameters and corre-
sponding changes in biotic integrity in regulated rivers
(Winget, 1984; Cazaubon and Giudicelli, 1999), these
observations were usually associated with tailwater
regions and typically the effects diminished with increas-
ing distance from intervening dams. We observed a stim-
ulation of microeukaryotes at stations more than 5 km
downstream of Harsha Lake; a response previously not
associated with regulated ow.
The river continuum hypothesis predicts that benthic
periphyton production (and presumably microeukaryotic
production within the sediment community) increases rap-
idly with increasing stream order caused by increasing
light availability due to decreasing canopy cover. Per cent
canopy at the six stations ranged from 20 to 80% with the
greatest coverage at the rst station downstream of the
reservoir. Statistical analysis indicated that there were no
signicant differences between stations above and below
the reservoir. As such, we reject the hypothesis that
increasing light availability due to decreasing canopy
cover caused the above and below reservoir differences
in microbial community structure.
Greater predominance of algal biomass in stream sed-
iments below the reservoir could result from a combina-
tion of increased water clarity (reduced total suspended
solids) and the subsidy effects of moderate increases in
stream ow velocity discussed below. However, two other
factors cannot be eliminated as possible proximal causes
for the observed above/below reservoir pattern. Harsha
Lake may decrease the severity of peak ow events within
the regulated reaches of the river decreasing scour-
induced changes in community structure. Stations down-
stream of Harsha Lake were located in an area of intense
property development and land use patterns may have
contributed to the observed pattern.
Seasonal patterns of microbial community structure
Analysis of microbial community structure indicated that
the primary pattern of variation in East Fork sedimentary
microbial community composition resulted from a shift
between a high biomass community dominated by
microeukaryotes (with a large proportion of phototrophs)
late in the calendar year to a predominantly prokaryotic,
low biomass community present early in the calendar
year. This shift was correlated with seasonal changes in
streamow and water temperature. Samples collected
after very high discharge events in February 1997 and
June 1997 (mean 7-day discharge >50 m
3
s
-1
; subse-
quently referred to as peak ow events) had very low
microbial biomass (mean, 9 nmol PLP g
-1
d.w) which
was predominantly prokaryotic (mean per cent prokaryotic
biomass, 66%). It has been proposed that hydraulic dis-
turbance coinciding with large rain events is the primary
determinant of periphyton biomass and community com-
position among different streams (Tett et al., 1978; Biggs
and Close, 1989; Biggs, 1995). Increases in water velocity
and sheer stress during large rain events may lead to
sloughing of biomass from benthic substrata. Similarly,
sediment movement and suspension during high stream-
Fig. 5. Temporal variation in sedimentary microbial community struc-
ture and stream discharge at stations upstream and downstream of
Harsha Lake.
A. Mean PC1 scores for samples collected upstream (solid squares)
and downstream (open squares) of Harsha Lake.
B. Mean discharge for the 7-day period before sample collection,
monitored at USGS station # 03247500 in Perintown, Ohio. Horizontal
axes indicate day of the calendar year (Julian Day).
262 S. D. Sutton and R. H. Findlay
2003 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 5, 256266
ow can lead to abrasion resulting in a net loss of biom-
ass. Our sampling regime did not allow for the monitoring
of biomass loss and accrual surrounding individual storm
events, but our results show that there was disproportion-
ate loss of microeukaryote biomass between October to
December and January to April. During these intervals,
there was no signicant change in prokaryotic carbon
biomass while total microeukaryote biomass dropped 3.5-
fold. We conclude that this drastic community shift
resulted from repeated disturbances associated with win-
ter and spring storms. This is corroborated by long-term
discharge records (196897) for stream gaging stations
in the Little Miami River basin which indicate that dis-
charge is lowest in September and October and highest
between December and June (Debrewer et al., 2000).
Interestingly, East Fork sedimentary microbial communi-
ties exhibited relatively low cyclopropyl/monoenoic and
trans/cis ratios during interval 1 (Fig. 4, panel 1 and 2)
indicating that at least part of the Gram-negative prokary-
otic community was healthy and actively growing (White
et al., 1997).
