Impact of sideways and bottom-up control factors on bacterial community succession over a tidal cycle Ashvini Chauhan1, Jennifer Cherrier, and Henry N. Williams Environmental Sciences Institute, Florida A&M University, Frederick S. Humphries Science Research Building, Suite 305-D, Tallahassee, FL 32307
In aquatic systems, bacterial community succession is a function of top-down and bottom-up factors, but little information exists on ‘‘sideways’’ controls, such as bacterial predation by Bdellovibriolike organisms (BLOs), which likely impacts nutrient cycling within the microbial loop and eventual export to higher trophic groups. Here we report transient response of estuarine microbiota and BLO spp. to tidal-associated dissolved organic matter supply in a riverdominated estuary, Apalachicola Bay, Florida. Both dissolved organic carbon and dissolved organic nitrogen concentrations oscillated over the course of the tidal cycle with relatively higher concentrations observed at low tide. Concurrent with the shift in dissolved organic matter (DOM) supply at low tide, a synchronous increase in numbers of bacteria and predatorial BLOs were observed. PCR-restriction fragment length polymorphism of small subunit rDNA, cloning, and sequence analyses revealed distinct shifts such that, at low tide, significantly higher phylotype abundances were observed from ␥-Proteobacteria, ␦-Proteobacteria, Bacteroidetes, and high GⴙC Gram-positive bacteria. Conversely, diversity of ␣-Proteobacteria, -Proteobacteria, and ChlamydialesVerrucomicrobia group increased at high tides. To identify metabolically active BLO guilds, tidal microcosms were spiked with six 13C-labeled bacteria as potential prey and studied using an adaptation of stable isotope probing. At low tide, representative of higher DOM and increased prey but lower salinity, BLO community also shifted such that mesohaline clusters I and VI were more active; with an increased salinity at high tide, halotolerant clusters III, V, and X were predominant. Eventually, 13C label was identified from higher micropredators, indicating that trophic interactions within the estuarine microbial food web are potentially far more complex than previously thought. Bdellovibrio-like organisms (BLOs) 兩 dissolved organic matter 兩 predator-prey interactions 兩 stable isotope probing 兩 tidal microbiota
M
arine dissolved organic matter (DOM) is one of the largest active reservoirs of reduced carbon at the earth’s surface and, to a large extent, as the primary consumers of this DOM, bacteria control its fate via assimilation and/or remineralization processes (1, 2). The fate of DOM is a also a function of physiologic status and taxonomic composition of the autochthanous microbiota as well as the relative DOM lability supplied to the system, all of which vary both spatially and temporally in response to physiochemical conditions (1, 3, 4). DOM that is assimilated into bacterial biomass is potentially available for trophic transfer via the microbial loop (5) and as such must be accounted for in estimates of marine carbon flux. Bacterial groups that mineralize DOM are taxonomically diverse (2, 3, 6), which is often a function of niche variability (1–3). Specifically, estuarine systems exhibit high spatiotemporal and physiochemical variability, often resulting in short-lived blooms of some bacterial spp. (7). Among other factors, salinity has been found to typically drive bacterial succession in estuarine systems (3) such that in Chesapeake Bay, ␣-Proteobacteria were predominant in the saltwater regions and -Proteobacteria in the freshwater regions (8). Bacterial succession is also a function of bottom-up substrate supply (i.e., dissolved and particulate organic and inorganic nutrients) from autochthanous and allochthanous sources (4, 9). In www.pnas.org兾cgi兾doi兾10.1073兾pnas.0809671106
tandem, top-down factors shape bacterial community structure through a variety of processes, including protistan grazing (10) and viral lysis (11). Therefore, in all likelihood at any given time, combinations of these factors drive bacterial community succession in aquatic ecosystems (12–14). Recently, Mou et al. (2) proposed that in marine systems, transient changes in the dissolved organic carbon (DOC) pool are less critical in structuring bacterial communities than those that result from viral lysis, protistan grazing, or even physicochemical conditions. Both grazing and viral lyses are selective, such that factors including nonsusceptibility, morphology, size, and motility offer protection to certain bacterial groups (12). However, other ‘‘sideways’’ factors, which are only beginning to be understood, also likely contribute to shifts in the bacterial composition through processes that exert both positive (syntrophy) and/or negative (allelopathy) effects (15). In this context, one of the largely ignored trophic links within the microbial loop processes is obligate and relatively nonspecific predation by Bdellovibrio-like organisms (BLOs), resulting in potential structural and functional successions of susceptible prey microbiota. BLOs can lyse a variety of Gram-negative bacteria (16, 17) and are characterized by a motile free-living attack form and an intraperiplasmic growth phase. Our recent study indicated that BLOs are more diverse than previously thought (18); most marine bacteria are susceptible to lysis by these predators (16, 18) and hence their sideways trophic interactions would likely result in successions within the microbial food web processes. This study was conducted in Apalachicola Bay, a river-dominated subtropical estuary located in the Florida Panhandle (Fig. S1 A). A combination of riverine discharge, gulf tides, and winds keep this system well mixed, likely resulting in dynamic cycling of DOM and inorganic nutrients from both allochthanous (river, wetlands) and autochthanous (in situ production) sources. The overall goal of the work presented here was to evaluate how transient changes in both bottom-up factors (supply of bulk DOM, i.e., both DOC and dissolved organic nitrogen [DON]) and sideways factors transiently influence bacterial community composition and associated functional changes within the predacious BLO guilds. An improved understanding of these tightly coupled predator-prey interactions and the effects of DOM supply will lead to a better understanding of the trophic links within the microbial loop and recycling of nutrients in coastal systems. Results Environmental Parameters and Nutrients. Salinity at our study site indicated vertical stratification of water at both high tides but was
Author contributions: A.C., J.C., and H.N.W. designed research; A.C. and J.C. performed research; A.C. contributed new reagents/analytic tools; A.C. and J.C. analyzed data; and A.C., J.C., and H.N.W. wrote the paper. The authors declare no conflict of interest. This article is a PNAS Direct Submission. Data deposition: The 16S rRNA gene sequences reported in this paper have been deposited in GenBank under accession numbers FJ160298 –FJ160358 (bacteria) and FJ160359 – FJ160412 (BLOs). 13C prey bacteria are included under FJ160294 –FJ160297, M59161, and DQ912807. 1To
whom correspondence should be addressed. E-mail:
[email protected].
