The 2009–2010 period was marked by an episode of intense drought known as
the El Niño Modoki event. Sampling of the Términos Lagoon (Mexico) was
carried out in November 2009 in order to understand the influence of these
particular environmental conditions on organic matter fluxes within the
lagoon's pelagic ecosystem and, more specifically, on the relationship between
phyto- and bacterioplankton communities. The measurements presented here
concern biogeochemical parameters (nutrients, dissolved and particulate
organic matter [POM], and dissolved polycyclic aromatic hydrocarbons [PAHs]),
phytoplankton (biomass and photosynthesis), and bacteria (diversity and
abundance, including PAH degradation bacteria and ectoenzymatic activities).
During the studied period, the water column of the Términos Lagoon functioned
globally as a sink and, more precisely, as a “nitrogen assimilator”. This was due to
the high production of particulate and dissolved organic matter (DOM), even though
exportation of autochthonous matter to the Gulf of Mexico was weak. We found
that “bottom-up” control accounted for a large portion of the variability of
phytoplankton productivity. Nitrogen and phosphorus stoichiometry mostly
accounted for the heterogeneity in phytoplankton and free-living prokaryote
distribution in the lagoon. In the eastern part, we found a clear decoupling
between areas enriched in dissolved inorganic nitrogen near the
Puerto Real coastal inlet and areas enriched in phosphate (PO
Coastal lagoons are complex environments, combining features of shallow
inland waterbodies wholly or partly sealed off from the adjacent coastal
oceans. They are influenced by tides, river inputs, the precipitation versus
evaporation balance, and the surface heat balance. Interactions between
freshwater and marine sources generate strong gradients of salinity, light,
and nutrient availability (Hauenstein and Ramírez, 1986).
Biological diversity is generally high in these environments
(Milessi et al., 2010). Located in the southern Gulf of Mexico
near Campeche Sound, the Términos Lagoon is one of the largest tropical coastal
lagoons worldwide, and its recognized environmental importance and protected
status are potentially threatened by petroleum-related industrial activities
inshore and offshore (García-Ríos et al., 2013). A first
tentative budget of salt and nutrients concluded that the Términos Lagoon was
slightly autotrophic on a yearly basis (David, 1999), but
this assessment was clearly based on scarce environmental data.
Chlorophyll
In aquatic ecosystems, bacteria utilize a large fraction (up to 90 %) of the primary production, since algal carbon exudates can be the principal source for bacterial production (Cole et al., 1988; Conan et al., 1999). Besides the utilization of a considerable part of the available organic matter, bacterioplankton communities also absorb inorganic nutrients, thus competing with phytoplankton communities (Conan et al., 2007; Hobbie, 1988). The bulk of the organic matter is a highly heterogeneous matrix which is primarily composed of complex and refractory substrates (Hoppe et al., 2002), but which also contains labile substrates such as proteins or peptides, oligosaccharides, and fatty acids. Extracellular enzymes are thus essential to aquatic microorganisms as they allow for the partitioning of complex organic substrates, including high molecular weight compounds which cannot pass through the cell membrane (Arnosti and Steen, 2013). As a function of genetic diversity, the capacity to produce extracellular enzymes is differently distributed in the bacterial community, directly impacting the range of substrates metabolized (Zimmerman et al., 2013). This phenomenon has global-scale implications since several meta-analyses have clearly evidenced differences in the metabolic capacities of microorganisms from temperate, tropical, or high latitude waters (Amado et al., 2013; Arnosti et al., 2011). At a local scale, alteration of the evaporation–precipitation balance due to climate change can be challenging, especially in the case of a coastal lagoon, as it is well known that changes in salinity may alter bacterial diversity and activity (Pedrós-Alió et al., 2000). Local anthropogenic inputs of organic pollutants such as polycyclic aromatic hydrocarbons (PAHs) may also affect bacterial diversity and activity (Aguayo et al., 2014; Jiménez et al., 2011; Rodríguez-Blanco et al., 2010). Indeed, PAHs, which can comprise as much as 25–35 % of total hydrocarbon content in crude oils (Head et al., 2006), are among the most abundant and ubiquitous pollutants in the coastal environment (González-Gaya et al., 2016). These compounds are recognized by the European and US environmental agencies as priority pollutants for the aquatic medium due to their toxicity, persistence, and ability to accumulate in the biota (Kennish, 1992). Hence, the presence of PAHs in the marine environment may induce an increase in the indigenous populations of marine bacteria that can break down and utilize these chemicals as a carbon source, the so-called “PAH-degrading bacteria” or “PAH degraders”. These bacteria are generally strongly selected in oil-impacted ecosystems, where they may account for 70 to 90 % of the total bacterial community (Gutierrez et al., 2014; Head et al., 2006).
