Download Biogeography of planktonic and benthic

RESEARCH ARTICLE
Biogeography of planktonic and benthic cyanobacteria in
coastal waters of the Big Island, Hawai’i
Samuel D. Chamberlain1, Katherine A. Kaplan2, Maria Modanu3, Katherine M. Sirianni1,
Senifa Annandale4 & Ian Hewson5
1
Department of Ecology and Evolutionary Biology, Cornell University, Ithaca, NY, USA; 2Department of Natural Resources, Cornell University,
Ithaca, NY, USA; 3Department of Neurobiology and Behavior, Cornell University, Ithaca, NY, USA; 4Natural Science Division, University of Hawai’i
at Hilo, Hilo, HI, USA; and 5Department of Microbiology, Cornell University, Ithaca, NY, USA
Correspondence: Ian Hewson, Department
of Microbiology, Cornell University, Ithaca,
NY 14853, USA.
Tel.: 607 255 0151;
fax: 607 255 3904;
e-mail: [email protected]
Received 8 December 2013; revised 18
March 2014; accepted 25 March 2014. Final
version published online 29 April 2014.
DOI: 10.1111/1574-6941.12337
MICROBIOLOGY ECOLOGY
Editor: Riks Laanbroek
Keywords
biogeography; cyanobacteria; community;
Hawaii.
Abstract
Cyanobacteria are biogeochemically significant constituents of coral reef ecosystems; however, little is known about biotic and abiotic factors influencing the
abundance and composition of cyanobacterial communities in fringing coral
reef waters. To understand the patterns of cyanobacterial biogeography in relation to coastal environmental factors, we examined the diversity of planktonic
and benthic cyanobacteria at 12 sites along the west coast of Hawaii’s Big
Island. We found distinct cyanobacterial communities in sediments compared
to the water column. In both sediments and water, community structure was
strongly related to overall biomass (chlorophyll a concentration), although
both these communities corresponded to different sets of biotic/abiotic variables. To examine the influence of freshwater input on planktonic cyanobacterial communities, we conducted a mesocosm experiment where seawater was
amended with freshwater from two sources representing high- and low-human
population influence. Planktonic cyanobacterial abundance decreased over time
in mesocosms, although chlorophyll a concentration significantly increased
with time, indicating cyanobacteria were likely outcompeted by other phytoplankton in incubations. Our results show that cyanobacterial community
structure may be affected by runoff from terrestrial habitats, but that the composition of cyanobacterial communities inhabiting these locations is also structured by factors not measured in this study.
Introduction
Marine cyanobacteria play a critical role in global biogeochemistry and nutrient cycling and are responsible for
nearly half of global primary production (Field, 1998).
Cyanobacteria produce labile organic matter (OM)
through photosynthesis, which is respired as carbon dioxide by either heterotrophic bacteria or archaea (Azam &
Malfatti, 2007). Cyanobacteria are therefore important to
global carbon cycling and marine food webs. Most
research on marine cyanobacteria in oligotrophic waters
focuses on ubiquitous picocyanobacteria, for example
Prochlorococcus and Synechococcus, or upon globally significant nitrogen-fixing cyanobacteria, although cyanobacteria are also an important component of coral reef
ecosystems (Charpy et al., 2012) and are the dominant
ª 2014 Federation of European Microbiological Societies.
Published by John Wiley & Sons Ltd. All rights reserved
picophytoplankton in Pacific coral reefs (Charpy, 1996,
2005; Charpy & Blanchot, 1998, 1999). In coral reefs,
Synechococcus are the dominant cyanobacteria by abundance and biomass, while Prochlorococcus are dominant in
adjacent open-ocean pelagic waters (Charpy & Blanchot,
1998, 1999). Coral reef cyanobacteria inhabit both benthic
and planktonic habitats and play important roles in reef
biogeochemistry and ecology. Benthic cyanobacterial mats
are sometimes responsible for nitrogen fixation in coral
reefs, supplying up to c. 25% of the nitrogen required by
primary producers (Charpy-Roubaud et al., 2001;
Charpy-Roubaud & Larkum, 2005; Kayanne et al., 2005;
Dong et al., 2008; Charpy et al., 2009). Planktonic cyanobacteria have also been observed to fix nitrogen at ecologically significant rates in coral ecosystems (Dong et al.,
2008). Cyanobacteria are also an important component of
FEMS Microbiol Ecol 89 (2014) 80–88
81
Biogeography of cyanobacteria in coastal Hawaii
microbial food webs and grazed biomass in coral ecosystems (Gonzalez et al., 1998; Charpy, 2005). Reef geomorphology and hydrology influence cyanobacterial
community structure, and these dynamics impact species
abundance, diversity, and nutrient limitation status
(Charpy & Blanchot, 1998; Dufour et al., 2001). Freshwater input to marine ecosystems influences cyanobacterial
diversity and community structure (Hewson & Fuhrman,
2004; Hewson et al., 2006a). Despite a growing appreciation for the biogeochemical and ecological roles of cyanobacteria in coral reef habitats, there is little information
on factors influencing their composition, especially in
fringing reef ecosystems that receive substantial inputs of
freshwater terrestrial runoff and groundwater discharge.