This article contains supporting information online at www.pnas.org/cgi/content/full/ 0809671106/DCSupplemental.
PNAS 兩 March 17, 2009 兩 vol. 106 兩 no. 11 兩 4301– 4306
ECOLOGY
Edited by David M. Karl, University of Hawaii, Honolulu, HI, and approved January 22, 2009 (received for review October 21, 2008)
193.00
10.00
Table 1. Relative bacterial and BLO phylotype abundances from samples collected over the tidal cycle at Dry Bar in Apalachicola Bay, Fl, in December 2006
8.00
Closest phylogenetic taxa/specie from NCBI
12.00
DOC Concentration ( μM)
189.00 187.00 185.00 183.00
6.00 181.00 179.00
DON Concentration ( μM)
191.00
4.00
177.00 175.00
2.00 23:20
2:15
3:50
5:20
8:18
10:50
12:20
14:20
Fig. 1. Changes in DOC (diamonds) and DON (triangles) concentrations vs. time over the course of the 12-h tidal cycle at Dry Bar in Apalachicola Bay, FL. Time points taken at 2320, 0350, 0818, 1220, and 1420 represent HT-I, OT, LT, IT, and HT-II, respectively. Error bars represent ⫾1 SD of duplicate samples.
uniformly mixed due to strong wind mixing at incoming tide (IT) and low tide (LT), respectively (Fig. S1 B and C). Small changes in salinity (19.5–21.8 ppt) were observed at 0.5 m over the course of the tidal cycle (Fig. S1C) and, for the most part, closely followed the changes in tide with the sharpest increase observed between LT and HT-II, when the winds subsided and the water became stratified again. No significant changes in temperature (15.4–16.5 °C) or dissolved oxygen (8.7–9.4 mg/L) were observed. DOC and DON concentrations oscillated over the course of the tidal cycle with relatively higher concentrations at LT (187 ⫾ 0.2 M C and 10 ⫾ 0.4 M N) than at outgoing tide (OT) or IT (Fig. 1). NO3 was between 2 and 5 M with lowest at low tide. NH4 concentrations remained below detection limit (data not shown). The DOC:TDN remained fairly constant over the 12-h sampling period between 15 and 17. The DOC:DON [or C:Norg], however, showed more variability, being appreciably lower at LT [19] than that observed at high tide [average of first and second high tides [HT-I and HT-II) ⫽ 24], OT [26], and IT [23]. Most Probable Number Estimates. Estimation of total bacteria and
BLOs indicated that bacteria in LT were ⬇20-fold greater than those at other time points (Fig. S1C). BLO community showed the same trend, such that higher numbers were observed at low tide. Phylogenetic Analyses of Microbiota over the Tidal Cycle. Microbial community structure over the course of the tidal cycle was assessed by polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) of the 16S rDNA. RFLP phylotypes were grouped into operational taxonomic units (OTUs) based on restriction patterns; 12 OTUs were identified at HT-I, 14 at OT, 15 at LT, and 10 each at IT and HT-II, respectively. BLO clone libraries consisted of 2 OTUs in HT-1, 1 each in OT and LT, 3 in IT, and 2 in HT-II. Rarefaction curves indicated that sufficient numbers of clones were sequenced to represent bacterial diversity (data not shown). Relative distribution of sequences within individual tidal clone libraries is presented in Table 1. At least 2 clones from each OTU were sequenced and taxonomically characterized by Basic Local Alignment Search Tool (BLAST). Dynamic bacterial community shifts were observed over the course of the tidal cycle mainly within ␣-, -, ␥-, and ␦Proteobacteria, Bacteroidetes, Chlamydiales-Verrucomicrobia group and Gram-positive bacteria, as shown in Fig. 2. Phylogenetic tree of microbiota identified over the tidal cycle is shown in Fig. S2. Among Proteobacteria, the most dominant group was ␣Proteobacteria that contributed from 50% to 79% of the total bacterial compositional makeup (Fig. 2 and Table 1). ␣Proteobacteria significantly increased at both high tides; species identified clustered mostly with Pelagibacter/uncultured SAR116 clade, Roseobacter spp., and Rhodobacteraceae family. An in4302 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0809671106
HT-I (%)
OT (%)
LT (%)
IT (%)
HT-II (%)
Uncultured Alpha-proteobacteria 12 3 Pelagibacter sp. 45 35 SAR116- like sp. 8 18 Uncultured Rhodobacteraceae sp. 6 0 Uncultured Roseobacter sp. 8 5 Rhizobium sp. 0 0 -Proteobacteria
6 29 0 4 11 0
0 42 0 17 0 6
0 54 2 0 15 0
Bacterial (prey) Communitya:
␣-Proteobacteria
Uncultured Comamonadaceae Burkholderia sp. Limnobacter sp. Ralstonia sp.
8 0 0 3 ␥-Proteobacteria
2 0 0 0
0 0 0 0
0 0 6 0
0 2 0 8
Uncultured Gamma-proteobacteria 3 Oleiphilus sp. 1 ␦-Proteobacteria
18 0
22 3
21 0
4 0
0
4
0
0
Uncultured Flavobacterium sp. 4 12 7 Uncultured Bacteroides sp. 0 5 12 Chlamydiales-Verrucomicrobia group
4 4
0 11
Uncultured Verrucomicrobia sp. High- G ⫹ C Gram-positive bacteria: Uncultured Actinobacterium sp.
Uncultured Desulfobulbus sp.
