Microbial Ecology

, Volume 68, Issue 1, pp 47–59 | Cite as

Diversity of Benthic Biofilms Along a Land Use Gradient in Tropical Headwater Streams, Puerto Rico

  • Sofía Burgos-Caraballo
  • Sharon A. Cantrell
  • Alonso Ramírez
Microbiology of Aquatic Systems

Abstract

The properties of freshwater ecosystems can be altered, directly or indirectly, by different land uses (e.g., urbanization and agriculture). Streams heavily influenced by high nutrient concentrations associated with agriculture or urbanization may present conditions that can be intolerable for many aquatic species such as macroinvertebrates and fishes. However, information with respect to how benthic microbial communities may respond to changes in stream ecosystem properties in relation to agricultural or urban land uses is limited, in particular for tropical ecosystems. In this study, diversity of benthic biofilms was evaluated in 16 streams along a gradient of land use at the Turabo watershed in Puerto Rico using terminal restriction fragment length polymorphism. Diversity indices and community structure descriptors (species richness, Shannon diversity, dominance and evenness) were calculated for both bacteria and eukaryotes for each stream. Diversity of both groups, bacteria and eukaryotes, did not show a consistent pattern with land use, since it could be high or low at streams dominated by different land uses. This suggests that diversity of biofilms may be more related to site-specific conditions rather than watershed scale factors. To assess this contention, the relationship between biofilm diversity and reach-scale parameters (i.e., nutrient concentrations, canopy cover, conductivity, and dissolved oxygen) was determined using the Akaike Information Criterion (AICc) for small sample size. Results indicated that nitrate was the variable that best explained variations in biofilm diversity. Since nitrate concentrations tend to increase with urban land use, our results suggest that urbanization may indeed increase microbial diversity indirectly by increasing nutrients in stream water.

Introduction

Natural landscapes are transformed by land use activities that are necessarily to satisfy human needs (e.g., agriculture and urbanization) [15]. These landscape modifications may affect stream ecosystems due to the close relationship that streams have with their surrounding landscapes [1]. In streams heavily influenced by agricultural or urban land uses, elevated nutrient concentrations and conductivity could be expected, while oxygen concentrations decrease [1]. Light availability may increase as canopy cover of riparian vegetation is eliminated, but can decrease with depth due to increased turbidity in altered streams. Other factors associated to land use, such as leaky pipes and combined sewer overflow entering streams, can have similar results [1, 37, 41, 48]. These degraded conditions may be intolerable for many aquatic organisms, such as macroinvertebrates and fishes, as well as other organisms with life cycles linked to freshwaters ecosystems (e.g., amphibians and reptiles) [7]. Therefore, sensitive species could be eliminated, while more tolerant species may persist and become more abundant causing a shift in community composition [26]. Likewise, previous studies from temperate regions demonstrated that microbial communities can change in composition as well as in diversity in streams influenced by urban land use [8, 39, 49]. For example, Dopheide et al. found that streams more impacted by urbanization showed higher diversity of ciliates than undisturbed forested areas [11]. A similar pattern was reported by Carrino-Kyker et al. for different functional groups of eukaryotes (algae, fungi, and protist) in urban vernal pools [8]. However, it is not clear if the same patterns observed in microbial communities from temperate regions will be similar in benthic biofilms from tropical streams. This is of major concern, as changes in species composition can result in alterations in ecosystem function [19].

The stream benthic environment is inhabited by complex assemblages of microbial communities, including bacteria and eukaryotes such as algae, fungi, and other protists. This conglomerate of microbial organisms, known as biofilms, is responsible for many essential ecosystem processes such as organic matter decomposition and nutrient cycling [3, 29, 37]. In streams on Caribbean islands, food webs are primarily based on biofilms and are the main food source for different species of fish and shrimp, which in turn represent an important protein source for humans [24]. However, the quality of these food sources can be affected by land uses [5]. Therefore, changes in the community composition of benthic biofilm, as a result of human impact to watershed and reach conditions, may result in the alteration in streams ecosystem function [2].

The purpose of this study was to analyze benthic biofilm community composition in tropical streams along a land use gradient. At the same time, we explored the relationship between reach and watershed factors in order to determine how these variables influenced community composition and diversity of benthic biofilms. We expected to find differences in community composition by considering forest and urbanization as extremes of the land use gradient. However, we did not predict a direction of change in diversity since previous studies indicate that microbial community richness can be positively or negatively influenced by urban land use [8, 11, 50].

Materials and Methods

Watershed and Reach-Scale Parameters

The Turabo watershed is located in the municipality of Caguas, in the central northeastern region of Puerto Rico (Fig. 1). The geology of the zone is dominated by volcanic and volcano clastic rocks from the late Cretaceous together with some fragments of plutonic and alluvium deposits from the Quaternary [47]. In general, the watershed is dominated by forest land use, with agriculture and urbanization as secondary land uses. Forest cover consists of montane, submontane, and lower montane evergreen forest combined with abandoned coffee plantations [18]. Agricultural activities consist of pastures with small fragments used for small-scale farming. The lower portions of the watershed are close to downtown Caguas where urbanization is concentrated. The areas of high elevation in the watershed consist mainly of suburban development. A closer examination of data revealed that no relationship exists between elevation and urban land use.
Fig. 1

Study site location. The shaded regions represent the subwatershed associated to each stream. Each number corresponds to the each study site as is indicated in the text and figures

Land Use Analyses

For our analyses, the term “land use” refers to land use/land cover data. Land use classifications were made for each study stream using geographic information systems (GIS, ESRI ver. 9.1). The subcatchment of each study stream was delimited by identifying the topographic divide following contour lines on topographic maps (scale 1:10,000) to estimate the area. GIS tools were used to classify, delineate, and digitize land uses. Photo interpretation mapping scale was 1:10,000 and was performed by creating polygons of areas no smaller than 4,000 m2. Dominant land uses were identified by using 2007 aerial images provided by the Department of Environmental and Natural Resources of the Puerto Rico Commonwealth. Each polygon was categorized into one of four general land use types: urban, forest, agriculture, and other (e.g., bare soil; Table 1).
Table 1

Land use percent at each study site

Study stream

Forest

Urban

Agriculture

Other

Catchment area (km2)

Elevation (m)

1

100

0

0

0

0.23

100

2

58

0

42

0

1.01

210

3

59

0

40

1

0.37

290

4

57

0

43

0

1.92

330

5

74

2

24

0

1.97

140

6

76

2

22

1

2.13

90

7

86

7

7

0

0.20

120

8

64

13

18

5

1.02

120

9

55

25

19

0

1.36

110

10

66

27

7

0

2.51

90

11

45

29

22

4

0.69

110

12

35

34

27

4

1.30

330

13

23

49

7

0

0.17

330

14

23

67

3

7

0.30

410

15

24

70

0

6

0.28

90

16

14

86

0

0

0.21

420

Each number corresponds to the each study site as is indicated in the text and figures

Physicochemical Parameters

At each stream, we measured water temperature, dissolved oxygen (O2) (using a luminescent dissolved oxygen probe, HQ10–HQ20 meters, Hach Company, Colorado), pH (using a Hanna Instrument, HI 98150 model), conductivity, total dissolved solids, and sodium chloride (all using a Hanna Instrument, HI 9835). Incident light was measured with a LI-Cor quantum sensor (LI-193 Spherical Quantum Sensor), and canopy cover was estimated taking four measurements with a concave spherical densiometer.