One area where coral reef ecosystems experience significant terrestrial influence is the Hawaiian Island archipelago. The Hawaiian Island archipelago comprises eight
major islands in the central North Pacific gyre. The
islands range in age from more than 27 million years in
the northwestern section of the archipelago to the most
recent and growing Island of Hawaii, which contains several active and inactive volcanoes (Clague et al., 1975). As
a consequence of island geology, Hawaii experiences
widespread groundwater discharge, which forms submarine freshwater plumes in coastal waters (Peterson et al.,
2009) and influences coastal marine ecosystems (Parsons
et al., 2008). Hawaii is currently experiencing rapid
human development on its drier west coast. Golf courses,
beach resorts, and residential developments have been
constructed to cater to a growing tourist economy (U.S.
Census Bureau, 2010), although agricultural development
in catchments is minimal due to extensive lava fields.
Combined terrestrial runoff and groundwater discharge
on Hawaii’s west coast are responsible for significant
nutrient input to the coastal marine zone (Street et al.,
2008). Nutrient input from terrestrial sources has been
linked to human land-use and development on Hawaii’s
west coast. Specifically, fertilized golf courses are responsible for significant nutrient loading to the coastal marine
zone (Dollar & Atkinson, 1992; Knee et al., 2010), while
less prevalent agricultural and urban land-use has not
been directly linked to nutrient loading in western Hawaii
(Knee et al., 2010). These inputs negatively impact coral
ecosystem health (Parsons et al., 2008; Becker et al.,
2013), and on the more developed island of Oahu,
human-impacted terrestrial runoff has led to significant
community restructuring in fringing reef ecosystems
(Lapointe & Bedford, 2011). Unlike more densely populated islands, human development is diffuse along
Hawaii’s western shore and pockets of undeveloped coastline still exist (Dollar & Atkinson, 1992). Due to the nature of human development on Hawaii, fringing reef
ecosystems are likely receiving freshwater input of variable
FEMS Microbiol Ecol 89 (2014) 80–88
quantity and quality, and the impact of freshwater inputs
on Hawaiian cyanobacterial communities is not well
understood.
In this study, we examined the biogeography of fringing reef cyanobacterial communities in the context of
several abiotic and biotic variables at 12 sites along the
west coast of Hawaii. Sites were chosen to encompass a
gradient of terrestrial influence, in order to examine its
impacts on cyanobacterial abundance and community
structure. We also examined the influence of freshwater
amendment of variable quality on cyanobacterial diversity and abundance in experimental mesocosms. In both
field and mesocosm studies, we quantified cyanobacterial
abundance by autofluorescence microscopy and examined cyanobacterial community assemblage composition
and diversity using automated ribosomal intergenic
spacer analysis (ARISA). Combining experimental and
observational methods enhanced our ability to understand patterns of cyanobacterial biogeography in the
region and to understand the impact of terrestrial runoff
and groundwater discharge on marine cyanobacterial
communities.
Materials and methods
Field sampling
We sampled 12 coastal sites to describe the biogeography
of cyanobacterial communities and examine their
response to terrestrial runoff in fringing coral reefs
(Table 1). We chose our sites based on adjacent human
population, freshwater input via groundwater discharge
or aboveground runoff, and extent of mixing with openocean seawater. At each site (c. 2 m total water depth,
which was 10–30 m offshore), we collected 4 L of surface
seawater into seawater-rinsed, 2-L polycarbonate bottles.
Snorkelers collected surface (0–2 cm) and deep sediments
(12–14 cm) using a 60-mL syringe corer. When collecting
sediments, the open syringe ends were plugged with rubber stoppers and kept vertical to ensure that no mixing
occurred between surface and deep sediment profiles.
Onshore, surface and deep sediments were separated and
placed in sterile 50-mL tubes for storage. A YSI probe
was used to measure dissolved oxygen content (mg L1),
conductivity (lS cm1), salinity, temperature (°C), and
pH at all sites. All samples were placed in a cooler for
transport to the laboratory and processed within 4–6 h of
collection.
Chlorophyll a concentration
Between 0.7 and 1 L of seawater was filtered through 47mm GF/F filters (Whatman) to determine chlorophyll a
ª 2014 Federation of European Microbiological Societies.
Published by John Wiley & Sons Ltd. All rights reserved
S.D. Chamberlain et al.
82
Table 1. Field site locations, salinity, temperature (°C), pH, dissolved oxygen (DO, mg L1), percentage organic matter in sediment (OM, % loss
on ignition), total suspended matter (TSM, mg L1), and chlorophyll a concentration (Chl a, lg L1)
Site
Lat
Long
Salinity
Temperature
pH
DO
OM
TSM
Chl a
Kiholo A
Kiholo B
Honaunau A
Honaunau B
Puako A
Puako B
A-Bay A
A-Bay B
Spencer
Kohala
Mahukona
Pololu
19.849
19.857
19.424
19.422
19.961
19.969
19.910
19.913
20.025
20.162
20.183
20.205
155.936
155.922
155.911
155.911
155.859
155.845
155.896
155.890
155.824
155.899
155.901
155.731
33.3
32.4
32.2
29.5
30.7
31.6
32.8
28.2
28
33.1
33.1
31.8
25.4
25.3
24.6
24.9
25.4
26.2
25.5
25.2
26.1
25.7
25
24.5
8.06
7.98
7.96
7.72
7.98
8.29
7.98
8.09
8.02
8.23
8.25
8.14
8.42
8.30
8.7
11.24
9.73
19.45
19.97
26.31
8.62
15.95
11.32
8.41
1.64
0.11
1.73
0.82
2.44
5.57
3.12
8.10
4.91
1.46
2.07
NA
3.4
28.4
1.0
2.9
8.6
5.9
4.1
19.5
8.0
3.8
56.8
26.1
29.0
53.0
20.5
87.4
49.5
91.1
55.0
125.7
135.3
32.2
30.7
42.0
A-Bay, Anaehoomalu Bay.