0 Bacteroidetes
2
2
0
0
4
0
0
2
0
0
0 0 0 100 0
0 0 100 0 0
0 0 36 16 48
0 76 24 0 0
Predator (BLO) Communityb:
␦-Proteobacteria Bacteriovorax Cluster III Bacteriovorax Cluster V Bdellovibrio Cluster VI Bacteriovorax Cluster X Bacteriovorax Cluster XII
28 0 0 72 0
aForty-eight
bacterial clones with positive inserts were compared from each library. bTwenty-five BLO clones with positive inserts were compared from each library. All libraries were subjected to Rarefaction analyses to establish that sufficient numbers of clones were sequenced to represent microbial diversity over the tidal cycle. SSU rDNA from at least two representatives of each phylotype were sequenced and compared to the closest cultivable phylogenetic relative from NCBI database.
crease, albeit at low levels, was also shown by -Proteobacteria and Chlamydiales-Verrucomicrobia groups at HT. At LT, an increased representation of ␥-Proteobacteria (25%) and Bacteroidetes (19%) were observed with decline in species belonging to ␣-Proteobacteria (50%); -Proteobacteria remained undetectable at this time. Unique groups identified only at LT clustered with ␦-Proteobacteria (uncultured Desulfobulbus spp., 4%), and high G⫹C Gram-positive bacteria (uncultured Actinobacterium spp., 2%); Chlamydiales-Verrucomicrobia group was absent. Community shifts were also evident within the predacious BLOs over the tidal cycle (Table 1), such that clusters III (28%) and X (72%) were identified in HT-I, cluster X (100%) in OT, and cluster VI (100%) at LT. At incoming tide, BLOs were more variable, with representations from clusters VI (36%), X (16%), and XII (48%). HT-II consisted of clusters V (76%) and VI (24%), respectively. Identification of Functionally Active BLO Guilds. Simultaneous to
assessing the bacterial diversity, tidal microcosms were set up from HT-I, OT, LT, IT, and HT-II and spiked with six 13C-labeled Chauhan et al.
Alpha-proteobacteria
Beta-proteobacteria
Gamma-proteobacteria
Delta-proteobacteria
Bacteroidetes
Chlamydiales-Verrucomicrobia Group
14
35
12
30
10
25 8 20 6 15 4 10 2
5
100 0
0 4:00 (OT )
8:30 (LT )
11:30 (IT )
14:30 (HT -II)
Fig. 2. Bacterial community shifts representing major phyla/taxa identified over the course of the 12-h tidal cycle at Dry Bar in Apalachicola Bay, FL, analyzed by PCR-RFLP of small subunit rDNA followed by sequence analyses. Time points at 2320, 0350, 0818, 1220, and 1420 represent HT-I, OT, LT, IT, and HT-II, respectively.
␥-Proteobacterial species as potential prey to identify metabolically active predatory BLOs over the tidal cycle. Prior to establishment of 13C-microcosms, labeled and unlabeled prey bacteria were separately diluted to extinction and DNA analyzed by ultracentrifugation to confirm sufficient labeling of cells had occurred such that DNA from terminally diluted series of each labeled prey was identified in the heavier bands but not in the lighter bands (data not shown). Species from ␥-Proteobacteria were chosen as prey for stable isotope probing (SIP) studies due in part to being preferred by Bacteriovorax (Bx) spp. (16, 18, 19). Moreover, ␥-Proteobacterial numbers were also found to be significantly higher at LT (Fig. 2 and Table 1). Other prey tested included Vibrio, Pseudoalteromonas, and Marinomonas, identified as potential prey from our previous studies in Apalachicola Bay (19), and 3 laboratory strains— Pseudomonas sp., E. coli sp., and Vibrio parahaemolyticus P5. Initial numbers and viability of the combined 13C-labeled prey in each microcosm was ⬇1.5 ⫻ 104/mL, which was within the range measured in the tidal samples (Fig. S1C). Most probable numbers (MPNs) measured every 24 h from the 13C-microcosms indicated rapid predation such that numbers of prey steadily declined with concomitant increase in BLO numbers (data not shown). Samples collected from 24 to 144 h were studied by SIP. Concomitant with the depletion of RFLPs representing each of the 13C-labeled prey spp., ‘‘new’’ OTUs were observed in the ‘‘heavier’’ clone libraries, with significant differences shown by LT and HT-II microcosms (Fig. S3 A and B). Sequencing of the new OTUs led to identification of BLOs that had predated and assimilated DNA from the 13Clabeled prey, providing clues on both BLO structure and function along the tidal cycle. Specifically, in LT samples, BLO clusters I, II, IV, V, and VI were found in the labeled fraction; HT-II samples contained clusters III, V, VI, and X (Fig. 3). Further, cluster VI in LT and cluster V in HT-II were the most metabolically actively BLOs based on the dominance of their specific RFLPs (Fig. S3 A and B). Moreover, this adaptation of SIP confirmed that BLOs likely have feeding preferences for certain bacteria, as discussed hereafter. Statistical Comparisons of the Tidal Microbiota. Principal coordinates
analyses (PCA) was performed on bacterial and BLO species identified from HT, OT, LT, IT, and HT-II to determine statistical differences. Bacteria identified from HT-I, OT, and IT clustered on the same axis, whereas LT and HT-II microbiota separated out on different axes (Fig. S4A). Similarly, BLOs from HT-I, IT, and LT clustered on different axes, but OT and HT-II appeared on the same axis (Fig. S4B). For bacteria, PCA axis 1 explained 46.93%, and axis 2, 24.33% of variability, with a cumulative percentage of 71.26%. For BLOs, PCA axis 1 explained 45.94%, and axis 2, Chauhan et al.
Cluster I
Cluster XII
10 changes
Fig. 3. Phylogenetic tree of partial 16S rRNA gene sequences of predatory Bdellovibrio-like organisms (BLOs)/Bx spp., over the course of the 12-h tidal cycle at Dry Bar in Apalachicola Bay, FL. The phylotree was constructed with PAUP v. 4.0b8 using maximum parsimony algorithm. Clones marked in boldface represent species that were identified from the ‘‘heavier’’ DNA from the 13C prey studies. Numbers at nodes represent bootstrap values (100⫻ resampling analysis); only values ⬎50 are presented. Geobacter metallireducens was used as outgroup. In this tree, BLO clusters I and VI represent the freshwater/terrestrial BLO species, and clusters III, IV, V, X, and XII represent marine/estuarine relatives. Recently, clusters I and VII have been proposed to belong to Peredibacteraceae family.