Despite their importance, few studies have considered the bacterial
communities of tropical inland aquatic ecosystems (Roland et
al., 2010) or coastal lagoons (Abreu et al., 1992; Hsieh et al., 2012;
MacCord et al., 2013; They et al., 2013) and almost none have dealt with
tropical coastal lagoons (Scofield et al., 2015). Among the existing
studies, very few have been conducted on bacterial communities and most of
them have been based on culture-dependent methods
(Lizárraga-Partida et al., 1987, 1986).
However, cultivable bacteria represent a very small fraction of the total
present bacteria (
Our study aims to evaluate the link between (i) biogeochemical (nutrients, dissolved and particulate organic matter [POM]), (ii) phytoplanktonic (biomass and photosynthetic activity), and (iii) free-living prokaryote (diversity, including PAH-degrading bacteria, and ectoenzymatic activities) parameters in the water column of the Términos Lagoon (Mexico) after a sustained period of minimum river discharge relative to the 2009–2010 El Niño Modoki episode. After having identified the main sources of nutrients in the lagoon (focused on nitrogen and phosphorus), we propose a geographical organization of the ecosystem to explain the distribution of the microbial pelagic communities across the lagoon.
Study site location and distribution of the 35 sampled stations in the lagoon.
The Términos Lagoon is a large (1936 km
Samples were collected at a 0.2 m depth at 35 stations distributed over the
whole lagoon (Fig. 1) from 21 to the 27 October 2009. In 2009, a
yearly cumulative discharge of
4.83
A vertical profile of temperature, salinity, and fluorescence was carried out
at each of the 35 stations using a Sea–Bird CTD probe (SBE 19) with a
precision of 0.01
As soon as the Niskin sampler was retrieved on board, a previously acid-washed 40 mL Schott® glass vial was rinsed with sampled water, filled, immediately injected with the fluorometric detection reagent for ammonia determination (as described in Holmes et al., 1999), sealed, and stored in the dark for later analysis in the laboratory. Following this, two 30 mL and one 150 mL plastic acid-washed vials were then rinsed with sampled water, filled, and stored in a specifically dedicated and refrigerated ice cooler, to be later deep-frozen in the laboratory while awaiting analysis of dissolved inorganic and organic nutrients, as follows:
Nitrate (NO
Samples for dissolved organic matter (DOM) were filtered through two
precombusted (24 h, 450
A 4 L acid-washed plastic container was also used for subsampling from the
Niskin bottle. This container was rinsed with sampled water, filled, and
stored in a dedicated ice cooler while awaiting filtration in the
laboratory as follows: 250 mL of seawater was filtered through a
precombusted (24 h, 450
For chlorophyll, 250 mL samples were filtered using 25 mm diameter Whatman® GF/F filters and immediately stored in liquid nitrogen. Chl and phaeopigment (Phaeo) were later extracted from the filters with 100 % methanol (Marker, 1972). Concentrations were determined by using the fluorometric technique (Lorenzen, 1966) on a Turner Design Trilogy fluorometer.
Photosynthetic–irradiance parameters (
Dissolved total PAH concentrations were determined by using the EnviroFlu-HC
submersible UV fluorometer (TriOS Optical Sensors, Germany), a commercially
available instrument dedicated to the in situ and real-time
quantification of PAHs in water. The sensor was calibrated in the laboratory
before the cruises following Tedetti et al. (2010) and Sauret et
al. (2016). In this work, the mean dissolved total PAH concentrations
derived from the sensor are given in ng L
Free-living prokaryotes were determined by flow cytometry
(Mével et al., 2008). Two-milliliter seawater samples were fixed
with 2 % formaldehyde for 1 h at 4
Nucleic acids were extracted on 0.2
Aminopeptidase,
The quantification of PAH-degrading bacteria was performed using the
most probable number (MPN) method. A total of 100
The comparative analysis of 16S rDNA- or 16S rRNA-based CE-SSCP fingerprints was
carried out with the PRIMER 6 software (PRIMER-E, Ltd., UK) using
Bray-Curtis similarities. We used the similarity profile test (SIMPROF)
(PRIMER 6) to check whether a specific sub-cluster can be recreated by randomly
permuting the entry ribotypes and samples when using hierarchical
agglomerative clustering. The significant branch (SIMPROF,
Canonical correspondence analysis (CCA) was used to investigate the
variations in the CE-SSCP profiles under the constraint of our set of
environmental variables using CANOCO software (version 5.0) as previously
described in Berdjeb et al. (2011). Significant variables (i.e.