Whereas JanuaryApril sediments were dominated by
prokaryotes, the community also included diatoms. The
emergence and success of this group of phototrophic
microeukaryotes during periods of increased streamow
was not surprising since these organisms often grow as
coherent mucilaginous communities and may be more
resistant to sheer stress (Biggs et al., 1998). Similarly,
these organisms (diatoms and phototrophic microeukary-
otes in general) often ourish at low water temperatures
that may inhibit heterotrophic microeukaryotes and inver-
tebrate grazers.
MayJune and JulySeptember showed waning dia-
toms, increased importance of heterotrophic microeukary-
otes, the persistence or emergence of Gram-positive
bacteria or facultative Gram-negative bacteria and
evidence of stress within the Gram-negative bacterial
community. Although we did not quantify grazing activity
in this study, we speculate that temperature-dependent
increases in grazing pressure depressed total microbial
biomass during this period. The increased importance of
Gram-positive bacteria and Gram-negative anaerobic
bacteria during these intervals is consistent with high
grazing pressure (Findlay et al., 1990b; Findlay and
Watling, 1998).
The range of cyclopropyl and trans/cis ratios were com-
parable to those reported for a channelized riverine sys-
tem in central Ohio (Langworthy et al., 2002). The
patterns of seasonal change were similar to those
reported by Smoot and Findlay (2001) as they detected
annual maxima between May and August for both cyclo-
propy/monoenoic and trans/cis ratios. The effect of water
temperature on physiologic status of bacteria in aquatic
sediments is not straightforward and the observed
response is likely related to abiotic and biotic parameters
that vary with temperature. It has been reported that
increasing water temperatures (between 7 and 20C) typ-
ically stimulate bacterial productivity, which requires
increased substrate supply (Shiah and Ducklow, 1994).
Oxygen solubility decreases as water temperatures
increase. The net result may be either oxidative stress (if
substrate supply exceeds oxygen supply) or starvation
conditions (substrate demand exceeds substrate supply),
both of which are known to induce the observed changes
in fatty acid ratios (Guckert et al., 1986; Wilkinson, 1988).
During October December microeukaryotes double
their biomass, phototrophic microeukaryotes returned to
importance and metabolic stress decreased within the
prokaryotic community. There are several possible proxi-
mal causes for these shifts within the microbial commu-
nity. During this period streamow typically increases in
the East Fork (Debrewer et al., 2000) but peak ow events
were absent. Water temperature decreased during this
period, increasing oxygen solubility and possibly decreas-
ing grazing rates. Conductivity also increased during this
period, indicating a likely increase in nutrients (Biggs,
1995; nutrient data were not available for this period).
Biggs et al. (1998) observed a subsidy response by micro-
bial communities to moderate increases in streamow
velocity. Microbial production was stimulated by increas-
ing rates of nutrient delivery and toxic waste product
export. Increased nutrient and oxygen concentrations
would likely have similar stimulatory effects. The absence
of peak ow events and a possible decrease in grazing
rates would allow the accumulation of biomass.
This eld study was conducted over a three-year period
and the seasonal trends developed were based on sam-
ples taken in at least two, and most often three, calendar
years. Together, these data provide a clearer understand-
ing of the microbial community dynamics in East Fork
sediments. There is an annual cycle of biomass loss and
accrual that is paralleled by complex shifts in community
structure. The process was driven predominantly by
changes in streamow and water temperature, which
mediate factors such as abrasion, sloughing, oxygen
solubility and availability and grazer/bacterial metabolic
activity.