30.29% of the variability, with a cumulative percentage of 76.23%, indicating a statistical distinction of microbiota as a function of tidal cycle. Further, significant differences (P ⬍ 0.001) of clone library sequences for both bacteria and BLOs indicated that predator-prey dynamics were significantly different at LT. Discussion In aquatic systems, both top-down (protistan grazing, viral lyses) and bottom-up (nutrient supply) factors significantly influence bacterial succession. Little information exists, however, on trophic interactions of predacious Bdellovibrio-like organisms, resulting in a potential sideways control on the structure and functions of the predated microbiota. In all likelihood, biotic factors work in tandem with abiotic factors such as salinity, temperature, water resident time, and hydrology, resulting in bacterial successions in aquatic systems (4, 8, 11–15). However, thus far, studies have not accounted for the cumulative influences of transient physicochemical changes along with bottom-up and sideways controls on short-term bacterial community succession. To this end, we conducted a 12-h tidal study in Apalachicola Bay, FL, a river-dominated estuary. Estuarine systems are unique such that coexistence between freshwater and marine bacterial ecotypes has been reported (3, 8), with dramatic shifts within these assemblages (8, 20). Bacterial groups identified at our study site during the 12-h period clustered with Proteobacteria (␣-, -, ␥-, and ␦-Proteobacteria), Bacteroidetes, Verrucomicrobia, and high G⫹C Gram-positive bacteria (Fig. 2 and PNAS 兩 March 17, 2009 兩 vol. 106 兩 no. 11 兩 4303
ECOLOGY
40
Phylotype Distribution of other Bacterial Groups
Alpha-Proteobacterial Phylotype Distribution
High G+C Gram-positive Bacteria
23:30 (HT -I)
Bacteriovorax sp. MNZ4 Bdellovibrio sp. SJ HT-I-8 FJ160360 HT-II-Labeled1 FJ160361 Cluster III HT-II-Labeled2 FJ160362 HT-I-19 FJ160367 70 IT-8 FJ160396 IT-16 FJ160399 65 HT1-100 FJ160392 HT-II-Labeled17 FJ160400 HT-II-101 FJ160359 HT-II-Labeled4 FJ160365 HT-II-Labeled5 FJ160366 HT-II-118 FJ160364 HT-II-115 FJ160363 100 Cluster V HT-II-6 FJ160405 HT-II-Labeled19 FJ160409 HT-II-15 FJ160406 HT-II-Labeled3 FJ160407 100 LT-Labeled7 FJ160408 Bacteriovorax sp. PS23S Bdellovibrio sp. JS10 OT-1 FJ160368 HT-1-95 FJ160370 75 OT-14 FJ160371 OT-17 FJ160373 Cluster X OT-3 FJ160369 HT-1-90 FJ160372 HT-II-Labeled7 FJ160412 Bacteriovorax sp. GSL41 IT-14 FJ160398 100 HT-II-Labeled100 FJ160393 IT-3 FJ160394 100 Bdellovibrio sp. clone YE-3E 95 Bacteriovorax sp. OC91 Cluster IV Bdellovibrio sp. clone CA-1F LT-Labeled5 FJ160383 LTLabeled6 FJ160384 LT-8 FJ160382 LT-1 FJ160374 LT-10 FJ160387 LT-Labeled17 FJ160389 80 IT-6 FJ160395 IT-12 FJ160397 LT-2 FJ160375 90 LT-Labeled18 FJ160390 LT-4 FJ160378 100 HT-II-1 FJ160402 Cluster VI HT-II-2 FJ160403 100 HT-II-Labeled6 FJ160404 LT-Labeled8 FJ160377 LT-3 FJ160376 LT-Labeled9 FJ160385 LT-9 FJ160386 LT-14 FJ160388 Bacteriovorax sp. EEC 100 Bacteriovorax sp. ETC 100 Bacteriovorax sp. F2 LT-Labeled3 FJ160410 Cluster II LT-Labeled4 FJ160411 LT-Labeled2 FJ160380 100 Bdellovibrio sp. Gunpowder 100 LT-Labeled1 FJ160379 IT-19 FJ160401 100 IT-1 FJ160391 Bacteriovorax sp. 16SPR3 Geobacter metallireducens 60
100
Table 1). For the most part, various groups did coexist throughout the course of the tidal cycle, but clear shifts in relative contribution of each group(s) to the compositional makeup of the community were observed concurrent with oscillations in DOM supply, and salinity, to a lesser extent. We did not directly measure protistan grazing and viral lyses but, obviously, both factors also likely contributed to some of the bacterial shifts observed over the tidal cycle. This study has potential implications on short-term processing of carbon, through the microbial food web in estuarine systems. Data collected from other stations and seasons in Apalachicola Bay (Carrabelle River, St. Vincent Sound, and Platform Bar) supported the conclusions drawn from this study such that regardless of spatiotemporal effects, lower tides consisted of substantially higher bacterial numbers and diversity, indicated by RFLP analyses (data not shown). Of major interest to our findings were the distinct differences of bacterial communities observed at LT (Table 1, Fig. 2, and Fig. S4A). At low tide, an increased representation of species belonging to ␥-Proteobacteria (25%) and Bacteroidetes (19%) were observed with significant decline of species representing ␣-Proteobacteria (50%). Surprisingly, at LT, -Proteobacteria, known to dominate in freshwater end-members of estuarine systems (8, 9), were absent. The wind mixing and subsequent loss of vertical stratification at LT (Fig. S1 B) may have diminished the -Proteobacterial species; we would have otherwise observed had the water column not been well mixed. Conversely, at both high tidal points, ␣-Proteobacteria was the dominant group, comprising greater than 70% of the community concomitant with increased salinity levels (Fig. 2 and Fig. S1B). Our data are in good agreement with previous findings that ␣-Proteobacteria have a propensity to thrive in lower-nutrient conditions, as indicative in the higher-tide events. Conversely, ␥-Proteobacteria species, being opportunistic, can rapidly use pulses of nutrients such as those after bloom events (21, 22) and the low tidal event in this study (Figs. 1 and 2). Additionally, at LT, Bacteroidetes and high G⫹C Gram-positive bacteria also increased (Table 1 and Fig. 2); Bacteroidetes are also known to respond rapidly and remineralize complex and labile DOM (1, 6, 23). Low phylotype abundances of Desulfobulbus sp., belonging to ␦-Proteobacteria were also observed at LT (Fig. 2 and Table 1). Wind-mixing likely resulted in vertical immigration of Desulfobulbus sp. from the sediments rather than allochthanously from the river plume because sediments are well-known reservoirs of such anaerobes (24). Bacterial species identified from HT samples clustered mainly with