Nutrients and major ions were measured by collecting a filtered water sample (Whatman® glass microfiber filters, GF/F 47 mm) at each stream. Samples were frozen until analysis at the University of New Hampshire Water Quality Analysis Laboratory. Samples were analyzed for chloride (Cl), nitrate (NO3), sulfate (SO4−3), sodium (Na+), potassium (K+), magnesium (Mg+2), and calcium (Ca+2) using a Dionex ICS1000 ion chromatograph. Nutrients analysis included ammonium (NH4+) and phosphate (PO4−3) (Westco Scientific SmartChem 200 discrete automated analyzer), total dissolved nitrogen (TDN) and dissolved organic carbon (DOC) (Shimadzu TOCV with total nitrogen module), and dissolved organic nitrogen (DON) (calculated as the differences between TDN and NO3-N + NH4+-N).

Benthic Biofilm Sampling

Samples were collected from September to November in 2008 during the rainy season. The northeastern region of Puerto Rico, where these studies were conducted, lacks strong climatic seasonality; therefore, it is probable that the time of year played a minor role. In each stream, biofilms were collected from rocks at four different riffles separated by at least one sequence of pools and riffle to reduce influence between them. Collected rocks were placed in individual sterile Whirl Packs bags and refrigerated during transport. Once in the laboratory, all benthic material was removed from the rocks by scraping the surface using a sterile metal brush and rinsed with autoclaved distillated water. From all removed material, approximately 1.5 ml was collected and centrifuged at 1,000 rpm velocity to create pellets which were frozen at −20° until DNA extraction.

DNA Extraction, PCR, and TRFLP’s Analysis

In order to determine benthic community structure, terminal restriction fragment length polymorphism (TRFLP) was analyzed for bacteria 16S recombinant DNA (rDNA) genes and Eukarya 18S rDNA; a fingerprint approach that is useful to differentiate communities, compare phylotype richness, and community structure among sites [14, 28, 36].

Biofilm DNA was extracted using MoBio UltraClean soil kit (MoBio Laboratories, Solana Beach, CA) following manufacturer protocols. DNA concentrations were quantified by fluorometry. Amplifications were conducted using 12.5 μg REDTaq Polymerase (Sigma-Aldrich, St. Louis, MO), 2–4 μg of DNA template, 0.625 μg of primers, and 0.5 μg of bovine serum albumin (BSA; Fermentas, Vilnius, Lithuania) to reduce interference with contaminants. Bacteria 16S rDNA genes were amplified using the following set of primers 16S-27F (5′AGAGTTTGATCMTGGCTCAG) and 16S-1525R (5′AAGGAGGTGWTCCARCC). Eukarya 18S rDNA was amplified using universal set of primers 18S-82F (5′GAAACTGCGAATGGCTC) and 18S-1520R (5′CYGCAGGTTCACCTAC). For each set of primers, only the forward primer was fluorescently labeled at the 5′ end with FAM dye (Sigma-Aldrich, Atlanta, GA). Polymerase chain reactions (PCR) were performed with a model 2720 thermal cycler (ABI, USA) using the following program: 1-min hot start at 80 °C, 94 °C for 5 min followed by 30 cycles of denaturation at 94 °C for 30 s, followed by annealing at 52 °C for 30 s, at 72 °C for 1 min 30 s, with a final extension step at 72 °C for 10 min. Amplified DNA was verified by electrophoresis of PCR mixtures in 1.0 % agarose in 1X TAE buffer.

Fluorescently labeled amplicons were enzymatically digested with BsuRI (HaeIII), Rsa, and MnlI (Fermentas Life Science, Maryland, USA) to find the enzymes with the higher resolution for both, bacteria and eukarya. MnlI was chosen as it provided the greatest numbers of terminal restriction fragments (T-RFs) for bacteria, as well as for eukarya. Digestions were performed at 37 °C for 120 min with 2.0 μl of buffer, 0.2 μl of enzyme, and 5.0 μl of amplified product and water to a final volume of 20 μl.

Samples were precipitated in ethanol to eliminate impurities, dried, and resuspended in Hi-Di formamide with GeneScan Liz 500 (ABI, USA) using a size standard that range between 50 and 500 bp. Those fragments that differed by 1 bp were considered different. Samples were run on ABI 3130 genetic analyzer (ABI, USA). Data were normalized by using fragments with a fluorescence >1 % of the total fluorescence. The area of each peak was standardized with respect to the sum of all peak areas of the sample. We considered the fragments that were present at three of the four replicate analyzed at each study site. Restriction fragments were analyzed using the web page MICA (Microbial community analysis [45]). The program APLAUS was selected to infer the community structure based on the TRFLP data of communities present in the biofilms [45], but since one TRFLP could represent different organisms, bacteria were identified at the Phyla level, unless they could be identified to a lower level. Databases of Eukarya were more limited and therefore no eukaryote could be identified.

Statistical Analysis

To reduce the influence of most common Operational Taxonomic Units (OTUs), a square root transformation was used for data obtained at each study site [17]. All OTUs present in a sample were sorted by abundance to identified shared OTUs (by considering those that were present in 10 of the 16 study sites) unique OTUs (by considering the phylotypes that were present in 1 of the 16 study sites). OTUs richness (S) was determined by counting the total number of OTUs present at each study site. Shannon diversity index (H′), Simpson’s diversity index (D), and Pielou’s evenness index (E) were determined using the average abundance of OTUs per study site in PC-Ord version 4.25 [4]. In addition, Smith and Wilson evenness (Evar), previously recommended as a measure of diversity for TRFLP’s data, was calculated [4].