Experimental procedures
Site description and sampling scheme
A three-year eld study was conducted in the East Fork of
the Little Miami River (East Fork) drainage basin in south-
west Ohio. The watershed, drains 1292 km
2
and lies primarily
in Clermont County but portions extend into Brown, Clinton,
Highland and Warren counties (Debrewer et al., 2000; for
map of study site, see Tetra Tech, 2001). Our study area
spanned 60 river kilometres along both the East Fork and a
Microbial community dynamics in stream sediments 263
2003 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 5, 256266
small tributary, Pleasant Run Creek. The intervening reser-
voir, William H. Harsha Lake [1800 ha, positioned at East
Fork river mile (EFRM) 20.4], was constructed in 1973 and
is used for ood control, water supply and recreational activ-
ities. Discharge from this reservoir is tightly regulated by the
U.S. Army Corps of Engineers. The drainage area upstream
of Harsha Lake is primarily agricultural and forested. In recent
years there has been intense residential and some urban
development in areas downstream of the reservoir in western
Clermont County. The countys population has nearly doubled
since 1970 and now exceeds 177 000. Six sampling stations
were established along the East Fork (three upstream and
three downstream of reservoir). In general, they were located
in shallow pool and run segments. Stations 1 and 3, located
upstream of Harsha Lake, were established at EFRM 44.1
and 42.8 respectively. Station 2 was established on a tribu-
tary, Pleasant Run Creek, that drains into the mainstem at
EFRM 42.6. Stations 4 and 5 (EFRM 12.7 and 12.3) were
adjacent to the Middle East Fork wastewater treatment facility
downstream of Harsha Lake. One was located upstream
(approximately 150 m) and the other was immediately down-
stream (approximately 65 m) of the wastewater efuent out-
fall. Station 6 was located at EFRM 4.8.
Total sedimentary microbial biomass and microbial com-
munity composition were determined for samples collected
in October 1996; February, June, and August 1997; February,
September, and December 1998; and March, June, and
October 1999. In October 1997 only total microbial biomass
was determined. Water temperature was measured at all
stations in August 1997, February, September and December
1998, and March, June and October 1999. Water depth, total
suspended solids, conductivity, and instantaneous stream-
ow velocity were measured in September and December
1998, and March, June and October 1999. Dissolved organic
carbon, sediment particle size and sediment organic matter
(per cent sediment dry weight) were determined for the last
four sampling periods. Stream discharge data for East Fork,
stations 4, 5 and 6 was obtained from the U.S. Geological
Survey gaging station for all sampling intervals. Total sus-
pended solids, and oxygen, nitrate/nitrite, total Kjeldahl nitro-
gen, dissolved ortho-phosphorus and total phosphorus
concentration data were obtained from Clermont Count
Ambient Water Quality Monitoring Program, East Fork of the
Little Miami River, Five-Year Status and Trends Report (Tetra
Tech, 2001). Data were collected May through October for
the years 19962000, although the data set is not complete
for all stations for all years.
Microbial biomass and community composition.
Triplicate sediment core samples (01 cm horizon, 1.75 cm
3
)
were collected for microbial biomass determinations. Micro-
bial biomass and community structure were determined by
phospholipid analysis as previously described (Findlay, 1996;
Findlay et al., 1989). Sediment cores were added directly to
a modied Bligh and Dyer organic solvent extraction mixture
(Findlay, 1996) consisting of methanol, dichloromethane, and
phosphate buffer (2:1:0.8) in the eld. After 24 h, lipids were
partitioned from water-soluble components after adding
dichloromethane and water to obtain a nal solvent ratio of
1:1:0.9 (methanol:dichloromethane:water/buffer). Total viable
biomass was quantied by measuring phospholipid-bound
phosphate (PLP) according to the procedures described by
Findlay et al. (1989). Sedimentary microbial community
structure was determined from ester-linked phospholipid fatty
acid (PLFA) proles (Findlay, 1996). Briey, the polar lipids
were recovered from the total lipid extract by silicic acid
column chromatography. The PLFA were derivitized to form
fatty acid methyl esters (FAMEs) by mild alkaline methanol-
ysis. Resultant fatty acid methyl esters were puried on solid
phase extraction columns packed with C18 resin and sepa-
rated and quantied by gas chromatography with ame ion-
ization detection. For this study, 31 FAMEs were quantied
and identied. Structural identication was based on relative
retention times, co-elution with standards and verication by
gas chromatography/mass spectrometry. FAME nomencla-
ture was based on Findlay and Dobbs (1993).
Environmental parameters
At each station, water temperature was measured with a
hand-held mercury thermometer and conductivity was mea-
sured with a digital conductivity meter (Fisher Scientic, Pitts-
burgh, PA). Water depth was measured with a tape measure.