Pelagibacter sp./SAR11 clade, which are one of the most abundant bacteria in marine systems (25). Interestingly, Roseobacter spp., because of their propensity to respond rapidly to patchiness of nutrient pulses in coastal environments (25, 26), showed very tight coupling to the concentrations of oscillating DOC at high tides (Fig. 1 and Table 1). Further, concurrent with the distinct bacterial community shifts, we also measured a 7- to 9-M C net increase of N-rich DOM over OT and IT, which likely led to a 20-fold increase in bacterial numbers at LT (Fig. S1C). DOM quantity alone, however, is not sufficient to explain the bacterial response, as LT values (187 ⫾ 0.2 MC) were not significantly different to those observed at HT-I or HT-II (187 ⫾ 4.4 and 184 ⫾ 0.1 MC, respectively). Therefore, in all likelihood, the observed bacterial community response is also a function of both the quantity and quality of DOM supply to the Bay. This is further supported by previous studies where -Proteobacteria and Bacteriodetes were predominantly identified from waters with variable DOM concentrations and complexity (1, 27). Conversly ␣-Proteobacteria dominated in both less complex and lower concentrations of DOM, such as those from algal-derived substrates (1, 9). Further, in estuarine systems, several source terms, both autochthanous and allochthanous, can contribute to the ambient DOM pool. Over the short term (i.e., on the order of hours), however, the relative contribution of each of these sources will change as a function of tidal stage (i.e., high-tide autochthanous, 4304 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0809671106
low-tide allochthanous). This change in DOM substrate supply is also likely to influence the bacterial community composition shifts observed at LT. Because DOM produced autochthanously via in situ plankton production is generally thought to be more bioavailable than that originating from terrestrial sources, due in part to its higher nitrogen content as well as its lower degree of complexity (28), changes in the carbon-to-nitrogen ratio of ambient DOM, specifically the C:Norg, could potentially be indicative of shifts in the relative contribution of the autochthanous and allochthanous DOM source terms over the course of a tidal cycle. In addition to evaluating the C:Norg of the ambient DOM over the course of the tidal cycle, we also evaluated the C:Norg of DOM for the marine and freshwater DOM source terms to Apalachicola Bay. For the marine source terms, samples were collected from West Pass (Fig. S1 A), one of the primary conduits of exchange between Gulf of Mexico and Apalachicola Bay waters, both at incoming tide (i.e., marine end-member, C:Norg ⫽ 18) and outgoing tide (i.e., Apalachicola Bay signature, C:Norg ⫽ 25). For the freshwater source term, a water sample was collected upriver and adjacent to the Apalachicola marsh system (Fig. S1 A), and this DOM was found to be essentially deplete in nitrogen (C:Norg ⫽ 338). The average C:Norg of DOM at HT was 24, which was approximately equal to that observed at West Pass during outgoing tide ⫽ 25, and, as would be expected in an estuarine system, is indicative of some degree of mixing between autochthanous (C:Norg ⫽ 18) and allochthanous (C:Norg ⫽ 338) DOM source terms. However, if the DOM shaping bacterial community structure at low tide was coming primarily from a freshwater source (i.e., river and marsh outwelling), then we would expect the LT C:Norg of the DOM to be at least higher than the average HT C:Norg of 24 rather than lower as was observed (LTC:Norg ⫽ 19). Furthermore, the taxonomic composition of the bacterial community during LT, though distinctly different from other tidal points, was not necessarily exclusively indicative of a freshwater source, which would be expected to be dominated by -Proteobacteria (1). In contrast, C:Norg of the sediment porewater (0–10 cm from data collected from a previous study) from the same study site was found to be 15. The observed C:Norg decrease of DOM from 26 at IT to 19 at LT and the increased representation of the ␥-Proteobacteria, ␦-Proteobacteria, Bacteroidetes, and high G⫹C Gram-positive bacteria identified at this time (Fig. 2 and Table 1) might therefore be due to, among other factors, the combined effects of winds and tidal pumping and mixing of N-rich sediment DOM into the overlying water column. Therefore, bottom-up processes are likely fueled by cumulative impacts at low tide in shallow estuarine systems, which appear to be significant drivers for surface-water bacterial productivity and community composition. The same might not hold true, however, for a stratified water column where freshwater allochthanous DOM sources may be more important. In light of the recent evidence from marine systems that topdown factors (predation/grazing) are significant drivers of bacterial succession, as opposed to the bottom-up factors (transient changes in carbon pool) (2), we next sought to study the sideways trophic interactions of predacious BLO guilds on tidal microbiota. Specifically, at low tide, an increase in the numbers and diversity of potential prey bacteria resulted in a surprisingly synchronous response in BLO numbers (Fig. S1C). Higher bacterial and virioplankton abundance in lower tides compared with high tides has previously been reported (29), and our study suggests that a similar relationship exists between bacterial predators and prey. An increased predatorial response at LT more than likely is a function of increased numbers of prey species, especially because BLOs thrive at higher prey densities of 105⫺106/mL (16). Although BLO numbers represent only those guilds able to predate V. parahaemolyticus P5 used in our assay, it did indicate an active sideways control mechanism over the tidal cycle, which was further studied. As expected in an estuary, BLO community also varied from a combination of saltwater/halotolerant clusters that were identified Chauhan et al.
Chauhan et al.