concentration in field and mesocosm samples. Filters were
folded, covered in aluminum foil, placed in whirlpaks,
and frozen at 20 °C. Chlorophyll a was then analyzed
according to protocols in Parsons et al. (1985), and sediment chlorophyll a was analyzed using the protocols of
Hewson et al. (2001).
Total suspended matter
One liter of seawater was filtered through preweighed and
desiccated GF/F filters. Filters were placed in foil packets
and frozen prior to analysis at Cornell University. Filters
were oven-dried at 60 °C for 7 days and reweighed to
quantify total suspended matter (TSM).
Sediment analyses
Sediments were combusted at 525 °C for 4 h and weighed
before and after combustion to determine percentage
organic matter (OM) via the loss-on-ignition method
(Dean, 1974). Dried sediments (c. 5 g) were sorted using a
phi-scale sieve set to determine sediment size. Dried sediments were placed in sieves and shaken for 1 min, and size
fractions were decanted into clean vessels. The mass of each
fraction was determined by fine balance.
Stable isotope analysis
All sediment and water samples were analyzed for their
nitrogen stable isotope values. Anthropogenic nutrient
loading into coastal waters can be identified by interpretation of nitrogen stable isotope values and has been used
extensively to identify pollutant inputs to Hawaiian reefs
(Parsons et al., 2008; Dailer et al., 2010; Lapointe & Bedford, 2011). d15N values were calculated relative to atmospheric nitrogen as follows:
ª 2014 Federation of European Microbiological Societies.
Published by John Wiley & Sons Ltd. All rights reserved
d15 N ð&Þ ¼ ððRsample =Rstandard Þ 1Þ 1000;
where R is the ratio of heavy to light isotopes (15N/14N)
of the sample or the reference standard of atmospheric
nitrogen. For seawater, 1 L of water was filtered through
47-mm GF/F filters (Whatman). Filters were folded, covered in aluminum foil, placed in whirlpaks, and frozen at
20 °C prior to isotope analysis. Sediments were milled
to a fine powder on a SPEX SamplePrep Mixer/Mill. One
milligram ( 0.1 mg) of each sediment and water filter
was encapsulated in tin and analyzed for d15N on a
Thermo Finnigan Delta V Advantage isotope ratio mass
spectrometer and NC2500 elemental analyzer at Cornell
University’s Stable Isotope laboratory. A laboratory standard was run for every 10 samples producing an accuracy
of 0.3&.
Mesocosm experiment
Freshwater input samples were collected from two locations representing variable terrestrial influence (Spencer
Beach; N 20.0252, W 155.8236, and an anchialine pond
located near the Anaehoomalu Resort; N 19.9089, W
155.8946), and seawater from Kiholo Bay (N 19.84899, W
155.93616). Freshwater input sources were filtered with a
0.2-lm filter to remove particulate matter. Anchialine
pond freshwater was used to represent low-human population influence, and Spencer Beach runoff water was
used to represent high-human population influence.
Anchialine pond freshwater is similar in composition to
natural groundwater inputs, whereas the freshwater collected at Spencer Beach, is influenced by a large urban/
agricultural catchment and the town of Kawaihae and
flows to the ocean as aboveground runoff. Four replicate
mesocosms (4-L acid-washed cubitainers) were each
FEMS Microbiol Ecol 89 (2014) 80–88
83
Biogeography of cyanobacteria in coastal Hawaii
amended with 80 mL Spencer Beach water (high-impact
treatment) and 80 mL anchialine pond water (low-impact
treatment), and four containers served as controls (no
amendment). Mesocosms were maintained at ambient
light and temperature in a flow-through open-top outdoor aquarium. Mesocosms were sampled after 1, 3, 7,
and 10 days of incubation, during which two replicates of
each treatment were sacrificed after day 3 and the
remaining two sacrificed after day 7. At each time point,
two mesocosms per treatment were sampled for cyanobacterial abundance, cyanobacterial diversity, and chlorophyll a concentration.
Cyanobacterial abundance
Cyanobacterial abundance was quantified by epifluorescence microscopy. Duplicate samples of seawater (10 mL)
from field and mesocosm samples were diafiltered
through 0.22-lm black Isopore (Millipore) filters. Filters
were mounted on glass slides in 30 lL of PBS/glycerol
(50% : 50%). All slides were immediately frozen prior to
analysis. Autofluorescent cyanobacteria were counted
under green light excitation with an Olympus BX-51 epifluorescence microscope. More than 200 autofluorescent
cells were counted in 100 fields of view.