Total suspended solids were determined according to stan-
dard procedures (American Public Health Association, 1995)
from triplicate depth-integrated water samples (1 l) collected
at each station, stored on ice and transported to the labora-
tory for processing. Dissolved organic carbon was deter-
mined by Pt-catalysed persulphate oxidation (Kaplan, 1992).
Particulate matter was removed in the eld from triplicate
40 ml samples with 25-mm GF/F glass microbre lters
(Whatman, Ann Arbor, MI) as described by Kaplan (1994).
Sediment scoop samples (approximately the top 2 cm) were
collected at each station and transported to the laboratory for
determination of particle size and sediment organic matter.
Sediments were air-dried for 47 days, plant debris was
removed and samples were sieved to collect standard size
fractions between 2 and 25 mm. Sub-samples of sediment
that passed through a 2-mm sieve (pan fraction) were oven-
dried and grain size distribution was determined using
Stokes Law approximations (Day, 1965). Additional subsam-
ples of the pan fraction (six) were baked at 450C for one
hour to determine per cent weight loss on combustion. This
measurement (reported as per cent of sediment dry weight)
represented an approximation of sediment organic matter.
Instantaneous streamow velocity measurements were
recorded at the time of sampling using a portable water ow
meter (Model 201D Marsh-McBirney, Frederick MD). To esti-
mate streamow history prior to sample collection, we
obtained discharge data for the East Fork from the U.S.
Geological Survey (USGS). The USGS stream gaging station
in Perintown, Ohio (station # 03247500 is located 10.2 river
kilometers from the conuence with the Little Miami River)
and provisional daily mean discharge data were available
online (U.S. Geological Survey, 2000).
Data analysis and statistical methods
Minitab Statistical software (version 12.5 for Windows 95,
Minitab Inc., State College, PA) and SPSS (version 8.0 for
264 S. D. Sutton and R. H. Findlay
2003 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 5, 256266
Windows 95, SPSS Inc., Chicago, IL) were used for statistical
tests. Data were a posteriori grouped according to seasonal
changes in water temperature and streamow (Table 2).
These groupings were used to determine seasonal trends in
total microbial biomass, microbial community structure and
measured environmental parameter (water temperature, con-
ductivity, water depth, total suspended solids, dissolved
organic carbon, sediment grain size, sediment organic matter
content, instantaneous streamow velocity and streamow).
Data that were not normally distributed were natural log
transformed for determination of main effects. A one-way
ANOVA and Tukeys Honestly Signicant Difference (HSD)
multiple comparison procedure (a = 0.05) were used to
determine seasonal variations.
To identify patterns of variation among PLFA proles,
FAME abundances were converted to weight percentage val-
ues, transformed (ln [wt percentage + 1]) and analysed by
principal component analysis (PCA). This ordination tech-
nique identies the primary or principal patterns of variation
in complex data sets. Samples with similar principal compo-
nent scores have similar PLFA proles and samples with
dissimilar principal component scores have dissimilar PLFA
proles. Individual fatty acids with high component loads
(either positive or negative) contribute strongly to the similar-
ity/dissimilarity of the proles. Principal component scores
are numerical descriptors of the communities and suitable for
further statistical analysis (as above).
Analysis of variance was used to determine spatial differ-
ences in community structure with respect to relative position
upstream or downstream of the articial reservoir. Samples
collected upstream of Harsha Lake (stations 1, 2 and 3) were
grouped separately from samples collected downstream of
the reservoir (stations 4, 5 and 6) and differences in commu-
nity structure (in terms of principal component sample score
data) were determined. Samples were also grouped by sea-
son and relative position to reservoir and a two-way ANOVA
was used to test if microbial community structure was differ-
ent with respect to season and relative spatial position. An
interaction term between season and relative position to res-
ervoir was also included in the two-way ANOVA model.
Further community structure information was determined
from the weight per cent ratios of specic PLFA (Guckert
et al., 1986; White et al., 1997). The relative importance of
phototrophic microeukaryotes was determined from the ratio
of w3/w6 polyenoic PLFA. To evaluate physiological status of
certain prokaryotic members of these communities, we cal-
culated the ratio of cyclopropyl PLFA, cy19:0, to its monenoic
precursor, 18:1w7c and the ratio of 16:1w7t to 16:1w7c. Cor-
responding cyclopropyl ratios for 16-carbon monounsat-
urated fatty acids and tran/cis ratios for 18-carbon
monounsaturates were not analysed because 18:1w7t was
not consistently detected in stream sediments and cy17:0
often coeluted with other PLFA.