A nonmetric multidimensional scaling (NMS) ordination was used to analyze the profiles of OTUs among study streams, using a square root transformation of the relative height for each OTUs. We choose this ordination for various reasons. It is an efficient ordination method where data does not have to be linearly related [32]. NMS also alleviates problems associated to data sets containing zeros [32]. In addition, the user can choose the distance measure appropriate for the data [32]. We run the NMS considering the following parameters: 6 axes as maximum, 500 iterations, 0.20 steps in length, and 50 runs with real data and randomized data. The dissimilarity distance measure selected for this ordination was Bray-Curtis because it considers the presence of a species as more informative than its absence, as the absence of species from the data is not necessarily associated with environmental factors [25]. It also detects differences in community composition between study sites and has been recommended for TRFLP’s data [43]. The stability criterion used was 0.0005. The final model selected had a stress value less than 20 and was run in PC-Ord. Three NMS analyses were performed, one for bacteria, other for eukarya, and a third one including bacteria and eukarya together.

We considered a set of candidate models consisting of watershed (e.g., land uses) and reach-scale variables to determine how these variables influenced bacteria, eukarya, and the complete biofilm community. Three main competing hypotheses were considered as follows: (a) microbial communities are influenced by land use, (b) microbial communities are influenced by reach-scale parameters, and (c) microbial communities are influenced by the interactions that exist between land uses and reach-scale parameters. We calculated the variance inflation factor to eliminate redundant variables from the analysis. Interactions among variables were also considered. Model selection was used to identify the hypothesis that best supported benthic biofilm diversity data. We used the Akaike Information Criterion corrected for small sampling size (AICc). Variables were checked for normality and transformed when needed. All possible models, included the intercept and error, were obtained from R (version 15.2.0) using the AICcmodavq package to compare and select the best models. Delta AIC (ΔAIC) was determined as the difference between one possible model and the model with the lowest AIC. Akaike weights (wI) were determined as the ratio that one model is best among other possible models using the likelihood of one model in comparison to the likelihood of all possible models. Evidence ratio was determined by dividing the Akaike weights of one model divided by the Akaike weights of the best possible model. Models were selected based on AICc and Akaike weights (wI) [21]. Simple linear regressions were performed to test for relationships suggested by the model.

Results

Subcatchment Land Use

In general, the Turabo catchment was dominated by forest cover. The area of 16 subcatchments considered in this study ranged from 0.21 to 2.50 km2 and formed a continuous gradient from forested to urban streams. Secondary forest was the dominant land cover in 10 of the 16 study streams with >50 % of forest cover (Table 1). Forest cover ranged from 14 to 100 % cover among all study sites (Table 1). Urban land use was variable and ranged from 0 to 86 %. Urbanization was characterized by being dense in the lower parts of the catchment, in particular near the city of Caguas, and more disperse in the upper areas, consisting of small clusters of houses and residential developments. Subcatchment area ranges from 0.17 to 2.51 km2, and no relationship was found with urban land use (R2 = 0.17, p = 0.07). Elevation ranges from 90 to 420 m. Agricultural activities in the Turabo catchment were variable and consisted of small crop patches and low-intensity cattle pastures dispersed throughout the catchment, reaching a maximum of 43 %. Exposed soils did not appear to be associated with agricultural or urban land use and comprised <7 % of the entire catchment (Table 1).

Water Physicochemistry

Most streams along the land use gradient were well-oxygenated (4.36–8.10 mg/l), had warm water temperatures (24.0–28.5 °C), and near-neutral pH values (7.04–8.00) (Table 2). Nutrient concentrations, TDS, and conductivity were highly variable among streams, with some streams highly nutrient enriched and with high water conductivity (Table 2). The concentrations of major anions and cations, such as Cl, Na+, K+, Mg+2, and Ca+2, were variable among study streams (Table 2). Elevated concentrations of Cl, Mg+2, and PO4−3 were found at study site 1, which was 100 % forest cover. Other cations like Ca+2 and K+ tended to be similar along the gradient, while Na+ varied among study sites. However, no trend was observed between these ions and land use. Comparing with the World Health Organization (WHO) and the Environmental Protection Agency (EPA) standards for drinking water with the physicochemistry of the streams, most of them did not reach maximum levels established by these agencies.
Table 2

Physical and chemical properties measured at each stream

Study stream

Temp. (°C)

DO (mg/L)

% O2 saturation

pH

Cond. (μS/s)

TDS (mg/L)

Cl (mg-Cl/L)

Ca (mg-Ca/L)

Na (mg-Na/L)

K (mg-K/L)

Mg (mg-Mg/L)

1

24.68

6.60

80.76

7.61

549.0

256.4

35.16

26.48

29.18

1.41

21.44

2

24.04

8.10

100.48

7.93

153.1

76.6

6.58

9.42

7.68

0.14

2.38

3

24.48

7.73

95.88

7.77

226.3

113.5

18.02

10.28

18.73

1.25

6.36

4

26.42

7.64

99.78

7.80

167.1

83.6

8.09

12.35

10.69

0.91

2.33

5

25.88

7.14

89.92

7.04

168.8

84.2

11.58

9.92

9.71

0.49

3.18

6

25.92

7.23

87.14

7.94

433.4

226.4

16.73

13.05

20.50

1.06

13.02

7

25.06

6.35

79.46

7.85

479.8

238.3

10.87

15.85

9.58

0.67

7.93

8

25.26

5.10

62.70

7.32

348.4

174.4

6.91

16.82

8.11

0.70

4.62

9

25.92

6.77

85.16

7.65

399.2

202.6

28.21

17.28

28.62

2.90

13.69

10

28.40

6.32

83.12

7.61

577.2

288.6

22.19

13.84

25.33

1.07

17.01

11

24.82

7.26

88.72

7.65

517.4

257.8

26.61

28.55

30.82

1.63

18.88

12

24.14

7.83

97.08

8.00

265.7

132.7

16.45

15.66

15.20

0.87

8.60

13

23.84

7.24

89.38

7.87

260.7

130.2

10.34

11.43

10.85

0.38

9.07

14

24.06

7.28

91.16

7.61

283.8

142.1

13.35

13.49

11.08

0.64

6.09

15

28.46

4.36

56.44

7.68

707.8

353.8

20.54

14.01

25.29

1.06

17.27

16

25.20

5.36

68.46

7.04

304.1

152.8

9.89

17.15

7.96

0.55

4.30

Study stream

SO4−2 (mg-S/L)

PO43− (μg-P/L)

NO3 (mg-N/L)

NH4 (μg-N/L)

TDN (mg-N/L)

DON (mg-N/L)

DOC (mg-C/L)

% Canopy cover

PAR (moles/m2)