soils and wetland sediments (34, 35); it is possible that these groups represent a secondary trophic tier of predatory bacteria within the microbial loop that rely on gliding mechanisms for predation. Rigorous confirmation must wait before the role(s) of such secondary predators can be established in marine systems. Because BLOs are host dependent and cannot replicate in extracellular environments, we do not expect cross-feeding on labeled by-products or dead labeled biomass, resulting in labeling of nonactive BLOs as a limitation to this adaptation of SIP. However, a priori selection of prey is likely a limitation, and results must be interpreted cautiously. Additionally, the latter incubations may represent enrichments on the 13C biomass. Our attempts to avoid these limitations included spiking 13C biomass that was well within the range estimated over the tidal cycle and establishment of several SIP microcosms that resulted in a time-dependent identification of metabolically active BLO/Bx guilds over the tidal cycle. Collectively, the findings presented here suggest a dynamic interplay between short-term changes to both bottom-up and sideways trophic interactions within the estuarine microbial food webs, complexities of which are only beginning to be understood. Materials and Methods Site Description and Sample Collection. The sampling site (Dry Bar, 29° 40.425⬘ N, 85° 03.406⬘ W), is located within Apalachicola River’s hydrologic discharge channel just southwest of the river mouth (Fig. S1 A). The total water-column depth at the study site was 3 m. At each tidal time point, temperature, salinity, and dissolved oxygen were measured by YSI probe (YSI Inc.) at 0.25-m depth intervals to evaluate the degree of stratification. Simultaneously, discrete and replicate samples were collected using a Geotech peristaltic pump with acidleached (10% HCl) Teflon tubing from 0.5 m depth to monitor for changes in dissolved organic carbon and nitrogen, nitrate, nitrite, and ammonium. For microbial analyses, 2 L of sample was collected over the tidal cycle, and 50-mL samples were processed through 0.8-m filters for MPNs. Further analyses are described in SI Text. Biogeochemical Methods. Dissolved organic carbon (DOC) concentrations were measured using a Shimadzu TOC-VCPH with modifications (36), as described in SI Text. Total dissolved nitrogen (TDN) was measured using a Shimadzu TNM-1, as described (37). DOC and TDN results were referenced against the materials obtained from the Rosenstiel School of Marine and Atmospheric Sciences (SI Text). Ammonium concentrations were determined colorimetrically (38). Dissolved organic nitrogen (DON) was determined by subtracting the sum of inorganic nitrogen constituents from TDN. Most Probable Numbers (MPNs) of Total Bacteria and BLOs. Three-tube dilution MPNs assays were performed as described (SI Text). DNA Extraction, Purification, and PCR Amplifications. DNA was extracted using UltraClean kit (Mo Bio), with final elution in sterile PCR-grade water (SI Text). Primers 27F (5⬘-AGAGTTTGATCCTGGCTCAG-3⬘)-1492R (5⬘- GGCTACCTTGTTACGACTT -3⬘) (39) were used to amplify total bacteria; Bac676F (5⬘-ATTTCGCATGTAGGGGTA-3⬘)-Bac1442R (5⬘-GCCACGGCTTCAGGTAAG-3⬘) and Bd529Fd (5⬘GGTAAGACGAGGGATCCT-3⬘)-Bd1007R (5⬘-TCTTCCAGTACATGTCAAG-3⬘) were used for BLO diversity (30). For samples that failed after PCR or gave weak amplicons, a seminested approach was followed (SI Text). Cloning of 16S rDNA and RFLP Analyses. Cloning of the 16S rDNA was performed using fresh PCR products ligated into pCRII-TOPO vector and transformed into E. coli TOP10F⬘ (Invitrogen) (refs. 19 and 35; SI Text). RFLPs consisted of separately digesting 10 L of amplicons using HhaI, MsPI, and HaeIII as previously reported (30) and confirmed by in silico analysis using CloneMap v2.11 software (CGC Scientific Inc.). RFLPs were run in 2.5% agarose gels. Clone libraries were analyzed by aRarefactWin 1.3 (http://www.uga.edu/⬃strata/software/) to confirm that sufficient clones were sequenced to identify tidal cycle diversity. Establishment of 13C Microcosms. Six potential prey included Vibrio sp., Pseudoalteromonas sp., Marinomonas sp., Pseudomonas putida, E. coli ML35, and Vibrio parahaemolyticus P5. Prey cells were isotopically labeled by growth in ISOGRO-13C Powder-Growth Medium 99 atom % 13C (Isotec). For estuarine prey, media was formulated at a salinity of 15 ppt, which was within the range measured in the samples. 13C-labeled prey were harvested at late log phase, PNAS 兩 March 17, 2009 兩 vol. 106 兩 no. 11 兩 4305
ECOLOGY
at other tidal points to a single freshwater cluster in low tide concomitant with a salinity drop (Fig. S1B, Fig. 3, and Table 1). However, using an adaptation of SIP, a better correlation of the structure to the function of predacious BLO community over the tidal cycle was achieved such that BLO sequences from the ‘‘heavier’’ clone library were significantly more diverse than those identified directly from the environmental DNA (Fig. 3 and Table 1). Because the numbers of native BLO populations in the environment are low (30, 31), most previous studies on BLO diversity using a culture based method or environmental DNA are likely flawed. This study therefore successfully demonstrated the use of an adaptation of SIP to trace the trophic flow of carbon from prey into metabolically active bacterial predator guilds in the estuarine microbial food web. To date, taxonomy of BLOs remains to be fully resolved. For the most part, clusters III, IV, V, IX, X, XI, XII, and XIII have been assigned to marine/estuarine Bx species, and clusters I, II, VI, VII, and VIII to the freshwater/terrestrial BLOs with 3 outlier isolates (GSL371, NZ7, and IP1) (18, 30); recently, clusters I and VII have been proposed to belong to Peredibacteraceae (32). Using SIP, we observed significant differences between metabolically active BLO clusters at LT and HT-II (Fig. S3 A and B), such that at lower salinity, a mix of freshwater clusters I and VI were found in the labeled fraction along with halotolerant clusters II, IV, and V (Fig. 3). Concomitant with an increase in salinity at HT-II, BLO community was representative of saltwater/halotolerant Bx clusters III, V, and X along with the freshwater cluster VI. Further, timedependent analyses indicated that most abundant OTU in LT heavier library was the freshwater cluster VI, whereas in HT-II saltwater Bx cluster V predominated, which has thus far been recovered only from low-salinity regions of Chesapeake Bay and Apalachicola Bay systems (16, 18, 19). Therefore, functional successions within the predatorial guilds must be driven by specific niches, including diversity of prey community. In all likelihood, an increase in the preferred ␥-Proteobacterial prey during lower tide resulted in the synchronous increase in BLO diversity. This was further supported statistically; both bacteria (prey) and BLO communities identified were significantly different at LT (P ⬍ 0.001; Fig. S4). The degree to which BLO