Variables that showed a similar pattern of seasonal change
as principal component samples scores were further analysed
using Pearson correlation analysis (environmental parame-
ters and corresponding PC sample scores as variables).
Calculation of prokaryote and microeukaryote biomass
To estimate the contribution of bacteria and microeukaryotes
to total sedimentary biomass, we used algorithms estab-
lished by Findlay and Dobbs (1993). For determination of
microeukaryote biomass, weight per cent values of all poly-
enoic FAMEs were summed and multiplied by two. The later
calculation accounted for the presence of monoenoic and
saturated PLFA (the sum of which are present, on average,
in equal proportion to polyenoic PLFA) in microeukaryote
membranes. The corresponding weight ratio of total microeu-
karyotic PLFA (weight per cent divided by 100) was then
multiplied by the total viable biomass (nmol PLP) to obtain
nanomoles microeukaryotic PLP. The percentage of total
PLFA derived from prokaryotes was estimated as 100 minus
the weight percentage of PLFA derived from eukaryotes.
Prokaryotic biomass (in the units of nanomoles prokaryotic
PLP) was calculated from total viable biomass as for microeu-
karyotes. Prokaryote and microeukaryote carbon biomass
was estimated from PLP using the following conversion fac-
tors: 100 nmols PLP mg
-1
bacterial carbon and 50 nmols
PLP

mg
-1
microeukaryote carbon (Dobbs and Findlay, 1993).
Acknowledgements
We gratefully acknowledge the Miami University Graduate
School for supporting S.D.S. as a Dissertation Scholar and
for a dissertation improvement grant awarded in 1999. This
research was supported in part by the Clermont County
Water District. We are grateful to L. A. Kaplan (Stroud Water
Research Center) for performing DOC analyses, R. J. Veley
(United States Geological Survey) for providing discharge
data and to P. Braasch (Clermont County Water District) for
sharing water quality data. We are in debt to many individuals
who donated their time and assistance for eld sampling
throughout the three-year study. Special thanks to Andrea
Aberegg, Amy Wells, Jason Porter, Al Christian and Jim
Smoot for help with eld and laboratory experiments. Thanks
also to W. H. Renwick, Department of Geography, Miami
University for use of the portable ow meter.
References
American Public Health Association (1995) Suspended sol-
ids. In Standard Methods for the Examination of Water and
Wastewater, 19th edn. Washington, DC: American Public
Health Association, Section 2540A and B.
Biggs, B.J.F. (1995) The contribution of ood disturbance,
catchment geology and land use to the habitat template of
periphyton in stream ecosystems. Freshwater Biol 33:
419438.
Biggs, B.J.F. (1996) Patterns in benthic algae of streams. In
Algal Eclology, Freshwater Benthic Ecosystems.
Stevenson, R.J., Bothwell, M.L, and Lowe, R.L. (eds).
San Diego, CA: Academic Press, pp. 3156.
Biggs, B.J.F. (2000) Eutrophication of streams and rivers:
dissolved nutrient-chlorophyll relationships for benthic
algae. J N Am Benthol Soc 19: 1731.
Biggs, B.J.F., and Close, M.E. (1989) Periphyton biomass
dynamics in gravel bed rivers: the relative effects of ows
and nutrients. Freshwater Biol 22: 209231.
Biggs, B.J.F., Goring, D.G., and Nikora, V.I. (1998) Subsidy
stress responses of stream periphyton to gradients in water
Microbial community dynamics in stream sediments 265
2003 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 5, 256266
velocity as a function of community growth form. J Phycol
34: 598607.
Bott, T.L., and Kaplan, L.A. (1985) Bacterial biomass, meta-
bolic state and activity in stream sediments: relation to
environmental variables and multiple assay comparisons.
Appl Environ Microbiol 50: 508522.