variables that significantly explained changes in 16S rDNA- and 16S
rRNA-based fingerprints) in our dataset were chosen using a
forward-selection procedure. Explanatory variables were added until further
addition of variables failed to contribute significantly (
In the studied period, the Términos Lagoon was characterized by a northwest–southeast
positive temperature gradient from
Mapped distribution of the physico–chemical parameters
measured in the Términos Lagoon in October 2009 for
Nitrate and ammonium concentrations (Fig. 2c and d) were at maximum near
the Palizada river outlet (16.6 and 0.3
The distribution pattern for PO
The distributions of dissolved organic carbon (DOC; Fig. 2f), nitrogen (DON;
Fig. 2g), and phosphorus (DOP; Fig. 2h) concentrations followed a pattern
comparable to that of PO
The three rivers were clearly the main source of particulate organic nitrogen
and phosphorus in the lagoon (Fig. 3). PON reached a maximum
concentration of 9.3
Same as Fig. 2 but for
Chlorophyll (Chl) and phaeopigment followed a convergent
distribution pattern (Fig. 3c and d) with maximum concentrations close to
or in the vicinity of the Palizada River mouth (
The maximum rate of carbon production per unit of chlorophyll at light
saturation (
Free-living prokaryote abundance ranged from 1.0 to 4.
Cell-specific aminopeptidase (Leu-MCA) and phosphatase (MUF-P) activities
reached maximum values near the mouths of the Palizada and Chumpan
rivers at 33, and 131.9 fmol L
Same as Fig. 2 but for
Dissolved total PAH concentrations (Fig. 5a) were higher near the El
Carmen Inlet (332 ng L
Same as Fig. 2 but for
Bacterial community structure defined as a function of 16S rDNA-based fingerprints from each sample singled out three individual stations (Palizada River, El Carmen Inlet and Candelaria River) and aggregated five groups of stations (Fig. 6a). Three of these groups included a large number of samples: nine stations located in the northeastern part of the lagoon near the Puerto Real Inlet; nine stations positioned in the middle of the lagoon north from Chumpan River; and eight stations situated to the southwest of Carmen Island. Two other groups with fewer stations identified intermediate communities found between the El Carmen Inlet and the Palizada River in the western part of the lagoon (stations 2, 4, and 6) and the Candelaria River in the middle of the lagoon (stations 22, 24, and 27).
Non-metric multidimensional scaling (NMDS) plot of the
Metabolically active bacterial communities as a function of 16S rRNA-based fingerprints singled out two stations (Palizada River and El Carmen Inlet) and aggregated five groups of stations that were slightly different from the DNA-based clusters (Fig. 6b). Three of these groups included a large number of samples: 15 stations located in the eastern part of the lagoon; 9 stations in the middle of the lagoon north of the Chumpan River; and 5 stations in the northwestern part of the lagoon near the El Carmen Inlet. Two other groups with fewer stations showed intermediate communities found near the Palizada River mouth (stations 6 and 8) and further east (stations 9 and 12).
To analyze the main environmental factors controlling the spatial distribution of total (Fig. 7a) and active (Fig. 7b) prokaryote communities, we performed a canonical correspondence analysis. In both DNA- and RNA-based analyses, the cumulative percentage of variance of the species–environment relationship indicated that the first and second canonical axes explained 48 and 24 % of the total variance for DNA and 45 and 31 % for RNA, respectively. The remaining axes accounted for less than 14 % of the total variance each and thus were not considered as significant enough.
Canonical correspondence analysis of
In the DNA-based CCA, the first canonical axis was positively correlated
with NO
With a contribution of about 76 % to the river inputs in the lagoon
(Fichez et al., 2016; Jensen et al., 1989), the Palizada River delivers most
of the new nitrogen input as nitrate and ammonium. High concentrations of
nitrogen were also measured in the Puerto Real Inlet, suggesting a second
nitrogen source from coastal seawater. These two sources have clearly
different impacts on primary producer development and activity as shown by
the Phaeo : Chl ratio (
Moreover, Day et al. (1982) demonstrated that small additions of filtered mangrove water had a stimulatory effect on pelagic primary production in the Términos Lagoon. This observation was later confirmed by Rivera-Monroy et al. (1998) who also evidenced a large temporal variability in the stimulating effect and a rapid inhibition due to variable humic substance concentrations. The relative decrease of productivity close to the Palizada plume could be due to humic matter, as we also found relatively high concentrations in dissolved PAHs (see hereafter Sect. 4.4). Finally, it is clear that bottom-up control of the system (by nutrients and/or humic substances) drove the differential responses of phytoplankton productivity in the eastern and western part of the lagoon, probably in conjunction with grazing activity (top-down control).