Cyanobacterial community analysis
Seawater (0.7–1.0 L) was diafiltered through 0.22-lm Durapore filters (47 mm diameter), and filters were frozen
in Whirl-Pak sample bags. Sediment samples were also
frozen at 20 °C prior to analysis. DNA on filters was
extracted using the Zymo Bacterial & Fungal DNA Mini
kit, while DNA from 200 mg sediment samples was
extracted using the Zymo Soil DNA Mini kit, following
manufacturer’s protocols.
The composition of benthic and planktonic cyanobacterial assemblages was assessed by ARISA (Fisher & Triplett, 1999) with the modifications of Hewson & Fuhrman
(2004). Two microliters of extracted DNA from each
sample was used as a template in PCRs. Each 50 lL reaction contained 19 PCR buffer, 2.5 mM MgCl2, 1 mM
dNTPs (Promega Nucleotide Mix), 10 pmol each of
primers 1392F and cyanobacterial-specific 23S-225R-TET
(labeled with 50 TET fluorochrome; Rocap et al., 2002),
and 0.2 ng lL1 BSA (Sigma 7030). PCR thermal cycling
started with a 5-min heating step at 94 °C, followed by
30 cycles of denaturing (94 °C for 30 s), annealing
(55 °C for 30 s), and extension (71 °C for 45 s), and
(following thermal cycling) a final extension step at 71 °C
for 7 min. PCRs were purified using Zymo Clean and
Concentrator
columns
following
manufacturer’s
recommendations to remove salts and unincorporated
FEMS Microbiol Ecol 89 (2014) 80–88
nucleotides and eluted in 10 lL of nuclease-free water.
Purified amplified DNA was quantified using Pico Green
fluorescence and diluted to 5 ng lL1. The products were
then run on an ABI 377XL capillary sequencer for 4.5 h
at 6 V, using a custom-made ROX-labeled size standard,
which contained fragments, every 50 bp from 200 to
1500 bp (Bioventures Inc.).
Outputs from fragment analysis were analyzed using
the PEAK SCANNER v 1.0 program (Applied Biosystems),
and then outputs on relative fluorescence intensity and
size were transferred to Microsoft Excel. ARISA assemblage fingerprints were analyzed according to protocols in
Hewson & Fuhrman (2006b) following the shifting bin
window technique. Peaks < 0.1% of total amplified DNA
fluorescence and peaks < 400 and > 1500 bp were discarded before binning 3 bp from 400 to 700 bp,
5 bp from 700 to 1000 bp, and 10 bp from 1000 to
1500 bp (Hewson & Fuhrman, 2004).
Statistical analyses
For the mesocosm experiment, Tukey’s HSD analysis was
used to determine significant differences between treatments within each time point, and generalized additive
models (GAM) were used to determine whether response
variables varied significantly with time and across treatments.
The Bray–Curtis dissimilarity index was applied to
ARISA assemblage fingerprints to determine pairwise differences between samples. To visualize similarities
between cyanobacterial communities across sites, multiple
hierarchical cluster analyses with single, complete, or
average linkage were conducted on community fingerprints from all water and sediment samples. The resulting
clusters were confirmed by running an analysis of similarity (ANOSIM) comparing pelagic to benthic communities.
Further analyses on community structure and influences were conducted separately because water and sediment communities showed consistent differences. We
examined abiotic predictor variables using pairwise linear
regressions to detect collinearity and only included one of
a pair of variables with Pearson correlation values of r
greater than 0.9. This resulted in water cyanobacterial
communities being ordinated in a canonical correspondence analysis (CCA) with seven variables (temperature,
salinity, DO, chlorophyll a, pH, TSM, and d15N), while
sediment communities were ordinated using nine variables (temperature, salinity, DO, chlorophyll a, pH, TSM,
OM, mean sediment size, and d15N). Deep-sediment ARISA assemblage fingerprints were omitted from the sediment CCA to avoid pseudo-replication because only one
set of environmental data was available for both depth
profiles. CCAs were conducted in R using the ‘vegan’
ª 2014 Federation of European Microbiological Societies.
Published by John Wiley & Sons Ltd. All rights reserved
S.D. Chamberlain et al.
84
package. All statistical analyses were conducted in
(R Core Team, 2013).
R
3.0.1
Results
Field analyses
Mesocosm analyses
Chlorophyll a concentration varied significantly with time
in mesocosm incubations (GAM, R2 = 0.697, P < 0.001),
but changes in concentration were not significantly different between treatments and control. Chlorophyll a concentrations were relatively low at the beginning of all
incubations (< 0.1 lg L1), peaked between days 3 and
6, and decreased to intermediate levels on day 10. On day
1, chlorophyll a concentrations in low-impact treatment
were significantly higher than those of control treatments
(Tukey’s HSD, P = 0.0153) and trended higher than the
high-impact treatment (Tukey’s HSD, P = 0.0608). Differences between treatments became pronounced on day
3 when chlorophyll a concentration in low-impact treatment spiked to a level significantly higher than the highimpact treatment (Tukey’s HSD, P = 0.0276). On days 7
and 10, concentrations converged between treatments and
decreased by day 10 (Fig. 3).
Cyanobacterial abundance did not vary significantly
with time or across treatment. Overall, cyanobacterial
abundance decreased over time, and mean abundance
dropped from between 12 300 and 15 200 cells mL1
to < 4500 cells mL1 at the end of all incubations.