Brmmer, I.H., Fehr, W., and Wagner-Dbler, I. (2000) Bio-
lm community structure in polluted rivers: abundance of
dominant phylogenetic groups over a complete annual
cycle. Appl Environ Microbiol 66: 30783082.
Cazaubon, A., and Giudicelli, J. (1999) Impact of the residual
ow on the physical characteristics and benthic community
(algae, invertebrates) of a regulated Mediterranean river:
the Durance, France. Regulated Rivers 15: 441461.
Day, P.R. (1965) Particle fractionation and particle-size anal-
ysis. In Methods of Soil Analysis, Part 1. Black, C.A. (ed.).
Madison, WI: American Society of Agronomy, pp. 545
567.
Debrewer, L.M., Rowe, G.L., Reutter, D.C., Moore, R.C.,
Hambrook, J.A., and Baker, N.T. (2000) Environmental
setting and effects on water quality in the Great and Little
Miami River basins, Ohio and Indiana. In National Water-
Quality Assessment Program Water Resources Investiga-
tion Report, 994201. Columbus, OH: U.S. Geological
Survey, pp. 6276.
Dobbs, F.C., and Findlay, R.H. (1993) Analysis of microbial
lipids to determine biomass and detect the response of
sedimentary microorganisms to disturbance. In Handbook
of Methods in Aquatic Microbial Ecology. Kemp, P.F.,
Sherr, B.F., Sherr, E.B, and Cole, J.J. (eds). Boca Raton,
FL: Lewis Publishers, pp. 347358.
Dobbs, F.C., and Guckert, J.B. (1988) Callianassa trilobata
(Crustcea: Thalassinidea) inuences abundance of meio-
fauna and biomass composition, and physiologic state of
microbial communities within its burrow. Mar Eco Prog
Series 45: 6979.
Dyer, S.D., White-Hull, C.E., Wang, X., Johnson, T.D., and
Carr, G.J. (1998) Determining the inuences of habitat and
chemical factors on instream biotic integrity for a Southern
Ohio watershed. J Aquat Ecosys Stress Recov 6: 91110.
Edlund, A., Nichols, P.D., Roffey, R., and White, D.C. (1985)
Extractable and lipopolysaccharide fatty acid proles from
Desulfovibrio species. J Lipid Res 26: 982988.
Findlay, R.H. (1996) The use of phospholipid fatty acids to
determine microbial community structure. In Molecular
Microbial Ecology Manual. Akkermans, A.D.L., van Elsas,
J.D, and de Bruijn, F.J. (eds). Netherlands, Kluwer:
Academic Publishers, pp. 4.1.4: 117.
Findlay, R.H., and Dobbs, F.C. (1993) Quantitative descrip-
tion of microbial communities using lipid analysis. In Hand-
book of Methods in Aquatic Microbial Ecology. Kemp, P.F.,
Sherr, B.F., Sherr, E.B, and Cole, J.J. (eds). Boca Raton,
FL: Lewis Publishers, pp. 271284.
Findlay, R.H., and Watling, L. (1998) Seasonal variation in
the structure of a marine benthic microbial community.
Microbial Ecol 36: 2330.
Findlay, R.H., King, G.M., and Watling, L. (1989) Efcacy of
phospholipid analysis in determining microbial biomass in
sediments. Appl Envion Microbiol 55: 28882893.
Findlay, R.H., Trexler, M.B., Guckert, J.B., and White, D.C.
(1990a) Laboratory study of disturbance in marine sedi-
ments: response of a microbial community. Mar Ecol Prog
Series 62: 121133.
Findlay, R.H., Trexler, M.B., and White, D.C. (1990b)
Response of a benthic microbial community to biotic dis-
turbance. Mar Ecol Prog Series 62: 135148.
Ford, J. (1989) The effects of chemical stress on aquatic
species composition and community structure. In Ecotoxi-
cology: Problems and Approaches. Levin, S.A., Harwell,
M.A., Kelly, J.R, and Kimbal, K.D. (eds). New York, NY:
Springer-Verlag, pp. 99144.
Guckert, J.B., Antworth, C.P., Nichols, P.D., and White, D.C.
(1985) Phospholipid, ester-linked fatty acid proles as
reproducible assays for changes in prokaryotic community
structure of estuarine sediments. FEMS Microbiol Ecol 31:
147158.