predation rates may have been influenced by protistan grazing or viral lysis over the tidal cycle remains tentative. Among a suite of predatory arsenals, motility speeds of up to 160 m/s and small size of ⬇0.2–0.5 wide, 0.5–2 m long (16) should lead to feeding failure by protists (10, 12). Also, unlike phage/protist predation, where prey size and physiology can influence mortality efficiency, BLO predation appears to be rather nonspecific (16, 18, 30), with the caveat that smaller prey yield lower numbers of progeny cells (16, 17). Thus BLO predatory rates are more likely to be a function of prey cell numbers, as high metabolic activity of free-living BLOs results in rapid starvation if preys are not encountered. The SIP studies also corroborate previous culture-based studies that BLOs have predatorial preferences for Vibrio spp. (16, 18) based on the comparatively rapid decline of Vibrio RFLPs, even more so for Vibrio sp. isolated from the bay (Fig. S3 A and B), indicating that autochthanous bacteria may serve as more lucrative prey than laboratory strains. Sequencing performed on the RFLP phylotypes further confirmed the taxonomic affiliations of depleting prey and OTUs belonging to predacious BLOs. It appears that cell wall surfaces of susceptible prey may contain motifs or receptor sites that are recognized by BLOs early in the predatorial response. Conversely, Marinomonas remained resistant to predation, likely due to mechanisms such as phenotypic plasticity of prey (33). Notably, after a week, Myxococcales sp., Bacteroides sp., and Stenotrophomonas sp., were also found in the 13C-DNA, albeit at lower numbers (data not shown). These genera have been identified as bacterial micropredators in previous SIP studies carried out with
washed with ASW (18), diluted, and immediately spiked into microcosms at ⬇2.5 ⫻ 103/mL (direct counts and viability confirmed by MPNs) containing samples collected at HT-I, OT, IT, LT, and HT-II. All incubations were at ambient temperature. At 0, 24, 48, 72, 96, 120, and 144 h, MPNs were performed to confirm depletion of prey and increase of BLOs. Sample (50 mL) was also collected at every 24 h and studied by SIP. Separation of 13C-DNA from 12C-DNA. To negate the possibility of 12C-DNA carryover into the 13C-DNA following ultracentrifugation, 200 ng of unlabeled archaeal DNA isolated from M. thermophila TM-1 (ATCC 43570) was added to the environmental DNA and subjected to CsCl-ethidium bromide density gradient centrifugation in a VTI 65.2 rotor at 55,000 rpm for 18 h at 20 °C, as previously described (34, 35). Bands were visualized with UV lamp (365 nm); separation was observed between lighter and heavier DNA bands. The lower bands were extracted and recentrifuged for additional purification. CsCl/EtBr were removed by standard methods; DNA was concentrated by Centricon (Millipore Corp.) and resuspended in 100 L of PCR grade water. Purity of labeled DNA was checked by the presence of spiked archaeal DNA, which was not detected in the heavier 13C-DNA fractions, but detected in all of the lighter 12C-DNA fractions (data not shown), indicating purity of labeled DNA.
1. Cottrell MT, Kirchman DL (2003) Contribution of major bacterial groups to bacterial biomass production (thymidine and leucine incorporation) in the Delaware estuary. Limnol Oceanogr 48:168 –178. 2. Mou X, Sun S, Edwards RA, Hodson RE, Moran MA (2008) Bacterial carbon processing by generalist species in the coastal ocean. Nature 451(7179):708 –711. 3. Rappe´ MS, Vergin K, Giovannoni SJ (2000) Phylogenetic comparisons of a coastal bacterioplankton community with its counterparts in open ocean and freshwater systems. FEMS Microbiol Ecol 33:219 –232. 4. Cherrier J, Bauer JE (2004) Bacterial utilization of transient plankton-derived dissolved organic carbon and nitrogen inputs in surface ocean waters. Aquat Microb Ecol 35:229 –241. 5. Azam F (1998) Microbial control of oceanic carbon flux: The plot thickens. Science 280:694 – 696. 6. Cottrell MT, Kirchman DL (2000) Natural assemblages of marine proteobacteria and members of the Cytophaga-Flavobacter cluster consuming low- and high-molecularweight dissolved organic matter. Appl Environ Microbiol 66:1692–1697. 7. Piccini C, Conde D, Alonso C, Sommaruga R, Pernthaler J (2006) Blooms of single bacterial species in a coastal lagoon of the southwestern Atlantic Ocean. Appl Environ Microbiol 72:6560 – 6568. 8. Bouvier TC, del Giorgio PA (2002) Compositional changes in free-living bacterial communities along a salinity gradient in two temperate estuaries. Limnol Oceanogr 47:453– 470. 9. Kirchman DL, Dittel AI, Findlay SEG, Fisher D (2004) Changes in bacterial activity and community structure in response to dissolved organic matter in the Hudson River, New York. Aquat Microb Ecol 35:243–257. 10. Ju¨rgens K, Matz C (2002) Predation as a shaping force for the phenotypic and genotypic composition of planktonic bacteria. Antonie Van Leeuwenhoek 81:413– 434. 11. Bouvier T, del Giorgio PA (2007) Key role of selective viral-induced mortality in determining marine bacterial community composition. Environ Microbiol 9:287–297. 12. Pernthaler J (2005) Predation on prokaryotes in the water column and its ecological implications. Nat Rev Microbiol 3:537–546. 13. Zhang R, Weinbauer MG, Qian PY (2007) Viruses and flagellates sustain apparent richness and reduce biomass accumulation of bacterioplankton in coastal marine waters. Environ Microbiol 9:3008 –3018. 14. Strom SL (2008) Microbial ecology of ocean biogeochemistry: A community perspective. Science 320:1043–1045. 15. Fuhrman JA, Hagstrom A (2008) Bacterial and archaeal community structure and its pattern. Microbial Ecology of the Oceans, ed Kirchman DL (Wiley, New York), 2nd Ed. 16. Williams HN, Pin˜eiro S (2006) Ecology of the predatory Bdellovibrio and like organisms. Predatory Prokaryotes, ed Jurkevitch E (Springer, New York), pp 214 –244. 17. Jurkevitch E (2000) The genus Bdellovibrio. The Prokaryotes: An Evolving Electronic Resource for the Microbiological Community, ed Dworkin M, et al. (Springer, New York), 3rd Ed, release 3.7, November 2, 2001. 18. Pin˜eiro SA, et al. (2007) Global survey of diversity among environmental saltwater Bacteriovoracaceae. Environ Microbiol 9:2441–2450. 19. Chauhan A, Williams HN (2006) Response of Bdellovibrio and like organisms (BALOs) to the migration of naturally occurring bacteria to chemo attractants. Curr Microbiol 53:516 –522. 20. Kisand V, Andersson N, Wikner J (2005) Bacterial freshwater species successfully immigrate to the brackish water environment in the northern Baltic. Limnol Oceanogr 50:945–956. 21. Fandino LB, Riemann L, Steward GF, Long RA, Azam F (2001) Variations in bacterial community structure during a dinoflagellate bloom analyzed by DGGE and 16S rDNA sequencing. Aquatic Microbial Ecology 23:119 –130. 22. Stoica E, Herndl GJ (2007) Bacterioplankton community composition in nearshore waters of the NW Black Sea during consecutive diatom and coccolithophorid blooms. Aquat Sci 69:413– 418.