1

0.00

189

0.74

124

1.06

0.19

2.50

81.28

99.28

2

1.46

25

0.12

5

0.19

0.06

1.15

86.74

78.05

3

0.00

12

0.31

12

0.48

0.16

3.77

82.42

251.31

4

1.60

27

0.35

8

0.49

0.13

1.39

67.97

201.00

5

2.58

31

0.32

5

0.48

0.16

2.51

83.98

155.30

6

0.00

14

0.21

12

0.32

0.09

2.17

85.49

257.31

7

3.63

26

0.28

5

0.38

0.09

1.05

93.50

89.20

8

3.19

54

0.05

11

0.11

0.05

1.79

80.71

76.21

9

0.00

19

0.59

40

0.91

0.28

4.72

75.35

85.91

10

0.00

26

0.37

12

0.35

0.00

2.21

70.31

230.69

11

0.00

41

0.26

30

0.54

0.25

4.28

69.11

8.63

12

1.30

18

1.01

5

1.13

0.12

1.17

77.95

224.57

13

2.32

20

1.09

4

1.30

0.21

1.05

88.09

64.20

14

1.27

6

0.84

4

0.96

0.12

1.36

88.35

3.17

15

0.00

50

0.41

61

0.61

0.14

1.91

53.04

118.84

16

0.72

3

0.97

16

1.09

0.11

1.06

85.80

28.46

Each number corresponds to the each study site as is indicated in the text and figures

Benthic Biofilm Community Composition

A total of 204 different OTUs were found for bacteria. Study sites differed in the numbers of unique OTUs present in a sample. For bacteria, a total of 53 OTUs were unique, present only at one study site (Fig. 2a). Study sites 2, 5, and 14 had 1 unique OTU, 7 and 13 had up to 12 unique OTUs, while the remaining sites had no more than 7 unique OTUs (Fig. 2a). Microbial community analyses (MICA) identified bp158 as Lactobacillales (order) and bp209 was identified as Proteobacteria (phylum), while other OTUs could not be assigned to any group. OTUs bp247 and bp250 were identified as members of the order Clostridiales.
Fig. 2

Number of unique OTUs (a) and total richness (b) of bacteria and eukarya present in biofilm communities among study sites. Black bacteria, gray eukarya

Bacterial richness (S) among study sites ranged from 5 to 88, with the highest richness at sites 7, 12, and 13 (Fig. 2b, Table 3). Shannon index (H′) was highest at sites S12 and S13 and ranged from 1.04 to 3.90 (Table 3). Dominance (D) ranged from 0.55 to 0.97, and study sites S7, S12, S13, and S16 had a D equal or higher than 0.95 (Table 3). Evenness among sites ranged from 0.61to 0.88 and was highest in sites S9, S12, and S13. Evar ranged from 0.31 to 0.52, being highest in 4, 9, and 11 (Table 3). No pattern was observed with shared OTUs and urban land use.

For eukarya, up to 414 different OTUs were observed, and similar to bacteria, the sites differed in the number of unique OTUs (Fig. 2a). We found 51 OTUs that were unique. At study sites 15 and 16, streams that had the highest urban cover, 15 and 13 unique OTUs were found, respectively (Fig. 2a). Other study sites had up to five unique OTUs (Fig. 2a). Eukarya OTUs could not be assigned to any group. Eukaryotes S ranged from 42 to 255, with highest richness at the more urbanized sites 15 and 16 (Fig. 2b, Table 1). H′ ranged from 2.58 to 4.37 and was highest at study sites 15 and 16 (Table 3). D ranged from 0.80 to 0.97, with D value equal or higher than 0.95 at study sites 1, 7, 13, 14, 15, and 16 (Table 3). E ranged from 0.55 to 0.87, and sites 7 and 13 had highest values. Evar ranged from 0.22 to 0.47, and highest values were found at sites 7 and 13 (Table 3).

Ordination of the Benthic Microbial Communities

Three NMS ordinations were generated using OTUs relative abundance of bacteria, eukarya, and biofilm communities (i.e., both groups combined, Fig. 3). The NMS ordination of bacteria had two axes (final stress 13.60). Axis 1 explained 10 % of the variation while axis 2 explained 73 % of the variation of the community. Visual inspection of the ordination showed a gradient in community composition, with only site 4 differing from the others (Fig. 3a). The NMS of eukarya resulted in a three-dimensional ordination (final stress 11.68). Axis 1 explained 17 % of the variation while axes 2 and 3 explained an additional 37 and 25 %, respectively. Graphing axes 2 and 3 showed that study sites 3, 9, 10, 11, and 13 were separated from the main cluster (Fig. 3b). The NMS for biofilm communities formed three axes (Fig. 3c). Axis 1 explained 6 % of the variation while axes 2 and 3 explained 43 and 31 % of the variation, respectively (final stress 11.26), and study sites 3, 4, 9, 10, and 11 were apart from the main cluster (Fig. 3c).
Fig. 3

NMS developed for benthic biofilms, (a) bacteria, (b) eukarya, and (c) NMS including both groups. The axes explaining the largest percent of the variance were selected for graphing

Effects of Watershed and Reach-Scale Factors

Models were performed to detect relationships between biofilm communities and watershed and reach-scale variables. Land use cover of forest, agriculture, and urbanization were considered as watershed scale variables, while reach-scale variables included dissolved oxygen concentration, percent oxygen saturation (relates the proportion of oxygen dissolved in water to the maximum amount that could be present at the same temperature), conductivity, magnesium cation (Mg+2), nitrate (NO3), total dissolved nitrogen (TDN), dissolved organic nitrogen (DON), canopy cover, and photosynthetic active radiation (PAR), as indicated by previous studies [6]. The variance inflation factor indicated that the percent of saturation and Mg+2 were redundant variables and therefore were excluded from further analysis.

Bacterial S was best predicted by a model that only included NO3 (Table 2). Similarly, bacteria Shannon diversity index (H′) was also explained by a model that included only NO3 (Table 2). For bacteria Simpson’s diversity index (D), no models were generated, while for bacteria Pielou’s evenness index (E), the ∆AIC value suggested that there was substantial evidence for the first three models considered (∆AIC <2). The evidence ratio indicated that the model that included only DON was 1.12 times as probable as the model that contained O2 + DON.