W: Puako
W: Mahukona
W: Pololu
W: Honaunau A
W: Kiholo S
W: Honaunau B
W: A−bay A
W: A−bay B
W: Kiholo N
W: Spencer
S: Spencer Shallow
S: Puako Deep B
S: Puako Shallow A
S: Puako Deep A
S: Kohala Deep
S: Puako Shallow B
S: Kiholo S Shallow
S: Kohala Shallow
S: Kiholo N Deep
S: Kiholo N Shallow
S: Honaunau Shallow B
S: Honaunau Deep B
S: Kiholo S Deep
S: Honaunau Shallow A
S: Mahukona Deep
S: Mahukona Shallow
W: Kohala
S: Honaunau Deep A
W: Puako B
S: A−bay Shallow B
S: Spencer Deep
S: A−bay Deep A
S: A−bay Shallow A
S: A−bay Deep B
0.8
0.4
0.0
Bray-Curtis Index
Cluster analyses revealed structuring among cyanobacterial communities, forming a distinct cluster between water
and sediments (Fig. 1). This was confirmed by a significant difference between water and sediment communities
(ANOSIM, R = 0.167, P = 0.001, 999 permutations). Species
richness and Simpson diversity indices in sediment were
significantly higher than in the water column (ANOVA,
F = 3.73, P = 0.02 and F = 9.02, P < 0.001, respectively),
and no differences were found in diversity indices or
community structuring between shallow and deep sediments (Fig. 1).
Relationships between biotic/abiotic variables and cyanobacterial community structure were visualized using
CCA. The total number of cyanobacterial operational taxonomic units (OTUs; peaks in ARISA electropherograms
which are approximately genera; Brown et al., 2005)
present in all water samples was lower (n = 31) than in
sediment samples (n = 99). In the water column, communities assembled into three broad groups, each best
explained by a different combination of variables. Chlorophyll a was strongly related to OTU composition in four
sites (Anaehoomalu Bay A, Anaehoomalu Bay B, Kiholo
North, and Spencer Beach). OTU composition in another
four sites (Pololu, Puako A, Mahukona, and Kohala) was
strongly related to pH, and to a lesser degree salinity and
TSM. Cyanobacterial communities at Kiholo South and
Honaunau A were similar to one another, but did not
correspond to measured environmental variables (Fig. 2).
Cyanobacterial communities in sediment were structured
in three groups, but these groupings did not match
groups in the water column. One group (Kohala, Kiholo
South, Puako B, and Honaunau B) was strongly associated with chlorophyll a concentrations, and d15N
displayed an equally strong, but opposing, effect. OTU
composition at three sites (Mahukona, Honaunau A, and
Anaehoomalu B) formed a group associated with TSM,
and to a lesser degree pH, OM, and sediment size. Sites
Spencer and Kiholo North formed a group primarily
associated with temperature (Fig. 2). Overall, in water
and sediment analyses, chlorophyll a concentration was
strongly associated with cyanobacterial community composition in at least four sites. Beyond this similarity, variables exerting dominant influence differed between water
and sediments. In sediments, chlorophyll a, d15N, TSM,
and temperature were the strongest predictors of community structure, whereas in the water column, chlorophyll
a and pH were the strongest predictors of community
structure (Fig. 2).
ª 2014 Federation of European Microbiological Societies.
Published by John Wiley & Sons Ltd. All rights reserved
Fig. 1. Cyanobacterial community cluster
using Bray–Curtis dissimilarity and complete
linkage for all sediment (S) and water (W)
samples. Water column samples generally
cluster away from sediments, suggesting
distinct cyanobacterial communities in these
two environments.
FEMS Microbiol Ecol 89 (2014) 80–88
85
Sediment
2
j
1
CCA2
δ15N
TSM
h
0
pH
l e
Salinity
Temp
g
b
a
f
k
d
−1
OM
Size
c
DO
Chl a
−3
−2
−1
0
1
2
0.8
Chlorophyll a concentration (μg L–1)
3
Biogeography of cyanobacteria in coastal Hawaii
Control
High-impact
Low-impact
b
0.6
ab
0.4
a
0.2
a
ab
b
0.0
3
0
3
6
9
pH
DO
Chl a
TSM
Temp
j ih
g
0
δ15N
Salinity
a
b
c
d
e
f
g
h
i
j
k
l
cf
−1
CCA2
1
ab
l
e
−2
d
k
−2
−1
Water
Column
0
1
2
Anaehoomalu Bay A
Anaehoomalu Bay B
Honaunau A
Honaunau B
Kiholo North
Kiholo South
Kohala
Mahukona
Pololu
Puako A
Puako B
Spencer
3
4
CCA1
Fig. 2. CCA of sediment and water column cyanobacterial
communities fitted to biotic and abiotic predictor variables. Vector
length and direction indicate relative influence on community
structure; letters represent weighted averages for species at each site.
Cyanobacterial abundance in the low-impact treatment
increased to 43 920 2935 cells mL1 on day 3, consistent with the chlorophyll a spike observed in the lowimpact treatment on day 3 (Fig. 4).
Species richness and Simpson diversity index did not
vary significantly with time or across treatments in mesocosm incubations. Increases in species richness and Simpson diversity index occurred in all treatments between
day 1 and 3. Increased levels of diversity were lost in control and low-impact treatments, and increased diversity
was maintained in the high-impact treatment.