Guckert, J.B., Hood, M.A., and White, D.C. (1986) Phospho-
lipid ester-linked fatty acid prole changes during nutrient
deprivation of Vibrio cholerae: increases in trans/cis ratio
and proportions of cyclopropyl fatty acids. Appl Environ
Microbiol 52: 794801.
Guckert, J.B., Nold, S.C., Boston, H.L., and White, D.C.
(1992) Periphyton response in an industrial receiving
stream: lipid-based physiological stress analysis and pat-
tern recognition of microbial community structure. Can J
Fish Aquat Sci 49: 25792587.
Kaneda, T. (1977) Fatty acids of the genus Bacillus: an
example of branched-chain preference. Bacteriol Rev 41:
391418.
Kaplan, L.A. (1992) Comparison of high-temperature and
persulfate oxidation methods for determination of dissolved
organic carbon in freshwaters. Limnol Oceanogr 37: 1119
1125.
Kaplan, L.A. (1994) A eld and laboratory procedure to col-
lect, process, and preserve freshwater sample for dis-
solved organic carbon analysis. Limnol Oceanogr 39:
14701476.
Kaplan, L.A., and Newbold, D. (1993) Biogeochemistry of
dissolved organic carbon entering streams. In Aquatic
Microbiology and Ecological Approach. Ford, T.E. (ed).
Boston, MA: Blackwell Science Publishers, pp. 139166.
Langworthy, D.E., Stapleton, R.D., Sayler, G.S., and Findlay,
R.H. (2002) Lipid analysis of the response of a microbial
community to polycyclic aromatic hydrocarbons. Microb
Ecol 43: 189198.
Leff, L.G. (1994) Stream bacterial ecology: a neglected eld?
Am Soc Microbiol News 60: 135138.
Lemke, M.J., and Leff, L.G. (1999) Bacterial populations in
an anthropogenically disturbed stream: comparison of dif-
ferent seasons. Microb Ecol 38: 234243.
Parkes, R.J., and Taylor, J. (1983) The relationship between
fatty acid distributions and bacterial respiration types in
contemporary marine sediments. Estuar Coast Shelf Sci
16: 173189.
Shiah, F.-K., and Ducklow, H.W. (1994) Temperature and
substrate regulation of bacterial abundance, production,
and specic growth rate in Chesapeake Bay, USA. Mar
Ecol Prog Series 103: 297308.
Smoot, J.C., and Findlay, R.H. (2001) Spatial and seasonal
variation in a reservoir sedimentary microbial community
as determined by phospholipid analysis. Microb Ecol 42:
350358.
266 S. D. Sutton and R. H. Findlay
2003 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 5, 256266
Tetra Tech (2001) Clermont county ambient water quality
monitoring program East Fork of the Little Miami River:
ve-year status and trends 19962000. [WWW document].
URL http://www.oeq.net/default.php?section=reports
Tett, P., Gallegos, C., Kelly, M.G., Hornberger, G.M., and
Cosby, B.J. (1978) Relationships among substrate, ow,
and benthic microalgal pigment density in Mechums River,
Virginia. Limnol Oceanogr 23: 785797.
United States Geological Survey. (2000) Surface-water daily
streamow statistics for E F L River, station 03247500.
[WWW document] URL http://waterdata.usgs.gov/oh/nwis/
uv/?site_no=03247500&PARAmeter_cd=00065,00060
White, D.C., Pinkart, H.C., and Ringelberg, D.B. (1997) Bio-
mass measurements: biochemical approaches. In Manual
of Environmental Microbiology. Hurst, C.J., Knudsen, G.R.,
McInerney, M.J., Stetzenbach, L.D., and Walter, M.V.
(eds). Washington, DC: American Society for Microbiology
Press, pp. 91101.
Wilkinson, S.G. (1988) Gram-negative bacteria. In Microbial
Lipids, Vol. 1. Ratledge, C., and Wilkinson, S.G. (eds). San
Diego, CA: Academic Press, pp. 299488.
Winget, R.N. (1984) Ecological studies of a regulated stream:
Huntington River, Emery County, Utah. Great Basin Nat
44: 231256.