No model was generated for eukarya richness, indicating that variability could not be explained by measured parameters. The ∆AIC value for Shannon diversity index (H′) for eukarya presented substantial evidence for the first three models. Models with reach-scale parameters competed very closely. The model containing dissolved oxygen and NO3 was 2.66 times as probable as the model that contained O2 and conductivity. For Simpson’s diversity index (D) of eukarya, the best model (∆AIC <2) included reach scale. The other models were not as efficient explaining the data (Table 4). For eukarya Pielou’s evenness index (E), substantial evidence was presented for the model included PAR, which was 3.13 times as probable as the most likely model which included PAR and NO3.
Table 3

Species richness, Shannon diversity index, Simpson diversity, Pilou equitability index, and Evar per study site

Sites

Bacteria

Eukarya

S

H′

D

E

Evar

S

H′

D

E

Evar

1

50

3.07

0.91

0.79

0.45

110

3.71

0.96

0.79

0.33

2

21

2.38

0.87

0.78

0.41

194

3.66

0.89

0.70

0.40

3

66

3.00

0.89

0.72

0.33

77

2.73

0.89

0.63

0.22

4

5

1.10

0.55

0.69

0.52

108

3.34

0.91

0.71

0.32

5

43

2.93

0.91

0.78

0.43

177

3.67

0.93

0.71

0.36

6

39

2.85

0.90

0.78

0.40

153

2.76

0.80

0.55

0.32

7

87

3.60

0.95

0.81

0.36

81

3.57

0.95

0.81

0.40

8

25

2.68

0.91

0.83

0.44

163

3.46

0.91

0.68

0.31

9

40

3.10

0.93

0.84

0.51

55

3.14

0.92

0.78

0.37

10

35

2.15

0.81

0.61

0.34

56

3.08

0.92

0.77

0.39

11

47

3.21

0.93

0.83

0.51

42

2.59

0.87

0.69

0.35

12

88

3.81

0.96

0.85

0.50

175

3.61

0.92

0.70

0.36

13

84

3.91

0.97

0.88

0.45

90

3.89

0.97

0.87

0.47

14

37

2.55

0.85

0.71

0.46

94

3.61

0.95

0.80

0.38

15

48

3.24

0.94

0.84

0.45

255

4.20

0.97

0.76

0.28

16

70

3.51

0.96

0.83

0.31

239

4.37

0.97

0.80

0.30

S species richness, H Shannon diversity index, D Simpson diversity, E Pilou evenness index, and Evar Smith and Wilson evenness

Table 4

Best approximating models for predicting bacteria and eukarya diversity index

Metric

Model

n

Df

LL

AICc

ΔAICc

wi

R2

Bacteria richness (S)

NO3

3

14

−70.3

148.6

0

0.90

0.27

Intercept

2

15

−73.16

151.25

3.15

0.10

Global

9

8

−66.88

181.75

33.65

0

0.15

Bacteria Shannon index (H′)

NO3

3

14

−14.15

36.29

0

0.82

0.18

Intercept

2

15

−16.32

37.57

2.03

0

 

Global

9

 

−9.62

67.23

31.69

0

0.19

Bacteria Pielou evenness (E)

DON

3

14

22.37

−36.74

0

0.52

0.23

O2+ DON

4

13

23.63

−35.62

1.12

0.30

0.29

Intercept

2

15

19.76

−34.59

2.15

0.18

Global

9

8

26.37

−4.75

31.99

0

0.18

Eukaryote Shannon index (H′)

O2+NO3

4

13

−5.76

23.16

0

0.49

0.42

O2 +Conductivity

4

13

−7.09

25.82

2.66

0.13

0.31

NO3

3

14

−8.96

25.91

2.75

0.12

0.19

O2+NO3+DON

5

12

−5.16

26.31

3.15

0.1

0.41

O2 + Cond + NO3 + DON

6

11

−2.64

26.6

3.44

0.09

0.53

Intercept

2

15

−11.23

27.38

4.22

0.06

Global

9

8

−2.18

52.36

29.2

0

0.39

Eukaryote Simpson diversity (D)

O2+NO3

4

13

34.21

−56.78

0

0.53

0.49

Agr

3

14

31.13

−54.27

2.51

0.15

0.3

NO3

3

14

31.03

−54.05

2.73

0.13

0.29

O2+NO3+DON

5

12

34.21

−52.42

4.36

0.06

0.44

Bosque + Agr

4

13

31.82

−52

4.78

0.05

0.31

Intercept

2

15

27.75

−50.58

6.2

0.02

 

Global

9

8

36.86

−25.71

31.07

0

0.4

Eukaryote Pielou evenness (E)

PAR + NO3

4

13

25.22

−38.81

0

0.7

0.5

PAR

3

14

21.84

−35.69

3.13

0.15

0.29

Forest +Agriculture

4

13

22.95

−34.27

4.54

0.07

0.34

O2+NO3

4

13

22.09

−32.54

6.27

0.03

0.37

Intercept

2

15

18.54

−32.16

6.65

0.03

 

Agriculture

3

14

19.25

−30.51

8.3

0.01

0.35

NO3

3

14

19.18

−30.36

8.45

0.01

0.31

 

Global

9

8

27.74

−7.48

31.33

0

0.41

Models were selected based on AICc scores. Models with the lowest AICc and higher wi were considered the best models

A closer examination of the parameters included in the models using regressions demonstrated that NO3 concentrations played an important role in the diversity indexes of bacteria and eukarya. Positive significant relationships were only found for bacteria S and NO3 concentrations (S: r2 = 0.32, p = 0.02; Fig. 4a). A similar trend was observed for bacteria H′, which increased at higher NO3 concentrations (H′: r2 = 0.24, p = 0.05; Fig. 4b). For eukaryotes, similar patterns were observed, were eukarya D and H′ increased with NO3; however, positive significant relationships were only found D (D: r2 = 0.33, p = 0.01 and H′: r2 = 0.25, p = 0.05; Fig. 4c, d). From all shared bacteria and eukarya, bp247 of bacteria and bp101 of eukarya were significantly related to NO3 concentrations (Fig. 5).
Fig. 4

Lineal regressions of bacteria and eukarya diversity and richness with NO3 concentrations. (a) Bacteria richness, (b) bacteria Shannon index, (c) eukarya Simpson diversity index, and (d) eukarya Shannon index. All diversity index assessed showed similar patterns

Fig. 5

Increases in abundances of (a) bp 247 of bacteria and (b) bp 101 as NO3 concentration increased

Discussion

Community structure analysis at our study sites showed that the number of species detected for eukarya was higher than for bacteria at most sites (14 of 16 sites). This may be related to the fact that eukarya comprise many photosynthetic organisms, such as filamentous algae, diatoms, as well as heterotrophic microbes such as fungi, protozoans, and rotifers. However, caution must be taken when interpreting these results. It is possible that the resolution of the technique used limits the differentiation of all microbial organisms present as we used a general primer to amplify for bacteria (16S) and eukarya (18S) domains. In addition, different species may be represented by the same OTUs. The diversity can increase when the number of unique species is high or when species are evenly distributed [31]. We could argue that differences in H′ among sites can be related to differences in the number of rare species inhabiting different streams because D was similar between bacteria and eukaryotes, ranging between 0.80 and 0.97. Although these results suggest that species composition varied among study sites, results from the NMS analyses showed that diversity was relatively similar among study sites. However, it is important to take into consideration that results of the NMS analysis are based on the species shared among sites and not necessarily in the number of species present/absent [32]. In other words, not all species are present in all study sites, and those that are shared might have similar abundances.