Discussion
Bray–Curtis dissimilarity indices suggest that compositionally distinct cyanobacterial communities exist between
the water column and sediments (Fig. 1). Differences in
bacterial community composition between the water
FEMS Microbiol Ecol 89 (2014) 80–88
Fig. 3. Chlorophyll a concentration (lg L1) at different sampling
time points in mesocosm treatments. High-impact treatment was
amended with freshwater from Spencer Beach, low-impact treatment
was amended with freshwater from an anchialine pool, and control
received no freshwater addition. Letters indicate significant
differences within each time point determined by Tukey’s HSD
(P < 0.05). Means and standard error are presented for each time
point (n = 2).
Cyanobacterial abundance (cells mL–1)
2
Time (Days)
Control
High-impact
Low-impact
40 000
30 000
20 000
10 000
0
0
3
6
9
Time (Days)
Fig. 4. Cyanobacterial abundance (cells mL1) at different sampling
time points in mesocosm treatments. High-impact treatment was
amended with freshwater from Spencer Beach, low-impact treatment
was amended with freshwater from an anchialine pool, and control
received no freshwater addition. Means and standard error are
presented for each time point (n = 2).
column and sediments have been previously observed in
marine ecosystems (Stevens et al., 2005; Hewson & Fuhrman, 2006a; Lozupone & Knight, 2007). These observations were further supported by CCAs, which show clear
differences in water column and sediment community
structure (Fig. 2). Increased species richness and diversity
ª 2014 Federation of European Microbiological Societies.
Published by John Wiley & Sons Ltd. All rights reserved
86
were found in sediments, which follows previous observations of high levels of microbial diversity in sediment
ecosystems (Lozupone & Knight, 2007). Distinct communities were not observed between surface (0–2 cm) and
deep (12–14 cm) sediment profiles (Fig. 1), although
community structuring between surface and deep sediments is a feature observed in other marine and reef environments (Hewson & Fuhrman, 2006a; Hewson et al.,
2007). Here, a lack of community structure between sediment profiles potentially indicates homogenization
between surface and deep sediment profiles. This is likely
due to mixing between these profiles, as sediments were
collected in c. 2 m of water in areas of high surf and tidal
action. Bioturbation by burrowing organisms may have
caused homogenization of deep and shallow sediment
horizons.
Based on our observations, cyanobacterial community
composition corresponds most strongly with overall biomass (inferred from chlorophyll a concentration) and less
strongly with other measured environmental parameters.
CCA results also indicate that sediment and water column communities may respond differently to biotic/abiotic variables and environmental conditions. In
sediments, chlorophyll a, d15N, TSM, and temperature
were the strongest predictors of community structure,
while in the water column, chlorophyll a and pH were
the strongest predictors of community structure (Fig. 2).
The correspondence of d15N and TSM to sediment OTUs
may indicate the influence of anthropogenic nutrient
loading on sediment cyanobacterial communities.
Chlorophyll a concentration (overall biomass) was the
strongest predictor of community composition in both
sediment and water column, and it corresponded strongly
to unique community assemblages in both environments
(Fig. 2). Productive nearshore environments may favor
particular community assemblages of cyanobacteria that
are not dominant in oligotrophic marine systems. This
may reflect the transition between oligotrophic openocean cyanobacterial communities and those more typical
of mesotrophic coastal conditions present in fringing
coral reef habitats. These results are in line with the general observation of transitions from Prochlorococcus-dominant open-ocean waters to Synechococcus-dominant coral
reef ecosystems (Charpy & Blanchot, 1998, 1999) and
other observations of distinct cyanobacterial community
structuring at the freshwater–seawater interface (Hewson
et al., 2006a, 2007). Bacterial community richness and
diversity have been shown to increase with productivity
in a number of marine ecosystems (Horner-Devine et al.,
2004; Hewson et al., 2006a, b, 2007; Hewson & Fuhrman,
2007).
Our mesocosm results indicate that as water column
chlorophyll a concentration and overall biomass increase
ª 2014 Federation of European Microbiological Societies.
Published by John Wiley & Sons Ltd. All rights reserved
S.D. Chamberlain et al.
(Fig. 3), cyanobacterial abundance decreases (Fig. 4). It is
therefore likely that other phytoplankton outcompete cyanobacteria in experimental mesocosms. Variable
response between treatments indicates that input from
sources of variable quality may have differential effects on
reef cyanobacterial communities. Specifically, we observed
a large spike in cyanobacterial abundance on day 3 of the
low-impact freshwater-amended treatment (Fig. 4),
although in all treatments cyanobacterial abundance
decreased to similar levels by the end of incubation.
Extending our results to coastal phytoplankton, it is possible that when a coastal zone receives an initial pulse of
groundwater discharge, cyanobacterial populations will
increase, but over longer periods or with chronic loading,
more competitive phytoplankton populations replace cyanobacteria. The similar response of cyanobacteria to
both inputs suggests a general response of marine cyanobacteria to coastal freshwater input.
Comparison of mesocosm observation with water column CCA shows that while community structuring is primarily related to chlorophyll a concentration and
productivity (Fig. 2), a particular water column cyanobacterial community assemblage may be favored in
coastal systems where cyanobacteria are outcompeted by
eukaryotic phytoplankton. These observations support the
idea that oligotrophic open-ocean cyanobacterial communities may be replaced by unique communities when in
contact with coastal zones of groundwater discharge, such
as Hawaii’s west coast.