At the time of our study, the Palizada River and Puerto Real Inlet were major
sources of nitrogen for the lagoon. Sediments are generally considered to be
a significant internal source of nutrients in shallow coastal ecosystems,
but they may also be a net sink of dissolved nitrogen, at least during
certain times of the year (Sundbäck et al., 2000; Tyler et al.,
2003). Rivera-Monroy et al. (1995a) measured nitrogen fluxes
between Estero Pargo (an unpolluted tidal creek), and a fringe mangrove
forest in the Términos Lagoon. They reported that mangrove sediments were a sink
of NO
Same as Fig. 2 but for the NOP : PP ratio.
Our large-scale study considering the whole lagoon provides some information
about the potential origin of phosphorus in the water column. It is clear
from our measurements that the phosphate distribution in the lagoon is
disconnected from nitrogen. This impacts the stoichiometry of particulate
organic matter (N : P ratio) through the whole lagoon, as shown by the
surprisingly and relatively low values of the PON : PP ratio (
Our analysis of biogeochemical trends in the Términos Lagoon has been combined
with the study of the spatial distribution of prokaryotic extracellular
activity. Bacterial aminopeptidase and lipase extracellular activities play
a key role in the transformation of biopolymers into small monomers,
since a large part of organic matter is in the form of large molecules
whereas small molecules (
Phosphatase activity is well known to be controlled by the availability of
soluble reactive phosphorus (Van Wambeke et al., 2009). This
activity was essentially observed in the vicinity of the Palizada River,
which is the main source of PP in the lagoon. However, in the Puerto Real Inlet,
which includes the two PO
Molecular fingerprinting (such as CE-SSCP) and next-generation sequencing technologies generally yield converging results (Ghiglione et al., 2005; Ghiglione and Murray, 2012; Ortega-Retuerta et al., 2012; Sauret et al., 2015), evidencing clear shifts in bacterial community structure as a function of changes in biogeochemical characteristics (Ghiglione et al., 2005). Numerous factors can regulate microorganism population dynamics, often simultaneously, and the literature contains evidence (Berdjeb et al., 2011; Fuhrman et al., 2013; Ghiglione et al., 2008) underlining the importance of relevant statistical analyses to investigate the relative importance of environmental factors in predicting bacterial community dynamics. It is generally recognized that the expression of ectoenzyme activities could result from species selection and population dynamics (Martinez et al., 1996), and the zonation of prokaryotic community structure in the eastern, middle, and western parts of the lagoon agree with such a paradigm. The community composition in the eastern part could be divided into two sub-clusters corresponding to the respective influences of the Palizada River mouth and El Carmen Inlet. Both DNA- and RNA-based fingerprinting show that the Palizada River and El Carmen Inlet hosted distinct prokaryotic communities, as previously observed in transition zones such as rivers (Ortega-Retuerta et al., 2012) or lagoon mouths (Rappé et al., 2000). The relationship between community composition and ectoenzyme activities was particularly evident when the lipase and aminopeptidase rates were considered. Lipase activity was magnified in the middle of the lagoon, with a south to north increasing gradient from the Chumpan River to Carmen Island that coincided with specific communities (cluster II in both DNA- and RNA-based fingerprinting). Other communities were found in the western part under the influence of the Palizada River where higher aminopeptidase activity was measured.