4306 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0809671106
DNA Sequencing and Phylogenetic Analysis. RFLPs were tentatively assigned to operational taxonomic units (OTUs); 2 clones from each OTU were sequenced at Florida State University with 27F/Bac676F/Bd529Fd primers. Chimera evaluation was performed via Bellerophon (40). Sequences were compared by BLAST (41) and aligned with ClustalX v. 1.8 (42). Evolutionary relationships among taxa were inferred using maximum parsimony by PAUP v. 4.0b8 (Sinauer Associates). Bootstrap resampling analysis for 100 replicates was done to estimate confidence of tree topologies. Statistical Analyses. Bacterial and BLO sequences generated from HT, OT, LT, IT, and HT-II were statistically analyzed using UniFrac (ref. 43 and SI Text). Comparative analyses were run to test which environments significantly differed using P test, UniFrac metric test, and PCA with the scatter plot option. ACKNOWLEDGMENTS. We thank G. Fortenberry and P. Jasrotia for technical assistance; Apalachicola National Estuarine Research Reserve (ANERR) staff for help with sample collection; and Dr. A. Ogram (University of Florida, Gainesville) for M. thermophila TM-1. This study was funded by Grants HRD-0531523 (HBCURISE) from the National Science Foundation and NA17AE1624 (EPP) from the National Oceanographic and Atmospheric Administration with partial support from the Title III Program at FAMU.
23. Bauer M, et al. (2006) Whole genome analysis of the marine Bacteroidetes ’Gramella forsetii’ reveals adaptations to degradation of polymeric organic matter. Environ Microbiol 8:2201–2213. 24. Purdy KJ, Embley TM, Nedwell DB (2002) The distribution and activity of sulphate reducing bacteria in estuarine and coastal marine sediments. Antonie Van Leeuwenhoek 81:181–187. 25. Alonso C, Pernthaler J (2006) Roseobacter and SAR11 dominate microbial glucose uptake in coastal North Sea waters. Environ Microbiol 8:2022–2030. 26. Moran MA, et al. (2007) Ecological genomics of marine roseobacters. Appl Environ Microbiol 73:4559 – 4569. 27. Eiler A, Langenheder S, Bertilsson S, Tranvik LJ (2003) Heterotrophic bacterial growth efficiency and community structure at different natural organic carbon concentrations. Appl Environ Microbiol 69:3701–3709. 28. Benner R (2003) Molecular indicators of the bioavailability of dissolved organic matter. Aquatic Ecosystems: Interactivity of Dissolved Organic Matter, eds Findlay SEG, Sinsabaugh RL (Academic, New York), pp 121–137. 29. Almeida MA, Cunha MA, Alcantara F (2001) Loss of estuarine bacteria by viral infection and predation in microcosm conditions. Microb Ecol 42:562–571. 30. Davidov Y, Friedjung A, Jurkevitch E (2006) Structure analysis of a soil community of predatory bacteria using culture-dependent and culture-independent methods reveals a hitherto undetected diversity of Bdellovibrio-and-like organisms. Environ Microbiol 8:1667–1673. 31. Zheng G, Wang C, Williams HN, Pineiro SA (2008) Development and evaluation of a quantitative real-time PCR assay for the detection of saltwater Bacteriovorax. Environ Microbiol 10:2515–2526. 32. Pin˜eiro SA, Williams HN, Stine OC (2008) Phylogenetic relationships amongst the 435 saltwater members of the genus Bacteriovorax using rpoB sequences and reclassification of 436 Bacteriovorax stolpii as Bacteriolyticum stolpii gen. nov., comb.nov. Int J Syst Evol Microbiol 58:1203–1209. 33. Shemesh Y, Jurkevitch E (2004) Plastic phenotypic resistance to predation by Bdellovibrio and like organisms in bacterial prey. Environ Microbiol 6:12–18. 34. Lueders T, Kindler R, Miltner A, Friedrich MW, Kaestner M (2006) Identification of bacterial micropredators distinctively active in a soil microbial food web. Appl Environ Microbiol 72:5342–5348. 35. Chauhan A, Ogram A (2006) Phylogeny of acetate-utilizing microorganisms in soils along a nutrient gradient in the Florida Everglades. Appl Environ Microbiol 72:6837– 6840. 36. Suzuki Y, Tanoue E, Ito H (1992) A high-temperature catalytic oxidation method for the determination of dissolved organic carbon in seawater: Analysis and improvement. Deep Sea Res 39:185–198. 37. Braman RS, Hendrix SA (1989) Nanogram nitrite and nitrate determination in environmental and biological materials by vanadium (iii) reduction with chemiluminescence detection. Anal Chem 61:2715–2718. 38. Solorzano L (1969) Determination of ammonia in natural waters by the phenolhypochlorite method. Limnol Oceanogr 14:799 – 801. 39. Lane DJ (1991) 16S/23S rRNA sequencing. Nucleic Acid Techniques in Bacterial Systematics, eds Stackebrandt E, Goodfellow M (Wiley, New York), pp 115–175. 40. Cole JR, et al. (2003) The Ribosomal Database Project (RDP-II): Previewing a new autoaligner that allows regular updates and the new prokaryotic taxonomy. Nucleic Acids Res 31:442– 443. 41. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ (1990) Basic local alignment search tool. J Mol Biol 215:403– 410. 42. Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F, Higgins DG (1997) The ClustalX windows interface: Flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res 24:4876 – 4882. 43. Lozupone C, Knight R (2005) UniFrac: A new phylogenetic method for comparing microbial communities. Appl Environ Microbiol 71:8228 – 8235.
Chauhan et al.