There is little information available about benthic biofilm community structure in either temperate or tropical streams. Moreover, the lack of consensus on how to describe microbial diversity impairs comparison of results among studies. Lyautey et al. found that bacteria richness in biofilms did not vary significantly among streams with different degrees of human impact [30]. They reported S values ranging from 32 to 41 phylotypes which lies within the intermediate species richness found in our study (5–88 phylotypes). This observation implies that benthic biofilm composition in stream ecosystems can vary markedly between regions, as they can share some species while others may be region specific [20]. In the case of eukarya, most of the studies have focused on single groups (e.g., algae and ciliates). This prevents us to make a direct comparison as we did not focus in any particular group, but rather considered all the different groups within eukarya. However, it can be expected that similar to bacteria, eukarya diversity varies significantly between temperate and tropical streams. The highest number of OTUs reported in previous studies, for ciliates, a total of 61 OTUs [11], and for algae, a full amount of 88 OTUs [46], was larger than the total found at several of our studied sites (e.g., sites 9, 10, and 11).

In general, the literature point out that urban streams tend to have species poor communities which often lacked native and pollution-sensitive organisms [38]. This pattern has been well-documented for macroinvertebrates in tropical and temperate streams [9, 34]. In contrast, microbial responses to urban land use are not as consistent. In a gradient of urbanization, bacterial communities can be similar [23] or contrasting among study sites [39]. Our results indicate that diversity of benthic bacteria and eukarya in tropical stream biofilms is not necessary directly related to land use gradient. For example, sites 1 and 15, which are at opposite sides of the land use gradient, had similar number of shared species of bacteria; whereas for eukarya, the amount of shared OTUs was similar between study sites 3 and 16, sites that represent the extremes of the gradient. Similarly, richness of both groups was either low or high at sites dominated by urbanization or forest cover. The lack of relationship is also apparent in terms of H′ and D. This may imply that reach-scale factors such as light and water physicochemistry, rather than watershed land use type, may be structuring benthic biofilm communities.

Nutrients, in particular nitrogen, can have an important role regulating microbial communities [22, 28]. According to the AICc analysis performed, variations of bacterial communities may be a function of nitrogen, as both S and H′ (but not D) tended to increase with nitrogen concentrations. In other work, using nutrient diffusing substrates, we have shown that nitrogen, instead of other nutrients (e.g., PO4−3), causes significant shift in diversity in tropical benthic biofilm [6]. These results support previous studies which have demonstrated that nitrogen is an important factor regulating microbial communities, such as Alphaproteobacteria and the Cytophaga-Flavobacteria cluster in aquatic ecosystems [22, 28, 33, 44]. Nitrate can also be used for microbial respiration (i.e., denitrification) which might be important in streams with low oxygen concentrations. In addition, nitrogen is an essential macronutrient that microbial organisms use for the protein and nucleic acid formation [33]. Therefore, it could be expected that in the presence of higher nitrate concentrations, microbial diversity is enhanced as more substrate (nitrogen) is available, possibly reducing interspecific competition.

No significant AICc models were generated to explain variations in species richness of eukarya. This indicates that the presence or absence of any given species within the eukarya domain does not depend directly upon any of the physicochemical variables evaluated in this study. In contrast, oxygen may have a significant influence in H and D as explained by the best AICc model. Despite these findings, the role of nitrogen in structuring eukaryote communities cannot be diminished, since best AICc models explaining H and D variation included different forms of nitrogen (e.g., NO3 and DON). Moreover, further simple regression analyses showed that no significant relationship exists between oxygen and H′ and D (H′: r = −0.22, p = 0.07; D: r = −0.19, p = 0.09). Conversely, these diversity estimators tend to increase as nitrate concentration rises. This is consistent with previous studies that have indicated that heterotrophic eukarya microbiota is more diverse in streams with elevated nutrient concentrations [11]. A high supply of nitrogen may enhance eukarya diversity in benthic biofilms for the same reason explained above for bacteria.

Agriculture and urban land uses commonly result in high nitrate concentrations in streams [1]. Consequently, it could be expected that streams with higher agriculture and urban cover will have higher diversity of microbes. However, this was not the case in our study. This may be due to the fact that in many instances, streams characterized by higher forest cover presented characteristics of impacted streams. For example, we have found that forested watersheds can have high nitrate levels when streams are influenced by point sources, such as illegal pipe discharge, causing them to resemble the physicochemistry of urban streams. We observed that at several of our study sites (in urban and nonurban areas), nearby houses were discharging waste water directly into streams. Consequently, nitrogen concentrations may be providing substrate and nutrients that favor the establishment of alien microbial species that under other conditions may not establish. For example, it is known that nitrate may promote the establishment of photosynthetic microbes whose efficiency could be limited under low concentration of nutrients [12]. At the same time, there is the possibility that bacteria growing in leaky sewage systems or septic tanks are reaching stream ecosystems by means of contaminated water. The presence of bp 247 (order of Clostridiales), an indicator of fecal contamination, supports this argument (e.g., Clostridium perfringens [16, 35]). This observation indicates that although nitrogen derived from urban areas may influence benthic biofilm communities by increasing diversity, it may be erroneous to conclude that more urbanized streams will always present higher diversity compared to streams associated with other land uses. This is because increases in nitrate concentration may not come only as a result of direct input, but may also be dependent upon local erosion and weathering process, atmospheric deposition, and the prevailing vegetation [40]. Therefore, it can be argued that reach-scale conditions are as important as watershed factors in structuring benthic communities, as has been demonstrated for benthic biofilm metabolism [6]. We propose that reach-scale factors should be considered when assessing the effect of different land uses upon the functioning and diversity of tropical streams ecosystem.

According to the urban stream syndrome, it is predicted that nutrient concentrations increase in streams highly impacted by urbanization [48]. Given the expected global increase in urban cover, similar increases in stream nitrogen might result in alterations in the diversity of benthic biofilms. Therefore, future efforts should focus on assessing how modifications associated to urban areas (e.g., nutrients enrichment and geomorphological alterations) alter the function of benthic biofilms and eventually stream ecosystem function (e.g., primary production and nitrogen fixation). Impacts on biofilms are likely to affect (positively or negatively) other aquatic organisms, as biofilms represent an important component of food webs in tropical stream ecosystems [24], and may reveal bottom-up effects associated to modifications in benthic biofilm dynamics, such as quality of food available for other organisms dependent of biofilms as a food source.