Our results suggest that cyanobacterial communities in
the water column and sediments of the fringing coral reef
are distinct and that their community structure is related
to the overall biomass and productivity of their ecosystems. However, our results also highlight strong site-specific heterogeneity in composition that is related to
localized abiotic and biotic factors. Benthic and planktonic cyanobacterial communities correspond to separate
sets of environmental variables, and as a result, these
communities likely respond differentially to environmental change. Furthermore, our results provide evidence that
terrestrial inputs may cause changes to the composition
and abundance of cyanobacterial communities, leading to
communities that are distinct from open-ocean ecosystems. Hence, as expected, increased human development
in catchments that influence freshwater discharge into
coastal zones (Dollar & Atkinson, 1992; Knee et al., 2010)
may restructure pelagic, oligotrophic cyanobacterial communities advected from the open ocean. Moreover, our
results suggest that cyanobacteria inhabiting the fringing
reefs of Hawaii are sensitive to freshwater inputs, and
planktonic and benthic cyanobacterial communities may
exhibit distinct community responses to environmental
change.
FEMS Microbiol Ecol 89 (2014) 80–88
Biogeography of cyanobacteria in coastal Hawaii
Acknowledgements
We thank the Department of Ecology and Evolutionary
Biology at Cornell University for funding and support.
The authors thank N. Hairston, J. Sparks, and the students of Cornell Graduate Tropical Field Ecology, January
2013, for assistance with sample collection, transport, and
processing. This work was supported by NSF grant OCE
0961894 awarded to IH.
References
Azam F & Malfatti F (2007) Microbial structuring of marine
ecosystems. Nat Rev Microbiol 5: 782–791.
Brown MV, Schwalbach MS, Hewson I & Fuhrman JA (2005)
Coupling 16S-ITS rDNA clone libraries and ARISA to show
marine microbial diversity: development and application to
a time series. Environ Microbiol 7: 1466–1479.
Bureau UC (2010) Statewide Population Growth Rates. Hawai’i
State Department of Business, Economic Development &
Tourism, Suitland, MD, USA.
Charpy L (1996) Phytoplankton biomass and production in
two Tuamotu atoll lagoons (French Polynesia). Mar Ecol
Prog Ser 145: 133–142.
Charpy L (2005) Importance of photosynthetic picoplankton
in coral reef ecosystems. Vie Milieu 55: 217–224.
Charpy L & Blanchot J (1998) Photosynthetic picoplankton in
French Polynesian atoll lagoons: estimation of taxa
contribution to biomass and production by flow cytometry.
Mar Ecol Prog Ser 162: 57–70.
Charpy L & Blanchot J (1999) Picophytoplankton biomass,
community structure and productivity in the Great
Astrolabe Lagoon, Fiji. Coral Reefs 18: 255–262.
Charpy L, Palinska KA, Casareto B, Langlade MJ, Suzuki Y,
Abed RMM & Golubic S (2009) Dinitrogen-fixing
cyanobacteria in microbial mats of two shallow coral reef
ecosystems. Microb Ecol 59: 174–186.
Charpy L, Casareto BE, Langlade MJ & Suzuki Y (2012)
Cyanobacteria in coral reef ecosystems: a review. J Mar Biol
2012: 1–9.
Charpy-Roubaud C & Larkum AWD (2005) Dinitrogen
fixation by exposed communities on the rim of Tikehau
atoll (Tuamotu Archipelago, French Polynesia). Coral Reefs
24: 622–628.
Charpy-Roubaud C, Charpy L & Larkum A (2001)
Atmospheric dinitrogen fixation by benthic communities of
Tikehau Lagoon (Tuamotu Archipelago, French Polynesia)
and its contribution to benthic primary production. Mar
Biol 139: 991–998.
Clague DA, Dalrymple GB & Moberly R (1975) Petrography
and K-Ar ages of dredged volcanic rocks from the western
Hawaiian Ridge and the southern Emperor Seamount chain.
Geol Soc Am Bull 86: 991–998.
Dailer ML, Knox RS, Smith JE, Napier M & Smith CM (2010)
Using d15N values in algal tissue to map locations and
FEMS Microbiol Ecol 89 (2014) 80–88
87
potential sources of anthropogenic nutrient inputs on the
island of Maui, Hawai’i, USA. Mar Pollut Bull 60: 655–671.
Dean WE (1974) Determination of carbonate and organic
matter in calcareous sediments and sedimentary rocks by
loss on ignition; comparison with other methods. J Sediment
Res 44: 242–248.
Dollar SJ & Atkinson MJ (1992) Effects of nutrient subsidies
from groundwater to nearshore marine ecosystems off the
island of Hawaii. Estuar Coast Shelf Sci 35: 409–424.
Dong J, Zhang Y, Wang Y, Zhang S & Wang H (2008) Spatial
and seasonal variations of cyanobacteria and their nitrogen
fixation rates in Sanya Bay, South China Sea. Sci Mar 72:
239–251.
Dufour P, Andrefou€et S, Charpy L & Garcia N (2001) Atoll
morphometry controls lagoon nutrient regime. Limnol
Oceanogr 46: 456–461.
Field CB (1998) Primary production of the biosphere:
integrating terrestrial and oceanic components. Science 281:
237–240.