The combination of DNA and RNA strengthens our observations as DNA-based
analysis alone would have failed to distinguish between active, dormant,
senescent, or dead cells and would thus prevent the assessment of the level of
activity of each detected bacterial population
(Rodríguez-Blanco et al., 2010). Even though the
abundance of bacteria in the sea is high, only a small fraction is
considered to be metabolically active (Del Giorgio and Bouvier,
2002). Bacterial growth rate has been shown to correlate with cellular rRNA
content (Kemp et al., 1993); therefore, information on
cellular activity may be obtained by tracking reverse-transcribed 16S rRNA
(Lami et al., 2009). In the present study, we focused on the free-living
prokaryotes and disregarded the particle-attached fraction by pre-filtering
the water by 3
Through the use of direct gradient multivariate ordination analyses, we
demonstrate that a complex array of biogeochemical parameters was the
driving force behind prokaryotic community structure shifts in the Términos
Lagoon. Physico–chemical parameters such as nitrate, oxygen, dissolved
organic matter (DOC, DON, DOP), and chlorophyll
The concentration of dissolved total PAHs was also a significant explanatory
variable of the metabolically active bacterial community structure. PAHs are
considered the most toxic component of crude oil to marine life and are
ubiquitous pollutants in the coastal environment (Kennish, 1992).
Our study was performed just before the 2010 Deepwater Horizon blowout
in the Gulf of Mexico, but several offshore oil platforms exist in the
shallow waters of the Campeche Bank in the southern part of the Gulf of Mexico,
for example in the Campeche field (Cheek-1) which is only 60 km north of
the Términos Lagoon (Warr et al., 2013). The coast of Campeche
itself was also impacted by the 1979
This study provides a new original set of biogeochemical characteristics for
one of the largest shallow tropical coastal lagoons. Due to the 2009–2010 El
Niño Modoki episode, climatic conditions in the Términos Lagoon region were
exceptionally dry at the time of our sampling, hence potentially indicative
of future environmental conditions resulting from the predicted trends in
climate change in the Central American region. We evidenced a clear
distinction in ecosystem functioning between the eastern and western parts
of the lagoon. Most of the oceanic water entering through the inlets spread
toward the southeast where dissolved organic matter accumulated. This area
did not support significant phytoplankton development. In the west, we
hypothesized a balance shift between top-down and bottom-up control to
explain the different responses in terms of phytoplankton productivity. The
decoupling between nitrogen inputs brought by oceanic waters, the
Palizada River, and phosphate inputs from the Chumpan River did not allow
for phytoplankton C fixation. Most of the phytoplankton biomass was
aggregated around the Palizada River mouth (which carried most of the
freshwater into the lagoon) in a P-depleted area (low phosphate
concentration and high bacterial phosphatase activity). Bacterial ectoenzyme
activities were mainly observed in the middle of the lagoon along a south
to north cross section stretching from the Chumpan River up to Carmen
Island. Maximum mineralization activities were found in this area, which
coincided with high extracellular lipase and aminopeptidase activities and
low DOC and O
Another significant outcome of our study has been to (i) link the spatial distribution of ectoenzymatic activities with changes in prokaryotic community structure and (ii) show that a combination of a complex set of physical and biogeochemical parameters was necessary to explain the changes in prokaryotic community structure. This study also emphasizes the use of direct multivariate statistical analyses to keep the influence of pollutants in perspective without denying the role of other physico–chemical variables to explain the dynamics of prokaryotic community structure in polluted areas.
Our study provides an extensive dataset efficiently mixing biogeochemical status with information on phytoplankton and prokaryotic structure and dynamics. This has never before been measured in the Términos Lagoon and the outcome offers a strong base of information and reflection for future studies on this essential coastal system and into the potential environmental conditions which might prevail as a consequence of future climate change. Further studies are needed to compare our dataset with high river input regime conditions and asses both how this might affect the observed uncoupling between nitrogen and phosphate and the dominant source of phosphorus and its consequences on primary production and prokaryotic activities. Finally, the role of top-down control should also be investigated in order to better understand the variability of the observed responses.
Data for this paper are now referenced in
The authors declare that they have no conflict of interest.
The present work was conducted within the frame of the Joint Environmental Study of the Términos Lagoon (JEST) and jointly financed by the French National Program EC2CO-DRIL, the Institut de Recherche pour le Développement (IRD), the Centre National de la Recherche Scientifique (CNRS), the University of “Université de Lille-I”, and the Universidad Autónoma Metropolitana-Iztapalapa (UAM-I). The authors are extremely grateful to the Instituto de Ciencias del Mar y Limnologia and the Universidad Nacional Autónoma de México (ICML-UNAM) for providing full access to their field station in Ciudad del Carmen. We also acknowledge P. A. and M. V. Ghighi for carefully proofreading. Thanks also to Julia, Joan, and Vivien for their support. M. Origel-Moreno was financially supported by Bonafont S. A De .C.V. and CONACyT during her PhD work. Edited by: C. Woulds Reviewed by: two anonymous referees