Notes

Acknowledgments

This work was supported by the Puerto Rico Alliances for Minority Participation (AMP grant no. HRD-0114586), GK-12 Fellow (NSF 0841338), and the Luquillo Long-Term Ecological Research Program (NSF DEB-0620910). Thanks to Rita Cáceres, Angel Santiago, Augustin Engman, Emma Caraballo, and Samuel Burgos that provided field assistance, and Alex E. Mercado-Molina for providing useful comments on the earlier draft. In addition, we thank Sean Kelly and two anonymous reviewers for comments and suggestions. We thank W. H. McDowell, University of New Hampshire, for analyzing water and for providing input on their interpretation. Special thanks to the personnel at the Biotechnology lab, at Universidad del Turabo for training and technical assistance, in particular to Dr. Jose Perez-Jimenez, Lisabeth Duval, Carmen Bonilla, and Diana Liz Laureano. Special thanks are given to Miguel Acevedo and Claudia Patricia Ruíz Díaz for their guidance in model construction and evaluation.

References

  1. 1.
    Allan D (2004) Landscapes and riverscapes: the influence of land use on stream ecosystems. Annu Rev Ecol Evol Syst 35:257–284CrossRefGoogle Scholar
  2. 2.
    Allison S, Martiny JBH (2008) Resistance, resilience, and redundancy in microbial communities. Proc Natl Acad Sci 105(Supplement 1):11512–11519PubMedCentralCrossRefPubMedGoogle Scholar
  3. 3.
    Battin TJ, Kaplan LA, Newbold JD, Cheng X, Hansen C (2003) Effects of current velocity on the nascent architecture of stream microbial biofilms. Appl Environ Microbiol 69(9):5443–5452PubMedCentralCrossRefPubMedGoogle Scholar
  4. 4.
    Blackwood CB, Hudleston D, Zak DR, Buyer J (2007) Interpreting ecological diversity indices applied to terminal restriction fragment length polymorphism data: insights from simulated microbial communities. Appl Environ Microbiol 73(16):5276–5283PubMedCentralCrossRefPubMedGoogle Scholar
  5. 5.
    Boëchat IG, Krüger A, Giani A, Figueredo CC, Gücker B (2011) Agricultural land-use affects the nutritional quality of stream microbial communities. FEMS Microbiol Ecol 77(3):568–576CrossRefPubMedGoogle Scholar
  6. 6.
    Burgos-Caraballo S (2013) Function and diversity of benthic biofilm communities in tropical stream ecosystems: the relevance of land use and reach factors. Dissertation, University of Puerto Rico, Rio Piedras (Puerto Rico). ProQuest Dissertations and Theses, 110. Retrieved from http://search.proquest.com/docview/1286729784?accountid=44825. (1286729784)
  7. 7.
    Camargo JA, Alonso A, Salamanca A (2005) Nitrate toxicity to aquatic animals: a review with new data for freshwater invertebrates. Chemosphere 58(9):1255–1267Google Scholar
  8. 8.
    Carrino-Kyker SR, Swanson AK, Burke DJ (2011) Changes in eukaryotic microbial communities of vernal pools along an urban–rural land use gradient. Aquat Microb Ecol 62(1):13–24CrossRefGoogle Scholar
  9. 9.
    De Jesus-Crespo R, Ramirez A (2011) Effects of urbanization on stream physicochemistry and macroinvertebrate assemblages in a tropical urban watershed in Puerto Rico. J N Am Benthol Soc 30(3):739–750CrossRefGoogle Scholar
  10. 10.
    Dopheide A, Lear G, Stott R, Lewis G (2009) Relative diversity and community structure of ciliates in stream biofilms according to molecular and microscopy methods. Appl Environ Microbiol 75(16):5261–5272PubMedCentralCrossRefPubMedGoogle Scholar
  11. 12.
    Falkowski R (2007) Aquatic photosynthesis, 2nd edn. Princeton University Press, PrincetonGoogle Scholar
  12. 13.
    Fierer N, Morse JL, Berthrong ST, Bernhardt BS, Jackson RB (2007) Environmental controls on the landscape-scale biogeography of stream bacterial communities. Ecology 88(9):2162–2173CrossRefPubMedGoogle Scholar
  13. 15.
    Foley JA, DeFries R, Asner GP, Barford C, Bonan G, Carpenter SR, Chapin FS, Coe GC et al (2005) Global consequences of land use. Science 309(5743):570–574CrossRefPubMedGoogle Scholar
  14. 16.
    Fujioka RS, Shizumura LK (1985) Clostridium perfringens, a reliable indicator of stream water quality. J Water Pollut Control Fed 57(10):986–992Google Scholar
  15. 17.
    Grant A, Ogilvie LA (2003) Terminal restriction fragment length polymorphism data analysis. Appl Environ Microbiol 69(10):6342–6343PubMedCentralCrossRefPubMedGoogle Scholar
  16. 18.
    Helmer EH, Ramos O, Del M López T, Quiñones M, Díaz W (2002) Mapping the forest type and land cover of Puerto Rico, a component of the Caribbean biodiversity hotspot. Caribb J Sci 38(3–4):165–183Google Scholar
  17. 19.
    Hooper DU, Chapin FS III, Ewel JJ, Hector A, Inchausti P, Lavorel S, Lawton JH, Lodge D, Loreau M, Naeem S, Schmid B, Setälä H, Symstad AJ, Vandermeer J, Wardle DA (2005) Effects of biodiversity on ecosystem functioning: a consensus of current knowledge. Ecol Monogr 75(1):3–35CrossRefGoogle Scholar
  18. 20.
    Humber JF, Dorigo U, Cecchi P, Le Berre B, Debrosas D, Bouvy M (2009) Comparison of the structure and composition of bacterial communities from temperate and tropical freshwater ecosystems. Environ Microbiol 11(9):2339–2350CrossRefGoogle Scholar
  19. 21.
    Johnson JB, Omland KS (2004) Model selection in ecology and evolution. Trends Ecol Evol 19(2):101–108CrossRefPubMedGoogle Scholar
  20. 22.
    Kobayashi Y, Kim C, Yoshimizu C, Kohzu A, Tayasu I, Nagata T (2009) Longitudinal changes in bacterial community composition in river epilithic biofilms: influence of nutrients and organic matter. Aquat Microb Ecol 54(2):135–152CrossRefGoogle Scholar
  21. 23.
    Lear G, Boothroyd IKG, Turner SJ, Roberts K, Lewis SGD (2009) A comparison of bacteria and benthic invertebrates as indicators of ecological health in streams. Freshw Biol 54(7):1532–1543CrossRefGoogle Scholar