Fisher MM & Triplett EW (1999) Automated approach for
ribosomal intergenic spacer analysis of microbial diversity
and its application to freshwater bacterial communities.
Appl Environ Microbiol 65: 4630–4636.
Gonzalez JM, Torreton J-P, Dufour P & Charpy L (1998)
Temporal and spatial dynamics of the pelagic microbial
food web in an atoll lagoon. Aquat Microb Ecol 16: 53–64.
Guilherme Becker C, Dalziel BD, Kersch Becker MF, Park MG
& Mouchka M (2013) Indirect effects of human
development along the coast on coral health. Biotropica 45:
401–407.
Hewson I & Fuhrman JA (2004) Richness and diversity of
bacterioplankton species along an estuarine gradient in
Moreton Bay, Australia. Appl Environ Microbiol 70: 3425–
3433.
Hewson I & Fuhrman JA (2006a) Spatial and vertical
biogeography of coral reef sediment bacterial and
diazotroph communities. Mar Ecol Prog Ser 306: 79–86.
Hewson I & Fuhrman JA (2006b) Viral impacts upon marine
bacterioplankton assemblage structure. J Mar Biol Assoc UK
86: 577–589.
Hewson I & Fuhrman JA (2007) Covariation of viral
parameters with bacterial assemblage richness and diversity
in the water column and sediments. Deep Sea Res Part I 54:
811–830.
Hewson I, O’Neil JM, Fuhrman JA & Dennison WC (2001)
Virus-like particle distribution and abundance in sediments
and overlying waters along eutrophication gradients in two
subtropical estuaries. Limnol Oceanogr 46: 1734–1746.
Hewson I, Capone DG, Steele JA & Fuhrman JA (2006a)
Influence of Amazon and Orinoco offshore surface water
plumes on oligotrophic bacterioplankton diversity in the
west tropical Atlantic. Aquat Microb Ecol 43: 11–22.
Hewson I, Steele JA, Capone DG & Fuhrman JA (2006b)
Remarkable heterogeneity in meso- and bathypelagic
bacterioplankton assemblage composition. Limnol Oceanogr
51: 1274–1283.
ª 2014 Federation of European Microbiological Societies.
Published by John Wiley & Sons Ltd. All rights reserved
88
Hewson I, Jacobson Meyers ME & Fuhrman JA (2007)
Diversity and biogeography of bacterial assemblages in
surface sediments across the San Pedro Basin, Southern
California Borderlands. Environ Microbiol 9: 923–933.
Horner-Devine MC, Carney KM & Bohannan BJM (2004) An
ecological perspective on bacterial biodiversity. Proc R Soc
Lond B Biol Sci 271: 113–122.
Kayanne H, Hirota M, Yamamuro M & Koike I (2005)
Nitrogen fixation of filamentous cyanobacteria in a coral
reef measured using three different methods. Coral Reefs 24:
197–200.
Knee K, Street JH, Grossman EG & Paytan A (2010) Nutrient
inputs to the coastal ocean from submarine groundwater
discharge in a groundwater-dominated system: relation to
land use (Kona coast, Hawaii, U.S.A.). Limnol Oceanogr 55:
1105–1122.
Lapointe BE & Bedford BJ (2011) Stormwater nutrient inputs
favor growth of non-native macroalgae (Rhodophyta) on
O’ahu, Hawaiian Islands. Harmful Algae 10: 310–318.
Lozupone CA & Knight R (2007) Global patterns in bacterial
diversity. P Natl Acad Sci USA 104: 11436–11440.
Parsons TR, Maita Y & Lalli CM (1985) A Manual of Chemical
and Biological Methods for Seawater Analysis. Pergamon
Press, Oxford.
ª 2014 Federation of European Microbiological Societies.
Published by John Wiley & Sons Ltd. All rights reserved
S.D. Chamberlain et al.
Parsons ML, Walsh WJ, Settlemier CJ, White DJ, Ballauer JM,
Ayotte PM, Osada KM & Carman B (2008) A multivariate
assessment of the coral ecosystem health of two
embayments on the lee of the island of Hawai’i. Mar Pollut
Bull 56: 1138–1149.
Peterson RN, Burnett WC, Glenn CR & Johnson A (2009)
Quantification of point-source groundwater discharges to
the ocean from the shoreline of the Big Island, Hawaii.
Limnol Oceanogr 54: 890–904.
R Core Team (2013) R: A Language and Environment for
Statistical Computing. R Foundation for Statistical
Computing, Vienna, Austria.
Rocap G, Distel DL, Waterbury JB & Chisholm SW (2002)
Resolution of Prochlorococcus and Synechococcus ecotypes by
using 16S–23S ribosomal DNA internal transcribed spacer
sequences. Appl Environ Microbiol 68: 1180–1191.
Stevens H, Brinkhoff T & Simon M (2005) Composition of
free-living, aggregate-associated and sediment
surface-associated bacterial communities in the German
Wadden Sea. Aquat Microb Ecol 38: 15–30.
Street JH, Knee KL, Grossman EE & Paytan A (2008)
Submarine groundwater discharge and nutrient addition to
the coastal zone and coral reefs of leeward Hawai’i. Mar
Chem 109: 355–376.
FEMS Microbiol Ecol 89 (2014) 80–88