  22. 24.
    Lefrançois E, Coat S, Lepoint G, Vachiéry N, Gros O, Monti D (2011) Epilithic biofilm as a key factor for small-scale river fisheries on Caribbean islands. Fish Manag Ecol 18(3):211–220CrossRefGoogle Scholar
  23. 25.
    Legendre P, Legendre L (1998) Numerical ecology. 2nd English edn. Elsevier Science BV, Amsterdam, p XVGoogle Scholar
  24. 26.
    Lenat DR, Crawford JK (1994) Effects of land use on water quality and aquatic biota of three North Carolina piedmont streams. Hydrobiology 294(3):185–199CrossRefGoogle Scholar
  25. 27.
    Liu WT, Marsh TL, Cheng H, Forney LJ (1997) Characterization of microbial diversity by determining terminal restriction fragment length polymorphisms of genes encoding 16S rRNA. Appl Environ Microbiol 63(11):4516–4522PubMedCentralPubMedGoogle Scholar
  26. 29.
    Lock MA (1993) Attached microbial communities in rivers. In: Ford TE (ed) Aquatic microbiology: an ecological approach. Blackwell Scientific Publication, Oxford, pp 113–138Google Scholar
  27. 30.
    Lyautey E, Teissier S, Charcosset JY, Rols JL, Garabetian F (2003) Bacterial diversity of epilithic biofilm assemblages of an anthropised river section, assessed by DGGE analysis of a 16S rDNA fragment. Aquat Microb Ecol 33(3):217–224CrossRefGoogle Scholar
  28. 31.
    Magurran AE (2004) Measuring biological diversity. Blackwell Science, OxfordGoogle Scholar
  29. 32.
    McCune B, Grace JB (2002) Analysis of ecological communities. MjM Software, OregonGoogle Scholar
  30. 33.
    Mills MM, Moore CM, Langlois R, Milne A, Achterberg E, Nachtigall K, Lochte K, Geider RJ, La Roche J (2008) Nitrogen and phosphorus co-limitation of bacterial productivity and growth in the oligotrophic subtropical North Atlantic. Limnol Oceanogr 53(2):824–834CrossRefGoogle Scholar
  31. 34.
    Morgan RP, Cushman SF (2005) Urbanization effects on stream fish assemblages in Maryland, USA. J N Am Benthol Soc 24(3):643–655CrossRefGoogle Scholar
  32. 35.
    Mueller-Spitz SR, Stewart LB, Val-Kulmp J, McLellan SL (2010) Freshwater suspended sediments and sewage are reservoirs for enterotoxin-positive Clostridium perfringens. Appl Environ Microbiol 76(16):5556–5562PubMedCentralCrossRefPubMedGoogle Scholar
  33. 36.
    Osborn AM, Moore ERB, Timmis KN (2000) An evaluation of terminal-restriction fragment length polymorphism (T-RFLP) analysis for the study of microbial community structure and dynamics. Environ Microbiol 2(1):39–50CrossRefPubMedGoogle Scholar
  34. 37.
    Paul BJ, Duthie HC, Taylor WD (1991) Nutrient cycling by biofilms in running waters of differing phosphorus status. J N Am Benthol Soc 10(1):31–41CrossRefGoogle Scholar
  35. 38.
    Paul M, Meyer JL (2001) Streams in the urban landscape. Annu Rev Ecol Syst 32:333–365CrossRefGoogle Scholar
  36. 39.
    Perryman SE, Rees GN, Walsh CJ (2008) Analysis of denitrifying communities in streams from an urban and non-urban catchment. Aquat Ecol 42(1):95–101CrossRefGoogle Scholar
  37. 40.
    Puckett LJ (1994) Nonpoint and point sources of nitrogen in major watersheds of the United States. US Geological Survey. Water resources investigation Report 94–4001Google Scholar
  38. 41.
    Ramirez A, De Jesus-Crespo R, Martino-Cardona DM, Martinez-Rivera N, Burgos-Caraballo S (2009) Urban streams in Puerto Rico: what can we learn from the tropics? J N Am Benthol Soc 28(4):1070–1079CrossRefGoogle Scholar
  39. 42.
    Rees GN, Baldwin DS, Watson GO, Perryman S, Nielsen D (2004) Ordination and significance testing of microbial community composition derived from terminal restriction fragment length polymorphisms: application of multivariate statistics. Antonie Van Leeuwenhoek 86(4):339–347CrossRefPubMedGoogle Scholar
  40. 44.
    Rubin MA, Leff LG (2007) Nutrients and other abiotic factors affecting bacterial communities in an Ohio River (USA). Microb Ecol 54(2):374–383CrossRefPubMedGoogle Scholar
  41. 45.
    Shyu C, Soule T, Bent SJ, Foster JA, Forney LJ (2007) MiCA: a web-based tool for the analysis of microbial communities based on terminal-restriction fragment length polymorphisms of 16S and 18S rRNA genes. J Microb Ecol 53(4):562–570CrossRefGoogle Scholar
  42. 46.
    Szabo KE, Makk J, Kiss TK, Eiler A, Acs E, Toth B, Kiss AK, Bertilsson S (2008) Sequential colonization by river peryphyton analysed by microscopy and molecular fingerprinting. Freshw Biol 53(7):1359–1371CrossRefGoogle Scholar
  43. 47.
    Veve T, Taggart B (1996) Atlas of ground-water resources in Puerto Rico and the US Virgin Islands. U. S. Geological Survey, San JuanGoogle Scholar
  44. 48.
    Walsh CJ, Roy AH, Feminella JW, Cottingham PD, Groffman PM, Morgan RP II (2005) The urban stream syndrome: current knowledge and the search for a cure. J N Am Benthol Soc 24(3):706–723CrossRefGoogle Scholar
  45. 49.
    Wang SY, Sudduth EB, Wallenstein MD, Wright JP, Bernhardt ES (2011) Watershed urbanization alters the composition and function of stream bacterial communities. PLoS ONE 6(8):e22972PubMedCentralCrossRefPubMedGoogle Scholar
  46. 50.
    Washington VJ, Lear L, Neale MW, Lewis GD (2013) Environmental effects on biofilm bacterial communities: a comparison of natural and anthropogenic factors in New Zealand streams. Freshw Biol 58(11):2277–2286Google Scholar

Copyright information

© Springer Science+Business Media New York 2014

Authors and Affiliations

  • Sofía Burgos-Caraballo
    • 1
  • Sharon A. Cantrell
    • 2
  • Alonso Ramírez
    • 3
  1. 1.Department of BiologyUniversity of Puerto RicoSan JuanUSA
  2. 2.Department of BiologyUniversidad del TuraboGuraboUSA
  3. 3.Department of Environmental SciencesUniversity of Puerto RicoSan JuanUSA

Personalised recommendations